Binding of serum response factor to cystic fibrosis transmembrane cond

http://www.100md.com 《核酸研究医学期刊》

     Laboratoire de Génétique Moléculaire et Chromosomique, Institut Universitaire de Recherche Clinique Montpellier, France 1Laboratoire de Prolifération et Différentiation Cellulaire, Institut de Génétique Humaine Montpellier, France 2Laboratoire de Génétique Moléculaire, Institut de Recherches Cliniques de Montréal Quebec, Canada 3Centre de Séquen?age Génomique, Institut de Génétique Humaine Montpellier, France

    *To whom correspondence should be addressed. Tel: +33 4 67 41 53 60; Fax: +33 4 67 41 53 65; Email: Marie-Catherine.Romey@igh.cnrs.fr

    Revised August 25, 2005. Accepted August 30, 2005.

    ABSTRACT

    CFTR expression is tightly controlled by a complex network of ubiquitous and tissue-specific cis-elements and trans-factors. To better understand mechanisms that regulate transcription of CFTR, we examined transcription factors that specifically bind a CFTR CArG-like motif we have previously shown to modulate CFTR expression. Gel mobility shift assays and chromatin immunoprecipitation analyses demonstrated the CFTR CArG-like motif binds serum response factor both in vitro and in vivo. Transient co-transfections with various SRF expression vector, including dominant-negative forms and small interfering RNA, demonstrated that SRF significantly increases CFTR transcriptional activity in bronchial epithelial cells. Mutagenesis studies suggested that in addition to SRF other co-factors, such as Yin Yang 1 (YY1) previously shown to bind the CFTR promoter, are potentially involved in the CFTR regulation. Here, we show that functional interplay between SRF and YY1 might provide interesting perspectives to further characterize the underlying molecular mechanism of the basal CFTR transcriptional activity. Furthermore, the identification of multiple CArG binding sites in highly conserved CFTR untranslated regions, which form specific SRF complexes, provides direct evidence for a considerable role of SRF in the CFTR transcriptional regulation into specialized epithelial lung cells.

    INTRODUCTION

    Expression of the cystic fibrosis transmembrane conductance regulator (CFTR) gene occurs in a subset of specialized cells of epithelial origin (1–3) and is tightly regulated both temporally and spatially (4,5). However, no clear mechanism responsible for this regulation has yet been reported in part due to the complexity of the non-coding regions structure of the CFTR gene.

    Despite the absence of a TATA element, the transcription of the CFTR gene may be initiated through several discrete start sites (6,7), including those that are tissue-specific (8). An intact CCAAT consensus is also required for accurate transcript initiation (9). Both basal and cAMP-mediated regulation of CFTR transcription involve a weak cAMP response element (CRE) nucleotide consensus (10,11) in tandem with a consensus inverted CCAAT element or Y box (9). In addition, CFTR transcription may be modulated by additional overlapping cis-acting elements, including a polymorphic YY1 site, located in the human minimal CFTR promoter (12). CFTR promoter activity may be enhanced by exogenous transfected NF-B via its binding to the –1103 B element (13). CFTR expression is also regulated through interactions with factors involved in remodeling of chromatin structure, such as CDP/cut and hGCN5 (14).

    Because none of these cis-acting elements confers tissue-specific control of CFTR expression, several studies focused on identifying potential regulatory elements that lie mainly outside the coding region. Extensive analyses of the chromatin structure identified multiple clusters of DNase I-hypersensitive sites (DHSs) (15,16), some of which contain tissue-specific enhancer elements (17,18). Multiple binding sites for a tissue-specific transcription factor, called HNF1, recently identified in various DHS core, modulate likewise CFTR expression (19). Despite extensive studies, the transcriptional regulatory networks of CFTR expression remain to be unraveled.

    Our current model for the regulation of human CFTR expression proposes that multiple transcription factors interact with previously reported polymorphic and composite cis-acting elements encompassing a CArG-like motif (12,20). The CArG box (CC(A/T)6GG), originally defined as the core component of the serum response element (SRE), is found in the 5' region of immediate early response genes, such as c-fos (21), and in many muscle-specific gene promoters (22). This DNA element binds serum response factor (SRF), highly regulated and interactive transcription factor, which is phylogenetically conserved and belongs to the MADS box family of proteins, with MCM1, Agamous, Deficiens and SRF (23). Although SRF is particularly enriched in skeletal, cardiac and smooth muscle lineages (24,25), it is also expressed in a variety of human cell lines from distinct tissue origins (26), and, to a lesser extent, in ectoderm-derived tissue such as gastric epithelium (27). Interestingly, it has been shown that Drosophila SRF homolog is expressed in a subset of tracheal cells and regulates cytoplasmic outgrowth during terminal branching of the tracheal system (28). Hence, it has been suggested that SRF plays an essential role in the formation of the respiratory system. Moreover, it has been recently shown that SRF is involved in myofibroblast differentiation during lung damage (29). Therefore, we hypothesized that trans-acting factors interacting with CArG-like motif, such as SRF, could be important in CFTR expression regulation. We provided evidence that SRF is expressed in bronchial epithelial cell expressing endogenous CFTR protein. Through the use of combined gel mobility shift–chromatin immunoprecipitation (ChIP) assays, we first demonstrated that SRF protein binds, both in vitro and in vivo, the CFTR minimal promoter region spanning the CArG-like motif. Mutagenic studies and transient co-transfections experiments showed that SRF alone is not sufficient to transactivate the CFTR gene. Upon some mutations of the CFTR-CArG-like element, promoter activity is severely decreased, potentially implying YY1 as a co-repressor. Given the YY1 well-defined role in the transcriptional regulation of some SRE-dependent promoters (30–33), we explored the putative role of YY1 on the CFTR promoter activity and demonstrated functional interplay between this factor and SRF. Finally, by stringent in silico structural analysis, we identified additional multiple near consensus CArG sites within the highly conserved promoter and intronic regions of the CFTR gene. We also documented the existence of SRF-containing protein complexes on numerous predicted CFTR CArG-like elements. Taken together, the data obtained classify CFTR as a novel SRF target gene, subject to modest but significant SRF activation, and open the gate to further investigate how SRF may be involved in the regulation of CFTR gene expression.

    MATERIALS AND METHODS

    Cell culture

    Human bronchial epithelial cells, Beas2B and A549 expressing endogenous CFTR were kindly provided by Dr Marc Chanson, Department of Pediatrics, Laboratory of Clinical Investigation, Switzerland. Beas2B and A549 cells were maintained in high glucose DMEM (4500 mg/l D-Glucose, Invitrogen Corporation) supplemented with 10% fetal bovine serum (FBS) (Eurobio, France) and 5 mM L-glutamine. C2C12 myoblasts were grown in DMEM supplemented with 10% FBS and 5 mM L-glutamine. All cell cultures, supplemented with 100 U ml–1 penicillin and 100 mg ml–1 streptomycin, were maintained at 37°C under 5% CO2. As positive or negative controls for the different assays, other cells used in this study included colon adenocarcinoma Caco-2 cell line (ECACC) with functional CFTR (34,35) and normal monkey kidney fibroblast-like COS-7 cell line (ATCC) which does not express CFTR (36,37); all were maintained in DMEM supplemented with FBS and were incubated in the same conditions as above.

    Plasmids constructs

    The original luciferase expression vector pGL3-Basic containing the wild-type human CFTR minimal promoter was described previously (12). To further study the putative composite cis-acting element located at –108 from the major transcription initiation site of the human CFTR promoter (38), the CFTR-CArG-like motif was either mutated in a consensus CArG box or in a degenerated motif. Mutations were introduced into the WT-pGL3 plasmid using an oligonucleotide-directed mutagenesis system (QuickChangeTMSite-Directed Mutagenesis Kit, Stratagene) according to the manufacturer's instructions. Mutagenesis primers are depicted in Table 1. The presence of mutations and sequence fidelity were verified by dideoxynucleotide sequencing. For each introduced mutation the loss or the improvement of transcription factor binding activity was checked by in vitro assays. The 3xSRE-fos TATA-luciferase reporter construct was kindly provided by R. A. Hipskind, Institut de Genetique Moleculaire, Montpellier, France. Expression vectors encoding either human wild-type SRF, the pHiv-SRF or SRF dominant-negative form deleted of an essential part of DNA-binding domain (DBD), the pHiv-DN, were generously given by D. Trouche, Laboratoire de Biologie Moleculaire Eucaryote, Toulouse, France. Another dominant-negative SRF mutant, pHiv-DBD, was constructed by PCR amplification of SRF fragment spanning the DBD flanked by HindIII and HincII sites to facilitate cloning into expression vector. The PCR was performed using pHiv-DBD-SRF primers (listed in Table 1), 10 ng of SRF expression vector, 200 μM of each dNTP, 0.5 μM of each primer and 1.5 U Taq DNA polymerase (Perkin Elmer, Inc.) in a 1x reaction buffer provided by the manufacturer. TranSilent human control siRNA vector and TranSilentTM human SRF siRNA vector (catalog no. SR1038) were purchased from Panomics. Expression vectors encoding either pCMV-YY1 or pcDNA3-YY1 were generously provided by E. Seto, E. Bonnefoy and A. Moustakas, respectively.

    Table 1 Sequences of oligonucleotides

    Transient transfections

    Cells were seeded at a density of 200 000 cells/2 ml of medium and plated in 6-well dishes (NUNClone?, Merck-Eurolab, Inc.). After a culturing period of 24 h, cells were transfected with PolyFect? transfection reagent (Qiagen Corporation, France) according to the manufacturer's recommendations. For transient transfection, we used 1.8 μg of plasmid reporter and 0.2 μg of internal control pRL-SV40 containing Renilla luciferase (Promega Corporation). The pGL3-Basic vector lacking both eukaryotic promoter and enhancer sequences was used as a negative control. After 48 h, the luciferase activity was evaluated with the Dual Luciferase Reporter Assay System (Promega Corporation), as described previously (12). Luminescence measurements were performed on a Luminoskan Ascent luminometer (ThermoLabsystem Corporation, France). Firefly luciferase activity was normalized to Renilla luciferase activity. All luciferase activities represent at least three independent experiments with each construct tested in triplicate per experiment. Luciferase activity data are expressed as the means ± SD computed from the results obtained from each set of transfection experiments. To minimize the possibility of errors in DNA amplification, at least two independently constructed clones were tested. For co-transfection assays, unless otherwise indicated, 1.8 μgof luciferase reporter and 1–2 μg of each expression vector were used. Transfection with siRNA was performed using Lipofectamine 2000 (Invitrogen) with 2.5 μg of siRNA. The cells were transfected with CFTR promoter plasmid reporter, and luciferase activity was determined as described above. Aliquots of cell lysates were used for western blotting to confirm the specificity and level of SRF protein knockdown.

    Preparation of whole cell proteins, nuclear extracts and recombinant protein production

    For the CFTR immunodetection analyses, Beas2B, A549, C2C12, COS-7 and Caco-2 cells were treated with 5 mM sodium phenylbutyrate, which has been previously reported to increase CFTR (39) and scraped into lysis buffer (50 mM Tris–HCl, 1% NP-40, 40 mM beta-glycerophosphate and 120 mM NaCl) supplemented with protease inhibitors . Nuclear extracts were prepared using the Nuclear extraction kit according to the procedure recommended by the supplier (Panomics, Ozyme, France). Protein concentrations were determined using the Bio-Rad Protein Assay kit with BSA as a standard, based on the classical Bradford method. The pGEX-SRF plasmid, a generous gift from D. Trouche (Laboratoire de Biologie Moleculaire Eucaryote), was used for protein expression. GST–SRF fusion protein was purified from Escherichia coli as described previously (40). The purity and concentration of protein samples were estimated by SDS–PAGE.

    Electrophoretic mobility shift assay (EMSA) and supershifts

    Oligonucleotides probes were synthesized corresponding to CFTR CArG-like element and mutant versions in which the CArG motif was either mutated in a consensus box or disrupted. The oligonucleotide sequences were listed in Table 2. Single-stranded complementary oligonucleotides containing the studied sequences were annealed, end-labeled with ATP (5000 Ci/mmol) and purified as described previously (12). Mobility assay conditions were specifically optimized for binding of SRF. The binding reaction was carried out in a total volume of 20 μl containing 2 ng (30 000 c.p.m.) of the labeled probe, 10 μl of binding buffer (2x), 0.25 μg poly(dI–dC) (Amersham Pharmacia) used as a non-specific competitor, 5–12 μg of nuclear extracts and 100–1000 ng of GST–SRF fusion chimera protein. For EMSA competition and antibody interference assays, proteins were incubated with either cold specific competitors (WT and SRE oligonucleotides, see Table 2), unspecific competitor (GATA oligonucleotide, see Table 2) or purified antibodies (anti-SRF: sc-335X; anti-YY1: sc-1703X or anti-HA, Santa Cruz, Biotechnology Inc., TEBU, France) for 20 min before addition of labeled probes. After incubation of 30 min at room temperature, complexes were resolved on 5% acrylamide:bisacrylamide (29:1) native gel containing 0.5x TBE at 26 mA for 1 h.

    Table 2 Ologonuc leotides used in EMSA analysis

    Formaldehyde cross-linking and chromatin immunoprecipitation

    Beas2B and A549 cells proteins were cross-linked to DNA by adding formaldehyde to a final concentration of 1% for 10 min at 37°C. The fixed cells were prepared for immunoprecipitation using the protocol of ChIP assay kit (Upstate Biotechnology Inc., Euromedex, France) with minor modifications. Cells were resuspended in 1 ml cold phosphate-buffered saline (PBS), pelleted by centrifugation and lysed in 800 μl of SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris–HCl, pH 8.1) containing protease inhibitors (1 mM PMSF, 1 μg/mlaprotinin and 1 μg/ml pepstatinA). After incubation on ice for 10 min, the cross-linked chromatin was sheared by sonication at the following conditions eight times for 10 s at 40% of maximum power and diluted 10-fold in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl, pH 8.1 and 167 mM NaCl) with protease inhibitors. Aliquots containing 20 μl of each aliquot were used as a control to show total input DNA. The chromatin solution was precleared with 80 μl of salmon sperm DNA/protein A agarose-50% slurry (Upstate Biotechnology Inc.) for 1 h at 4°C under agitation, and subsequently incubated at 4°C on a rotating stand overnight with either 3 μl of anti-SRF antibody (sc-335X, Santa Cruz, Biotechnology Inc.) or an irrelevant antibody (anti-HA). Immune complexes were precipitated by the addition of 60 μl of salmon sperm DNA/protein A agarose and incubated at 4°C for 1 h and low speed spinning. Precipitates were washed once with 1 ml of low salt buffer, once with high salt buffer, once with LiCl buffer and twice with 1x TE, according to the manufacturer's recommendations (Upstate Biotechnology Inc.). The cross-links were reversed by the addition of 20 μl of 5 M NaCl and incubation at 65°C overnight. The samples were treated with proteinase K, phenol–chloroform extracted, ethanol precipitated and resuspended in 30 μl of sterile water. PCR analysis was carried out using primers from different regions of the human CFTR promoter and as additional controls from promoter regions of a number of genes that are either silent (?-globin) or having at least one functional CArG element (?-actin). The sequences of PCR primers are listed in Table 1. After 45 cycles of amplification, PCR products were separated on 1.8% agarose gels and visualized by ethidium bromide staining and UV transillumination.

    Reverse transcriptase–PCR

    Total RNA from epithelial cells (Beas2B and A549), myoblasts (C2C12) and fibroblasts (COS-7) was isolated using the SV Total RNA Isolation System? (Promega, France). Reverse transcription was performed with 3 μg of total RNA, 300 ng of random hexamers (Invitrogen Corporation), 10 mM dNTPs, 1 μl of RNasin RNase inhibitor (Promega), 2 μl of DTT at 0.1 M, 4 μl of first strand buffer (Invitrogen), 200 U of M-MLV RT (Invitrogen) and nuclease-free water (Promega) in a final reaction volume of 20 μl. For each RNA template, a control reaction without RT was included. For each RT reaction condition, an H2O no-template reaction was included as an additional negative control. The reaction mixture was incubated first for 10 min at 25°C, then 40 min at 42°C; this was followed by a heat inactivation step of 72°C for 3 min. Aliquots containing 1 μl of the cDNA synthesis reaction mixture were used for PCR analyses. The expression level of SRF mRNA was analyzed by PCR amplification of the endogenous Hypoxanthine PhosphoRybosyl-Transferase (HPRT) gene used as internal control. Amplification mixture included either 10 pmol of each SRF-specific primer or 10 pmol of each HPRT-specific primer, 1.5 U of AmpliTaq polymerase (Applied Biosystem Corporation) and nuclease-free water (Promega) in a 25 μl volume. Amplification was performed with an initial denaturation step at 95°C for 5 min, followed by 35 cycles of 30 s of 95°C denaturation, 60 s of 52°C annealing and 2 min extension at 72°C. Amplification of CFTR was performed as described previously (41). The RT–PCR products were separated on an 8% non-denaturing polyacrylamide gel, stained with ethidium bromide. To ensure that the sequence of cDNA products derived from both SRF and CFTR RT–PCR analyses correspond to respective published sequences, the cDNA bands were purified and prepared for dideoxy DNA sequencing. Controls included RNA preparations from monkey kidney fibroblasts (COS-7), which are negative for CFTR expression (36,37).

    Western blot analysis

    Twenty-five microgram protein extracts (except where otherwise indicated) were resolved on a 10% (except where otherwise indicated) SDS–polyacrylamide gel and transferred onto a nitrocellulose membrane (except where otherwise indicated). All membranes were Ponceau S stained to ensure sample integrity. Briefly, the membranes were blocked with 5% skim milk in PBS supplemented with 0.01% Tween-20 and incubated with diluted primary antibodies in 5% skim milk (1:5000 diluted anti-SRF Abs: sc-335X and sc-25290X purchased from Santa Cruz; 1:1000 diluted anti-CFTR Ab: 24.1 from R&D System; 1:2500 diluted anti-?-actin antibody: AC-15 from Sigma, 1:2500 diluted anti-?-tubulin antibody: T6074 from Sigma) overnight. The membranes were then washed and incubated with an appropriate horseradish peroxidase (HRP) conjugated secondary antibody at 1:15 000 in PBS-5% milk. The membranes were reacted with chemiluminescence reagent ECL (Roche Diagnostic) as described by the manufacturer and subsequently exposed to Biomax photographic film (Kodak Corporation). The protein levels of the actin housekeeping gene (except where otherwise stated) were assayed for internal control of protein loading.

    Indirect immunofluorescence

    A549, Beas2B, C2C12 and COS-7 cells were plated on Lab-TekTMII CC2TM Chamber SlideTM systems (Nalge Nunc) 1 day before immunohistochemical analysis. The cells were fixed and permeabilized in 4% paraformaldehyde and 0.1% Triton X-100, respectively, for 20 min durations at room temperature. Non-specific binding sites were blocked with 1% BSA in PBS for 1 h. The cells were subsequently incubated with primary antibodies diluted in 1:100 in PBS–BSA. Antibodies used included, the clone L12B4 (Upstate biotechnology), an anti-CFTR antibody known to recognize human, rat, mouse and bovine CFTRs (42) and the H-300 or the A11 anti-SRF antibody (sc-13029 and sc-25290, Santa Cruz). The specificity of the anti-CFTR antibody has been described previously (43,44) and was also verified by using a CFTR-positive human Caco-2 cell line (34,35). For nuclear counterstaining, DAPI (sc-3598, Santa Cruz) was applied at a concentration of 1:5000. Cells were then washed with PBS–BSA and incubated with either donkey anti-rabbit fluorescein isothiocyanate-conjugated (FITC) or donkey anti-mouse Cy3TM conjugated secondary antibody (Jackson ImmunoResearch Laboratory Inc.) diluted 1:200 for 1 h at room temperature. Cells were washed twice with PBS–BSA, once with PBS, and mounted with Dakocytomation fluorescent mounting medium (Dakocytomation). Indirect immunofluorescence was examined with an inverted Leica DMR microscope (Leica Microsystemes SAS). Final images were further processed using Adobe Photoshop 7.0 software.

    Statistical analyses

    Transfection data are expressed as the mean ± SE. Paired comparisons were made using student's t-test. Data were considered statistically significant at P < 0.05. All graphical data and statistical analyses were generated with GraphPAD Prism software (Version 3.0). Densitometric analysis was performed using computerized densitometry and ImageQuantTL v.3.0 software (Amersham Biosciences). Densitometric values for both transcript and protein bands were determined in areas of equal size and are reported in normalized arbitrary units relatives to the expression of either HPRT gene or actin protein set to 1 (except where otherwise indicated).

    Computational analyses

    The computer program ConSite (http://www.phylofoot.org/), particularly suited to the analysis of large genomic sequences, was used to identify putative CArG binding sites within the promoter and intronic regions of the human CFTR gene. The algorithm AliBaba 2.1 (http://www.alibaba2.com/) was then used to verify these predictions and minimize false negatives or positives. To prevent a possible bias introduced by palindromic or internally repetitive cis-regulatory elements, overlapping matches, including on opposite DNA strands, were defined as a single match.

    To assess the importance of the putative CArG elements identified in the non-coding regions of the human CFTR gene, we performed multiple sequence alignments. The human, cynomolgus, gibbon, squirrel monkey, rabbit, pig, sheep and cow sequences were aligned using the ClustalW global alignment tool (http://clustalw.genome.ad.jp/) (45). The gap open and extension penalties were set at 15 and 6.66, respectively, and sequence delay divergence of 30% with a DNA transition weight of 0.5 was used.

    RESULTS

    CFTR minimal promoter region encompassing the CFTR-CArG-like element is remarkably conserved

    To search the cis-acting elements involved in the CFTR gene expression regulation, we previously performed an in silico search for transcription factor binding sites (46), and reported that the human minimal CFTR promoter contains a CArG-like motif, which is localized –108 bp upstream of the major transcription initiation start site (38). Because so far all described CArG-box elements were found in the regulatory regions of immediate early genes (47) and in muscle-specific genes (22), we confirmed our first binding site prediction by using two other algorithms, AliBaba (http://www.alibaba2.com/) and ConSite (http://clustalw.genome.ad.jp/). As shown in Figure 1A, the CFTR-CArG-like element sequence (GC(A/T)6GG) diverges slightly at the 5' from the published SRF consensus sequence (CC(A/T)6GG) (22,37). However, this divergence is similar to that observed in the smooth muscle-gamma-actin promoter which has been demonstrated to bind SRF (48).

    Figure 1 Cross-species comparison of the CFTR-CArG-like element. (A) Visualization of consensus CArG binding site: pop-up window containing information about SRF transcription factor obtained from in silico analysis performed with the ConSite tool. (B) Detailed alignment view of the CArG-like element-including region of human CFTR minimal promoter with seven orthologous sequences: Cynomolgus, Gibbon, Squirrel Monkey, Rabbit, Pig, Sheep and Cow. Dashes indicate gaps that were introduced to maximize alignment. Each sequence is represented by a classical text representation with each nucleotide assigned its own color (A is colored in red, C in blue, G in green and T in black). The numberings above the Cynomolgus sequence are indicated according to the major human transcription initiation site. In the left panel, the phylogenetic relationships between the different species are illustrated in a rooted tree, directly derived from ClustalW program. In the right panel, human/species homologies are indicated. The accession number of each CFTR genomic sequence is indicated below each species name, in brackets and bold characters. The human CFTR-CArG-like motif is noted above the Cynomolgus sequence.

    To assess the putative functional importance of CFTR-CArG-like binding site, we used an orthogonal approach of deducing regulatory elements by considering orthologous CFTR minimal promoter regions from several mammalian species representing the Primates, Artiodactylia and Lagomorpha orders. Comparison of human/species homologies was carried out using ClustalW multiple CFTR sequence alignments according to the numerous alignment parameters described previously (49). Consistent with previous studies (50,51), we showed that the human minimal CFTR promoter is remarkably conserved across the orders studied (Figure 1B, right panel). We found that the CFTR-CArG-like element is conserved within all the Primates species studied and also within the pig while, within the sheep, cow and rabbit species, some substitutions occur in the middle of the CArG-like motif (Figure 1B). These data are consistent with previous CFTR promoter phylogenetic work that identified two types of regulatory elements: some are conserved between species, such as a non-consensus CRE at positions –0.1 kb relative to ATG and some are species-specific elements, such as a 300 bp purine.pyrimidine (Pu.Py) stretch that is present only in rodents (50).

    Bronchial epithelial expression of SRF

    Because we could not previously identify SRF binding to the CFTR CArG-like element, even though an anti-SRF antibody had been used (12), we asked whether SRF, mainly described as highly expressed in muscle cells, was also present in epithelial cell lines. To measure the expression levels of SRF mRNAs, we used a RT–PCR-based procedure previously described by Davis et al. (52). As reverse-transcribed control, the Hypoxanthine PhosphoRybosyl-Transferase (HPRT) housekeeping gene was amplified. Amplification of SRF from bronchial epithelial cells and myoblasts yielded a fragment of the same size, 651 bp (Figure 2A), as that of the full-length SRF transcript produced by RT–PCR from human hearts (52). The 453 bp band corresponding to SRF lacking exon 5 (SRF5) was only detected in the C2C12 myoblast line (Figure 2A). To ensure that these bands were not the result of a PCR artifact or genomic contamination of samples, negative controls such as exclusion of RT or RNA during cDNA synthesis were included in each assay. As expected, no PCR product appeared when either RT or the template was omitted (Figure 2A). Densitometric analysis revealed that bronchial epithelial cells contained an appreciable amount of SRF mRNA. Nuclear extracts from A549, Beas2B and C2C12 cells were electrophoresed and then assayed by western blotting with anti-SRF antibody reported to recognize all four isoforms of the protein (52). As shown in Figure 2B, we found that robust expression of a 67 kDa band, that based on molecular mass, corresponds to full-length SRF (SRF-FL) in all of the cell lines examined. Consistent wit RT–PCR analysis, a 57 kDa band corresponding to SRF5 was only detected in the C2C12 myoblast line. As a control, we checked the ?-actin levels (Figure 2B). Densitometric analysis of western blot revealed no apparent difference in the expression level of SRF proteins between the myoblasts and epithelial cells (Figure 2B). In subsequent western blot analyses in which parallel blots were probed with only the secondary antibody, no band was detected (data not shown), thus confirming that these bands are indeed specific for SRF. As further evidence, A549, Beas2B and C2C12 cells express relatively large amounts of SRF, the cell lines were assessed by indirect immunofluorescence labeling. A similar staining pattern was noted in the bronchial epithelial cells and myoblasts (Figure 2C, rows A–C). The nuclei have been stained blue with DAPI. As negative control, no staining was detected when the primary antibody was omitted (Figure 2C, row D). Consistent with other studies reporting that SRF is ubiquitously expressed in a variety of cell types (26,27,53), our SRF expression studies demonstrated that human SRF gene is expressed in bronchial epithelial cells.

    Figure 2 Bronchial epithelial expression of SRF. (A) Detection of SRF and HPRT transcripts from both bronchial epithelial and myoblast cell lines by RT–PCR analysis. Total RNAs isolated from A549, Beas2B and C2C12 cells, as indicated, were reverse transcribed with random hexamers and PCR amplified with SRF or HPRT primers. Amplified products were directly stained with bromide ethidium on non-denaturing polyacrylamide gel. Expected size of PCR products is given at left of each transcript. Negative controls are noted as follow: No RT for the absence of reverse transcriptase and No RNA for an H2O no-template reaction. Densitometric analysis of SRF bands intensity was performed with relative HPRT mRNA expression in C2C12 myoblasts. (B) Western blot analysis of nuclear extracts prepared from both epithelial and myoblast lines. Each extract (25 μg) was separated by SDS/PAGE (10%) and western blotted using an anti-SRF Ab. Relative mass, in kilodaltons, is indicated alongside. As a control for proteins loading in SDS–PAGE, an anti ?-actin Ab was used in western blot. Numbers below the ?-actin panel indicate densitometric values of the total SRF bands normalized over the corresponding ?-actin bands and expressed relative to the C2C12 controls bands, which are set to 1. (C) The A549, Beas2B and C2C12 cells (rows A, B and C, respectively) were grown on coverslips, fixed and subjected to indirect immunofluorescence with anti-SRF Ab. Left panel, cells were stained with FITC-conjugated secondary antibody (green). Middle panel, cells were counterstained with DAPI (blue) for nuclear staining. Right panel, shown is merge of SRF immunostaining and DAPI. Row D, background immunofluorescence obtained in the absence of the primary antibody.

    SRF binds the CFTR-CArG-like element

    Because the CArG box is not only found in the regulatory regions of many muscle-specific genes but also located at the center of other SREs (54) and forms the core binding site for SRF (55), we have attempted to determine whether or not SRF binds to the CFTR-CArG-like element sequence. We performed two series of bandshift experiments. First, we incubated the 32P-labelLed WT oligonucleotide corresponding to CFTR-CArG-like element sequence (as indicated in Table 2) with bacterially expressed SRF. As shown in Figure 3A (lane 3), EMSA revealed a recombinant SRF–WT duplex. As shown in Figure 3A (lane 5), bands defined as the CFTR-CArG-like-SRF complex disappeared when 100-fold molar excess of specific competitor containing the consensus SRE motif, derived from human c-fos promoter (56), was added. In contrast, competition by an irrelevant oligonucleotide (noted UC) did not appear to significantly alter the complex suggesting specificity of binding (lane 6). To confirm precisely the identity of protein binding in the minimal CFTR promoter (from –121 to –83 bp), bacterially expressed SRF was preincubated with anti-SRF antibody before incubation with 32P-labelled probe. As shown in Figure 3A (lane 4), bands corresponding to CFTR-CArG-like-SRF complex disappeared, and more slowly migrating bands appeared when anti-SRF antibody was added. To verify the antibody specificity, control experiments were performed either in the absence of the SRF protein (Figure 3A, lane 7) or in the presence of an irrelevant antibody (Figure 3A, lane 8). These results demonstrate that the CFTR-CArG-like element is a bona fide SRF-binding site. Second, we performed additional EMSA with the WT oligonucleotide, as main binding target, and nuclear extracts from both epithelial and muscle cells (Figure 3B). While three nucleoprotein complexes were formed from nuclear extracts of Beas2B and C2C12 cells on the WT probe, five complexes were detected from A549 cells. Each nuclear protein complex was designated by an upper case relative to nuclear lysate followed by a Roman numeral indicating its position into the gels. The binding specificity of the different complexes was examined by competition analyses with both specific and non-specific unlabeled oligonucleotides (Figure 3B, lanes 2–4). Although a few non-specific binding activities were observed, complexes BII, BIII, AIII, AV, CI, CII and CIII were efficiently competed with an excess of both WT and c-fos SRE unlabeled specific competitors indicating binding specificity (Figure 3B, lanes 2 and 3, respectively). Control experiments including an irrelevant oligonucleotide did not reveal any change in electrophoretic profiles (Figure 3B, lanes 4). To evaluate whether the SRF protein contributes to the formation of nucleoprotein complexes interacting with the CFTR-derived SRF site, we used a specific antibody against the SRF protein. Although the incubation with the anti-SRF antibody apparently failed to give any discrete supershift from various sources of nuclear proteins, a significant decrease of specific complexes was observed (Figure 3B, lanes 6). This effect seemed to be specific because none of these nuclear protein complexes was supershifted or abolished when a non-specific antibody was used (Figure 3B, lanes 7). Control experiments containing only the anti-SRF antiserum and the radiolabeled WT oligonucleotide did not reveal any binding activities (Figure 3B, lanes 8). In addition, supershift assays performed with the radiolabeled c-fos SRE oligonucleotide instead of the WT probe showed that bands corresponding to SRE–SRF complex disappeared and more slowly migrating bands appeared (Figure 3B, lane 10). The absence of supershift for specific WT binding complexes (Figure 3D, lanes 6) may be explained if the epitope recognized by the anti-SRF antibody is part of (or located near) the CFTR-CArG-like DNA-binding site. Antibody binding would compete DNA binding, resulting in a significant decrease or total disappearance of CArG-like complexes rather than in supershifted complexes. Alternatively, if the interaction between the CFTR-CArG-like DNA-binding site and the SRF protein is insufficiently strong, it could be not as stable as the one observed with the SRE probe (Figure 3B, lane 10) to be detected as a supershift in gel retardation assays. On the one hand, other DNA-binding proteins that are potentially part of this complex could interfere with the antibody's access to SRF protein. Taken together, these EMSA studies establish SRF-binding CArG-like box within the minimal promoter of the human CFTR gene.

    Figure 3 SRF is a component of CFTR-CArG-like binding complexes. (A) EMSAs were performed with a radiolabeled oligonucleotide encompassing the studied CFTR-CArG-like motif (noted WT, see Table 2) and bacterially expressed SRF (lane 3). Competition assays were carried out using a 100-fold molar excess of two specific competitors, noted WT and SRE (lanes 2 and 5, respectively). UC (lane 6) indicates unspecific competitor that is used as negative control. Supershift assays were carried out using either a specific antibody (lane 4) or an irrelevant antibody (lane 8). The arrow SS indicates the Ab-supershifted complex. Neither the retarded nor the supershifted complexes were found when the 32P-labeled WT double-stranded oligonucleotide was incubated either only with the binding buffer (lane 1) or with specific anti-SRF antibody in the absence of purified SRF protein (lane 7). (B) EMSAs were performed with WT 32P-labeled probe in the absence of proteins (lanes 1) or with nuclear extracts (lanes 5) isolated from Beas2B (upper panel), A549 (middle panel) and C2C12 cells (lower panel). Three or five nuclear protein complexes were detected and designated by a cell type-dependent upper-case to the left of autoradiograms. Some binding reaction mixtures included a 100-fold molar excess of the indicated cold probes (lanes 2–4). Immunologic assays were performed with either anti-SRF Ab (lanes 6), or anti-HA as irrelevant antiserum (lanes 7). As a control, the 32P-labeled WT probe was incubated with anti-SRF Ab without nuclear extracts (lanes 8). In additional control experiments, radiolabeled SRE double-stranded oligonucleotide corresponding to the SRE binding site of the c-fos promoter was subjected to EMSAs with nuclear extracts from Beas2B cells (upper panel) in the absence (lane 9) or presence of anti-SRF antiserum (lane 10). Nuclear protein complexes that were significantly decreased or supershifted with anti-SRF Ab are indicated.

    SRF binds the CArG-like motif of the CFTR minimal promoter within intact chromatin under physiological conditions

    Although there is extensive evidence that SRF can bind to CArG-like elements in in vitro assays and that it is involved in transcriptional regulation through CArG elements in reporter assays in cultured cells with a typical epithelioid morphology (53), there is a lack of direct evidence for involvement of SRF in transcription of the endogenous CFTR gene within the context of intact chromatin. To directly address whether SRF also binds in vivo, ChIP was carried out in bronchial epithelial cells. For this purpose, DNA-binding proteins of Beas2B epithelial cells were covalently linked to genomic DNA by treatment of the cells with formaldehyde. Cross-linked chromatin was immunoprecipitated with either specific or irrelevant antibodies, anti-SRF and anti-HA, respectively. The precipitated chromatin DNA was then purified and amplified by PCR with specific primers of the target sequences (Table 1). The promoter of ?-globin gene, which lacks CArG elements, was used as negative control (Figure 4, row A). The skeletal ?-actin promoter sequence, which contains at least a consensus SRF-binding site, was used as positive control (Figure 4, row B). As expected, PCR signals were obtained when the DNA/protein adducts were immunoprecipitated with an anti-SRF antibody from both Beas2B (Figure 4, rows B and C, lane 1) and A549 (data not shown) epithelial cell lines. In contrast, the negative controls, in which immunoprecipitation was performed with an irrelevant antibody (Figure 4, rows A–C, lane 3) or without antibody (Figure 4, rows A–C, lane 2), did not show any PCR signal. These results were reproduced in several independently isolated ChIP populations. These data show that the CFTR-CArG-like binding site identified within the CFTR minimal promoter binds SRF in vivo, in bronchial cell lines.

    Figure 4 ChIP analysis of SRF binding to the endogenous CFTR minimal promoter. Chromatin immunoprecipitation was carried out as described in Materials and Methods. Sheared DNA/protein complexes were immunoprecipitated by using either an anti-SRF Ab or an irrelevant anti-HA Ab. Then, PCR was carried out to detect the endogenous CArG regions in immunoprecipitated chromatin fragments. Target promoter sequences are indicated as follow: row A, ?-globin; row B, ?-actin; row C, CFTR. Lane 1 shows amplification of target sequences in immunoprecipitated chromatin fragments with anti-SRF Ab. Lanes 2 and 3: PCR amplification of control samples, without Ab or with an irrelevant Ab. Lane 4 shows amplification of 1:100 dilution samples of total input DNA for immunoprecipitation.

    The human minimal CFTR promoter shows activity in both epithelial and muscular cells

    As the SRF transcription factor is thought to mainly mediate the tissue-specific transcription in muscular cells, we first assessed the activity of the minimal CFTR promoter in both epithelial and muscular cell types. We therefore addressed the ability of the WT-pGL3 construct, previously described (12) to drive the expression in Beas2B, A549 and C2C12 cells. As shown in Figure 5A, an efficient transcription of the CFTR promoter could be readily observed in all three cell lines compared with monkey kidney fibroblasts (COS-7), which are negative for CFTR expression (36,37). Surprisingly, the activity of the minimal CFTR promoter in C2C12 cells was 4 to 5.5 times higher than that observed in Beas2B and A549 cells, respectively. Although the expression of the CFTR protein was previously evidenced in some muscle cells (57–59), it has never been investigated in mouse C2C12 myoblasts. Therefore, we investigated the CFTR expression at both mRNA and protein levels. To examine CFTR mRNA expression in mouse myoblasts (C2C12), total RNA isolated from epithelial and non-epithelial cells was analyzed via RT–PCR with CFTR-specific primers (Table 1). As shown in Figure 5B, while no PCR product for CFTR was seen with COS-7 cells, the expected CFTR amplified product was obtained for both epithelial (A549 and Beas2B) and muscular (C2C12) cell lines. No product was detected in the absence of reverse transcriptase or RNA (noted No RT or No RNA, respectively). Although the HPRT PCR control showed non-equivalent amounts of cDNA used in the different PCRs, densitometric analysis revealed that C2C12 cells contained an appreciable amount of CFTR mRNA. Further determination of CFTR expression in C2C12 myoblasts was conducted by western blot analysis (Figure 5C). Caco-2 cells were used as positive control and COS-7 as negative control. As shown in Figure 5C, while mature fully glycosylated band C CFTR was observed in A549, Beas2B and C2C12 cells, the core glycosylated band B CFTR occurred at the same molecular mass only in a subset of cell lines, indicative of potential variations in glycosylation process. To rule out the possibility of antibodies cross-reactivity or non-specific labeling, indirect immunofluorescence analysis was performed with another anti-CFTR antibody, L12B4 (Figure 5D). As expected, the Caco-2 cell line used as a positive control showed strong immunoreactivity for CFTR protein (Figure 5D, row D). In contrast, and as expected, COS-7 cells used as negative control did not stain with this anti-CFTR antibody (Figure 5D, row E). No staining was detected when the primary antibody was omitted (Figure 5D, row F). The comparison of immunofluorescence images between epithelial (Figure 5D, row A, B and D) and muscular (Figure 5D, row C) cells provides evidence that CFTR is significantly expressed in C2C12 myoblasts. However, because the indirect immunofluorescence experiments were performed in permeabilized cells, they do not distinguish between CFTR that is located in the apical membrane and CFTR that is located in a vesicular pool just beneath the apical membrane. These findings, together with the transcriptional activities presented above, suggest that the minimal CFTR promoter might contain tissue-specific enhancer(s) as well as repressive cis-regulatory elements that have the ability to mitigate the activity of the CFTR promoter between epithelial and non-epithelial lineages. Because CF affects mainly epithelial tissues and Beas2B consistently yielded higher enhancer activity than A549, we restricted subsequent studies to this bronchial epithelial cell line.

    Figure 5 Analysis of CFTR minimal promoter activity and CFTR expression. (A) CFTR minimal promoter activities in Beas2B, A549, C2C12 and COS-7 cell lines. The amount of 1.8 μg of WT-pGL3 construct was co-transfected together with 0.2 μg of internal control pRL-SV40 containing Renilla luciferase. The empty pGL3-Basic vector was transfected as a negative control. Normalized luciferase values are illustrated as relative light units. The mean of two independent plasmids preps and each experiment with standard errors is shown. The experimental value for luciferase activity represented the mean ± SD of four separate experiments in triplicates. (B) RT–PCR analysis of CFTR mRNA expression. Total RNA was harvested from the indicated cell lines and subjected to PCR as described in ‘Materials and Methods’. cDNA derived from COS-7 was used as a negative control. Samples without reverse transcriptase (noted No RT) or with an H2O no-template reaction (noted No RNA) were used as additional negative controls. Samples were also amplified for HPRT to control for RNA degradation during DNase treatment and reverse transcription. Densitometric analysis of CFTR bands intensity was performed with relative HPRT mRNA expression in Beas2B bronchial epithelial cells. (C) Western blot analysis. Whole cell proteins extracts prepared from indicated cells were subjected to electrophoresis with 7% SDS–PAGE gel and transferred to supported PVDF membrane. CFTR immunoreactivity was detected using 24.1 anti-CFTR Ab as the primary antibody. Positions for the core glycosylated (band B) and mature fully glycosylated CFTR (band C) are indicated by the arrows at the right. Numbers below the ?-actin panel indicate intensity values of integrated optical densities of the mature and immature CFTR bands normalized over the corresponding ?-actin bands and expressed relative to the Beas2B controls bands, which are set to 1. (D) Indirect immunofluorescence analysis. The A549, Beas2B, C2C12, Caco-2 and COS-7 cells (rows A–E, respectively) were fixed, permeabilized, and then stained with the L12B4 anti-CFTR Ab, followed by the appropriate secondary antibody as described in ‘Materials and Methods’. Left panel, cells were stained with Cy3-conjugated secondary antibody (red). Middle panel, cells were counterstained with DAPI (blue) for nuclear staining. Right panel, shown is merge of CFTR immunostaining and DAPI. Row F, Background immunofluorescence obtained in the absence of the primary antibody.

    Effects of SRF on the proximal CFTR promoter

    To assess whether SRF protein might affect CFTR transcriptional regulation, we took two related approaches. First, we co-transfected either exogenous full-length SRF or truncated SRF mutants with the wild-type CFTR minimal promoter. Second, we utilized a siRNA approach to ask whether a reduction in endogenous SRF levels could affect basal CFTR activation. Luciferase assays were carried out in Beas2B cells using the indicated SRF expression vectors and the WT-pGL3luciferase reporter construct (Figure 6B, left panel). The 3XSRE-fos TATA-Luciferase reporter gene was used as positive control because its transcriptional activity is highly dependent on SRF binding (Figure 6B, right panel). As expected, while this synthetic promoter containing several consensus SREs was markedly and significantly activated (>2-fold) by forced expression of full-length SRF protein (pHiv-SRF) (Figure 6B, right panel), truncated SRF mutants either containing only the amino acids 133–264 corresponding to DNA-binding domain of SRF (pHiv-DBD-SRF), or deleted of amino acids 153–165, an essential part for the DNA binding (pHiv-DN-SRF) induced modest but significant decrease of SRE activity (10% and 30%, respectively). When these SRF expression vectors were transiently co-transfected with the WT-pGL3 luciferase reporter construct, the trans-activator effect of SRF was notably reduced. Indeed, overexpression of full-length and dominant-negative (pHiv-DN-SRF) SRF proteins resulted in 50% increased and 30% decreased of luciferase activity, respectively (Figure 6B, left panel). Unexpectedly and inconsistent with previous study (60), the pHiv-DBD-SRF construct induced 40% increase of the WT-pGL3 luciferase activity. This results disparity may be explained by the fact that the c-fos promoter study (60) assessed the DBD–SRF domain activity via microinjection of corresponding truncated SRF polypeptide. Perhaps, the best explanation for this increased activity is that the DNA-binding domain of SRF, also called MADS box, is sufficient for transcriptional activation of some SRF-dependent genes because it mediates interactions with accessory co-activators (61). These results suggest that neither of the regions outside the MADS box, N- or C-terminal, appear essential for transactivation of the CFTR promoter by SRF.

    Figure 6 Effect of SRF on the activity of CArG-containing CFTR minimal promoter. (A) Schematic representation of different SRF constructs analyzed. The SRF fragments shown were cloned into the pHiv expression vector as indicated in ‘Materials and Methods’. The position of the MADS box containing the DNA-binding domain and dimerization domain of SRF is depicted by black rectangle. WT SRF construct encodes a human full-length SRF open reading frame. DBD SRF construct encodes only the DNA-binding domain of SRF. DN SRF deletion construct encodes full-length SRF deleted of an essential part for the DNA binding. (B) Expression of exogenous SRF in Beas2B cells. Cells were co-transfected with the WT-pGL3 luciferase reporter plasmid (left panel) together with the empty pHiv vector, pHiv-SRF, pHiv-DBD-SRF or pHiv-DN-SRF constructs, as indicated below the bar diagram. As a positive control, 3XSRE luciferase reporter plasmid was used. In each experiment, a Renilla luciferase expression plasmid was included to normalize for transfection efficiency. For each reporter plasmid, luciferase activity (mean ± SE) is derived from 4 to 7 different transfection experiments performed in triplicate. The errors bars indicate the standard deviations. The luciferase activity value 100 was assigned to the samples co-transfected with the indicated reporter plasmid and the empty pHiv expression vector. Statistical significance is referred to differences with the activity obtained with empty expression vector. *, P < 0.05 versus WT or 3XSRE. (C) Effects of SRF siRNA on CFTR promoter activity and expression in Beas2B epithelial cells. Left panel: Beas2B cells were co-transfected with either human control siRNA vector or plasmid encoding SRF siRNA as indicated. Forty-eight hours following transfection, total proteins were harvested and analyzed by western blotting. Densitometric analysis showed a substantial decrease of endogenous SRF. Right panel: the luciferase activity value 100 was assigned to the samples co-transfected with the control siRNA. *, P < 0.05 versus WT or 3XSRE.

    To further establish the role of SRF in mediating the increase in CFTR gene transcription, we performed siRNA experiments (Figure 6C). To demonstrate the efficacy of specific SRF siRNA construct, Beas2B cells were co-transfected with either control or SRF siRNA plasmids and western blots were performed to analyze levels of SRF in these cells. Densitometric analysis of western blot showed that endogenous SRF levels were strongly reduced in the presence of the SRF siRNA plasmid (Figure 6C, left panel). Then, control and SRF siRNA plasmids were co-transfected into Beas2B cells with either the WT-pGL3 or the 3XSRE luciferase reporter construct. As shown in Figure 6C (right panel), reduction of endogenous SRF protein level decreased the activity of both wild-type CFTR promoter and 3XSRE-fos TATA-Luciferase reporter constructs by 25%.

    Taken together, these findings indicate that endogenous SRF into bronchial epithelial cells normally function as transcriptional coactivator for the CArG-mediated transactivation. However, we found that SRF activated the CFTR promoter less well than the SRE-containing synthetic promoter, suggesting that the CArG-like element mediated CFTR transcriptional activation probably required other bridging or co-activating factors.

    The CArG-like element contained in the human minimal CFTR promoter is important but not sufficient for basal transcriptional activity in Beas2B cells

    To assess the transcriptional relevance of the CArG-like binding sequence of CFTR minimal promoter, mutants were performed to create either more consensus CArG element or degenerated motif (Table 1). We used the D upper-case and the C upper-case to designate the Degenerated CArG motifs and the Consensus CArG elements, respectively. To ensure the efficacy of mutations in abrogating transcription factor binding, we first performed a series of EMSA experiments with each CArG mutant and the WT probe (Table 2). As expected, based on the sequence analysis, the mutations resulting in more degenerated CArG motifs (D1, D2, D3 and D4 mutants) completely abolished SRF-binding activity (Figure 7A, lanes 1–4), when we compared with SRF-WT complexes (Figure 7A, lane 5). In addition, the probes containing more consensus CArG boxes (C1, C2, C3, C4 and C5 mutants) bound SRF more strongly in nuclear extracts prepared from Beas2B (Figure 7A, lanes 6–10). Inclusion of an anti-SRF antibody partially abolished SRF-binding activity and resulted in a slower mobility band, confirming the identity of the binding protein (Figure 7A, lane 11). Then, minimal CFTR promoter constructs bearing the same consensus and disrupted CArG boxes were tested in a transient transfection assays in Beas2B cells (Figure 7B). Compared with luciferase activity resulting from the WT-pGL3 construct, variable increases in reporter activity are observed when D1, D2, D3 and D4 mutant constructs, with disrupted CArG motif, were transiently transfected. On the other hand, mutations that create more consensus CArG element result in either reducing or increasing in the transcriptional activity (Figure 7B). Consistent with the preceding gel-shift results revealing an uncharacterized protein/DNA complex with slightly slower mobility and variable intensity compared with the SRF–DNA complex, these finer mutagenesis studies suggest that both co-activators and co-repressors might interact with the studied promoter sequences. Taken together, these data demonstrated that the CFTR-CArG-like element alone is not sufficient to confer basal transcriptional activation of the CFTR gene.

    Figure 7 The CFTR-CArG-like element is not sufficient for basal transcriptional activity in Beas2B cells. (A) EMSA analysis with nuclear extracts from SRF protein-enriched Beas2B cells using mutated labeled oligonucleotide probes (sequences listed in Table 2). Ab against SRF was included as indicated (lane 10). The arrow SS indicates the supershifted complexes. (B) Basal transcriptional activity of CArG variants of the CFTR promoter. Luciferase activity obtained with the WT-pGL3 luciferase construct was defined as 100%, and relative luciferase activities from mutant constructs are expressed as a percentage of this value. Firefly luciferase activity was normalized with respect to Renilla luciferase activity. The errors bars indicate the standard deviations. *, P < 0.05 versus WT.

    Functional antagonism between SRF and YY1 through competition for the CFTR-CArG-like binding

    As a further means of establishing the role for SRF-binding CArG boxes in mediating either the CFTR-enhancer or silencer activity observed with the different minimal CFTR promoter mutants, we performed additional computational analyses. Sequence analysis of CArG boxes and their immediate flanking sequences showed that C1, C3 and C4 constructs contain consensus binding sites for a number of transcription factors, in addition to SRF, including the transcription factor YY1. These findings with previous work (12) have particularly evoked our interest to examine the role of YY1 in SRF-mediated CFTR transcriptional activity.

    In a first set of experiments, we evaluated whether the YY1 factor contributes to the formation of nucleoprotein complexes interacting with the WT, C1, C3 and C4 radiolabeled probes. As shown in Figure 8A, incubation with specific anti-YY1 antibody resulted in either a significant decrease or total disappearance of some complexes normally formed between the CArG boxes and a component of Beas2B nuclear extracts, establishing DNA binding between YY1 and the sequences tested. As expected, control experiment performed with radiolabeled probe, D2, devoid of any YY1 binding site and the anti-YY1 antiserum did not reveal any change in electrophoretic profiles indicating specificity of the YY1 binding. Thus, these EMSAs demonstrated that the CFTR-CArG-like (WT probe) and more consensus CArG elements (C1 < C3 < C4) served as binding site for at least two distinct nuclear factors, namely SRF and YY1 from bronchial epithelial cells.

    Figure 8 Functional interplay between SRF and YY1. (A) YY1 binding to CFTR-CArG elements. Nuclear extracts from Beas2B cells were incubated with 32P-labelled oligonucleotide probe containing either CArG-like (noted WT), consensus CArG box (noted C1, C3 and C4) in the presence or absence of specific antibody raised against YY1 as indicated. D2, which contain neither CArG element nor YY1 site, is included as negative control. Arrows indicate the position of YY1. (B) Mutually exclusive DNA-binding activity of YY1 versus SRF to the CFTR-CArG site. The same labeled CArG oligonucleotides that those used in panel A were incubated with YY1-transfected Beas2B nuclear extracts (5 μg) in the presence of increasing amounts of bacterial purified SRF (100–1000 ng). Arrows indicate the position of SRF and YY1. (C) Functional antagonism between SRF and YY1. Beas2B were transfected with constructs containing either the CFTR-CArG-like element (noted WT) or consensus CFTR-CArG-boxes (noted C1, C3 and C4), in the presence of SRF and/or YY1 expression vectors. Luciferase activities were normalized for transfection efficiency to the Renilla luciferase used as internal control. The luciferase activity value 100 was assigned to samples co-transfected with the indicated reporter plasmid and the empty expression vector. The mean ± SD from at least three independent experiments is represented. *, P < 0.05 versus WT, C1, C3 or C4.

    Then, to determine whether the interaction of these two factors with their target DNA (CFTR-CArG-like or more consensus CArG sequence) was either mutually inclusive or exclusive, EMSAs were conducted using bacterially expressed SRF and nuclear extracts prepared from Beas2B cells transfected with a YY1 expression vector as a source of YY1 DNA-binding activity. As shown in Figure 8B, the YY1-binding complex was significantly reduced by increasing SRF binding activity, suggesting their mutually exclusive binding to CArG sites.

    Consistent with other studies (32,62), these EMSAs results suggested as a working hypothesis that both SRF and YY1 likely act as functional competitors in the CFTR promoter activity. Since the minimal CFTR promoter is activated by SRF and knowing from previous studies (30–33) that YY1 negatively regulates some promoter containing CArG elements, it seemed logical to first assess whether YY1 could also repress the CFTR promoter. We asked whether CArG-like-binding site in the CFTR promoter acted as negative regulatory elements. Several CFTR promoter reporter constructs containing either CArG-like (namely, WT) or consensus CArG element (C1, C3 and C4) were co-transfected with SRF and/or YY1 expression vectors into Beas2B cells. As shown in Figure 8C, while co-transfections with SRF alone resulted in 1.5- to 2-fold luciferase activation, forced expression of YY1 protein caused a strong decrease in reporter activities (50–75% of the control luciferase value). Therefore, we wanted to determine the functional relationship of both factors in the CFTR activity regulation. Transient co-transfections analysis of combination of protein expression vectors showed that over-expression of SRF either did not restore (Figure 8C, panel with the WT and C4 reporter constructs) or only very slightly raised the YY1-repressed reporter activities (Figure 8C, panel with the C1 and C3 reporter constructs). These findings are consistent with results obtained from co-transfections performed with the YY1 expression vector alone and previous alpha-actin promoter studies (30–33), in which YY1 acted as a strong repressor. Taken together, these data suggested functional antagonism between SRF and YY1 through DNA-binding competition.

    Identification of multiple CArG binding sites in the non-coding regions of the human CFTR gene and assessment of their putative functional importance

    To assess the putative importance of CArG motifs in the human CFTR expression regulation, we performed an in silico inspection of additional sequences upstream of the minimal promoter and in the intronic regions. Four CArG binding sites were found in the CFTR promoter regions, and seven in the intronic regions. The distribution and the positions of these putative binding sites for SRF are shown in Figure 9A. Then, to confirm the computational predictions, we used several approaches. First, the nucleotide regions harboring these predicted cis-acting elements were analyzed for conservation in the cow, as physiologically relevant transcription factor binding sites are frequently conserved in the non-coding regions of orthologous genes (49,63). Second, the SRF-binding sites were tested by EMSAs and also for one of them by ChIP.

    Figure 9 The CFTR gene untranslated regions contain multiple SRF-binding sites (CArG-boxes). (A) Schematic drawing of CFTR gene shows the exons (circles) and the location of the putative CArG boxes in the non-coding regions, including that identified in the minimal CFTR promoter. (B) The CFTR gene non-coding regions encompassing the putative CArG elements were analyzed for conservation in the cow. Species homology and positions of the conserved sequence within BAC clones are indicated, as reported in a recent work (49). (C) Comparative binding of SRF to CFTR CArG boxes. Each promoter and intronic CArG box-containing oligonucleotide (see Table 2) was radiolabeled and purified as described in ‘Materials and Methods’. Upper panel, binding reaction was performed with bacterially expressed SRF. Ab against SRF was included as indicated. Arrow indicates specific DNA–SRF complexes and SS indicates the Ab-supershifted complexes. Lower panel, EMSA analysis with nuclear extracts from Beas2B cells using mutated labeled oligonucleotide probes. (D) ChIP analysis of SRF binding to the endogenous CArG regions. Lane 1 shows amplification of target sequence in immunoprecipitated chromatin fragments with anti-SRF Ab. The target sequence encompasses a CArG element, located to –2373 bp in the minimal CFTR promoter. Lane 2 shows PCR amplification of a control sample, without Ab. Lane 3 shows amplification of 1:100 dilution samples of total input DNA for immunoprecipitation.

    We showed that all the predicted CArG elements occur in highly conserved CFTR non-coding regions (Figure 9B), which may contain tissue-specific transcription factors (5). We provided evidence that the majority of in silico predicted CArG motifs form specific DNA–protein complexes with bacterially expressed SRF protein (Figure 9C, left panel). We also demonstrated by using Beas2B nuclear extracts and supershift assays that some of these CArG elements bound in vitro SRF protein (Figure 9C, right panel, lanes 1, 3, 7 and 9). Finally, additional ChIP experiments indicated that the only tested CArG element, located at –2373 bp in the CFTR promoter, bound SRF in chromatin from intact cultured Beas2B cells (Figure 9D, lane 1). Collectively, these findings support the concept that SRF may play a key role in the CFTR expression regulation via binding to multiple CFTR CArG elements.

    DISCUSSION

    While great strides have been made in CFTR research with emphasis on the structure and function of this protein, only incremental advances have been achieved in the field of transcriptional regulation. In our previous works (12,20), we have identified a polymorphic CArG-like site and showed that a naturally occurring sequence variation in this motif may enhance the basal CFTR transcriptional activity. The goal of the present study was to better characterize the human CFTR promoter region encompassing the CArG-like element.

    Comparisons of nucleotide sequences of this CFTR promoter region in eight mammalian species representing three different orders (Primates, Artiodactyla and Lagomorpha) revealed high levels of conservation of the minimal promoter region, as described previously (64). However, the CArG-like element (GC(T/A)6GG), which does not conform to the authentic CArG box (CC(T/A)6GG), is absent from three species (sheep, cow and rabbit). Even though this motif occurs in only a subset of the input sequences, it does not necessarily mean that it is of no functional importance. Instances of cis-regulatory elements being species-specific have been already reported for the CFTR gene (50). More generally, it is well known that some transcription factors can tolerate more than one type of nucleotide at a given position of the binding site. Instances of substitutions in both the central core of the CArG element and the highly conserved contact points for SRF, at the terminal ends, not impairing SRF binding have been previously reported (48,65–68). Taken together, these data support the notion that the CFTR-CArG-like element has possible important implication in the CFTR transcriptional regulation.

    Interestingly, analysis of SRF expression has revealed relatively high expression levels of SRF protein in the bronchial epithelial cells (Figure 2), as shown in other differentiated epithelial cells (27). Accumulating evidence supports the concept that SRF could contribute to the CFTR expression regulation. In addition to regulating growth-responsive genes and numerous muscle-specific genes (22,69), SRF is also described as a key trans-binding factor in various important physiological events. For example, SRF has been shown to regulate pulmonary development in Drosophila (28), to promote both re-epithelialization and muscular structure restoration during gastric ulcer healing (70) and more recently to be involved in the regulation of endoplasmic reticulum-targeted gene (67). Therefore, it was relevant to examine the potential role of SRF on the CFTR transcriptional activity.

    In the present study, we demonstrated that CFTR-CArG-like element forms specific DNA–protein complexes that include SRF protein as part of the complex, in vitro and in vivo (Figures 3 and 4). We have investigated the expression driven by the proximal human CFTR promoter in a pGL3 construct in both epithelial and myoblasts lines. Results of our transient transfections of 3XSRE-fos TATA and WT-pGL3 luciferase reporter genes showed that the human minimal CFTR promoter encompassing a CArG-like motif was sufficient to drive the expression in both cell types (Figure 5). Consistent with other studies showing CFTR expression in smooth muscle tissue (71,72), our results demonstrated that the human CFTR minimal promoter had the ability to drive the expression in muscle lineages. Our transient co-transfections analyses performed with full-length and dominant-negative SRF forms expression vectors, and also specific SRF siRNA constructs (Figure 6, panels B and C) provide the first evidence for a positive role of the SRF protein in the regulation of the human CFTR promoter. However, the SRF-induced CFTR transactivation, though significant, is relatively modest. These data with the results of finer mutagenesis studies presented in Figure 7 support the concept that SRF alone is not sufficient to drive the basal expression of the CFTR gene. As suggested by other previous studies (73–75), it is likely that additional factors might be required to optimally drive the CFTR promoter, as well as to specify the cell-type-restricted expression. Since YY1 was previously shown to bind the CFTR promoter (12) and a subset of SREs serve as binding sites for YY1 (30–33), we first evaluated this candidate in the SRF-mediated CFTR transcriptional activity. The data analysis suggests a functional antagonism between SRF and YY1 through binding competition for the CFTR-CArG-like box. However, the analysis of combinatorial co-transfection studies, presented in Figure 8C, suggests that the functional interplay between SRF and YY1, alone, is not sufficient to account for basal CFTR transcriptional activation. Further studies will be necessary to determine whether another player might be involved in this functional interplay and accounted for the basal transcriptional CFTR activity observed in the Beas2B bronchial epithelial cells.

    Interestingly, we showed that, in addition to the CArG-like element defined in the minimal promoter of the CFTR gene, there are three CArG motifs located at the 5' end of the gene and seven are present in the intronic regions (Figure 9). We also evidenced SRF binding to CArG elements of the endogenous CFTR gene in the context of intact chromatin (Figures 4 and 9D). Although the role of these more distal CArG boxes has yet to be assessed, these findings are of major importance, since it has been shown that optimal activity of the SRF-dependent target genes may require a number of CArG boxes, including those that are defined in the intronic region (66,67,76). Moreover, as already suggested for the -actin promoters (32,74), perhaps the presence multiple CArG elements in the CFTR gene with cooperative binding events might be essential for SRF to prevent YY1 from binding to the CFTR promoter. Clearly, further studies are needed to directly investigate the role of these multiple CArG elements in transcriptional regulation of the CFTR gene.

    Taken together, our results demonstrated for the first time that the human CFTR promoter is a novel SRF target gene, subject to modest but significant SRF activation partially due to functional antagonism between YY1 and SRF through mutually exclusive DNA-binding activities of YY1 versus SRF to the CFTR-CArG-like site.

    A critical question is, thus, how might SRF contribute to orchestrating cell-restricted and context-dependent programs of CFTR gene expression? Our findings and investigations to date clearly indicate the presence of several different underlying mechanisms of SRF-dependent transcriptional activation, as represented in Figure 10. These might include association of SRF with a variety of cell-restricted co-factors (noted in gray bubbles, Figure 10), such as MRTF-A highly expressed in epithelial cells of the lung, kidney, colon and testis, and MRTF-B also expressed in smooth muscle cells (77), ternary complex factors, such as p62 and SAP-1 (25,78) and remodeling chromatin co-factors in an HAT-dependent manner, such as the CREB-binding protein (79). In addition, it is well known that binding of SRF to the CArG box induces an acute bend in the DNA and that this bending may vary with changing base compositions across the CArG box (80). Such bending can probably facilitate interactions with other proteins already involved in the CFTR expression regulation, such as NF-B (13) and C/EBP (9) and described as interacting with SRF (81). Finally, the Ras-related GTPases family (rac, RhoA and cdc42) might have a critical role in the activation of SRF (82). Indeed, it is reported that RhoA-induced SRF activation occurs through a phenomenon of ‘actin treadmilling’ restricted to CArG-dependent genes that do not have adjacent ETS-binding sites (e.g. smooth muscle cell restricted CArG genes) (83,84). Moreover, it has been shown that the activation of SRF may be mediated by the NF-B and C/EBP transcription factors (81).

    Figure 10 Schematic model depicting potential mechanism that might contribute to epithelial-specific regulation of the CFTR gene. In this model, arrows and bubbles marked in gray indicate multiple potential SRF-mediated regulation pathways, such as protein–protein interaction and potential induction by the Ras-related GTPases cascades. Blue and pink bubbles indicate trans-factors identified in this study that contribute to the basal CFTR transcriptional activity. Black and double arrow shows the functional antagonism between YY1 and SRF identified in this study through the CFTR-CArG-like element characterization.

    These data provide the foundation for further studies on the regulation of CFTR and will unable the rational design of further functional studies, including various strategies aiming to further dissect the different regulation pathways.

    ACKNOWLEDGEMENTS

    This study was supported by grants from both French associations: Vaincre La Mucoviscidose (VLM) and Fondation pour la Recherche Medicale (FRM). J.D. was supported by PhD fellowship grants from the VLM. We are particularly indebted to A. Fernandez and N. Lamb for their continuous support to A.L. and for critical reading of the manuscript and. The authors thank G. Carnac for many valuable discussions and suggestions. The authors also thank the investigators mentioned in Materials and Methods for the kind gifts of crucial reagents. Funding to pay the Open Access publication charges for this article was provided by the Centre National de la Recherche Scientifique (CNRS).

    REFERENCES

    Crawford, I., Maloney, P.C., Zeitlin, P.L., Guggino, W.B., Hyde, S.C., Turley, H., Gatter, K.C., Harris, A., Higgins, C.F. (1991) Immunocytochemical localization of the cystic fibrosis gene product CFTR Proc. Natl Acad. Sci. USA, 88, 9262–9266 .

    Denning, G.M., Ostedgaard, L.S., Cheng, S.H., Smith, A.E., Welsh, M.J. (1992) Localization of cystic fibrosis transmembrane conductance regulator in chloride secretory epithelia J. Clin. Invest., 89, 339–349 .

    Engelhardt, J.F., Yankaskas, J.R., Ernst, S.A., Yang, Y., Marino, C.R., Boucher, R.C., Cohn, J.A., Wilson, J.M. (1992) Submucosal glands are the predominant site of CFTR expression in the human bronchus Nature Genet., 2, 240–248 .

    Trezise, A.E., Linder, C.C., Grieger, D., Thompson, E.W., Meunier, H., Griswold, M.D., Buchwald, M. (1993) CFTR expression is regulated during both the cycle of the seminiferous epithelium and the oestrous cycle of rodents Nature Genet., 3, 157–164 .

    Mouchel, N., Broackes-Carter, F., Harris, A. (2003) Alternative 5' exons of the CFTR gene show developmental regulation Hum. Mol. Genet., 12, 759–769 .

    Yoshimura, K., Nakamura, H., Trapnell, B.C., Dalemans, W., Pavirani, A., Lecocq, J.P., Crystal, R.G. (1991) The cystic fibrosis gene has a ‘housekeeping’-type promoter and is expressed at low levels in cells of epithelial origin J. Biol. Chem., 266, 9140–9144 .

    Koh, J., Sferra, T.J., Collins, F.S. (1993) Characterization of the cystic fibrosis transmembrane conductance regulator promoter region. Chromatin context and tissue-specificity J. Biol. Chem., 268, 15912–15921 .

    White, N.L., Higgins, C.F., Trezise, A.E. (1998) Tissue-specific in vivo transcription start sites of the human and murine cystic fibrosis genes Hum. Mol. Genet., 7, 363–369 .

    Pittman, N., Shue, G., LeLeiko, N.S., Walsh, M.J. (1995) Transcription of cystic fibrosis transmembrane conductance regulator requires a CCAAT-like element for both basal and cAMP-mediated regulation J. Biol. Chem., 270, 28848–28857 .

    Matthews, R.P. and McKnight, G.S. (1996) Characterization of the cAMP response element of the cystic fibrosis transmembrane conductance regulator gene promoter J. Biol. Chem., 271, 31869–31877 .

    McDonald, R.A., Matthews, R.P., Idzerda, R.L., McKnight, G.S. (1995) Basal expression of the cystic fibrosis transmembrane conductance regulator gene is dependent on protein kinase A activity Proc. Natl Acad. Sci. USA, 92, 7560–7564 .

    Romey, M.C., Pallares-Ruiz, N., Mange, A., Mettling, C., Peytavi, R., Demaille, J., Claustres, M. (2000) A naturally occurring sequence variation that creates a YY1 element is associated with increased cystic fibrosis transmembrane conductance regulator gene expression J. Biol. Chem., 275, 3561–3567 .

    Brouillard, F., Bouthier, M., Leclerc, T., Clement, A., Baudouin-Legros, M., Edelman, A. (2001) NF-kappa B mediates up-regulation of CFTR gene expression in Calu-3 cells by interleukin-1beta J. Biol. Chem., 276, 9486–9491 .

    Li, S., Moy, L., Pittman, N., Shue, G., Aufiero, B., Neufeld, E.J., LeLeiko, N.S., Walsh, M.J. (1999) Transcriptional repression of the cystic fibrosis transmembrane conductance regulator gene, mediated by CCAAT displacement protein/cut homolog, is associated with histone deacetylation J. Biol. Chem., 274, 7803–7815 .

    Smith, A.N., Barth, M.L., McDowell, T.L., Moulin, D.S., Nuthall, H.N., Hollingsworth, M.A., Harris, A. (1996) A regulatory element in intron 1 of the cystic fibrosis transmembrane conductance regulator gene J. Biol. Chem., 271, 9947–9954 .

    Smith, D.J., Nuthall, H.N., Majetti, M.E., Harris, A. (2000) Multiple potential intragenic regulatory elements in the CFTR gene Genomics, 64, 90–96 .

    Rowntree, R.K., Vassaux, G., McDowell, T.L., Howe, S., McGuigan, A., Phylactides, M., Huxley, C., Harris, A. (2001) An element in intron 1 of the CFTR gene augments intestinal expression in vivo Hum. Mol. Genet., 10, 1455–1464 .

    Phylactides, M., Rowntree, R., Nuthall, H., Ussery, D., Wheeler, A., Harris, A. (2002) Evaluation of potential regulatory elements identified as DNase I hypersensitive sites in the CFTR gene Eur. J. Biochem., 269, 553–559 .

    Mouchel, N., Henstra, S.A., McCarthy, V.A., Williams, S.H., Phylactides, M., Harris, A. (2004) HNF1alpha is involved in tissue-specific regulation of CFTR gene expression Biochem. J., 378, 909–918 .

    Romey, M.C., Guittard, C., Chazalette, J.P., Frossard, P., Dawson, K.P., Patton, M.A., Casals, T., Bazarbachi, T., Girodon, E., Rault, G., et al. (1999) Complex allele is associated with milder forms of cystic fibrosis than allele S549R(T>G) alone Hum. Genet., 105, 145–150 .

    Treisman, R. (1986) Identification of a protein-binding site that mediates transcriptional response of the c-fos gene to serum factors Cell, 46, 567–574 .

    Miano, J.M. (2003) Serum response factor: toggling between disparate programs of gene expression J. Mol. Cell Cardiol., 35, 577–593 .

    Shore, P. and Sharrocks, A.D. (1995) The MADS-box family of transcription factors Eur. J. Biochem., 229, 1–13 .

    Browning, C.L., Culberson, D.E., Aragon, I.V., Fillmore, R.A., Croissant, J.D., Schwartz, R.J., Zimmer, W.E. (1998) The developmentally regulated expression of serum response factor plays a key role in the control of smooth muscle-specific genes Dev. Biol., 194, 18–37 .

    Camoretti-Mercado, B., Dulin, N.O., Solway, J. (2003) Serum response factor function and dysfunction in smooth muscle Respir. Physiol. Neurobiol., 137, 223–235 .

    Magnaghi-Jaulin, L., Masutani, H., Lipinski, M., Harel-Bellan, A. (1996) Analysis of SRF, SAP-1 and ELK-1 transcripts and proteins in human cell lines FEBS Lett., 391, 247–251 .

    Chai, J., Baatar, D., Moon, W., Tarnawski, A. (2002) Expression of serum response factor in normal rat gastric mucosa J. Physiol. Pharmacol., 53, 289–294 .

    Guillemin, K., Groppe, J., Ducker, K., Treisman, R., Hafen, E., Affolter, M., Krasnow, M.A. (1996) The pruned gene encodes the Drosophila serum response factor and regulates cytoplasmic outgrowth during terminal branching of the tracheal system Development, 122, 1353–1362 .

    Yang, Y., Zhe, X., Phan, S.H., Ullenbruch, M., Schuger, L. (2003) Involvement of serum response factor isoforms in myofibroblast differentiation during bleomycin-induced lung injury Am. J. Respir. Cell Mol. Biol., 29, 583–590 .

    Liu, T., Wu, J., He, F. (2000) Evolution of cis-acting elements in 5' flanking regions of vertebrate actin genes J. Mol. Evol., 50, 22–30 .

    Natesan, S. and Gilman, M. (1995) YY1 facilitates the association of serum response factor with the c-fos serum response element Mol. Cell. Biol, 15, 5975–5982 .

    Chen, C.Y. and Schwartz, R.J. (1997) Competition between negative acting YY1 versus positive acting serum response factor and tinman homologue Nkx-2.5 regulates cardiac alpha-actin promoter activity Mol. Endocrinol., 11, 812–822 .

    Martin, K.A., Gualberto, A., Colman, M.F., Lowry, J., Walsh, K. (1997) A competitive mechanism of CArG element regulation by YY1 and SRF: implications for assessment of Phox1/MHox transcription factor interactions at CArG elements DNA Cell Biol, 16, 653–661 .

    Mohamed, A., Ferguson, D., Seibert, F.S., Cai, H.M., Kartner, N., Grinstein, S., Riordan, J.R., Lukacs, G.L. (1997) Functional expression and apical localization of the cystic fibrosis transmembrane conductance regulator in MDCK I cells Biochem. J, 322, 259–265 .

    Bradbury, N.A., Clark, J.A., Watkins, S.C., Widnell, C.C., Smith, H.S. (1999) Characterization of the internalization pathways for the cystic fibrosis transmembrane conductance regulator Am. J. Physiol., 276, L659–L668 .

    Tousson, A., Van Tine, B.A., Naren, A.P., Shaw, G.M., Schwiebert, L.M. (1998) Characterization of CFTR expression and chloride channel activity in human endothelia Am. J. Physiol, 275, C1555–C1564 .

    Peter, K., Varga, K., Bebok, Z., McNicholas-Bevensee, C.M., Schwiebert, L., Sorscher, E.J., Schwiebert, E.M., Collawn, J.F. (2002) Ablation of internalization signals in the carboxyl-terminal tail of the cystic fibrosis transmembrane conductance regulator enhances cell surface expression J. Biol. Chem., 277, 49952–49957 .

    Chou, J.L., Rozmahel, R., Tsui, L.C. (1991) Characterization of the promoter region of the cystic fibrosis transmembrane conductance regulator gene J. Biol. Chem., 266, 24471–24476 .

    Cheng, S.H., Fang, S.L., Zabner, J., Marshall, J., Piraino, S., Schiavi, S.C., Jefferson, D.M., Welsh, M.J., Smith, A.E. (1995) Functional activation of the cystic fibrosis trafficking mutant delta F508-CFTR by overexpression Am. J. Physiol, 268, L615–L624 .

    Shore, P. and Sharrocks, A.D. (1994) The transcription factors Elk-1 and serum response factor interact by direct protein-protein contacts mediated by a short region of Elk-1 Mol. Cell. Biol., 14, 3283–3291 .

    Romey, M.C., Tuffery, S., Desgeorges, M., Bienvenu, T., Demaille, J., Claustres, M. (1996) Transcript analysis of CFTR frameshift mutations in lymphocytes using the reverse transcription-polymerase chain reaction technique and the protein truncation test Hum. Genet., 98, 328–332 .

    Kartner, N., Augustinas, O., Jensen, T.J., Naismith, A.L., Riordan, J.R. (1992) Mislocalization of delta F508 CFTR in cystic fibrosis sweat gland Nature Genet., 1, 321–327 .

    Mendes, F., Farinha, C.M., Roxo-Rosa, M., Fanen, P., Edelman, A., Dormer, R., McPherson, M., Davidson, H., Puchelle, E., De Jonge, H., et al. (2004) Antibodies for CFTR studies J. Cyst. Fibros., Suppl. 2, 69–72 .

    Farinha, C.M., Mendes, F., Roxo-Rosa, M., Penque, D., Amaral, M.D. (2004) A comparison of 14 antibodies for the biochemical detection of the cystic fibrosis transmembrane conductance regulator protein Mol. Cell. Probes, 18, 235–242 .

    Thompson, J.D., Higgins, D.G., Gibson, T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice Nucleic Acids Res., 22, 4673–4680 .

    Romey, M.C., Guittard, C., Carles, S., Demaille, J., Claustres, M., Ramsay, M. (1999) First putative sequence alterations in the minimal CFTR promoter region J. Med. Genet., 36, 263–264 .

    Herschman, H.R. (1991) Primary response genes induced by growth factors and tumor promoters Annu. Rev. Biochem., 60, 281–319 .

    Hirschi, K.K., Lai, L., Belaguli, N.S., Dean, D.A., Schwartz, R.J., Zimmer, W.E. (2002) Transforming growth factor-beta induction of smooth muscle cell phenotype requires transcriptional and post-transcriptional control of serum response factor J. Biol. Chem., 277, 6287–6295 .

    Williams, S.H., Sahota, V., Palmai-Pallag, T., Tebbutt, S.J., Walker, J., Harris, A. (2003) Evaluation of gene targeting by homologous recombination in ovine somatic cells Mol. Reprod. Dev., 66, 115–125 .

    Vuillaumier, S., Dixmeras, I., Messai, H., Lapoumeroulie, C., Lallemand, D., Gekas, J., Chehab, F.F., Perret, C., Elion, J., Denamur, E. (1997) Cross-species characterization of the promoter region of the cystic fibrosis transmembrane conductance regulator gene reveals multiple levels of regulation Biochem. J., 327, 651–662 .

    Williams, S.H., Mouchel, N., Harris, A. (2003) A comparative genomic analysis of the cow, pig, and human CFTR genes identifies potential intronic regulatory elements Genomics, 81, 628–639 .

    Davis, F.J., Gupta, M., Pogwizd, S.M., Bacha, E., Jeevanandam, V., Gupta, M.P. (2002) Increased expression of alternatively spliced dominant-negative isoform of SRF in human failing hearts Am. J. Physiol. Heart. Circ. Physiol., 282, H1521–H1533 .

    Psichari, E., Balmain, A., Plows, D., Zoumpourlis, V., Pintzas, A. (2002) High activity of serum response factor in the mesenchymal transition of epithelial tumor cells is regulated by RhoA signaling J. Biol. Chem., 277, 29490–29495 .

    Spencer, J.A. and Misra, R.P. (1996) Expression of the serum response factor gene is regulated by serum response factor binding sites J. Biol. Chem., 271, 16535–16543 .

    Treisman, R. (1992) The serum response element Trends. Biochem. Sci., 17, 423–426 .

    Gualberto, A., LePage, D., Pons, G., Mader, S.L., Park, K., Atchison, M.L., Walsh, K. (1992) Functional antagonism between YY1 and the serum response factor Mol. Cell. Biol., 12, 4209–4214 .

    Gadsby, D.C., Nagel, G., Hwang, T.C. (1995) The CFTR chloride channel of mammalian heart Annu. Rev. Physiol., 57, 387–416 .

    Tilly, B.C., Bezstarosti, K., Boomaars, W.E., Marino, C.R., Lamers, J.M., de Jonge, H.R. (1996) Expression and regulation of chloride channels in neonatal rat cardiomyocytes Mol. Cell. Biochem, 157, 129–135 .

    Robert, R., Thoreau, V., Norez, C., Cantereau, A., Kitzis, A., Mettey, Y., Rogier, C., Becq, F. (2004) Regulation of the cystic fibrosis transmembrane conductance regulator channel by beta-adrenergic agonists and vasoactive intestinal peptide in rat smooth muscle cells and its role in vasorelaxation J. Biol. Chem, 279, 21160–21168 .

    Gauthier-Rouviere, C., Cai, Q.Q., Lautredou, N., Fernandez, A., Blanchard, J.M., Lamb, N.J. (1993) Expression and purification of the DNA-binding domain of SRF: SRF-DB, a part of a DNA-binding protein which can act as a dominant negative mutant in vivo Exp. Cell. Res., 209, 208–215 .

    Hill, C.S., Wynne, J., Treisman, R. (1994) Serum-regulated transcription by serum response factor (SRF): a novel role for the DNA binding domain EMBO J., 13, 5421–5432 .

    Ellis, P.D., Martin, K.M., Rickman, C., Metcalfe, J.C., Kemp, P.R. (2002) Increased actin polymerization reduces the inhibition of serum response factor activity by Yin Yang 1 Biochem. J, 364, 547–554 .

    Wasserman, W.W., Palumbo, M., Thompson, W., Fickett, J.W., Lawrence, C.E. (2000) Human-mouse genome comparisons to locate regulatory sites Nature Genet., 26, 225–228 .

    Denamur, E. and Chehab, F.F. (1994) Analysis of the mouse and rat CFTR promoter regions Hum. Mol. Genet., 3, 1089–1094 .

    Zilberman, A., Dave, V., Miano, J., Olson, E.N., Periasamy, M. (1998) Evolutionarily conserved promoter region containing CArG*-like elements is crucial for smooth muscle myosin heavy chain gene expression Circ. Res., 82, 566–575 .

    Miano, J.M., Carlson, M.J., Spencer, J.A., Misra, R.P. (2000) Serum response factor-dependent regulation of the smooth muscle calponin gene J. Biol. Chem., 275, 9814–9822 .

    Streb, J.W. and Miano, J.M. (2005) AKAP12alpha, an atypical serum response factor-dependent target gene J. Biol. Chem., 280, 4125–4134 .

    L'Honore, A., Lamb, N.J., Vandromme, M., Turowski, P., Carnac, G., Fernandez, A. (2003) MyoD distal regulatory region contains an SRF binding CArG element required for MyoD expression in skeletal myoblasts and during muscle regeneration Mol. Biol. Cell, 14, 2151–2162 .

    Treisman, R. (1995) Journey to the surface of the cell: Fos regulation and the SRE EMBO J., 14, 4905–4913 .

    Chai, J., Baatar, D., Tarnawski, A. (2004) Serum response factor promotes re-pithelialization and muscular structure restoration during gastric ulcer healing Gastroenterology, 126, 1809–1818 .

    Robert, R., Thoreau, V., Norez, C., Cantereau, A., Kitzis, A., Mettey, Y., Rogier, C., Becq, F. (2004) Regulation of the cystic fibrosis transmembrane conductance regulator channel by beta-adrenergic-agonists and VIP in rat smooth muscle cells and its role in vasorelaxation J. Biol. Chem, 279, 21160–21168 .

    Davies, W.L., Vandenberg, J.I., Sayeed, R.A., Trezise, A.E. (2004) Cardiac expression of the cystic fibrosis transmembrane conductance regulator involves novel exon 1 usage to produce an unique amino-terminal protein J. Biol. Chem., 279, 15877–15887 .

    Chow, K.L. and Schwartz, R.J. (1990) A combination of closely associated positive and negative cis-acting promoter elements regulates transcription of the skeletal alpha-actin gene Mol. Cell. Biol., 10, 528–538 .

    Lee, T.C., Chow, K.L., Fang, P., Schwartz, R.J. (1991) Activation of skeletal alpha-actin gene transcription: the cooperative formation of serum response factor-binding complexes over positive cis-acting promoter serum response elements displaces a negative-acting nuclear factor enriched in replicating myoblasts and nonmyogenic cells Mol. Cell. Biol., 11, 5090–5100 .

    Moss, J.B., McQuinn, T.C., Schwartz, R.J. (1994) The avian cardiac alpha-actin promoter is regulated through a pair of complex elements composed of E boxes and serum response elements that bind both positive- and negative-acting factors J. Biol. Chem, 269, 12731–12740 .

    Mack, C.P. and Owens, G.K. (1999) Regulation of smooth muscle alpha-actin expression in vivo is dependent on CArG elements within the 5' and first intron promoter regions Circ. Res., 84, 852–861 .

    Wang, D.Z., Li, S., Hockemeyer, D., Sutherland, L., Wang, Z., Schratt, G., Richardson, J.A., Nordheim, A., Olson, E.N. (2002) Potentiation of serum response factor activity by a family of myocardin-related transcription factors Proc. Natl Acad. Sci. USA, 99, 14855–14860 .

    Mo, Y., Ho, W., Johnston, K., Marmorstein, R. (2001) Crystal structure of a ternary SAP-1/SRF/c-fos SRE DNA complex J. Mol. Biol., 314, 495–506 .

    Qiu, P. and Li, L. (2002) Histone acetylation and recruitment of serum responsive factor and CREB-binding protein onto SM22 promoter during SM22 gene expression Circ. Res., 90, 858–865 .

    Pellegrini, L., Tan, S., Richmond, T.J. (1995) Structure of serum response factor core bound to DNA Nature, 376, 490–498 .

    Montaner, S., Perona, R., Saniger, L., Lacal, J.C. (1999) Activation of serum response factor by RhoA is mediated by the nuclear factor-kappaB and C/EBP transcription factors J. Biol. Chem., 274, 8506–8515 .

    Hill, C.S., Wynne, J., Treisman, R. (1995) The Rho family GTPases RhoA, Rac1, and CDC42Hs regulate transcriptional activation by SRF Cell, 81, 1159–1170 .

    Sotiropoulos, A., Gineitis, D., Copeland, J., Treisman, R. (1999) Signal-regulated activation of serum response factor is mediated by changes in actin dynamics Cell, 98, 159–169 .

    Gineitis, D. and Treisman, R. (2001) Differential usage of signal transduction pathways defines two types of serum response factor target gene J. Biol. Chem., 276, 24531–24539 .

    Wadman, I.A., Osada, H., Grutz, G.G., Agulnick, A.D., Westphal, H., Forster, A., Rabbitts, T.H. (1997) The LIM-only protein Lmo2 is a bridging molecule assembling an erythroid, DNA-binding complex which includes the TAL1, E47, GATA-1 and Ldb1/NLI proteins EMBO J., 16, 3145–3157 .(Céline René, Magali Taulan, Florence Ira)

    http://www.100md.com/html/DirDu/2007/02/17/37/00/82.htm