Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The bronchioles constitute the most distal part of the conducting airways, located before the transition to the functionally distinct respiratory portion of the lung. The predominant cells in the bronchiolar epithelium are the ciliated cells and the nonciliated Clara cells. The Clara cell represents a well-differentiated cell type with a high secretory activity. Numerous proteins have been shown to be secreted from the Clara cells, the major secretory product being the Clara cell secretory protein (CCSP) (1). Abundant expression of this small secretory protein serves as a marker for high differentiation of the Clara cell. CCSP, also called uteroglobin, polychlorinated biphenyl binding protein, or Clara cell 10-kD protein, is a homodimeric, low molecular weight protein, and it has been cloned from several mammalian species, including rat (2, 3), mouse (4, 5), rabbit (6, 7), and human (8). In lung, CCSP is specifically expressed in the Clara cells (9 and references therein).

Studies in transgenic mice have demonstrated that cis-acting elements directing lung tissue and Clara cell-specific expression of CCSP reside within 2.3 kb of the 5' flanking region of the rat gene (10, 11). For the rabbit and mouse CCSP genes, similar results have been presented (12). Transient transfection experiments in the human lung adenocarcinoma cell line NCI-H441, considered to be Clara cell-like, have demonstrated cis-acting elements important for Clara cell-specific expression of CCSP in the proximal mouse and rat CCSP promoter (11, 14, 15). Furthermore, studies in these cells have demonstrated interaction of hepatocyte nuclear factor (HNF)-3 and HNF-3 and thyroid transcription factor (TTF)-1 with the cis-acting elements in the proximal promoter (16). The expression of these transcription factors is high in the bronchiolar epithelium (20 and references therein), and they are involved in the regulation of the cell-specific expression of CCSP. However, the developmental expression of TTF-1, HNF-3, and HNF-3 does not correlate with the differentiation-dependent expression of CCSP in that these factors start being expressed early during embryonal development, whereas CCSP is turned on at a later stage and is further upregulated the last days before birth (3, 21). In contrast, during embryonal development, there is a correlation between the expression of the transcription factor CCAAT/enhancer binding protein (C/EBP) and CCSP (22, 24). In addition, recent studies from our laboratory have demonstrated a correlation between C/EBP and CCSP in in vitro cell culture models (25).

The C/EBP family of transcription factors belongs to the large family of basic region-leucine zipper (bZIP) transcription factors (26). C/EBP was the first identified member of this family of DNA binding proteins (27). The basic region of C/EBP factors is a highly positively charged domain that directly interacts with the DNA. The leucine zipper domain is involved in homodimerization and heterodimerization. All proteins from the C/EBP family have been shown to form homodimers and heterodimers. C/EBP, C/EBP, and C/EBP are highly similar in their C-terminal basic region and leucine zipper domains with a higher degree of diversity in their N-terminal transactivation domains (28, 29). As a consequence of the high similarity in the basic region, C/EBP, C/EBP, and C/EBP have been shown to interact with virtually identical DNA sequences (28). C/EBP factors play important roles in controlling differentiation and differentiation-dependent processes. In many tissues, most notably in liver, fat, and white blood cells of the myelomonocytic lineage, C/EBP factors are important regulators of proliferation, cell cycle arrest, and gene expression (31).

C/EBP factors are expressed in a number of different tissues, and previous studies have demonstrated expression of C/EBP, C/EBP, and C/EBP in the lung. Although expression of C/EBP in mice has been shown to be highest in the lung (28), little is known about the cellular localization of C/EBP expression in this tissue. In human fetal lung, C/EBP expression in alveolar epithelial cells is induced in tissue culture (36). In rat lung, immunofluorescence studies have demonstrated C/EBP expression in alveolar type II cells as well as a weaker expression in bronchiolar Clara cells (24).

In line with this, C/EBP has been shown to be expressed in isolated primary rat Clara cells (25). In the embryonic rat lung, C/EBP expression temporally reflects the differentiation-dependent expression pattern of CCSP in the bronchiolar epithelium and of surfactant protein A in alveolar epithelial cells (24). Furthermore, histologic examination of lungs from C/EBP (/) knockout mice demonstrates alveolar abnormalities with hyperproliferation of epithelial cells (37). Taken together, this suggests that C/EBP factors could be involved in controlling lung cell differentiation and differentiation-dependent gene expression.

In this study, we have described the expression of C/EBP and C/EBP in the bronchiolar epithelium of the murine lung. Transient transfection experiments were used to demonstrate that C/EBP and C/EBP regulate expression of the murine CCSP gene. DNase I footprint analysis and mutagenesis studies demonstrated interaction of C/EBP and C/EBP with two C/EBP-binding sequences, forming a compound response element in the proximal CCSP promoter. Electrophoretic mobility shift assays (EMSAs) showed that C/EBP and C/EBP both bind to the identified C/EBP-binding sites. Furthermore, the simultaneous expression of C/EBP and C/EBP in the Clara cell may play an important regulatory role, as C/EBP-C/EBP heterodimers were preferentially formed at both binding sites, and transient transfections with both C/EBP and C/EBP resulted in a superinduction of the CCSP promoter.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Immunofluorescence

C/EBP and C/EBP in lung sections were detected using polyclonal C/EBP and C/EBP antibodies (Santa Cruz Biotechnology, Santa Cruz, CA). Antibody-antigen complexes were detected by the use of a donkey antirabbit secondary antibody conjugated with fluorescein isothiocyanate (FITC). Tissues were prepared as follows: FVB/n mice were killed by cervical dislocation, and a cannula was tied in place at the trachea. The lungs were perfusion fixed with 4% paraformaldehyde during concomitant ventilation via the tracheal cannula, cryoprotected in sucrose, and frozen. Cryosections 12 µm thick were mounted on gelatin-chrome-alum-coated slides. Tissue sections were blocked with 2% bovine serum albumin (BSA), 0.3% Triton, and 1% normal goat serum in phosphate-buffered saline. The primary antibody was diluted at 1:100 in blocking solution, and the conjugated donkey antirabbit antibody was diluted at 1:50 in blocking solution. For control slides, the primary antibody was omitted. Sections were examined with a Zeiss Axioplan 2 microscope (Zeiss, Göttingen, Germany) with filters for FITC.

Cell Culture

A549 lung adenocarcinoma cells were cultured at 37°C in a humidified atmosphere with 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, penicillin (100 IU/ml), and streptomycin (0.1 mg/ml). Mammalian COS-1 cells were cultured in DMEM supplemented as above.

Transient Transfection Assays and Luciferase Assay

For transient transfection experiments, A549 cells were seeded in 30-mm-diameter cell culture wells 24 h before a transfection experiment and transfected at approximately 60% confluence. Plasmid DNAs were transiently transfected into A549 cells by liposome-mediated DNA transfer using the LipofectAMINE transfection reagent (Life Technologies, Rockville, MD), according to the protocol provided by the manufacturer. In each transfection experiment, 1.5 µg of reporter plasmid together with 0.01, 0.05, or 0.5 µg of pCMVC/EBP expression plasmid were used. The overall amount of DNA was kept constant at 2.0 µg by the addition of pUC18 plasmid. Transfected cells were subsequently harvested for luciferase assay in 150 µl lysis buffer (25 mM Tris-PO₄, pH 7.8, 15% glycerol, 2% 3-[3-cholamidopropyl] diemethyl-ammonio-1 propanesulfate [CHAPS], 1% lecithin, 1% BSA, and protease inhibitors), and the cell mix was centrifuged for 5 min to remove cell debris. Luciferase activity was monitored according to the GenGlow luciferase assay kit (Bio Orbit, Turku, Finland) using an Anthos Lucy 1 luminometer (Rosys-Anthos, Hombrechtikon, Switzerland). All experiments were performed in duplicates or triplicates. Neither C/EBP nor C/EBP caused any upregulation of the empty pUBTluc vector (data not shown).

Nuclear Extracts

Transient transfections of COS-1 cells for overexpression of C/EBP factors were carried out using the diethylaminoethyl (DEAE)- dextran method (38). Briefly, the cells were washed with serum-free DMEM before a 2-h incubation in serum-free DMEM transfection medium with 20 µg C/EBP or C/EBP expression plasmid and 500 µg DEAE dextran. Cells were harvested 72 h after transfection, and nuclear proteins were prepared as described previously (39). Nuclear protein concentrations were assayed by the bicinchoninic acid (BCA) method (Pierce, Rockford, IL).

DNase I Footprinting

All DNA fragments used in the DNase I footprinting assays were labeled using Klenow fragment (Amersham, Little Chalfont, UK). DNase I footprinting assays were performed by incubating labeled DNA probe (4 × 10⁶ cpm) together with nuclear proteins from COS-1 cells transfected with expression plasmids for C/EBP factors, as well as control extracts, in 20 µl of binding buffer (20 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], pH 7.5, 25 mM NaCl, 0.5 mM ethylenediaminetetraacetic acid [EDTA], 1 mM dithiothreitol, and 1 µg of poly[dIdC] as nonspecific competitor) for 30 min at room temperature. After incubation, the reaction mixture was treated with 0.1 U of DNase I (Boehringer-Mannheim, Mannheim, Germany) (DNase I was diluted to 5 µl in 5 mM CaCl₂ and 5 mM MgCl₂) for 30 s at room temperature. DNase I activity was terminated by addition of 30 µl stop reagent (1% sodium dodecyl sulfate, 50 mM EDTA, and 200 µg/ml yeast transfer RNA). The samples were extracted with phenol/chloroform, followed by chloroform extraction, and the DNA was precipitated with ethanol. DNA was dissolved in 10 µl of denaturing sample buffer (90% formamide, 0.5 × Tris-borate (TBE), 0.01% bromophenol blue, and 0.01% xylene cyanol). An equal amount of radioactivity was loaded into each lane and analyzed on precast 6% denaturing polyacrylamide sequencing gels (Stratagene, La Jolla, CA). Gels were air-dried and exposed to autoradiographic film.

Electrophoretic Mobility Shift Assays

Double-stranded synthetic oligonucleotides (described in Figure 1) were end-labeled using [-³²P] adenosine triphosphate and T4 polynucleotide kinase. EMSAs were performed by incubating 2 × 10⁶ cpm of oligonucleotide probe together with nuclear protein (1 to 8 µg) in 20 µl of binding buffer (20 mM Hepes, pH 7.6, 40 mM KCl, 2 mM MgCl₂, 1 mM dithiothreitol, 0.5 mM ethyleneglycol- bis-[-aminoethyl ether]-N,N',-tetraacetic acid, 4% Ficoll, and 2 µg of poly[dIdC] as nonspecific competitor) for 15 min at room temperature. Before addition of labeled probe, samples were preincubated in binding buffer for 15 min at room temperature. In some experiments, unlabeled, double-stranded oligonucleotide was included as a competitor in the preincubation step. In case of antibody supershift assays, one microliter of antiserum or preimmune serum was added in the preincubation step. To achieve equimolar amounts of C/EBP and C/EBP for the heterodimer experiments, C/EBP and C/EBP shifts were quantified using a FUJIX BAS 2000 PhosphorImager (Fujifilm, Tokyo, Japan), and subsequently, the nuclear extracts were diluted to an equimolar concentration with nuclear extract buffer. Polyclonal antibodies to C/EBP and C/EBP were from Santa Cruz Biotechnology. The resulting protein-DNA complexes were resolved on a pre-electrophoresed, nondenaturing 5% polyacrylamide gel with 0.5 × TBE as running buffer. Gels were vacuum-dried and exposed to autoradiographic film.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Nucleotide sequence of the proximal murine CCSP promoter (100 to 61) (numbers are based on Reference 15). The two underlined regions represent the two C/EBP-binding sites found in this study showing seven of ten (distal) and eight of ten (proximal) matches with a consensus C/EBP-binding site (30). Synthetic oligonucleotides corresponding to each C/EBP-binding site are shown. Dots denote the mutated bases in the promoter fragment used in the cotransfection assays and DNase I footprint analyses and in the oligonucleotides used in EMSAs.

Plasmids and Site-Directed Mutagenesis

The plasmid pSKmP4.4 (PstI), containing 4.4 kb of the mouse CCSP gene including the promoter region, was cut with HphI or HphI/SacI to obtain promoter fragments spanning nucleotides 2,100 to +7 or 172 to +7, respectively (all numbers are based on Reference 15). The fragments were blunted and subcloned into the SmaI site of the pUBTluc vector (40) to drive the firefly luciferase gene, and plasmids were designated pMCP2.1luc and pMCP0.17luc, respectively. The constructs were analyzed by DNA sequence analysis to verify their authenticity and correct orientation. The pCMVC/EBP expression plasmid has been described elsewhere (25). To yield the corresponding pCMVC/EBP expression plasmid, the C/EBP complementary DNA was subcloned into the BamHI, EcoRI sites of the pCMV5 plasmid (41). Plasmid pMCP0.17luc served as template for site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene), according to the protocol provided by the manufacturer. The oligonucleotides used for site-directed mutagenesis were as follows (complementary strands are not shown, mutated bases are underlined): distal site mutation, GAT GAC CAA GTA AAT AAT ACC GTC TCC TAA GTG GAG CGC; proximal site mutation, CTC CTA AGT GGA GCA CCG TCA CTG CCC TCT ACC. For the mutation of both sites, the plasmid with the distal site mutation was subsequently mutated with the proximal site mutation oligonucleotides. After site-directed mutagenesis, the mutations were verified by DNA sequence analysis.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

C/EBP and C/EBP Are Differentially Expressed in the Mouse Distal Airway

Previous studies by us and others have indicated a role for C/EBP in lung epithelial gene expression, including gene expression in Clara cells (24, 25). In addition to C/EBP, C/EBP has been demonstrated to be expressed at high levels in the lung (28, 29). To characterize the expression pattern of C/EBP and C/EBP in the mouse distal lung, we performed immunofluorescence studies using antibodies directed to C/EBP and C/EBP. As shown in Figure 2A, the labeling for C/EBP was weak in the epithelium of the distal airways. However, a stronger staining for C/EBP was observed in epithelial cells of the alveolar region, indicating that the expression of C/EBP is higher in this part of the lung. Intense, mainly nuclear labeling for C/EBP was localized to the bronchioles, and a weaker labeling was seen in cells of the alveolar region (Figure 2B). Thus, it seems that C/EBP expression is primarily localized to the airways, whereas C/EBP expression is higher in the alveolar region. Control sections with primary antibody omitted, exhibited no labeling in the distal airway nor in the alveolar region (Figure 2C). Clara cells constitute up to 70% of the epithelial cells in the mouse bronchiolar epithelium (42). Thus, the labeling of the majority of the cells in the airway epithelium strongly indicates the presence of C/EBP expression in the Clara cells and a weaker expression of C/ EBP. In contrast to liver, where C/EBP expression is very low under normal conditions and induced upon inflammatory stimuli, these results show a high-level constitutive expression of this protein in the Clara cells.

View larger version (32K): [in this window] [in a new window]   Figure 2.   Immunofluorescence localization of C/EBP and C/EBP in terminal airways and alveolar region of the mouse lung. Cryosections of adult mouse lung were immunostained for C/EBP (A) or C/EBP (B). In C, antibodies were omitted as a control. Insets represent phase contrast micrographs of corresponding sections. White bars indicate 50 µm.

C/EBP and C/EBP Transactivate the CCSP Gene in Lung Epithelial Cells via Response Element(s) in the Proximal Promoter

Both C/EBP and C/EBP are expressed in the Clara cells. To determine whether the mouse CCSP gene is directly regulated by C/EBP factors, the effects of C/EBP and C/EBP on CCSP transcription were examined by transient transfection assays. A luciferase reporter plasmid containing a 2.1-kb promoter region of the mouse CCSP gene was cotransfected into A549 cells together with C/EBP or C/EBP expression vectors. The continuous cell line A549 is of lung epithelial origin and does not express C/EBP (24) nor C/EBP (data not shown) and thus serves as a useful model to investigate the effects of C/EBP factors on lung gene expression (25). Cotransfection of C/EBP or C/EBP expression plasmid with the luciferase reporter plasmid resulted in the induction of CCSP promoter activity in a dose-dependent manner (Figure 3A). When increasing amounts of expression plasmid were transfected, the maximal observed induction of CCSP promoter activity by C/EBP was 5-fold, and higher amounts of C/EBP expression plasmid did not cause any further induction. Cotransfection of the C/EBP expression plasmid resulted in a maximal induction of CCSP promoter activity reaching 15-fold. Previous studies in lung cell lines and transgenic mice have demonstrated that elements necessary for cell-specific expression of CCSP reside within 166 bp upstream of the start site of transcription in the CCSP promoter (14). Therefore, to determine the location of the C/EBP responsive elements in the mouse CCSP promoter, the effects of C/EBP and C/EBP were further examined using a reporter plasmid containing a 172 to +7-bp promoter fragment of the mouse CCSP gene in transient transfections. Again, cotransfection of C/EBP or C/EBP expression plasmid resulted in a marked induction of CCSP promoter activity in a dose-dependent manner (Figure 3B). Cotransfection of the 172-bp CCSP promoter plasmid together with C/EBP expression plasmid demonstrated a maximal induction that was 12-fold, whereas similar experiments using C/EBP expression plasmid demonstrated a maximal induction reaching 28-fold, indicating that C/EBP is a stronger transactivator of the CCSP gene. The finding that the 172-bp promoter region is sufficient for the efficient transactivation by C/EBP and C/EBP indicates that the C/EBP response element(s) resides within this region of the mouse CCSP promoter.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Cotransfection studies to determine transactivation of the CCSP promoter by C/EBP and C/EBP. The reporter plasmids contained one copy of the CCSP promoter sequences 2.1 to +7 bp (A) and 172 to +7 bp (B) in front of a luciferase reporter gene (numbers are based on Reference 15). Increasing amounts of pCMVC/EBP or pCMVC/EBP expression plasmids were cotransfected in A549 lung epithelial cells with constant amounts of reporter plasmid. Error bars indicate ± SD (n = 6).

C/EBP and C/EBP Interact with Two C/EBP-Binding Sites within the Proximal CCSP Promoter Region

To determine the location of C/EBP-binding sites within the 172-bp CCSP promoter region, DNase I footprint analyses were performed using nuclear extracts from COS-1 cells transfected with expression plasmid for C/EBP or C/EBP or control extracts. The DNase I footprint analysis of the 172-bp CCSP promoter region together with C/EBP protein demonstrated DNA-protein interactions in sequences located between 100 to 78 bp in the proximal mouse CCSP promoter (Figure 4, lanes 1 to 3). As seen in Figure 1, analysis of this region of the promoter revealed a seven of 10 match with a C/EBP consensus site (30). In contrast, when performing the DNase I footprint analysis of the 172-bp CCSP promoter region together with C/EBP protein, a more extended footprint on the coding strand was demonstrated in sequences located between 100 to 62 bp in the proximal mouse CCSP promoter, indicating that further DNA-protein interaction was taking place (Figure 4, lanes 4 to 6). Analysis of the corresponding sequence revealed an additional putative C/EBP-binding site with an eight of 10 match to the C/EBP consensus site, approximately 9 bp downstream of the first demonstrated C/EBP-binding site in the CCSP promoter (Figure 1). C/EBP interaction with this second more proximal site was less prominent.

View larger version (54K): [in this window] [in a new window]   Figure 4.   Detection of C/EBP-binding sites within the proximal promoter of the murine CCSP gene by DNase I footprint analysis. The region between nucleotides 172 to +7 of the CCSP gene was subjected to DNase I footprint analysis by incubation with nuclear extracts from COS-1 cells transfected with C/EBP or C/EBP expression plasmids or control extracts. DNase I footprint analysis of the coding strand: lanes 1 to 3, increasing amounts of C/EBP containing nuclear extract; lanes 4 to 6, increasing amounts of C/EBP containing nuclear extract; lanes 7 to 9, increasing amounts of control extract; lane 10, G+A Maxam-Gilbert sequencing. The protected regions (78 to 100 and 62 to 100) are indicated to the left.

Both C/EBP-Binding Sites in the Compound Response Element Are Required for Full Transactivation by C/EBP and C/EBP

The finding that C/EBP factors interact with two putative C/EBP-binding sites in the proximal mouse CCSP promoter prompted us to investigate the importance of each site in site-directed mutagenesis studies. Three mutated constructs were made with either the proximal, distal, or both C/EBP-binding sites mutated as depicted in Figure 1. As seen in Figure 5A, the mutation of the distal C/EBP site, 93 to 84 bp, resulted in a total abolishment of transactivation potential of both C/EBP and C/EBP. When the proximal C/EBP site was mutated, i.e., the site located at 74 to 65 bp, the transactivating efficiency of C/EBP was notably decreased (Figure 5B), although some induction was still observed. Moreover, the transactivation of C/EBP was abolished, although this protein only weakly interacted with the proximal site in DNase I footprint analysis. When both sites were mutated, no transactivation of the reporter gene was detected (Figure 5C). The finding that both C/EBP-binding sites within the response element were needed for full transactivation of both C/EBP and C/EBP led us to further investigate the binding of C/EBP and C/EBP to the promoter. For this purpose we used the mutated CCSP promoter, with the sites in the response element altered, in DNase I footprint analyses. As seen in Figure 6, when the distal site was mutated, no binding of C/EBP or C/EBP was detected to this site. In addition, the binding of C/EBP and C/EBP to the proximal site was affected, possibly explaining the total abolishment of transactivation when the distal site was mutated (Figure 6, lanes 1 to 4). When performing DNase I footprint analysis with the promoter containing a mutation in the proximal site, no binding of C/EBP or C/EBP was detected in this region (Figure 6, lanes 5 to 7). These findings show that both C/EBP-binding sites are important for full transactivation of the CCSP gene by C/EBP and C/EBP. Thus, the two C/EBP-binding sites in the proximal CCSP promoter form a compound response element in that the integrity of both sites is necessary for its function.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Functional analysis of the C/EBP-binding sites within the C/EBP response element by cotransfection studies using promoter-reporter gene constructs altered by site-directed mutagenesis. The 172 to +7 CCSP promoter-luciferase construct was subjected to site-directed mutagenesis in order to alter the DNA sequences of the C/EBP-binding sites. Increasing amounts of pCMVC/EBP or pCMVC/EBP expression plasmids were cotransfected in A549 cells with constant amounts of promoter-reporter gene construct with the distal site altered (A), the proximal site altered (B), or both sites altered (C). Error bars indicate ± SD (n = 6).

View larger version (39K): [in this window] [in a new window]   Figure 6.   Analysis of C/EBP-binding activity of the sites within the compound C/EBP response element by DNase I footprint analysis of the proximal CCSP promoter with the C/EBP-binding sites altered by site-directed mutagenesis. The region between nucleotides 172 to +7 of the CCSP gene was cut out from the respective plasmids containing altered C/EBP-binding sites and labeled. The fragment with the distal C/EBP-binding site altered was incubated in the presence of nuclear extract from COS-1 cells transfected with expression plasmids for C/EBP (lane 2) or C/EBP (lane 3), respectively, or with control nuclear extract (lane 4). Lane 1 represents G+A Maxam-Gilbert sequencing of the fragment. The fragment with the proximal C/EBP-binding site altered was incubated in the presence of nuclear extract from COS-1 cells transfected with expression plasmids for C/EBP (lane 6) or C/EBP (lane 7), respectively. Lane 5 represents G+A Maxam-Gilbert sequencing of the fragment.

C/EBP and C/EBP Interact with Both C/EBP-Binding Sites in Electrophoretic Mobility Shift Assays

The weaker interaction of C/EBP with the proximal site in DNase I footprint analyses suggests a potential differential binding of C/EBP and C/EBP to the C/EBP-binding sites in the response element. This led us to determine the binding efficiency of C/EBP and C/EBP to the two identified C/EBP-binding sites in EMSAs. Two oligonucleotides were constructed to include the distal and the proximal C/EBP-binding sites, respectively (Figure 1). These oligonucleotides were incubated in the presence of nuclear extracts from COS-1 cells, transfected with expression plasmids for C/EBP or C/EBP. In EMSAs, using the oligonucleotide encompassing the distal site as probe, prominent bands were detected both in the presence of C/EBP and C/EBP (Figure 7A, lanes 1 and 5). These bands were efficiently abolished by competition with unlabeled homologous oligonucleotide, whereas no competition was observed when unlabeled, mutated oligonucleotide was included (Figure 7A, lanes 3, 4, 7, and 8). On inclusion of C/EBP and C/EBP antibodies, respectively, the bands were completely abolished and a supershift appeared (Figure 7A, lanes 2 and 6). A similar pattern could be seen for the oligonucleotide encompassing the proximal C/EBP-binding site (Figure 7B). To determine the relative binding efficiency of C/EBP and C/EBP to the C/EBP-binding sites, an EMSA was performed with each respective oligonucleotide in the presence of two different amounts of nuclear extract. As seen in Figure 7C, the binding efficiency of both C/EBP and C/EBP to the distal site was higher compared with the proximal site. Although the binding efficiencies of C/EBP and C/EBP were different between the two C/EBP-binding sites, the relative binding efficiencies of C/EBP and C/EBP to each individual C/EBP site were similar (Figure 7C, compare lanes 1 to 6 and 3 to 8). The finding that both C/EBP and C/EBP can bind to both sites is in line with earlier studies concerning the DNA binding specificity of C/EBP isoforms (28, 29). However, the results were somewhat in contrast to the findings from the DNase I footprint analyses, in which binding of C/EBP to the proximal site was less distinct.

View larger version (31K): [in this window] [in a new window]   Figure 7.   Determination of C/EBP and C/EBP binding activity to the C/EBP-binding sites in the response element by EMSAs. Oligonucleotides corresponding to the distal and proximal C/EBP-binding site, respectively, were used. (A, B) End-labeled oligonucleotides representing each C/EBP-binding site (see Figure 1) were incubated in the presence of nuclear extract from COS-1 cells transfected with expression plasmid for C/EBP or C/EBP. Oligonucleotides were incubated with nuclear extract containing C/EBP or C/EBP (lanes 1 and 5). In lanes 2 and 6, antibodies against C/EBP or C/EBP, respectively, were added. To establish specific binding, unlabeled homologous oligonucleotide was added in a 100-fold excess (lanes 3 and 7) or unlabeled mutated oligonucleotide in a 100-fold excess (lanes 4 and 8). (C) To determine binding activity of C/EBP and C/EBP to the C/EBP-binding sites, end-labeled oligonucleotides representing each site in the C/EBP response element were incubated in the presence of nuclear extract containing C/EBP or C/EBP. Oligonucleotides representing the distal and the proximal C/EBP-binding site were incubated in the presence of increasing amounts of nuclear extract containing C/EBP (distal: lanes 1 and 2; proximal: lanes 5 and 6) and C/EBP (distal: lanes 3 and 4; proximal: lanes 7 and 8).

Preferential Formation of C/EBP-C/EBP Heterodimers at Both C/EBP-Binding Sites

C/EBP factors have previously been shown to form heterodimers (28, 29, 43). The expression of both C/EBP and C/EBP in the Clara cell led us to examine whether C/EBP and C/EBP heterodimers could form at the C/EBP-binding sites in the promoter. When EMSAs were performed using the oligonucleotide encompassing the distal C/EBP-binding site, together with a titration series of nuclear extract from COS-1 cells containing C/EBP and C/EBP, a new complex of intermediate migration was noted between the slower migrating C/EBP complex and the faster migrating C/EBP complex (Figure 8A). This demonstrates the formation of a C/EBP-C/EBP heterodimer at the distal C/EBP site. In addition, at equimolar amounts of C/EBP and C/EBP, the homodimers of C/EBP and C/EBP virtually disappeared, indicating that the heterodimer is formed more efficiently than the respective homodimer (Figure 8A, lane 3). Experiments using the proximal site oligonucleotide similarly demonstrated preferential formation of the heterodimer at equimolar concentrations of C/EBP and C/EBP (Figure 8B). However, although no homodimers were formed at equimolar amounts, the binding of the heterodimer was relatively weaker compared with the binding of the homodimers seen with either C/EBP or C/EBP alone (Figure 8B, compare lanes 1 and 5 to lane 3). The finding that C/EBP and C/EBP demonstrate differences in their binding pattern to the two C/EBP-binding sites led us to investigate the interaction of the preferentially formed heterodimer with the sites in DNase I footprint analysis. By using a titration series of nuclear extract from COS-1 cells containing C/EBP and C/EBP, it was shown that the heterodimer interacts with the distal site, whereas interaction with the proximal site was much weaker (Figure 9). This is in line with the results from the EMSAs, where it was shown that the binding efficiency of the heterodimer to the proximal site was lower.

View larger version (60K): [in this window] [in a new window]   Figure 8.   EMSA to investigate heterodimer formation between C/EBP and C/EBP at the C/EBP-binding sites in the C/EBP response element of the mouse CCSP promoter. A titration series of C/EBP and C/EBP containing nuclear extracts was incubated in the presence of labeled oligonucleotide spanning the distal C/EBP-binding site (A) or the proximal C/EBP-binding site (B). The presence of the respective homodimer and the heterodimer is indicated to the left.

View larger version (57K): [in this window] [in a new window]   Figure 9.   Determination of C/EBP interaction with the CCSP promoter in the presence of both C/EBP and C/EBP. A titration series of nuclear extracts from COS-1 cells transfected with C/EBP or C/EBP expression plasmid were incubated with a fragment spanning sequences 172 to +7 of the mouse CCSP promoter in DNase I footprint analysis. DNase I footprint analysis of the coding strand: lane 1, G+A Maxam-Gilbert sequencing; lanes 2 to 8, titration series of C/EBP and C/EBP containing nuclear extract; lane 9, control nuclear extract; lane 10, G+A Maxam-Gilbert sequencing.

CCSP Promoter Activity Is Superinduced by Coexpression of C/EBP and C/EBP in A549 Cells

To determine the functional importance of C/EBP and C/EBP heterodimers in CCSP gene expression, transient transfection assays were performed with both C/EBP and C/EBP expression plasmids together. Because we had shown that the compound C/EBP response element resides within the proximal CCSP promoter, we used the plasmid containing the 172-bp sequence of the CCSP promoter together with a titration series of C/EBP and C/EBP expression plasmids. The titration was performed at the amount of reporter plasmid that had previously been shown to give the maximal transactivation (Figure 3B). In cotransfections, when both C/EBP and C/EBP expression plasmids were included together, a superinduction was observed with the maximal induction of CCSP promoter activity reaching 50-fold (Figure 10A). Transfection with C/EBP expression plasmid alone gave a 12-fold induction, whereas transfection with C/EBP expression plasmid alone resulted in a 25-fold induction of the reporter gene (Figure 10A), in line with the earlier results (Figure 3B). To investigate the importance of each C/EBP-binding site in the transactivation, we performed transfection studies using equal amounts of C/EBP and C/EBP expression plasmids together with the reporter construct containing mutations in the proximal, the distal, or both sites as above. Mutation of either or both site(s) resulted in a total abolishment of transactivation in the presence of C/EBP and C/EBP (Figure 10B). Together with the preferential formation of C/EBP-C/EBP heterodimers at the C/EBP-binding sites of the compound response element, the synergistic effects suggest an important role for these heterodimers in the regulation of the CCSP gene.

View larger version (20K): [in this window] [in a new window]   Figure 10.   Cotransfection assay to determine synergistic transcriptional activity by C/EBP and C/EBP on the mouse CCSP promoter. (A) Constant amount of reporter plasmid containing CCSP promoter sequences 172 to +7 was cotransfected with different amounts of pCMVC/EBP and pCMVC/EBP expression plasmids into lung epithelial A549 cells. (B) Constant amount of reporter plasmid containing CCSP promoter sequences 172 to +7 with either or both C/EBP-binding sites mutated was cotransfected with equal amounts of pCMVC/EBP and pCMVC/ EBP expression plasmids. Error bars indicate ± SD (n = 6).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Although extensive studies have been performed concerning the roles of C/EBP transcription factors in the liver (34) and in adipocytes (33), little is known about the potential role(s) of C/EBP isoforms in the lung. C/EBP, C/EBP, and C/EBP were shown early to be expressed in the lung, and the highest expression of C/EBP was seen in the lung (28, 44). In this study, we demonstrated high levels of C/EBP in the bronchiolar epithelium, whereas lower levels were seen in alveolar areas. In contrast, C/EBP expression was higher in the alveolar region. These findings offer an explanation of the findings from histologic examination of lungs from C/EBP (/) mice, which demonstrated alveolar abnormalities with hyperproliferation of type II cells, whereas the bronchiolar epithelium exhibited a normal phenotype (37). The high expression of C/EBP in the bronchioles could substitute for the absence of C/EBP here, but not in the alveolar epithelium where expression is lower. That such compensation can occur has been demonstrated in studies of phosphoenolpyruvate carboxykinase (PEPCK) gene regulation in liver where C/EBP can substitute for C/EBP in C/EBP (/) mice (45).

The differentiated Clara cell abundantly expresses CCSP. This protein has been estimated to constitute 40% of the total secreted protein from the Clara cell (46). As shown here, the Clara cell constitutively expresses both the C/EBP and C/EBP isoforms. In line with this, the present study demonstrates that the promoter of the mouse CCSP gene can be directly transactivated by both C/EBP and C/EBP. By carrying out a series of transient cotransfection assays in a lung epithelial cell line, we found that C/EBP and C/EBP stimulate transcription of CCSP with different efficiencies and that a compound response element containing two adjacent C/EBP-binding sites is situated in the proximal CCSP promoter.

The transactivation potential of C/EBP and C/EBP has previously been shown to be comparable to minor differences (28). A higher transactivation potential for C/EBP compared with C/EBP, has been demonstrated for some specific genes (47, 48). In the present study, we demonstrate differences in the transactivation potential of C/EBP and C/EBP paralleling a difference in binding pattern of these transcription factors to the two adjacent C/EBP-binding sites within the compound response element. The DNase I footprint analyses demonstrated a distinguished DNA-protein interaction spanning both C/EBP-binding sites in the presence of C/EBP, whereas binding of C/EBP to the proximal C/EBP-binding site was less prominent. The degree of the differential binding remains to be elucidated as differences in protein levels of C/EBP and C/EBP in the nuclear extract may exist, and C/EBP and C/EBP may differ in their binding kinetics, which could result in a divergent footprint pattern. However, the difference was still evident even as increasing amounts of C/EBP containing nuclear extract were used. A differential binding of C/EBP and C/EBP is surprising, as earlier studies have demonstrated only minor differences in DNA-binding specificity between these proteins (28). In line with these previous studies, it was shown in the EMSAs that both C/EBP and C/EBP can bind to each individual site with equal efficiency. A possible explanation for the differential binding seen in the context of the full promoter only, and thus in the DNase I footprint analyses but not in the EMSAs, is offered by previous studies concerning DNA-bending of C/EBP and C/EBP. It has been shown that C/EBP and C/EBP induce a directed bend of similar magnitude, but whereas C/EBP introduces a bend that is orientated toward the minor groove, C/EBP induces a directed bend toward the major groove (49). It has been proposed that DNA-bending by bZIP proteins may inhibit DNA binding of proteins that recognize overlapping or adjacent DNA sequence elements (50). That may be of significance as the two C/EBP-binding sites in the CCSP promoter are separated by only nine nucleotides. Furthermore, the nucleotide sequence of the two C/EBP-binding sites exhibits differences. The proximal site has nonconsensus nucleotides in positions 3 and 4 of the left half-site, positions that have previously been reported to be important for DNA-binding of C/EBP factors (30). In line with this, C/EBP and C/EBP bind less efficiently to the proximal C/EBP-binding site. Binding of C/EBP to the stronger distal site may induce a directed bend that inhibits binding of C/EBP to the weaker proximal site. A pattern similar to the effects seen on the CCSP promoter regarding the transactivation potential of C/EBP and C/ EBP has been described for the mouse serum amyloid A3 (SAA3) gene (47, 51). In these studies, it was described that the mouse SAA3 gene contains two adjacent C/EBP-binding sites situated in the proximal promoter similarly to what we have shown for the CCSP gene. The presence of two adjacent C/EBP-binding sites may be an important feature in the regulation of specific genes in the presence of C/EBP and C/EBP.

The full transactivation of C/EBP and C/EBP is dependent on the integrity of both C/EBP-binding sites in the compound response element in the proximal CCSP promoter. As was shown in the site-directed mutagenesis studies, mutation of the distal site resulted in abolishment of transactivation from C/EBP and C/EBP in transfection studies. Mutation of the proximal site resulted in a marked reduction of C/EBP activity but also abolished the activity of C/EBP, although C/EBP interaction was weak with this site, as shown in the DNase I footprint analysis. Furthermore, the binding of C/EBP to the distal site did not seem to be affected by the mutation of the proximal site. Thus, the absence of transactivation by C/EBP when the proximal site is mutated can hardly be explained by abolished interaction of C/EBP with this site only. An additional explanation is a potential involvement of Sp transcription factors. Members of the Sp transcription factor family have been shown to be important for the regulation of the rabbit uteroglobin/CCSP gene (52). Furthermore, the CCSP gene has been shown to interact with Sp1/ Sp3 in a region spanning the proximal site (53), i.e., at the site of mutation of the proximal site. One could speculate that Sp factors interact with a region close to the proximal site and synergize with C/EBP in the transactivation of the CCSP gene, as it previously has been shown that Sp1 and C/EBP act cooperatively in the transactivation of the Cyp2D5 gene (54). Consequently, it is possible that mutation of the proximal site, which may affect binding of Sp factors, leads to an abolishment of cooperativity between Sp factors and C/EBP and hence a lowered transactivation activity. In contrast, C/EBP is strongly interacting with both C/EBP-binding sites in the promoter and may not be dependent on cooperative activity with Sp factors for full transactivation. This is supported by the finding that the mutation of the proximal site resulted in a remaining weak transactivation potential of C/EBP.

In this study, we demonstrate the preferential formation of C/EBP-C/EBP heterodimers at both C/EBP-binding sites in the CCSP promoter. The property to efficiently form heterodimers has been shown for C/EBP, C/EBP, and C/EBP (28, 29). In addition, preferential formation of C/EBP heterodimers has been described for C/EBP and C/EBP (29, 43), and for truncated forms of C/EBP and C/EBP (28). In this study we describe the preferential formation of full length C/EBP-C/EBP heterodimers paralleling the enhanced transactivation of the CCSP promoter in the presence of both C/EBP and C/EBP. When performing cotransfection studies with C/EBP and C/EBP alone, 12- and 28-fold inductions of the reporter gene were seen, respectively. These levels were further induced reaching 50-fold in cotransfection studies with C/EBP and C/EBP plasmid together, indicating that these factors synergistically transactivate the promoter. Synergistic transactivation by C/EBP factors has previously been described for specific genes in the liver (43, 48). In the case of the CCSP promoter, it was demonstrated that the C/EBP-C/EBP heterodimers preferentially form at both binding sites, and mutation of either or both site(s) resulted in total abolishment of transactivation. This indicates that the integrity of both sites is necessary for transactivation by the heterodimer, similar to the results with the respective homodimer. In addition, results from the DNase I footprint analysis indicate that the main interaction of the C/EBP-C/EBP heterodimer with the CCSP promoter is taking place at the distal site. Although the mechanism underlying the synergistic transactivation remains to be established, these results suggest that the synergistic transactivation of CCSP by C/EBP and C/EBP involves C/EBP-C/EBP heterodimers. Previously, preferential formation of C/EBP-C/EBP heterodimers has been described at a C/EBP-binding site in the human interleukin-6 gene and was paralleled by a synergistic transcriptional activation in the presence of both C/EBP and C/EBP (43), much as we have demonstrated for C/EBP, C/EBP, and the CCSP gene. Moreover, it has been shown that C/EBP and C/EBP act synergistically in transactivation of the 1-acid glycoprotein gene (48). Thus, it seems plausible that through these mechanisms, the coexpression of C/EBP and C/EBP in the differentiated Clara cell under normal conditions is important for the high level expression of CCSP and that the levels of CCSP can be directly influenced by the relative expression levels of these transcription factors.

Expression of CCSP serves as a differentiation marker for the bronchiolar Clara cell during fetal development. Moreover, it has been shown in numerous studies that injury to the bronchiolar epithelium results in decreased levels of CCSP. This probably reflects dedifferentiation of the Clara cells, as these cells serve as progenitor cells in the repair after injury to the epithelium. In line with this, a full understanding of the regulation of the CCSP gene may provide further insight into differentiation-dependent processes of the Clara cell, both during development and after lung injury. Previous studies have demonstrated a correlation between C/EBP and CCSP expression in lung cellular differentiation, both during embryonal development (22, 24) and in in vitro cell culture models (25). The transcription factors TTF-1 and HNF-3, which have been shown to be important for CCSP gene regulation, correlate spatially but not temporally with CCSP expression during lung embryonal development as described previously. Thus, the present findings of activation of the murine CCSP promoter by C/EBP suggest that C/EBP accounts for the differentiation-dependent expression of CCSP, whereas TTF-1 and HNF-3 give Clara cell specificity. Further studies will have to investigate if C/EBP is acting cooperatively with TTF-1 and/or HNF-3 to give the high expression levels observed in the adult lung. The importance of C/EBP in differentiation-dependent processes is well documented in liver and fat, and a similar role for C/EBP is likely to account for the Clara cell as well. Certainly, C/EBP is important for differentiation of the alveolar epithelium, as demonstrated by the hyperproliferation of type II cells in C/EBP (/) mice (37). Moreover, the expression of C/EBP temporally reflects the differentiation-dependent expression of surfactant protein A in the rat alveolar epithelium (24). C/EBP may play important regulatory roles in controlling differentiation in several cell types of the lung.

In this study, we demonstrate the existence of C/EBP in addition to C/EBP in the bronchiolar epithelium. What might be the role of C/EBP in this context? The spatial expression pattern of C/EBP with the highest levels in the bronchiolar epithelium suggests that it plays an important role here. Specifically, its involvement in controlling the expression of the major secretory protein of the Clara cell, CCSP, further supports this notion. In other tissues, C/EBP is involved in controlling differentiation as well as C/EBP expression. In adipocytes, for instance, C/EBP is expressed early during differentiation, before C/EBP (33). Conversely, in cells of myelomonocytic lineage, C/EBP and C/EBP are expressed in the more differentiated cells (55). Studies of C/EBP in the developing rabbit lung have demonstrated a differentiation-dependent expression pattern, with C/EBP levels reaching maximum just before birth (36). This is paralleled by several differentiation-dependent processes in the alveolar region of the lung. However, it is also reflected by the differentiation-dependent expression pattern of uteroglobin/ CCSP during development (56, 57). This suggests that C/EBP as well as C/EBP are involved in controlling differentiation-dependent processes within the lung. A model including the action of these two factors together is further supported by the findings presented here, demonstrating a maximal induction of the CCSP gene when both C/EBP and C/EBP are present as they are in the mature differentiated Clara cell.

In lung, C/EBP expression is upregulated by glucocorticoids, as recently demonstrated in studies using explant cultures of fetal human lung (36). Glucocorticoids have several important roles in the lung, as was drastically demonstrated in the glucocorticoid receptor knockout mouse, which succumbed shortly after birth in respiratory failure owing to impaired embryonal development of the bronchioles and alveoli (58). The high level of CCSP expression in lung is further increased by glucocorticoids in both rabbit and rodents. Binding sites for the glucocorticoid receptor have been demonstrated in the upstream region of the rabbit CCSP promoter (59). However, studies in transgenic mice show that these elements are not necessary for the glucocorticoid effect (12), and this upstream region is absent in the rat and mouse promoters. So far, no additional glucocorticoid response elements have been described to account for the glucocorticoid stimulation of CCSP expression. Together this raises the possibility that the glucocorticoid effects on CCSP expression are indirect and mediated via C/EBP. A similar model has been proposed for differentiated adipocytes where C/EBP is induced by glucocorticoids (60) and has been suggested to mediate the induction of the obese gene by glucocorticoids in fat (61).

We have demonstrated that the lung-specific CCSP gene is regulated by C/EBP and C/EBP through interaction with two adjacent C/EBP-binding sites, forming a compound C/EBP response element. Moreover, C/EBP and C/EBP are constitutively expressed in the bronchiolar Clara cell. It is known that C/EBP factors are expressed in the lung. However, their respective roles have been little studied. The finding that C/EBP and C/EBP synergistically transactivate the CCSP gene and that there seems to be a correlation between CCSP expression and C/EBP factors during differentiation suggest that these C/EBP factors play a key role in differentiation-dependent processes in the bronchiolar epithelium of the lung.