In Vivo and In Vitro Analysis of Factor Binding Si
http://www.100md.com
病菌学杂志 2006年第1期
Department of Molecular Biology and Biochemistry and Cancer Research Institute, University of California, Irvine, California 92697-3905
ABSTRACT
Jaagsiekte sheep retrovirus (JSRV) is the causative agent of ovine pulmonary adenocarcinoma, a contagious lung cancer of sheep that arises from type II pneumocytes and Clara cells of the lung epithelium. Studies of the tropism of this virus have been hindered by the lack of an efficient system for viral replication in tissue culture. To map regulatory regions important for transcriptional activation, an in vivo footprinting method that couples dimethyl sulfate treatment and ligation-mediated PCR was performed in murine type II pneumocyte-derived MLE-15 cells infected with a chimeric Moloney murine leukemia virus driven by the JSRV enhancers (Mo+JS Mo-MuLV). In vivo footprints were found in the JSRV enhancers in two regions previously shown to be important for JSRV long terminal repeat (LTR) activity: a binding site for the lung-specific transcription factor HNF-3 and an E-box element in the distal enhancer adjacent to an NF-B-like binding site. In addition, in vivo footprints were detected in two downstream motifs likely to bind C/EBP and NF-I. Mutational analysis of a JSRV LTR reporter construct (pJS21luc) revealed that the C/EBP binding site is critical for LTR activity, while the putative NF-I binding element is less important; elimination of these sites resulted in 70% and 40% drops in LTR activity, respectively. Electrophoretic mobility shift assays using nuclear extracts from MLE-15 murine Clara cell-derived mtCC1-2 cells with probes corresponding to the NF-I or C/EBP sites revealed several complexes. Antiserum directed against NF-IA, C/EBP, or C/EBP supershifted the corresponding protein-DNA complexes, indicating that these isoforms, which are also important for the expression of several cellular lung-specific genes, may be important for JSRV expression in lung epithelial cells.
INTRODUCTION
Jaagsiekte sheep retrovirus (JSRV) is the etiological agent of ovine pulmonary adenocarcinoma (3, 11, 27, 30, 32, 45), a contagious lung cancer of sheep and goats that closely resembles human adenocarcinoma of the deep airways, including bronchioloalveolar carcinoma (34). Lung cancer is the leading cause of human cancer deaths worldwide (19); the incidence of human lung adenocarcinoma has steadily risen, and this neoplasm is now responsible for up to one quarter of lung cancer fatalities (4, 19). Thus, ovine pulmonary adenocarcinoma presents a unique opportunity to study the mechanisms of lung carcinogenesis in an animal model.
Retroviruses have historically given insights into the mechanisms of oncogenesis. However, most retroviruses induce neoplasms of hematopoietic cells, whereas the great majority of human cancers are carcinomas (originating from epithelial tissues). JSRV induces tumors of differentiated lung epithelial cells: type II pneumocytes (21, 41) and Clara cells (36). Both in vivo and in vitro, JSRV can infect several different cell types, as judged by the presence of JSRV DNA (17, 31, 33). However, high-level virus expression is observed only in tumor cells in the lung (28, 29). This highly restricted productive infection in vivo makes JSRV quite unique among retroviruses. We have previously reported that the JSRV tropism for lung epithelial cells is correlated with the transcriptional specificity of the viral long terminal repeat (LTR) for these cells (28). Retroviral LTRs contain the viral promoter and enhancer elements, and numerous studies with other retroviruses have indicated the importance of the LTR enhancers in determining pathogenicity and disease specificity.
Studies of cellular genes specifically expressed in lung epithelial cells have identified several important transcription factors. Both type II pneumocytes and Clara cells specifically express several proteins found in surfactant, including SP-A, SP-B, SP-C, SP-D, and the Clara cell secretory protein (CCSP). SP-C and CCSP are expressed exclusively in type II pneumocytes and Clara cells, respectively (21, 36, 41). The upstream promoter and enhancer regions of these genes have revealed several common regulatory elements (24, 42). Members of the winged helix family of transcription factors, such as HNF-3, HFH-4, and HFH-8 are expressed throughout lung development, and binding sites for these factors are found on several surfactant promoters. Other transcription factors important for surfactant gene expression include the homeodomain protein TTF-1, USF, NF-I, Ets-I, SP-I, and C/EBP (24, 42). In searching the JSRV LTR U3 sequences, we found that the JSRV LTR has one or more putative binding sites for all of these factors except for TTF-I (Fig. 1) (28). In further studies, we showed that the lung- and liver-specific transcription factor HNF-3 is important for JSRV LTR activity in the type II pneumocyte-derived cell line MLE-15 (22, 28). In this study, we explored the roles of other transcription factors in JSRV LTR activity in MLE-15 cells and Clara cell-derived mtCC1-2 cells by a combination of in vivo footprinting, site-specific mutations, and electrophoretic mobility shift assay (EMSA) supershifts.
MATERIALS AND METHODS
Cell lines. MLE-15 cells (provided by J. Whitsett, Cincinnati, OH) were grown in RPMI 1640 (Gibco BRL) containing 2% fetal bovine serum, 0.5% insulin-transferrin-sodium selenite (Sigma), 5 mg/liter transferrin, 10 mM HEPES, 10–8 M -estradiol, and 10–8 M hydrocortisone. NIH 3T3 (American Type Culture Collection) and mtCC1-2 cells (provided by F. DeMayo, Baylor College of Medicine, Houston, TX) were grown in Dulbecco's minimal essential medium (American Type Culture Collection) containing 10% fetal bovine serum. All three cell types were cultured at 37°C with 5% CO2.
In vivo DMS treatment of Mo+JS Mo-MuLV-infected cells. To achieve partial methylation of guanines in vivo, two 15-cm-diameter plates of subconfluent MLE-15 cultures infected with Mo+JS Mo-MuLV (a chimeric Moloney murine leukemia virus driven by the JSRV enhancers) were treated with 1% dimethyl sulfate (DMS) in growth medium at 37°C for 2 min. The treatment was stopped by aspiration of the DMS-containing medium followed by an immediate rinse with 25 ml of phosphate-buffered saline (PBS) prewarmed to 37°C and two successive 30-s washes with 25 ml of PBS at 37°C. Genomic DNA was harvested by the addition of 3 ml of cell lysis solution (300 mM NaCl, 50 mM Tris-Cl [pH 8.0], 25 mM EDTA [pH 8.0], 0.2% [vol/vol] sodium dodecyl sulfate, 0.2 mg of proteinase K per ml); this mixture was allowed to incubate for 5 min before the cell slurry was removed into a test tube by gentle scraping. As a control, genomic DNA was also harvested from untreated infected cells in parallel. Both samples were incubated at 37°C from 4 h to overnight, after which the genomic DNA was phenol-chloroform extracted and ethanol precipitated.
In vitro DMS treatment of DNA. Extracted genomic DNA from control cultures was subjected to DMS treatment in vitro by incubation with 1% DMS in H2O for 1 min at 25°C. The reaction was stopped by the addition of ice-cold DMS stop buffer (1.5 M sodium acetate [pH 7.0], 1 M -mercaptoethanol, 100 μg of Saccharomyces cerevisiae tRNA per ml), which was immediately followed by the addition of 2.5 volumes of ethanol on dry ice. Samples were precipitated by incubation for at least 30 min at –70°C and pelleted in a microcentrifuge for 15 min at 4°C. DNA pellets were allowed to air dry for 10 min and resuspended in 200 μl of 1 M piperidine in H2O for 15 min at room temperature prior to cleavage.
Piperidine cleavage. Following in vivo or in vitro DMS treatment, extracted genomic DNA from each cell type was cleaved at all methylated guanines by incubation in 200 μl of 1 M piperidine for 30 min at 90°C. The piperidine was removed by lyophilization, and the cleaved DNA pellets were resuspended in 360 μl of TE buffer (10 mM Tris-Cl, 1 mM EDTA; pH 7.5). Residual piperidine was removed by two successive ethanol precipitations. The first entailed addition of 40 μl of 3 M sodium acetate followed by 1 ml of 100% ethanol and incubation for 30 min at –70°C. DNA samples were pelleted in a microcentrifuge for 15 min at 4°C and resuspended in 500 μl of TE buffer. The DNA pellets were ethanol precipitated a second time by the addition of 170 μl of 8 M ammonium acetate and 670 μl of isopropanol and incubation for at least 30 min at –70°C. The precipitated samples were pelleted by microcentrifugation as described above, washed with 500 μl of 75% ethanol, and repelleted for 5 min at room temperature. The resulting DNA pellets were resuspended in double-distilled water to a final concentration of 0.4 μg/μl.
LMPCR. Two micrograms of DMS-treated and piperidine-cleaved genomic DNA was used for ligation-mediated PCR (LMPCR) as described previously (15, 16, 25, 35) with minor modifications. Single-stranded DNA fragments with guanine residues at both termini result from the DMS treatment and piperidine cleavage. To provide appropriate substrates for linker ligation, double-stranded, blunt-ended molecules were generated by primer extension from an Mo-MuLV- or JSRV-specific oligonucleotide (Table 1 and Fig. 2D). This first-strand primer extension was accomplished by incubation of 2 μg of DMS-treated and piperidine-cleaved DNA with 0.3 pmol of primer 1 (Mol 2E, Mol 1A, and JSRV 1B) in 30 μl of 1x Vent DNA polymerase buffer (New England Biolabs) containing 4 mM MgSO4, 0.25 mM concentrations of each deoxynucleoside triphosphate, and 0.5 U of Vent DNA polymerase (New England Biolabs). The DNA was denatured at 95°C for 5 min, annealed by incubation at 55°C for 20 min, and extended by a subsequent incubation for 10 min at 72°C. Ligation of the unidirectional linker was completed by the addition of 20 μl of 110 mM Tris-Cl (pH 7.5)-17.5 mM MgCl2-50 mM dithiothreitol, 25 μl of 10 mM MgCl2-20 mM dithiothreitol-3 mM ATP (pH 7.0)-4 μM unidirectional linker (in 50 mM Tris-Cl [pH 7.7]), and 3 U of T4 DNA ligase (Gibco/BRL). The unidirectional linker oligonucleotide sequences have been described by Mueller and Wold (25) and are as follows: LMPCR.1, 5'-GCGGTGACCCGGGAGATCTGAATTC-3', and LMPCR.2, 5'-GAATTCAGATC-3'. This mixture was incubated at 17°C overnight, after which the DNA was recovered by ethanol precipitation. The precipitated DNA pellet was resuspended in 50 μl of H2O, and PCR amplification was accomplished by the addition of 50 μl of 2x Vent buffer containing 8 mM MgSO4, 5 mM deoxynucleoside triphosphate mix, 1 pmol of Mo-MuLV or JSRV oligonucleotide 2 (Mol 2F, Mol 2A, JSRV 2D, and JSRV 2B) (Table 1 and Fig. 2D), 1 pmol of oligonucleotide LMPCR.1, and 1 U of Vent DNA polymerase. These samples were placed in a thermocycler and cycled 17 times with a profile of 95°C for 1 min, 66°C for 2 min, and 72°C for 1 min, with a final extension of 10 min at 72°C. Following amplification, PCR products were labeled by the addition of 5 μl of labeling buffer (2 mM concentrations of each deoxynucleoside triphosphate, 1x Vent polymerase buffer, 8 mM MgSO4, 1 U of Vent polymerase, 2.3 pmol of an Mo-MuLV- or JSRV-specific 32P-end-labeled oligonucleotide [oligonucleotide 3 {Mol 3F, Mol 3A, JSRV 3B, and JSRV 3D}; Table 1 and Fig. 2D]) and subjected to two rounds of 95°C for 1 min, 69°C for 2 min, and 72°C for 1 min. Each reaction mixture was then subjected to phenol-chloroform extraction and ethanol precipitation prior to electrophoresis on a 6% polyacrylamide sequencing gel. The reactions were visualized by autoradiography with Kodak BioMax MR film and also by phosphorimaging on a Molecular Dynamics 445 SI PhosphorImager.
Plasmids. pJS21luc was constructed by cloning the JSRV21 LTR directly upstream of the firefly luciferase gene in pGL3basic, as previously described (28). PCRs were performed using the PfuTurbo polymerase (Stratagene) as recommended by the manufacturers.
Deletions and alterations of the NF-I and C/EBP sites were generated by PCR mutagenesis (for schematic diagrams of each mutant, see Fig. 6).
All constructs were verified by nucleotide sequencing and/or restriction digestion.
Transient transfection and luciferase assays. Transient transfections were performed on 2 x 105 to 4 x 105 cells plated on six-well plates (Falcon) approximately 24 h before transfection. For each well, 700 ng of plasmid DNA (500 ng of reporter plasmid and 200 ng of pRLnull to ascertain transfection efficiency) and 6 μl of fugene (Roche) were used as recommended by the manufacturers. Experiments were performed using the dual luciferase reporter system (Promega). Activity of the reporter plasmids was calculated as the percent activity with respect to the wild-type pJS21luc plasmid.
After 48 h, transfected cells were washed with PBS and lysed with 500 μl/well of passive lysis buffer (Promega). Dual luciferase assays were performed on 20 μl of cleared lysate by rapid addition of luciferase assay reagent (Promega), and light output was integrated over 10 seconds at room temperature using a Moonlight 2010 luminometer (Analytical Luminescence Laboratory). Luciferase activity was normalized for transfection efficiency and cell extract preparation by sequentially measuring Renilla luciferase activity.
All of the experiments described above were performed in at least 12 independent transfections using two separate DNA preparations. Results were expressed as means ± standard errors (95% confidence interval). The linear range of the reaction was predetermined.
Nuclear extracts, oligonucleotides, and EMSA. Nuclear extracts were prepared from MLE-15, mtCC1-2, and NIH 3T3 cell lines as described previously (12). Protein concentrations were determined by Bradford assay (Bio-Rad).
For the electrophoretic mobility shift assays, the following double-stranded oligonucleotide probes were used: NFIdown EMSA (GCTTTTGGCACTGCTTCATAGAAA) and C/EBP EMSA (GGTGATTGTGTAAGAATCCG), corresponding to canonical NF-I and C/EBP binding sites at positions –135 to –112 and –82 to –63 of the U3 of JSRV21, respectively. Underlined nucleotides indicate core binding sequences for the respective transcription factor.
Oligonucleotide probes were end labeled with [-32P]ATP using T4 polynucleotide kinase. All binding reactions were performed at room temperature using buffers purchased from Geneka and under conditions recommended by the manufacturer. Nuclear extracts were incubated with the probe for 30 min before electrophoresis. Rabbit antisera raised against mouse NF-IA and C/EBP// were purchased from Geneka or Santa Cruz Biotechnologies, respectively. For supershift assays, 1 μl of normal serum or antibody was preincubated with nuclear extracts for 30 min. The anti-C/EBP antisera produced supershifts in control EMSAs using nuclear extracts from 293T cells transiently transfected with a C/EBP, , or expression plasmid. In competition gel shifts, the probe and cold competitor oligonucleotide were premixed prior to incubation with nuclear extracts. Each binding reaction was subjected to electrophoresis through a 5% nondenaturing polyacrylamide gel in 1x Tris-glycine-EDTA (NF-I) or 0.5x Tris-borate-EDTA (C/EBP) buffer. After 2 h, gels were dried and exposed to film at –70°C for approximately 5 h.
RESULTS
In vivo footprinting of the Mo+JS LTR in infected MLE-15 cells. Cells of the bronchiolar epithelium quickly lose their differentiated characteristics when cultured in vitro, and no ovine cell lines that maintain these properties exist. As a result, an efficient ovine tissue culture system for propagation of JSRV has not yet been established, although low-level productive infection is observed in most ovine cell lines (33). In the past, we have made use of the mouse type II pneumocyte-derived cell line MLE-15; this line was derived from lung tumors of transgenic mice expressing simian virus 40 large T antigen from the SP-C promoter (43). These cells maintain a differentiated phenotype through several passages and continue to express several type II pneumocyte markers, including surfactant proteins and HNF-3. We found that the JSRV LTR has a high relative level of activity in MLE-15 cells (28). However, MLE-15 cells cannot be infected by JSRV, since murine cells lack a functional JSRV receptor (37).
To further study the role of the JSRV LTR in determining lung tropism, a Moloney murine leukemia virus (Mo-MuLV) driven by a chimeric LTR, in which the Mo-MuLV U3 enhancer sequences (–342 to –150) were replaced with a large portion of the JSRV U3 region, was constructed (Fig. 2A) (23). This virus, termed Mo+JS Mo-MuLV, carries bases –266 to –37 of the JSRV LTR in an otherwise intact Mo-MuLV genome (Fig. 2B and C). We found that like wild-type Mo-MuLV, Mo+JS Mo-MuLV is able to infect most mouse cell lines, including MLE-15 (23). In transient transfection assays, the Mo+JS LTR was three- to fivefold more specific for lung epithelial cells than the wild-type Mo-MuLV LTR. Moreover, in infectivity assays, Mo+JS Mo-MuLV was 3 to 4 logs more specific than wild-type Mo-MuLV for MLE-15 infection versus nonlung epithelial cell lines (23). Mo+JS Mo-MuLV-infected MLE-15 cells therefore contain active proviruses in which transcription is driven by the JSRV LTR enhancers. As such, they provide a system for studying factors bound to the JSRV LTR in vivo.
We previously have made use of an in vivo footprinting method to investigate factor binding to the Mo-MuLV LTR in different cell types (16, 25). This method involves the limited in vivo methylation of guanine residues in the DNA of intact infected cells by treatment with DMS followed by extraction of the DNA and cleavage at methylated G residues with piperidine. A guanine residue bound by a factor may be protected from methylation and subsequent cleavage. Guanine residues also may be hypermethylated if the DNA is bent or if factor binding creates a hydrophobic pocket, and this makes the residues hypersensitive to cleavage. For comparison, genomic DNA from the same cells is first extracted and then methylated in vitro and analyzed in parallel. The region of interest is then selectively amplified and radiolabeled by LMPCR using nested strand-specific primers. The end products are resolved in a sequencing gel and analyzed in a phosphorimager. The resulting pattern is a "G-ladder" for the region adjacent to the PCR primers, and the relative intensities of each labeled band reflect the efficiency of methylation at that guanine residue. Comparison of the LMPCR reactions from in vivo- and in vitro-methylated DNA thus provides an indication of in vivo factor binding to those sequences. Bases that are hypersensitive to cleavage will have a more intense in vivo image than the relative band strength of the in vitro-treated DNA. Protected bases will likewise have a less intense corresponding band on the gel than in the control in vitro-treated DNA.
To map the JSRV regulatory regions important for Mo+JS Mo-MuLV transcriptional activity, infected MLE-15 cells were subjected to in vivo DMS treatment and LMPCR analysis. The primers used are shown in Table 1, and their relative positions are diagrammed in Fig. 2D. Two sets of primers for each strand were used to allow resolution of promoter-proximal and -distal JSRV sequences. JSRV residue positions in the Mo+JS LTR are designated relative to the JSRV LTR. The Mol 2E, JSRV 2D, and JSRV 3D oligonucleotide primers were designed for LMPCR analysis of the promoter-proximal end of the lower (antisense) strand. In vivo footprinting with these primers is shown in Fig. 3. Our previous mutational analysis of JSRV LTR-luciferase reporters indicated that the upstream HNF-3 site is important for JSRV LTR activity in MLE-15 cells (22), and this site showed evidence for factor binding in vivo. It contained a protected guanine residue on the antisense (lower) strand, just outside the core TGTTTG region (JSRV LTR nucleotide –134) (Fig. 3A). In addition, analysis of the antisense strand also revealed protection within the core sequence of an adjacent CAAT box that matches a consensus NF-I binding site. (This NF-I site is referred to as the downstream NF-I site, since there is an additional upstream consensus NF-I site in the JSRV LTR at nucleotides –227 to –210.)
Phosphorimager analysis was used to quantify the extent of protection or hypersensitivity, as shown in Fig. 3B. Protections where there was a >25% change in the intensity of the band were considered major, while changes of <25% were considered minor or weaker. The 3' Mo-MuLV/JSRV LTR junction is positioned in the middle of the downstream HNF-3 binding site, but the HNF-3 core binding sequence (TAAACA) is maintained in Mo+JS. Interestingly, this site also showed protection of the single guanine residue on the lower (antisense) strand (–37), even though our previous mutational analyses had indicated that the downstream HNF-3 site is not necessary for LTR activity in MLE-15 cells (22). Two weaker protections were detected on the same gel within the nearby putative Ets-I/Gfi-I sites at positions –108 and –109 (Fig. 3A). Footprints were also detected on the lower (antisense) strand in a large region lacking sequences that match canonical binding sites. Three guanine residues, including –153, –155, and –171, were protected, and –153 was strongly protected. Additionally, a stretch of hypersensitive thymidine and cytosine residues was detected from –168 to –166, which included a strong hypersensitivity at thymidine –167. This suggests that a novel MLE-15 cell factor(s) might be binding to the JSRV enhancer sequences upstream of the HNF-3 binding site.
When the upper (sense) strand of the JSRV sequences in Mo+JS Mo-MuLV-infected MLE-15 cells was analyzed, protected and hypersensitive bases were detected within a second CAAT box containing a consensus C/EBP binding site (Fig. 4A). Guanine and adenosine residues from –68 to –70 were hypermethylated. In addition, guanine –73 was protected. All three residues are contained within the core of the C/EBP binding sequence. A putative Gfi-I binding site overlaps with the C/EBP site, but the footprint was far from the core Gfi-I binding sequence. This suggested that C/EBP isoforms may bind to the JSRV enhancer in MLE-15 cells as well.
We previously showed that elimination of a putative NF-B site (–260 to –248) diminished activity of the JSRV LTR in MLE-15 cells by 50%; however, an NF-B antibody was not able to supershift an EMSA complex, indicating that a factor other than NF-B binds this site in MLE-15 cells (28). Indeed, a strong in vivo hypersensitive residue on the upper strand was detected directly outside of the putative NF-B site within an overlapping region that resembles an E-box (Fig. 4B). The footprinting pattern in the distal enhancer suggests that an E-box-binding protein, such as N-myc or upstream stimulating factor (USF), may bind to this region in MLE-15 cells, although this does not eliminate the possibility that a separate factor(s) occupies the closely apposed NF-B-like element.
A summary of the in vivo footprinting of the JSRV sequences in the Mo+JS LTR in infected MLE-15 cells is shown in Fig. 5. Underlined bases and triangles indicate protected (lowercase letters, triangles pointing away from the DNA) or hypersensitive (uppercase letters, arrowheads pointing toward the DNA) bases. Strong interactions are shown with bold lettering or filled arrowheads. Most of the stronger footprints were in the promoter-proximal half of the JSRV enhancers, including the adjacent upstream HNF-3 and downstream NF-I sites, as well as the C/EBP site. Several other sites, such as two putative Ets-I sites and the upstream NF-I site, did not show footprints.
While the in vivo DMS footprinting reported here allowed identification of footprints in the Mo+JS Mo-MuLV LTR, the relative protections and hypersensitivities observed were less intense than in our previous studies of cells infected with wild-type Mo-MuLV (16). One possible explanation could be that in the Mo+JS Mo-MuLV-infected MLE-15 cells, a significant percentage of proviruses might have been transcriptionally inactive. Such proviruses would not be expected to have footprints, and they would dilute the signals from active proviruses.
Transcription from the JSRV LTR in lung epithelial cells is dependent on NF-I and C/EBP binding sites. The in vivo footprinting indicated that in addition to HNF-3, factors are bound to the putative C/EBP and downstream NF-I binding sites in the JSRV sequences of the Mo+JS LTR in infected MLE-15 cells. To test whether these sites are important for activity of the native JSRV LTR in these cells, the downstream NF-I site (corresponding to residues –129 to –117) was deleted from a JSRV LTR-luciferase reporter plasmid (pJS21luc) and used in transient transfection assays (Fig. 6). In order to account for any distance or spatial effects, another plasmid was constructed, with the core NF-I binding region changed from TTGGCACTGCTTCAT to TTTTTACTGCCCCCC (mutated residues are in bold, and core sequences are underlined). Each plasmid was transfected into MLE-15 cells, and expression levels were compared to wild-type pJS21luc. As seen in Fig. 7A, LTR activity was reduced by 40% in MLE-15 cells when the downstream NF-I site was eliminated by nucleotide substitution, but there was no significant change in activity when the site was deleted. Thus, elimination of the putative NF-I binding site by itself had a relatively modest effect on JSRV LTR activity in MLE-15 cells. We similarly mutated or deleted (deletion of nucleotides –77 to –69; mutation of TGATTGTGTAAGAA to TGACCGTCTCCCAA [mutated residues are in bold, and core sequences are underlined]) the C/EBP site from pJS21luc (Fig. 6) and tested the activity in MLE-15 cells (Fig. 7A). Both of these mutants showed a 75% reduction in LTR activity, which indicated that the C/EBP site is important for JSRV LTR activity in MLE-15 cells, since elimination of this site by itself substantially reduced activity.
We studied the same NF-I and C/EBP mutations in the Clara cell-derived mtCC1-2 cells (Fig. 7B). Results similar to those for the MLE-15 cells were observed, indicating that the downstream NF-I site is of less importance and the C/EBP site of greater importance for JSRV LTR activity in these cells.
In light of reports describing cooperativity between HNF-3, NF-I, and C/EBP sites in the expression of some cellular genes (1, 5, 40), double mutations that eliminated any combination of the upstream HNF-3, downstream NF-I, or C/EBP sites were introduced into pJS21luc (Fig. 6). As reported previously (22), the upstream HNF-3 binding site is essential for transcriptional activation in MLE-15 but not mtCC1-2 cells. In MLE-15 cells, these double mutant reporter plasmids all exhibited a drastic reduction in LTR activity. In fact, the levels detected were essentially equal to reporter assay values observed in a vector carrying only the JSRV basal promoter [pJS21luc(–37)], indicating that the doubly mutated LTR enhancers were virtually nonfunctional (Fig. 7A). It should be noted that the LTR activities of reporter vectors containing single mutations in the C/EBP or HNF-3 sites were not statistically different from that of pJS21luc(–37).
As mentioned above, we previously reported that LTR activity of a reporter vector containing a mutated HNF-3 site is approximately equal to the wild-type LTR in mtCC1-2 (22) cells, and we showed here that single mutations in the NF-I site result in only a small decrease in LTR activity (Fig. 7B). Interestingly, reporter vectors with mutations in both the HNF-3 and NF-I sites showed extremely low activity, suggesting that JSRV expression is dependent on binding of either HNF-3 or NF-I (but not necessarily both) to the adjoining sites in Clara cells. All three pJS21luc double mutants exhibited reporter activities as low as or lower than pJS21luc(–37), indicating that the HNF-3, NF-I, and C/EBP sites are collectively critical for JSRV enhancer activity in mtCC1-2 cells.
For further study, we focused on the NF-I and C/EBP binding sites because of the footprint data described above and the results shown in Fig. 7. Also, previous reports indicated that isoforms of these factors contribute to tissue-specific gene expression in lung epithelial cells (2, 8, 9, 20, 26). However, several weak footprints were also detected in other consensus binding elements, as shown in Fig. 5. Therefore, we individually mutated each of those putative binding sites in the LTR reporter plasmid and measured the activities of the mutants. Although we previously reported that mutating the NF-B-like site reduced LTR activity by 50% (28), mutation of the adjacent E-box element did not significantly reduce enhancer activity. All other mutations in pJS21luc had little or no effect on reporter levels, with the exception that mutation in the putative SP-I site resulted in an approximately 50% reduction of activity in MLE-15 cells. A summary of all mutations in JSRV LTR reporter plasmids is shown in Table 2.
NF-IA and C/EBP/ bind to JSRV LTR sequences. NF-I is a transcription factor that plays a key role in activating the surfactant protein C promoter (2). Although NF-I expression can be detected in a wide variety of tissues, the differential expression of NF-I isoforms contributes to tissue-specific expression during and after development (10). There are at least four mammalian NF-I genes (A, B, C, and X), each of which is transcribed and further spliced to yield multiple isoforms (10, 38). NF-IA has been shown to transactivate the SP-C promoter in HeLa cells cotransfected with an NF-IA expression plasmid (2). In order to investigate whether NF-IA can bind to the JSRV downstream NF-I site, a series of EMSAs was performed using a labeled double-stranded oligonucleotide probe containing sequences from –134 to –115 of the JSRV LTR. As seen in Fig. 8A, a large protein-DNA complex was formed when this NF-I probe was incubated with nuclear extracts from MLE-15. A supershifted complex was observed when the nuclear extracts were preincubated with anti-NF-IA antiserum. This suggests that a factor (or complexes of factors) binds to the JSRV downstream NF-I site in these cells and that a component of this complex is NF-IA. Nuclear proteins from mtCC1-2 cells similarly complexed with the NF-I probe. The anti-NF-IA antiserum also supershifted the band, but with less efficiency than observed in the MLE-15 EMSA, suggesting that NF-IA is expressed in both cell lines but binds to this site in different ratios.
We have previously shown that the JSRV LTR drives transcription more efficiently in the lung epithelial cell lines MLE-15 and mtCC1-2 than in nonlung epithelial cell lines, such as the mouse fibroblast cell line NIH 3T3 (28). The poor levels of LTR activation in nonlung cells can be partly explained by the lack of HNF-3 and, potentially, cooperating coactivators in these cells (22, 28). In addition to the presence of tissue-specific transcriptional activators, certain isoforms of widely expressed transcription factors also facilitate gene expression in specific cell types. In order to compare the isoform expression of a ubiquitously expressed transcription factor (NF-I) in nonlung epithelial cells with that present in the lung cell lines MLE-15 and mtCC1-2, gel shift assays were performed using NIH 3T3 nuclear extracts. Although this gel contained a complex with a migration pattern similar to the MLE-15 EMSA, the anti-NF-IA antiserum did not supershift this band. Therefore, NIH 3T3 cells do not express NF-IA isoforms that can bind to the JSRV downstream NF-I site, but they express other NF-I isoform(s) or other transcription factors that can bind to this site.
C/EBP has three main isoforms (, , and ) that are expressed throughout the lung (6, 39). Isoforms and are highly expressed in the bronchiolar epithelium, while is mostly found in alveolar macrophages (39). Because C/EBP binds as a dimer (or heterodimer), one or more isoforms may bind to a C/EBP site. Although C/EBP is considered to be more important for expression in bronchiolar epithelium (7-9, 39), heterodimers of C/EBP and have been reported to work synergistically on some lung-specific promoters (8, 9). To identify the C/EBP isoform(s) that can bind to the JSRV element, a series of EMSAs was performed in the presence of isoform-specific anti-C/EBP antibodies (Fig. 8B). As expected, incubation of a probe encompassing the JSRV C/EBP site with nuclear extracts from MLE-15 or mtCC1-2 cells produced an EMSA pattern composed of at least three major bands. Addition of anti-C/EBP antiserum supershifted the slowest-migrating band, but no change in intensity was observed in the other complexes. Anti-C/EBP antiserum diminished the intensity of the faster-migrating complexes and caused a faint supershift. Interestingly, antiserum against C/EBP had no effect on any complex, even though this isoform had the highest transactivation potential in NIH 3T3 cells cotransfected with pJS21luc and C/EBP, , or expression plasmids (data not shown). Thus, it appears that C/EBP and C/EBP, but not C/EBP, are capable of binding to the JSRV C/EBP site in the lung epithelial-derived cell lines MLE-15 and mtCC1-2.
DISCUSSION
In this study, our goal was to identify transcription factors and binding sites that are responsible for the lung epithelial cell specificity of the JSRV LTR. Since there are multiple putative factor binding sites in the LTR, our approach was first to identify sites that show evidence of occupancy by in vivo footprinting (25). We first carried out in vivo footprinting of the JSRV enhancer in a mouse lung epithelial cell line (MLE-15) that supports a high level of transcriptional activity from the JSRV LTR (28) by studying MLE-15 cells infected with a chimeric Mo-MuLV (Mo+JS) containing the JSRV enhancers (23). The in vivo footprinting was consistent with our previous in vitro studies implicating HNF-3 in JSRV LTR activity (22), since a footprint was found in the JSRV upstream HNF-3 site in the Mo+JS LTR in the infected cells. In addition, footprints were found in other factor binding sites in the Mo+JS LTR, which suggested that factors binding to these sites might also be important for JSRV LTR activity in lung epithelial cells. This was addressed by mutating or deleting these sites in the context of the native JSRV LTR, followed by assessing LTR activity in transient transfection reporter assays.
In vivo footprints in the JSRV C/EBP and downstream NF-I binding sites were also found in the Mo+JS Mo-MuLV-infected cells. Mutation or deletion of the downstream NF-I site alone had relatively little effect on activity of the native JSRV LTR in MLE-15 cells (Fig. 7). This site is directly adjacent to the upstream (distal) HNF-3 binding site; we previously showed that the upstream HNF-3 site is important for JSRV LTR activity in MLE-15 cells. The close placement of these HNF-3 and NF-I sites suggests that these two factors may cooperate, but experiments to address this possibility have not yet indicated a synergistic interaction between the two sites. It is interesting that the enhancer of the liver-specific albumin gene also has closely positioned HNF-3 and NF-I sites (18). It appears that the two factors interact on the albumin enhancer, although the exact nature of this cooperation is unknown. EMSAs shown in Fig. 8A indicated that at least the NF-IA isoform is expressed in MLE-15 and mtCC1-2 cells. Antibodies against other NF-I isoforms suitable for EMSA supershifts were not available. Increasing evidence suggests that tissue-specific expression of genes is facilitated not only by binding of factors that are exclusively expressed in specific cell types but also by a complementing array of specific isoforms of ubiquitously expressed transcription factors. NF-IA is enriched in lung epithelial cells, and it likely has a critical role in activating the surfactant protein C promoter in type II pneumocytes (2). Similarly, NF-IA may enable high-level JSRV LTR transcriptional activation more efficiently than other NF-I isoforms.
Mutations or deletions of the C/EBP site in the JSRV LTR (downstream of the paired HNF-3 and NF-I sites) had a pronounced effect on transcriptional activity in MLE-15 or mtCC1-2 cells (>75% reduction). Thus, the C/EBP site appears to be important for LTR activity in both of these lung epithelial cell types. EMSAs indicated that C/EBP and C/EBP but not C/EBP are expressed in MLE-15 cells. This was somewhat surprising, since C/EBP has been reported to be important for the expression of some genes in lung epithelial cells (8, 9). Isoforms of C/EBP are expressed ubiquitously, but expression is highest in cells involved in fat metabolism (fat and liver cells and type II pneumocytes) and terminally differentiated cells (6, 7, 39). The C/EBP binding site is missing from the LTRs of endogenous JSRV-related proviruses in the sheep genome, which are not expressed in the lung (3). Thus, activation of the JSRV LTR by C/EBP parallels the lung-specific expression of the exogenous JSRV.
The in vivo footprinting of the Mo+JS LTR in infected MLE-15 cells also identified a footprint near an NF-B-like site (–260 to –248) in the distal enhancer occupied by an unknown protein. In transient transfection assays of JSRV LTR truncations, LTR activity decreased by 50% when nucleotides adjacent to the footprint were mutated (28). Although sequences in this region match a canonical NF-B binding site, an EMSA with an anti-NF-B antibody did not result in a supershift. The NF-B-like site overlaps with an E-box element that can potentially bind isoforms of N-myc or upstream stimulating factor (USF). USF was recently found to bind to and activate the surfactant protein A enhancer in fetal rabbit lung type II cells (13, 14). Although USF is expressed ubiquitously, the same study showed that USF 2 was enriched in type II cells compared to lung fibroblasts, suggesting that this isoform may be important for type II-specific gene expression. However, selective mutation of the E-box element in the JSRV LTR did not significantly affect transcriptional activity, although USF 1 and 2 were found to bind to this site in EMSAs (data not shown).
Finally, it should be mentioned that the factors involved in JSRV LTR expression in type II pneumocytes may differ from those important in Clara cells. Alteration of the C/EBP site diminished LTR activity in both MLE-15 and mtCC1-2 cells, indicating that a C/EBP isoform is important for expression in both of these cell types. However, mutation or deletion of the upstream HNF-3 site by itself had relatively little effect on JSRV LTR activity in Clara-derived mtCC1-2 cells, in contrast to the substantial loss of activity in type II-derived MLE-15 cells (22). As suggested in Fig. 7, one explanation could be that the HNF-3 and NF-I sites are functionally redundant in mtCC1-2 cells. Whereas the JSRV LTR may utilize either HNF-3 or NF-I for efficient expression in mtCC1-2 cells, NF-I may not effectively substitute for HNF-3 in MLE-15 cells. Alternatively, some other factor(s) also might be important for JSRV LTR activity in Clara cells. We have not yet identified a factor that is important for JSRV expression in mtCC1-2 but not MLE-15 cells. Although supershift assays indicate that MLE-15 and mtCC1-2 express identical NF-I and C/EBP isoforms (Fig. 8), it is important to note that in comparing the migration patterns between the two nuclear extracts, these cells might express different splice variants and/or contain isoforms at different ratios. In vivo footprinting of Mo+JS Mo-MuLV-infected mtCC1-2 cells might identify binding sites for other factors important for JSRV LTR expression in these cells, and such experiments are in progress.
ACKNOWLEDGMENTS
This work was supported by grant R01 CA82654 from the National Cancer Institute. K.M.-E. was supported by training grant T32 CA09054. Support of the UCI Cancer Research Institute and the DNA sequencing core of the Chao Family Comprehensive Cancer Center is acknowledged.
REFERENCES
Antes, T. J., and B. Levy-Wilson. 2001. HNF-3 beta, C/EBP beta, and HNF-4 act in synergy to enhance transcription of the human apolipoprotein B gene in intestinal cells. DNA Cell Biol. 20:67-74.
Bachurski, C. J., S. E. Kelly, S. W. Glasser, and T. A. Currier. 1997. Nuclear factor I family members regulate the transcription of surfactant protein-C. J. Biol. Chem. 272:32759-32766.
Bai, J., R. Y. Zhu, K. Stedman, C. Cousens, J. Carlson, J. M. Sharp, and J. C. DeMartini. 1996. Unique long terminal repeat U3 sequences distinguish exogenous jaagsiekte sheep retroviruses associated with ovine pulmonary carcinoma from endogenous loci in the sheep genome. J. Virol. 70:3159-3168.
Barsky, S. H., R. Cameron, K. E. Osann, D. Tomita, and E. C. Holmes. 1994. Rising incidence of bronchioloalveolar lung carcinoma and its unique clinicopathologic features. Cancer 73:1163-1170.
Beurton, F., U. Bandyopadhyay, B. Dieumegard, R. Barouki, and M. Aggerbeck. 1999. Delineation of the insulin-responsive sequence in the rat cytosolic aspartate aminotransferase gene: binding sites for hepatocyte nuclear factor-3 and nuclear factor I. Biochem. J. 343:687-695.
Birkenmeier, E. H., B. Gwynn, S. Howard, J. Jerry, J. I. Gordon, W. H. Landschulz, and S. L. McKnight. 1989. Tissue-specific expression, developmental regulation, and genetic mapping of the gene encoding CCAAT/enhancer binding protein. Genes Dev. 3:1146-1156.
Breed, D. R., L. R. Margraf, J. L. Alcorn, and C. R. Mendelson. 1997. Transcription factor C/EBPdelta in fetal lung: developmental regulation and effects of cyclic adenosine 3',5'-monophosphate and glucocorticoids. Endocrinology 138:5527-5534.
Cassel, T. N., J. A. Gustafsson, and M. Nord. 2000. CYP2B1 is regulated by C/EBP alpha and C/EBP delta in lung epithelial cells. Mol. Cell Biol. Res. Commun. 3:42-47.
Cassel, T. N., L. Nordlund-Moller, O. Andersson, J. A. Gustafsson, and M. Nord. 2000. C/EBPalpha and C/EBPdelta activate the Clara cell secretory protein gene through interaction with two adjacent C/EBP-binding sites. Am. J. Respir. Cell Mol. Biol. 22:469-480.
Chaudhry, A. Z., G. E. Lyons, and R. M. Gronostajski. 1997. Expression patterns of the four nuclear factor I genes during mouse embryogenesis indicate a potential role in development. Dev. Dyn. 208:313-325.
DeMartini, J. C., and D. F. York. 1997. Retrovirus-associated neoplasms of the respiratory system of sheep and goats. Ovine pulmonary carcinoma and enzootic nasal tumor. Vet. Clin. N. Am. Food Anim. Pract. 13:55-70.
Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11:1475-1489.
Gao, E., Y. Wang, J. L. Alcorn, and C. R. Mendelson. 1997. The basic helix-loop-helix-zipper transcription factor USF1 regulates expression of the surfactant protein-A gene. J. Biol. Chem. 272:23398-23406.
Gao, E., Y. Wang, J. L. Alcorn, and C. R. Mendelson. 2003. Transcription factor USF2 is developmentally regulated in fetal lung and acts together with USF1 to induce SP-A gene expression. Am. J. Physiol. Lung Cell Mol. Physiol. 284:L1027-L1036.
Garrity, P. A., and B. J. Wold. 1992. Effects of different DNA polymerases in ligation-mediated PCR: enhanced genomic sequencing and in vivo footprinting. Proc. Natl. Acad. Sci. USA 89:1021-1025.
Granger, S. W., and H. Fan. 1998. In vivo footprinting of the enhancer sequences in the upstream long terminal repeat of Moloney murine leukemia virus: differential binding of nuclear factors in different cell types. J. Virol. 72:8961-8970.
Holland, M. J., M. Palmarini, M. Garcia-Goti, L. Gonzalez, I. McKendrick, M. de las Heras, and J. M. Sharp. 1999. Jaagsiekte retrovirus is widely distributed both in T and B lymphocytes and in mononuclear phagocytes of sheep with naturally and experimentally acquired pulmonary adenomatosis. J. Virol. 73:4004-4008.
Jackson, D. A., K. E. Rowader, K. Stevens, C. Jiang, P. Milos, and K. S. Zaret. 1993. Modulation of liver-specific transcription by interactions between hepatocyte nuclear factor 3 and nuclear factor 1 binding DNA in close apposition. Mol. Cell. Biol. 13:2401-2410.
Landis, S. H., T. Murray, S. Bolden, and P. A. Wingo. 1998. Cancer statistics, 1998. CA Cancer J. Clin. 48:6-29.
Li, F., E. Rosenberg, C. I. Smith, K. Notarfrancesco, S. R. Reisher, H. Shuman, and S. I. Feinstein. 1995. Correlation of expression of transcription factor C/EBP alpha and surfactant protein genes in lung cells. Am. J. Physiol. 269:L241-L247.
Mason, R. J., and J. M. Shannon. 1997. Alveolar type II pneumocytes, p. 543-555. In J. B. West, R. G. Crystal, E. R. Weibel, and P. J. Barnes (ed.), The lung: scientific foundations, vol. 1. Lippincott-Raven, Philadelphia, Pa.
McGee-Estrada, K., M. Palmarini, and H. Fan. 2002. HNF-3beta is a critical factor for the expression of the Jaagsiekte sheep retrovirus long terminal repeat in type II pneumocytes but not in Clara cells. Virology 292:87-97.
McGee-Estrada, K., M. Palmarini, C. Hallwirth, and H. Fan. 2005. The Jaagsiekte sheep retrovirus enhancer increases Moloney murine leukemia virus genomic expression in lung epithelial cells. Virus Genes 31:251-263.
Mendelson, C. R. 2000. Role of transcription factors in fetal lung development and surfactant protein gene expression. Annu. Rev. Physiol. 62:875-915.
Mueller, P. R., and B. Wold. 1989. In vivo footprinting of a muscle specific enhancer by ligation mediated PCR. Science 246:780-786.
Nord, M., M. Lag, T. N. Cassel, M. Randmark, R. Becher, H. J. Barnes, P. E. Schwarze, J. A. Gustafsson, and J. Lund. 1998. Regulation of CCSP (PCB-BP/uteroglobin) expression in primary cultures of lung cells: involvement of C/EBP. DNA Cell Biol. 17:481-492.
Palmarini, M., C. Cousens, R. G. Dalziel, J. Bai, K. Stedman, J. C. DeMartini, and J. M. Sharp. 1996. The exogenous form of Jaagsiekte retrovirus is specifically associated with a contagious lung cancer of sheep. J. Virol. 70:1618-1623.
Palmarini, M., S. Datta, R. Omid, C. Murgia, and H. Fan. 2000. The long terminal repeat of Jaagsiekte sheep retrovirus is preferentially active in differentiated epithelial cells of the lungs. J. Virol. 74:5776-5787.
Palmarini, M., P. Dewar, M. De las Heras, N. F. Inglis, R. G. Dalziel, and J. M. Sharp. 1995. Epithelial tumour cells in the lungs of sheep with pulmonary adenomatosis are major sites of replication for Jaagsiekte retrovirus. J. Gen. Virol. 76:2731-2737.
Palmarini, M., and H. Fan. 2001. Retrovirus-induced ovine pulmonary adenocarcinoma, an animal model for lung cancer. J. Natl. Cancer Inst. 93:1603-1614.
Palmarini, M., M. J. Holland, C. Cousens, R. G. Dalziel, and J. M. Sharp. 1996. Jaagsiekte retrovirus establishes a disseminated infection of the lymphoid tissues of sheep affected by pulmonary adenomatosis. J. Gen. Virol. 77:2991-2998.
Palmarini, M., J. M. Sharp, M. de las Heras, and H. Fan. 1999. Jaagsiekte sheep retrovirus is necessary and sufficient to induce a contagious lung cancer in sheep. J. Virol. 73:6964-6972.
Palmarini, M., J. M. Sharp, C. Lee, and H. Fan. 1999. In vitro infection of ovine cell lines by Jaagsiekte sheep retrovirus. J. Virol. 73:10070-10078.
Perk, K., and I. Hod. 1982. Sheep lung carcinoma: an endemic analogue of a sporadic human neoplasm. J. Natl. Cancer Inst. 69:747-749.
Pfeifer, G. P., S. D. Steigerwald, P. R. Mueller, B. Wold, and A. D. Riggs. 1989. Genomic sequencing and methylation analysis by ligation mediated PCR. Science 246:810-813.
Plopper, C. G., M. H. Dallas, and A. R. Buckpitt. 1997. Clara cells, vol. 1. Lippincott-Raven, Philadelphia, Pa.
Rai, S. K., F. M. Duh, V. Vigdorovich, A. Danilkovitch-Miagkova, M. I. Lerman, and A. D. Miller. 2001. Candidate tumor suppressor HYAL2 is a glycosylphosphatidylinositol (GPI)-anchored cell-surface receptor for jaagsiekte sheep retrovirus, the envelope protein of which mediates oncogenic transformation. Proc. Natl. Acad. Sci. USA 98:4443-4448.
Rupp, R. A., U. Kruse, G. Multhaup, U. Gobel, K. Beyreuther, and A. E. Sippel. 1990. Chicken NFI/TGGCA proteins are encoded by at least three independent genes: NFI-A, NFI-B and NFI-C with homologues in mammalian genomes. Nucleic Acids Res. 18:2607-2616.
Sugahara, K., T. Sadohara, M. Sugita, K. Iyama, and M. Takiguchi. 1999. Differential expression of CCAAT enhancer binding protein family in rat alveolar epithelial cell proliferation and in acute lung injury. Cell Tissue Res. 297:261-270.
Tsay, W., Y. M. Lee, S. C. Lee, M. C. Shen, and P. J. Chen. 1997. Synergistic transactivation of HNF-1alpha, HNF-3, and NF-I contributes to the activation of the liver-specific protein C gene. DNA Cell Biol. 16:569-577.
Voelker, D. R., and R. J. Mason. 1989. Alveolar type II epithelial cells. Marcel Dekker, New York, N.Y.
Whitsett, J. A., and S. W. Glasser. 1998. Regulation of surfactant protein gene transcription. Biochim. Biophys. Acta 1408:303-311.
Wikenheiser, K. A., D. K. Vorbroker, W. R. Rice, J. C. Clark, C. J. Bachurski, H. K. Oie, and J. A. Whitsett. 1993. Production of immortalized distal respiratory epithelial cell lines from surfactant protein C/simian virus 40 large tumor antigen transgenic mice. Proc. Natl. Acad. Sci. USA 90:11029-11033.
Wingender, E., X. Chen, R. Hehl, H. Karas, I. Liebich, V. Matys, T. Meinhardt, M. Pruss, I. Reuter, and F. Schacherer. 2000. TRANSFAC: an integrated system for gene expression regulation. Nucleic Acids Res. 28:316-319.
York, D. F., R. Vigne, D. W. Verwoerd, and G. Querat. 1992. Nucleotide sequence of the jaagsiekte retrovirus, an exogenous and endogenous type D and B retrovirus of sheep and goats. J. Virol. 66:4930-4939.(Kathleen McGee-Estrada an)
ABSTRACT
Jaagsiekte sheep retrovirus (JSRV) is the causative agent of ovine pulmonary adenocarcinoma, a contagious lung cancer of sheep that arises from type II pneumocytes and Clara cells of the lung epithelium. Studies of the tropism of this virus have been hindered by the lack of an efficient system for viral replication in tissue culture. To map regulatory regions important for transcriptional activation, an in vivo footprinting method that couples dimethyl sulfate treatment and ligation-mediated PCR was performed in murine type II pneumocyte-derived MLE-15 cells infected with a chimeric Moloney murine leukemia virus driven by the JSRV enhancers (Mo+JS Mo-MuLV). In vivo footprints were found in the JSRV enhancers in two regions previously shown to be important for JSRV long terminal repeat (LTR) activity: a binding site for the lung-specific transcription factor HNF-3 and an E-box element in the distal enhancer adjacent to an NF-B-like binding site. In addition, in vivo footprints were detected in two downstream motifs likely to bind C/EBP and NF-I. Mutational analysis of a JSRV LTR reporter construct (pJS21luc) revealed that the C/EBP binding site is critical for LTR activity, while the putative NF-I binding element is less important; elimination of these sites resulted in 70% and 40% drops in LTR activity, respectively. Electrophoretic mobility shift assays using nuclear extracts from MLE-15 murine Clara cell-derived mtCC1-2 cells with probes corresponding to the NF-I or C/EBP sites revealed several complexes. Antiserum directed against NF-IA, C/EBP, or C/EBP supershifted the corresponding protein-DNA complexes, indicating that these isoforms, which are also important for the expression of several cellular lung-specific genes, may be important for JSRV expression in lung epithelial cells.
INTRODUCTION
Jaagsiekte sheep retrovirus (JSRV) is the etiological agent of ovine pulmonary adenocarcinoma (3, 11, 27, 30, 32, 45), a contagious lung cancer of sheep and goats that closely resembles human adenocarcinoma of the deep airways, including bronchioloalveolar carcinoma (34). Lung cancer is the leading cause of human cancer deaths worldwide (19); the incidence of human lung adenocarcinoma has steadily risen, and this neoplasm is now responsible for up to one quarter of lung cancer fatalities (4, 19). Thus, ovine pulmonary adenocarcinoma presents a unique opportunity to study the mechanisms of lung carcinogenesis in an animal model.
Retroviruses have historically given insights into the mechanisms of oncogenesis. However, most retroviruses induce neoplasms of hematopoietic cells, whereas the great majority of human cancers are carcinomas (originating from epithelial tissues). JSRV induces tumors of differentiated lung epithelial cells: type II pneumocytes (21, 41) and Clara cells (36). Both in vivo and in vitro, JSRV can infect several different cell types, as judged by the presence of JSRV DNA (17, 31, 33). However, high-level virus expression is observed only in tumor cells in the lung (28, 29). This highly restricted productive infection in vivo makes JSRV quite unique among retroviruses. We have previously reported that the JSRV tropism for lung epithelial cells is correlated with the transcriptional specificity of the viral long terminal repeat (LTR) for these cells (28). Retroviral LTRs contain the viral promoter and enhancer elements, and numerous studies with other retroviruses have indicated the importance of the LTR enhancers in determining pathogenicity and disease specificity.
Studies of cellular genes specifically expressed in lung epithelial cells have identified several important transcription factors. Both type II pneumocytes and Clara cells specifically express several proteins found in surfactant, including SP-A, SP-B, SP-C, SP-D, and the Clara cell secretory protein (CCSP). SP-C and CCSP are expressed exclusively in type II pneumocytes and Clara cells, respectively (21, 36, 41). The upstream promoter and enhancer regions of these genes have revealed several common regulatory elements (24, 42). Members of the winged helix family of transcription factors, such as HNF-3, HFH-4, and HFH-8 are expressed throughout lung development, and binding sites for these factors are found on several surfactant promoters. Other transcription factors important for surfactant gene expression include the homeodomain protein TTF-1, USF, NF-I, Ets-I, SP-I, and C/EBP (24, 42). In searching the JSRV LTR U3 sequences, we found that the JSRV LTR has one or more putative binding sites for all of these factors except for TTF-I (Fig. 1) (28). In further studies, we showed that the lung- and liver-specific transcription factor HNF-3 is important for JSRV LTR activity in the type II pneumocyte-derived cell line MLE-15 (22, 28). In this study, we explored the roles of other transcription factors in JSRV LTR activity in MLE-15 cells and Clara cell-derived mtCC1-2 cells by a combination of in vivo footprinting, site-specific mutations, and electrophoretic mobility shift assay (EMSA) supershifts.
MATERIALS AND METHODS
Cell lines. MLE-15 cells (provided by J. Whitsett, Cincinnati, OH) were grown in RPMI 1640 (Gibco BRL) containing 2% fetal bovine serum, 0.5% insulin-transferrin-sodium selenite (Sigma), 5 mg/liter transferrin, 10 mM HEPES, 10–8 M -estradiol, and 10–8 M hydrocortisone. NIH 3T3 (American Type Culture Collection) and mtCC1-2 cells (provided by F. DeMayo, Baylor College of Medicine, Houston, TX) were grown in Dulbecco's minimal essential medium (American Type Culture Collection) containing 10% fetal bovine serum. All three cell types were cultured at 37°C with 5% CO2.
In vivo DMS treatment of Mo+JS Mo-MuLV-infected cells. To achieve partial methylation of guanines in vivo, two 15-cm-diameter plates of subconfluent MLE-15 cultures infected with Mo+JS Mo-MuLV (a chimeric Moloney murine leukemia virus driven by the JSRV enhancers) were treated with 1% dimethyl sulfate (DMS) in growth medium at 37°C for 2 min. The treatment was stopped by aspiration of the DMS-containing medium followed by an immediate rinse with 25 ml of phosphate-buffered saline (PBS) prewarmed to 37°C and two successive 30-s washes with 25 ml of PBS at 37°C. Genomic DNA was harvested by the addition of 3 ml of cell lysis solution (300 mM NaCl, 50 mM Tris-Cl [pH 8.0], 25 mM EDTA [pH 8.0], 0.2% [vol/vol] sodium dodecyl sulfate, 0.2 mg of proteinase K per ml); this mixture was allowed to incubate for 5 min before the cell slurry was removed into a test tube by gentle scraping. As a control, genomic DNA was also harvested from untreated infected cells in parallel. Both samples were incubated at 37°C from 4 h to overnight, after which the genomic DNA was phenol-chloroform extracted and ethanol precipitated.
In vitro DMS treatment of DNA. Extracted genomic DNA from control cultures was subjected to DMS treatment in vitro by incubation with 1% DMS in H2O for 1 min at 25°C. The reaction was stopped by the addition of ice-cold DMS stop buffer (1.5 M sodium acetate [pH 7.0], 1 M -mercaptoethanol, 100 μg of Saccharomyces cerevisiae tRNA per ml), which was immediately followed by the addition of 2.5 volumes of ethanol on dry ice. Samples were precipitated by incubation for at least 30 min at –70°C and pelleted in a microcentrifuge for 15 min at 4°C. DNA pellets were allowed to air dry for 10 min and resuspended in 200 μl of 1 M piperidine in H2O for 15 min at room temperature prior to cleavage.
Piperidine cleavage. Following in vivo or in vitro DMS treatment, extracted genomic DNA from each cell type was cleaved at all methylated guanines by incubation in 200 μl of 1 M piperidine for 30 min at 90°C. The piperidine was removed by lyophilization, and the cleaved DNA pellets were resuspended in 360 μl of TE buffer (10 mM Tris-Cl, 1 mM EDTA; pH 7.5). Residual piperidine was removed by two successive ethanol precipitations. The first entailed addition of 40 μl of 3 M sodium acetate followed by 1 ml of 100% ethanol and incubation for 30 min at –70°C. DNA samples were pelleted in a microcentrifuge for 15 min at 4°C and resuspended in 500 μl of TE buffer. The DNA pellets were ethanol precipitated a second time by the addition of 170 μl of 8 M ammonium acetate and 670 μl of isopropanol and incubation for at least 30 min at –70°C. The precipitated samples were pelleted by microcentrifugation as described above, washed with 500 μl of 75% ethanol, and repelleted for 5 min at room temperature. The resulting DNA pellets were resuspended in double-distilled water to a final concentration of 0.4 μg/μl.
LMPCR. Two micrograms of DMS-treated and piperidine-cleaved genomic DNA was used for ligation-mediated PCR (LMPCR) as described previously (15, 16, 25, 35) with minor modifications. Single-stranded DNA fragments with guanine residues at both termini result from the DMS treatment and piperidine cleavage. To provide appropriate substrates for linker ligation, double-stranded, blunt-ended molecules were generated by primer extension from an Mo-MuLV- or JSRV-specific oligonucleotide (Table 1 and Fig. 2D). This first-strand primer extension was accomplished by incubation of 2 μg of DMS-treated and piperidine-cleaved DNA with 0.3 pmol of primer 1 (Mol 2E, Mol 1A, and JSRV 1B) in 30 μl of 1x Vent DNA polymerase buffer (New England Biolabs) containing 4 mM MgSO4, 0.25 mM concentrations of each deoxynucleoside triphosphate, and 0.5 U of Vent DNA polymerase (New England Biolabs). The DNA was denatured at 95°C for 5 min, annealed by incubation at 55°C for 20 min, and extended by a subsequent incubation for 10 min at 72°C. Ligation of the unidirectional linker was completed by the addition of 20 μl of 110 mM Tris-Cl (pH 7.5)-17.5 mM MgCl2-50 mM dithiothreitol, 25 μl of 10 mM MgCl2-20 mM dithiothreitol-3 mM ATP (pH 7.0)-4 μM unidirectional linker (in 50 mM Tris-Cl [pH 7.7]), and 3 U of T4 DNA ligase (Gibco/BRL). The unidirectional linker oligonucleotide sequences have been described by Mueller and Wold (25) and are as follows: LMPCR.1, 5'-GCGGTGACCCGGGAGATCTGAATTC-3', and LMPCR.2, 5'-GAATTCAGATC-3'. This mixture was incubated at 17°C overnight, after which the DNA was recovered by ethanol precipitation. The precipitated DNA pellet was resuspended in 50 μl of H2O, and PCR amplification was accomplished by the addition of 50 μl of 2x Vent buffer containing 8 mM MgSO4, 5 mM deoxynucleoside triphosphate mix, 1 pmol of Mo-MuLV or JSRV oligonucleotide 2 (Mol 2F, Mol 2A, JSRV 2D, and JSRV 2B) (Table 1 and Fig. 2D), 1 pmol of oligonucleotide LMPCR.1, and 1 U of Vent DNA polymerase. These samples were placed in a thermocycler and cycled 17 times with a profile of 95°C for 1 min, 66°C for 2 min, and 72°C for 1 min, with a final extension of 10 min at 72°C. Following amplification, PCR products were labeled by the addition of 5 μl of labeling buffer (2 mM concentrations of each deoxynucleoside triphosphate, 1x Vent polymerase buffer, 8 mM MgSO4, 1 U of Vent polymerase, 2.3 pmol of an Mo-MuLV- or JSRV-specific 32P-end-labeled oligonucleotide [oligonucleotide 3 {Mol 3F, Mol 3A, JSRV 3B, and JSRV 3D}; Table 1 and Fig. 2D]) and subjected to two rounds of 95°C for 1 min, 69°C for 2 min, and 72°C for 1 min. Each reaction mixture was then subjected to phenol-chloroform extraction and ethanol precipitation prior to electrophoresis on a 6% polyacrylamide sequencing gel. The reactions were visualized by autoradiography with Kodak BioMax MR film and also by phosphorimaging on a Molecular Dynamics 445 SI PhosphorImager.
Plasmids. pJS21luc was constructed by cloning the JSRV21 LTR directly upstream of the firefly luciferase gene in pGL3basic, as previously described (28). PCRs were performed using the PfuTurbo polymerase (Stratagene) as recommended by the manufacturers.
Deletions and alterations of the NF-I and C/EBP sites were generated by PCR mutagenesis (for schematic diagrams of each mutant, see Fig. 6).
All constructs were verified by nucleotide sequencing and/or restriction digestion.
Transient transfection and luciferase assays. Transient transfections were performed on 2 x 105 to 4 x 105 cells plated on six-well plates (Falcon) approximately 24 h before transfection. For each well, 700 ng of plasmid DNA (500 ng of reporter plasmid and 200 ng of pRLnull to ascertain transfection efficiency) and 6 μl of fugene (Roche) were used as recommended by the manufacturers. Experiments were performed using the dual luciferase reporter system (Promega). Activity of the reporter plasmids was calculated as the percent activity with respect to the wild-type pJS21luc plasmid.
After 48 h, transfected cells were washed with PBS and lysed with 500 μl/well of passive lysis buffer (Promega). Dual luciferase assays were performed on 20 μl of cleared lysate by rapid addition of luciferase assay reagent (Promega), and light output was integrated over 10 seconds at room temperature using a Moonlight 2010 luminometer (Analytical Luminescence Laboratory). Luciferase activity was normalized for transfection efficiency and cell extract preparation by sequentially measuring Renilla luciferase activity.
All of the experiments described above were performed in at least 12 independent transfections using two separate DNA preparations. Results were expressed as means ± standard errors (95% confidence interval). The linear range of the reaction was predetermined.
Nuclear extracts, oligonucleotides, and EMSA. Nuclear extracts were prepared from MLE-15, mtCC1-2, and NIH 3T3 cell lines as described previously (12). Protein concentrations were determined by Bradford assay (Bio-Rad).
For the electrophoretic mobility shift assays, the following double-stranded oligonucleotide probes were used: NFIdown EMSA (GCTTTTGGCACTGCTTCATAGAAA) and C/EBP EMSA (GGTGATTGTGTAAGAATCCG), corresponding to canonical NF-I and C/EBP binding sites at positions –135 to –112 and –82 to –63 of the U3 of JSRV21, respectively. Underlined nucleotides indicate core binding sequences for the respective transcription factor.
Oligonucleotide probes were end labeled with [-32P]ATP using T4 polynucleotide kinase. All binding reactions were performed at room temperature using buffers purchased from Geneka and under conditions recommended by the manufacturer. Nuclear extracts were incubated with the probe for 30 min before electrophoresis. Rabbit antisera raised against mouse NF-IA and C/EBP// were purchased from Geneka or Santa Cruz Biotechnologies, respectively. For supershift assays, 1 μl of normal serum or antibody was preincubated with nuclear extracts for 30 min. The anti-C/EBP antisera produced supershifts in control EMSAs using nuclear extracts from 293T cells transiently transfected with a C/EBP, , or expression plasmid. In competition gel shifts, the probe and cold competitor oligonucleotide were premixed prior to incubation with nuclear extracts. Each binding reaction was subjected to electrophoresis through a 5% nondenaturing polyacrylamide gel in 1x Tris-glycine-EDTA (NF-I) or 0.5x Tris-borate-EDTA (C/EBP) buffer. After 2 h, gels were dried and exposed to film at –70°C for approximately 5 h.
RESULTS
In vivo footprinting of the Mo+JS LTR in infected MLE-15 cells. Cells of the bronchiolar epithelium quickly lose their differentiated characteristics when cultured in vitro, and no ovine cell lines that maintain these properties exist. As a result, an efficient ovine tissue culture system for propagation of JSRV has not yet been established, although low-level productive infection is observed in most ovine cell lines (33). In the past, we have made use of the mouse type II pneumocyte-derived cell line MLE-15; this line was derived from lung tumors of transgenic mice expressing simian virus 40 large T antigen from the SP-C promoter (43). These cells maintain a differentiated phenotype through several passages and continue to express several type II pneumocyte markers, including surfactant proteins and HNF-3. We found that the JSRV LTR has a high relative level of activity in MLE-15 cells (28). However, MLE-15 cells cannot be infected by JSRV, since murine cells lack a functional JSRV receptor (37).
To further study the role of the JSRV LTR in determining lung tropism, a Moloney murine leukemia virus (Mo-MuLV) driven by a chimeric LTR, in which the Mo-MuLV U3 enhancer sequences (–342 to –150) were replaced with a large portion of the JSRV U3 region, was constructed (Fig. 2A) (23). This virus, termed Mo+JS Mo-MuLV, carries bases –266 to –37 of the JSRV LTR in an otherwise intact Mo-MuLV genome (Fig. 2B and C). We found that like wild-type Mo-MuLV, Mo+JS Mo-MuLV is able to infect most mouse cell lines, including MLE-15 (23). In transient transfection assays, the Mo+JS LTR was three- to fivefold more specific for lung epithelial cells than the wild-type Mo-MuLV LTR. Moreover, in infectivity assays, Mo+JS Mo-MuLV was 3 to 4 logs more specific than wild-type Mo-MuLV for MLE-15 infection versus nonlung epithelial cell lines (23). Mo+JS Mo-MuLV-infected MLE-15 cells therefore contain active proviruses in which transcription is driven by the JSRV LTR enhancers. As such, they provide a system for studying factors bound to the JSRV LTR in vivo.
We previously have made use of an in vivo footprinting method to investigate factor binding to the Mo-MuLV LTR in different cell types (16, 25). This method involves the limited in vivo methylation of guanine residues in the DNA of intact infected cells by treatment with DMS followed by extraction of the DNA and cleavage at methylated G residues with piperidine. A guanine residue bound by a factor may be protected from methylation and subsequent cleavage. Guanine residues also may be hypermethylated if the DNA is bent or if factor binding creates a hydrophobic pocket, and this makes the residues hypersensitive to cleavage. For comparison, genomic DNA from the same cells is first extracted and then methylated in vitro and analyzed in parallel. The region of interest is then selectively amplified and radiolabeled by LMPCR using nested strand-specific primers. The end products are resolved in a sequencing gel and analyzed in a phosphorimager. The resulting pattern is a "G-ladder" for the region adjacent to the PCR primers, and the relative intensities of each labeled band reflect the efficiency of methylation at that guanine residue. Comparison of the LMPCR reactions from in vivo- and in vitro-methylated DNA thus provides an indication of in vivo factor binding to those sequences. Bases that are hypersensitive to cleavage will have a more intense in vivo image than the relative band strength of the in vitro-treated DNA. Protected bases will likewise have a less intense corresponding band on the gel than in the control in vitro-treated DNA.
To map the JSRV regulatory regions important for Mo+JS Mo-MuLV transcriptional activity, infected MLE-15 cells were subjected to in vivo DMS treatment and LMPCR analysis. The primers used are shown in Table 1, and their relative positions are diagrammed in Fig. 2D. Two sets of primers for each strand were used to allow resolution of promoter-proximal and -distal JSRV sequences. JSRV residue positions in the Mo+JS LTR are designated relative to the JSRV LTR. The Mol 2E, JSRV 2D, and JSRV 3D oligonucleotide primers were designed for LMPCR analysis of the promoter-proximal end of the lower (antisense) strand. In vivo footprinting with these primers is shown in Fig. 3. Our previous mutational analysis of JSRV LTR-luciferase reporters indicated that the upstream HNF-3 site is important for JSRV LTR activity in MLE-15 cells (22), and this site showed evidence for factor binding in vivo. It contained a protected guanine residue on the antisense (lower) strand, just outside the core TGTTTG region (JSRV LTR nucleotide –134) (Fig. 3A). In addition, analysis of the antisense strand also revealed protection within the core sequence of an adjacent CAAT box that matches a consensus NF-I binding site. (This NF-I site is referred to as the downstream NF-I site, since there is an additional upstream consensus NF-I site in the JSRV LTR at nucleotides –227 to –210.)
Phosphorimager analysis was used to quantify the extent of protection or hypersensitivity, as shown in Fig. 3B. Protections where there was a >25% change in the intensity of the band were considered major, while changes of <25% were considered minor or weaker. The 3' Mo-MuLV/JSRV LTR junction is positioned in the middle of the downstream HNF-3 binding site, but the HNF-3 core binding sequence (TAAACA) is maintained in Mo+JS. Interestingly, this site also showed protection of the single guanine residue on the lower (antisense) strand (–37), even though our previous mutational analyses had indicated that the downstream HNF-3 site is not necessary for LTR activity in MLE-15 cells (22). Two weaker protections were detected on the same gel within the nearby putative Ets-I/Gfi-I sites at positions –108 and –109 (Fig. 3A). Footprints were also detected on the lower (antisense) strand in a large region lacking sequences that match canonical binding sites. Three guanine residues, including –153, –155, and –171, were protected, and –153 was strongly protected. Additionally, a stretch of hypersensitive thymidine and cytosine residues was detected from –168 to –166, which included a strong hypersensitivity at thymidine –167. This suggests that a novel MLE-15 cell factor(s) might be binding to the JSRV enhancer sequences upstream of the HNF-3 binding site.
When the upper (sense) strand of the JSRV sequences in Mo+JS Mo-MuLV-infected MLE-15 cells was analyzed, protected and hypersensitive bases were detected within a second CAAT box containing a consensus C/EBP binding site (Fig. 4A). Guanine and adenosine residues from –68 to –70 were hypermethylated. In addition, guanine –73 was protected. All three residues are contained within the core of the C/EBP binding sequence. A putative Gfi-I binding site overlaps with the C/EBP site, but the footprint was far from the core Gfi-I binding sequence. This suggested that C/EBP isoforms may bind to the JSRV enhancer in MLE-15 cells as well.
We previously showed that elimination of a putative NF-B site (–260 to –248) diminished activity of the JSRV LTR in MLE-15 cells by 50%; however, an NF-B antibody was not able to supershift an EMSA complex, indicating that a factor other than NF-B binds this site in MLE-15 cells (28). Indeed, a strong in vivo hypersensitive residue on the upper strand was detected directly outside of the putative NF-B site within an overlapping region that resembles an E-box (Fig. 4B). The footprinting pattern in the distal enhancer suggests that an E-box-binding protein, such as N-myc or upstream stimulating factor (USF), may bind to this region in MLE-15 cells, although this does not eliminate the possibility that a separate factor(s) occupies the closely apposed NF-B-like element.
A summary of the in vivo footprinting of the JSRV sequences in the Mo+JS LTR in infected MLE-15 cells is shown in Fig. 5. Underlined bases and triangles indicate protected (lowercase letters, triangles pointing away from the DNA) or hypersensitive (uppercase letters, arrowheads pointing toward the DNA) bases. Strong interactions are shown with bold lettering or filled arrowheads. Most of the stronger footprints were in the promoter-proximal half of the JSRV enhancers, including the adjacent upstream HNF-3 and downstream NF-I sites, as well as the C/EBP site. Several other sites, such as two putative Ets-I sites and the upstream NF-I site, did not show footprints.
While the in vivo DMS footprinting reported here allowed identification of footprints in the Mo+JS Mo-MuLV LTR, the relative protections and hypersensitivities observed were less intense than in our previous studies of cells infected with wild-type Mo-MuLV (16). One possible explanation could be that in the Mo+JS Mo-MuLV-infected MLE-15 cells, a significant percentage of proviruses might have been transcriptionally inactive. Such proviruses would not be expected to have footprints, and they would dilute the signals from active proviruses.
Transcription from the JSRV LTR in lung epithelial cells is dependent on NF-I and C/EBP binding sites. The in vivo footprinting indicated that in addition to HNF-3, factors are bound to the putative C/EBP and downstream NF-I binding sites in the JSRV sequences of the Mo+JS LTR in infected MLE-15 cells. To test whether these sites are important for activity of the native JSRV LTR in these cells, the downstream NF-I site (corresponding to residues –129 to –117) was deleted from a JSRV LTR-luciferase reporter plasmid (pJS21luc) and used in transient transfection assays (Fig. 6). In order to account for any distance or spatial effects, another plasmid was constructed, with the core NF-I binding region changed from TTGGCACTGCTTCAT to TTTTTACTGCCCCCC (mutated residues are in bold, and core sequences are underlined). Each plasmid was transfected into MLE-15 cells, and expression levels were compared to wild-type pJS21luc. As seen in Fig. 7A, LTR activity was reduced by 40% in MLE-15 cells when the downstream NF-I site was eliminated by nucleotide substitution, but there was no significant change in activity when the site was deleted. Thus, elimination of the putative NF-I binding site by itself had a relatively modest effect on JSRV LTR activity in MLE-15 cells. We similarly mutated or deleted (deletion of nucleotides –77 to –69; mutation of TGATTGTGTAAGAA to TGACCGTCTCCCAA [mutated residues are in bold, and core sequences are underlined]) the C/EBP site from pJS21luc (Fig. 6) and tested the activity in MLE-15 cells (Fig. 7A). Both of these mutants showed a 75% reduction in LTR activity, which indicated that the C/EBP site is important for JSRV LTR activity in MLE-15 cells, since elimination of this site by itself substantially reduced activity.
We studied the same NF-I and C/EBP mutations in the Clara cell-derived mtCC1-2 cells (Fig. 7B). Results similar to those for the MLE-15 cells were observed, indicating that the downstream NF-I site is of less importance and the C/EBP site of greater importance for JSRV LTR activity in these cells.
In light of reports describing cooperativity between HNF-3, NF-I, and C/EBP sites in the expression of some cellular genes (1, 5, 40), double mutations that eliminated any combination of the upstream HNF-3, downstream NF-I, or C/EBP sites were introduced into pJS21luc (Fig. 6). As reported previously (22), the upstream HNF-3 binding site is essential for transcriptional activation in MLE-15 but not mtCC1-2 cells. In MLE-15 cells, these double mutant reporter plasmids all exhibited a drastic reduction in LTR activity. In fact, the levels detected were essentially equal to reporter assay values observed in a vector carrying only the JSRV basal promoter [pJS21luc(–37)], indicating that the doubly mutated LTR enhancers were virtually nonfunctional (Fig. 7A). It should be noted that the LTR activities of reporter vectors containing single mutations in the C/EBP or HNF-3 sites were not statistically different from that of pJS21luc(–37).
As mentioned above, we previously reported that LTR activity of a reporter vector containing a mutated HNF-3 site is approximately equal to the wild-type LTR in mtCC1-2 (22) cells, and we showed here that single mutations in the NF-I site result in only a small decrease in LTR activity (Fig. 7B). Interestingly, reporter vectors with mutations in both the HNF-3 and NF-I sites showed extremely low activity, suggesting that JSRV expression is dependent on binding of either HNF-3 or NF-I (but not necessarily both) to the adjoining sites in Clara cells. All three pJS21luc double mutants exhibited reporter activities as low as or lower than pJS21luc(–37), indicating that the HNF-3, NF-I, and C/EBP sites are collectively critical for JSRV enhancer activity in mtCC1-2 cells.
For further study, we focused on the NF-I and C/EBP binding sites because of the footprint data described above and the results shown in Fig. 7. Also, previous reports indicated that isoforms of these factors contribute to tissue-specific gene expression in lung epithelial cells (2, 8, 9, 20, 26). However, several weak footprints were also detected in other consensus binding elements, as shown in Fig. 5. Therefore, we individually mutated each of those putative binding sites in the LTR reporter plasmid and measured the activities of the mutants. Although we previously reported that mutating the NF-B-like site reduced LTR activity by 50% (28), mutation of the adjacent E-box element did not significantly reduce enhancer activity. All other mutations in pJS21luc had little or no effect on reporter levels, with the exception that mutation in the putative SP-I site resulted in an approximately 50% reduction of activity in MLE-15 cells. A summary of all mutations in JSRV LTR reporter plasmids is shown in Table 2.
NF-IA and C/EBP/ bind to JSRV LTR sequences. NF-I is a transcription factor that plays a key role in activating the surfactant protein C promoter (2). Although NF-I expression can be detected in a wide variety of tissues, the differential expression of NF-I isoforms contributes to tissue-specific expression during and after development (10). There are at least four mammalian NF-I genes (A, B, C, and X), each of which is transcribed and further spliced to yield multiple isoforms (10, 38). NF-IA has been shown to transactivate the SP-C promoter in HeLa cells cotransfected with an NF-IA expression plasmid (2). In order to investigate whether NF-IA can bind to the JSRV downstream NF-I site, a series of EMSAs was performed using a labeled double-stranded oligonucleotide probe containing sequences from –134 to –115 of the JSRV LTR. As seen in Fig. 8A, a large protein-DNA complex was formed when this NF-I probe was incubated with nuclear extracts from MLE-15. A supershifted complex was observed when the nuclear extracts were preincubated with anti-NF-IA antiserum. This suggests that a factor (or complexes of factors) binds to the JSRV downstream NF-I site in these cells and that a component of this complex is NF-IA. Nuclear proteins from mtCC1-2 cells similarly complexed with the NF-I probe. The anti-NF-IA antiserum also supershifted the band, but with less efficiency than observed in the MLE-15 EMSA, suggesting that NF-IA is expressed in both cell lines but binds to this site in different ratios.
We have previously shown that the JSRV LTR drives transcription more efficiently in the lung epithelial cell lines MLE-15 and mtCC1-2 than in nonlung epithelial cell lines, such as the mouse fibroblast cell line NIH 3T3 (28). The poor levels of LTR activation in nonlung cells can be partly explained by the lack of HNF-3 and, potentially, cooperating coactivators in these cells (22, 28). In addition to the presence of tissue-specific transcriptional activators, certain isoforms of widely expressed transcription factors also facilitate gene expression in specific cell types. In order to compare the isoform expression of a ubiquitously expressed transcription factor (NF-I) in nonlung epithelial cells with that present in the lung cell lines MLE-15 and mtCC1-2, gel shift assays were performed using NIH 3T3 nuclear extracts. Although this gel contained a complex with a migration pattern similar to the MLE-15 EMSA, the anti-NF-IA antiserum did not supershift this band. Therefore, NIH 3T3 cells do not express NF-IA isoforms that can bind to the JSRV downstream NF-I site, but they express other NF-I isoform(s) or other transcription factors that can bind to this site.
C/EBP has three main isoforms (, , and ) that are expressed throughout the lung (6, 39). Isoforms and are highly expressed in the bronchiolar epithelium, while is mostly found in alveolar macrophages (39). Because C/EBP binds as a dimer (or heterodimer), one or more isoforms may bind to a C/EBP site. Although C/EBP is considered to be more important for expression in bronchiolar epithelium (7-9, 39), heterodimers of C/EBP and have been reported to work synergistically on some lung-specific promoters (8, 9). To identify the C/EBP isoform(s) that can bind to the JSRV element, a series of EMSAs was performed in the presence of isoform-specific anti-C/EBP antibodies (Fig. 8B). As expected, incubation of a probe encompassing the JSRV C/EBP site with nuclear extracts from MLE-15 or mtCC1-2 cells produced an EMSA pattern composed of at least three major bands. Addition of anti-C/EBP antiserum supershifted the slowest-migrating band, but no change in intensity was observed in the other complexes. Anti-C/EBP antiserum diminished the intensity of the faster-migrating complexes and caused a faint supershift. Interestingly, antiserum against C/EBP had no effect on any complex, even though this isoform had the highest transactivation potential in NIH 3T3 cells cotransfected with pJS21luc and C/EBP, , or expression plasmids (data not shown). Thus, it appears that C/EBP and C/EBP, but not C/EBP, are capable of binding to the JSRV C/EBP site in the lung epithelial-derived cell lines MLE-15 and mtCC1-2.
DISCUSSION
In this study, our goal was to identify transcription factors and binding sites that are responsible for the lung epithelial cell specificity of the JSRV LTR. Since there are multiple putative factor binding sites in the LTR, our approach was first to identify sites that show evidence of occupancy by in vivo footprinting (25). We first carried out in vivo footprinting of the JSRV enhancer in a mouse lung epithelial cell line (MLE-15) that supports a high level of transcriptional activity from the JSRV LTR (28) by studying MLE-15 cells infected with a chimeric Mo-MuLV (Mo+JS) containing the JSRV enhancers (23). The in vivo footprinting was consistent with our previous in vitro studies implicating HNF-3 in JSRV LTR activity (22), since a footprint was found in the JSRV upstream HNF-3 site in the Mo+JS LTR in the infected cells. In addition, footprints were found in other factor binding sites in the Mo+JS LTR, which suggested that factors binding to these sites might also be important for JSRV LTR activity in lung epithelial cells. This was addressed by mutating or deleting these sites in the context of the native JSRV LTR, followed by assessing LTR activity in transient transfection reporter assays.
In vivo footprints in the JSRV C/EBP and downstream NF-I binding sites were also found in the Mo+JS Mo-MuLV-infected cells. Mutation or deletion of the downstream NF-I site alone had relatively little effect on activity of the native JSRV LTR in MLE-15 cells (Fig. 7). This site is directly adjacent to the upstream (distal) HNF-3 binding site; we previously showed that the upstream HNF-3 site is important for JSRV LTR activity in MLE-15 cells. The close placement of these HNF-3 and NF-I sites suggests that these two factors may cooperate, but experiments to address this possibility have not yet indicated a synergistic interaction between the two sites. It is interesting that the enhancer of the liver-specific albumin gene also has closely positioned HNF-3 and NF-I sites (18). It appears that the two factors interact on the albumin enhancer, although the exact nature of this cooperation is unknown. EMSAs shown in Fig. 8A indicated that at least the NF-IA isoform is expressed in MLE-15 and mtCC1-2 cells. Antibodies against other NF-I isoforms suitable for EMSA supershifts were not available. Increasing evidence suggests that tissue-specific expression of genes is facilitated not only by binding of factors that are exclusively expressed in specific cell types but also by a complementing array of specific isoforms of ubiquitously expressed transcription factors. NF-IA is enriched in lung epithelial cells, and it likely has a critical role in activating the surfactant protein C promoter in type II pneumocytes (2). Similarly, NF-IA may enable high-level JSRV LTR transcriptional activation more efficiently than other NF-I isoforms.
Mutations or deletions of the C/EBP site in the JSRV LTR (downstream of the paired HNF-3 and NF-I sites) had a pronounced effect on transcriptional activity in MLE-15 or mtCC1-2 cells (>75% reduction). Thus, the C/EBP site appears to be important for LTR activity in both of these lung epithelial cell types. EMSAs indicated that C/EBP and C/EBP but not C/EBP are expressed in MLE-15 cells. This was somewhat surprising, since C/EBP has been reported to be important for the expression of some genes in lung epithelial cells (8, 9). Isoforms of C/EBP are expressed ubiquitously, but expression is highest in cells involved in fat metabolism (fat and liver cells and type II pneumocytes) and terminally differentiated cells (6, 7, 39). The C/EBP binding site is missing from the LTRs of endogenous JSRV-related proviruses in the sheep genome, which are not expressed in the lung (3). Thus, activation of the JSRV LTR by C/EBP parallels the lung-specific expression of the exogenous JSRV.
The in vivo footprinting of the Mo+JS LTR in infected MLE-15 cells also identified a footprint near an NF-B-like site (–260 to –248) in the distal enhancer occupied by an unknown protein. In transient transfection assays of JSRV LTR truncations, LTR activity decreased by 50% when nucleotides adjacent to the footprint were mutated (28). Although sequences in this region match a canonical NF-B binding site, an EMSA with an anti-NF-B antibody did not result in a supershift. The NF-B-like site overlaps with an E-box element that can potentially bind isoforms of N-myc or upstream stimulating factor (USF). USF was recently found to bind to and activate the surfactant protein A enhancer in fetal rabbit lung type II cells (13, 14). Although USF is expressed ubiquitously, the same study showed that USF 2 was enriched in type II cells compared to lung fibroblasts, suggesting that this isoform may be important for type II-specific gene expression. However, selective mutation of the E-box element in the JSRV LTR did not significantly affect transcriptional activity, although USF 1 and 2 were found to bind to this site in EMSAs (data not shown).
Finally, it should be mentioned that the factors involved in JSRV LTR expression in type II pneumocytes may differ from those important in Clara cells. Alteration of the C/EBP site diminished LTR activity in both MLE-15 and mtCC1-2 cells, indicating that a C/EBP isoform is important for expression in both of these cell types. However, mutation or deletion of the upstream HNF-3 site by itself had relatively little effect on JSRV LTR activity in Clara-derived mtCC1-2 cells, in contrast to the substantial loss of activity in type II-derived MLE-15 cells (22). As suggested in Fig. 7, one explanation could be that the HNF-3 and NF-I sites are functionally redundant in mtCC1-2 cells. Whereas the JSRV LTR may utilize either HNF-3 or NF-I for efficient expression in mtCC1-2 cells, NF-I may not effectively substitute for HNF-3 in MLE-15 cells. Alternatively, some other factor(s) also might be important for JSRV LTR activity in Clara cells. We have not yet identified a factor that is important for JSRV expression in mtCC1-2 but not MLE-15 cells. Although supershift assays indicate that MLE-15 and mtCC1-2 express identical NF-I and C/EBP isoforms (Fig. 8), it is important to note that in comparing the migration patterns between the two nuclear extracts, these cells might express different splice variants and/or contain isoforms at different ratios. In vivo footprinting of Mo+JS Mo-MuLV-infected mtCC1-2 cells might identify binding sites for other factors important for JSRV LTR expression in these cells, and such experiments are in progress.
ACKNOWLEDGMENTS
This work was supported by grant R01 CA82654 from the National Cancer Institute. K.M.-E. was supported by training grant T32 CA09054. Support of the UCI Cancer Research Institute and the DNA sequencing core of the Chao Family Comprehensive Cancer Center is acknowledged.
REFERENCES
Antes, T. J., and B. Levy-Wilson. 2001. HNF-3 beta, C/EBP beta, and HNF-4 act in synergy to enhance transcription of the human apolipoprotein B gene in intestinal cells. DNA Cell Biol. 20:67-74.
Bachurski, C. J., S. E. Kelly, S. W. Glasser, and T. A. Currier. 1997. Nuclear factor I family members regulate the transcription of surfactant protein-C. J. Biol. Chem. 272:32759-32766.
Bai, J., R. Y. Zhu, K. Stedman, C. Cousens, J. Carlson, J. M. Sharp, and J. C. DeMartini. 1996. Unique long terminal repeat U3 sequences distinguish exogenous jaagsiekte sheep retroviruses associated with ovine pulmonary carcinoma from endogenous loci in the sheep genome. J. Virol. 70:3159-3168.
Barsky, S. H., R. Cameron, K. E. Osann, D. Tomita, and E. C. Holmes. 1994. Rising incidence of bronchioloalveolar lung carcinoma and its unique clinicopathologic features. Cancer 73:1163-1170.
Beurton, F., U. Bandyopadhyay, B. Dieumegard, R. Barouki, and M. Aggerbeck. 1999. Delineation of the insulin-responsive sequence in the rat cytosolic aspartate aminotransferase gene: binding sites for hepatocyte nuclear factor-3 and nuclear factor I. Biochem. J. 343:687-695.
Birkenmeier, E. H., B. Gwynn, S. Howard, J. Jerry, J. I. Gordon, W. H. Landschulz, and S. L. McKnight. 1989. Tissue-specific expression, developmental regulation, and genetic mapping of the gene encoding CCAAT/enhancer binding protein. Genes Dev. 3:1146-1156.
Breed, D. R., L. R. Margraf, J. L. Alcorn, and C. R. Mendelson. 1997. Transcription factor C/EBPdelta in fetal lung: developmental regulation and effects of cyclic adenosine 3',5'-monophosphate and glucocorticoids. Endocrinology 138:5527-5534.
Cassel, T. N., J. A. Gustafsson, and M. Nord. 2000. CYP2B1 is regulated by C/EBP alpha and C/EBP delta in lung epithelial cells. Mol. Cell Biol. Res. Commun. 3:42-47.
Cassel, T. N., L. Nordlund-Moller, O. Andersson, J. A. Gustafsson, and M. Nord. 2000. C/EBPalpha and C/EBPdelta activate the Clara cell secretory protein gene through interaction with two adjacent C/EBP-binding sites. Am. J. Respir. Cell Mol. Biol. 22:469-480.
Chaudhry, A. Z., G. E. Lyons, and R. M. Gronostajski. 1997. Expression patterns of the four nuclear factor I genes during mouse embryogenesis indicate a potential role in development. Dev. Dyn. 208:313-325.
DeMartini, J. C., and D. F. York. 1997. Retrovirus-associated neoplasms of the respiratory system of sheep and goats. Ovine pulmonary carcinoma and enzootic nasal tumor. Vet. Clin. N. Am. Food Anim. Pract. 13:55-70.
Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11:1475-1489.
Gao, E., Y. Wang, J. L. Alcorn, and C. R. Mendelson. 1997. The basic helix-loop-helix-zipper transcription factor USF1 regulates expression of the surfactant protein-A gene. J. Biol. Chem. 272:23398-23406.
Gao, E., Y. Wang, J. L. Alcorn, and C. R. Mendelson. 2003. Transcription factor USF2 is developmentally regulated in fetal lung and acts together with USF1 to induce SP-A gene expression. Am. J. Physiol. Lung Cell Mol. Physiol. 284:L1027-L1036.
Garrity, P. A., and B. J. Wold. 1992. Effects of different DNA polymerases in ligation-mediated PCR: enhanced genomic sequencing and in vivo footprinting. Proc. Natl. Acad. Sci. USA 89:1021-1025.
Granger, S. W., and H. Fan. 1998. In vivo footprinting of the enhancer sequences in the upstream long terminal repeat of Moloney murine leukemia virus: differential binding of nuclear factors in different cell types. J. Virol. 72:8961-8970.
Holland, M. J., M. Palmarini, M. Garcia-Goti, L. Gonzalez, I. McKendrick, M. de las Heras, and J. M. Sharp. 1999. Jaagsiekte retrovirus is widely distributed both in T and B lymphocytes and in mononuclear phagocytes of sheep with naturally and experimentally acquired pulmonary adenomatosis. J. Virol. 73:4004-4008.
Jackson, D. A., K. E. Rowader, K. Stevens, C. Jiang, P. Milos, and K. S. Zaret. 1993. Modulation of liver-specific transcription by interactions between hepatocyte nuclear factor 3 and nuclear factor 1 binding DNA in close apposition. Mol. Cell. Biol. 13:2401-2410.
Landis, S. H., T. Murray, S. Bolden, and P. A. Wingo. 1998. Cancer statistics, 1998. CA Cancer J. Clin. 48:6-29.
Li, F., E. Rosenberg, C. I. Smith, K. Notarfrancesco, S. R. Reisher, H. Shuman, and S. I. Feinstein. 1995. Correlation of expression of transcription factor C/EBP alpha and surfactant protein genes in lung cells. Am. J. Physiol. 269:L241-L247.
Mason, R. J., and J. M. Shannon. 1997. Alveolar type II pneumocytes, p. 543-555. In J. B. West, R. G. Crystal, E. R. Weibel, and P. J. Barnes (ed.), The lung: scientific foundations, vol. 1. Lippincott-Raven, Philadelphia, Pa.
McGee-Estrada, K., M. Palmarini, and H. Fan. 2002. HNF-3beta is a critical factor for the expression of the Jaagsiekte sheep retrovirus long terminal repeat in type II pneumocytes but not in Clara cells. Virology 292:87-97.
McGee-Estrada, K., M. Palmarini, C. Hallwirth, and H. Fan. 2005. The Jaagsiekte sheep retrovirus enhancer increases Moloney murine leukemia virus genomic expression in lung epithelial cells. Virus Genes 31:251-263.
Mendelson, C. R. 2000. Role of transcription factors in fetal lung development and surfactant protein gene expression. Annu. Rev. Physiol. 62:875-915.
Mueller, P. R., and B. Wold. 1989. In vivo footprinting of a muscle specific enhancer by ligation mediated PCR. Science 246:780-786.
Nord, M., M. Lag, T. N. Cassel, M. Randmark, R. Becher, H. J. Barnes, P. E. Schwarze, J. A. Gustafsson, and J. Lund. 1998. Regulation of CCSP (PCB-BP/uteroglobin) expression in primary cultures of lung cells: involvement of C/EBP. DNA Cell Biol. 17:481-492.
Palmarini, M., C. Cousens, R. G. Dalziel, J. Bai, K. Stedman, J. C. DeMartini, and J. M. Sharp. 1996. The exogenous form of Jaagsiekte retrovirus is specifically associated with a contagious lung cancer of sheep. J. Virol. 70:1618-1623.
Palmarini, M., S. Datta, R. Omid, C. Murgia, and H. Fan. 2000. The long terminal repeat of Jaagsiekte sheep retrovirus is preferentially active in differentiated epithelial cells of the lungs. J. Virol. 74:5776-5787.
Palmarini, M., P. Dewar, M. De las Heras, N. F. Inglis, R. G. Dalziel, and J. M. Sharp. 1995. Epithelial tumour cells in the lungs of sheep with pulmonary adenomatosis are major sites of replication for Jaagsiekte retrovirus. J. Gen. Virol. 76:2731-2737.
Palmarini, M., and H. Fan. 2001. Retrovirus-induced ovine pulmonary adenocarcinoma, an animal model for lung cancer. J. Natl. Cancer Inst. 93:1603-1614.
Palmarini, M., M. J. Holland, C. Cousens, R. G. Dalziel, and J. M. Sharp. 1996. Jaagsiekte retrovirus establishes a disseminated infection of the lymphoid tissues of sheep affected by pulmonary adenomatosis. J. Gen. Virol. 77:2991-2998.
Palmarini, M., J. M. Sharp, M. de las Heras, and H. Fan. 1999. Jaagsiekte sheep retrovirus is necessary and sufficient to induce a contagious lung cancer in sheep. J. Virol. 73:6964-6972.
Palmarini, M., J. M. Sharp, C. Lee, and H. Fan. 1999. In vitro infection of ovine cell lines by Jaagsiekte sheep retrovirus. J. Virol. 73:10070-10078.
Perk, K., and I. Hod. 1982. Sheep lung carcinoma: an endemic analogue of a sporadic human neoplasm. J. Natl. Cancer Inst. 69:747-749.
Pfeifer, G. P., S. D. Steigerwald, P. R. Mueller, B. Wold, and A. D. Riggs. 1989. Genomic sequencing and methylation analysis by ligation mediated PCR. Science 246:810-813.
Plopper, C. G., M. H. Dallas, and A. R. Buckpitt. 1997. Clara cells, vol. 1. Lippincott-Raven, Philadelphia, Pa.
Rai, S. K., F. M. Duh, V. Vigdorovich, A. Danilkovitch-Miagkova, M. I. Lerman, and A. D. Miller. 2001. Candidate tumor suppressor HYAL2 is a glycosylphosphatidylinositol (GPI)-anchored cell-surface receptor for jaagsiekte sheep retrovirus, the envelope protein of which mediates oncogenic transformation. Proc. Natl. Acad. Sci. USA 98:4443-4448.
Rupp, R. A., U. Kruse, G. Multhaup, U. Gobel, K. Beyreuther, and A. E. Sippel. 1990. Chicken NFI/TGGCA proteins are encoded by at least three independent genes: NFI-A, NFI-B and NFI-C with homologues in mammalian genomes. Nucleic Acids Res. 18:2607-2616.
Sugahara, K., T. Sadohara, M. Sugita, K. Iyama, and M. Takiguchi. 1999. Differential expression of CCAAT enhancer binding protein family in rat alveolar epithelial cell proliferation and in acute lung injury. Cell Tissue Res. 297:261-270.
Tsay, W., Y. M. Lee, S. C. Lee, M. C. Shen, and P. J. Chen. 1997. Synergistic transactivation of HNF-1alpha, HNF-3, and NF-I contributes to the activation of the liver-specific protein C gene. DNA Cell Biol. 16:569-577.
Voelker, D. R., and R. J. Mason. 1989. Alveolar type II epithelial cells. Marcel Dekker, New York, N.Y.
Whitsett, J. A., and S. W. Glasser. 1998. Regulation of surfactant protein gene transcription. Biochim. Biophys. Acta 1408:303-311.
Wikenheiser, K. A., D. K. Vorbroker, W. R. Rice, J. C. Clark, C. J. Bachurski, H. K. Oie, and J. A. Whitsett. 1993. Production of immortalized distal respiratory epithelial cell lines from surfactant protein C/simian virus 40 large tumor antigen transgenic mice. Proc. Natl. Acad. Sci. USA 90:11029-11033.
Wingender, E., X. Chen, R. Hehl, H. Karas, I. Liebich, V. Matys, T. Meinhardt, M. Pruss, I. Reuter, and F. Schacherer. 2000. TRANSFAC: an integrated system for gene expression regulation. Nucleic Acids Res. 28:316-319.
York, D. F., R. Vigne, D. W. Verwoerd, and G. Querat. 1992. Nucleotide sequence of the jaagsiekte retrovirus, an exogenous and endogenous type D and B retrovirus of sheep and goats. J. Virol. 66:4930-4939.(Kathleen McGee-Estrada an)