Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Surfactant proteins (SPs) are under the control of a variety of potential developmental (1) and hormonal regulators (2) including phorbol ester (3), tumor necrosis factor- (4), epidermal growth factor (5), insulin (6), interleukin-1 (7), keratinocyte growth factor (8), and O₂ (9). However, a crucial role has been implicated for two transcription factors, thyroid transcription factor (TTF)-1 and hepatocyte nuclear factor (HNF)-3, to mediate the expression of SPs. The promoter region of the SP-B gene has been extensively studied (10). In the human SP-B gene, a region located at the immediate 5' flanking region of the basal promoter TATA box between 80 and 110 contains two cis-acting elements for TTF-1 and HNF-3 binding (10). These two elements are important for the specificity and the activation of gene expression in the lung. Farther upstream of the promoter is an enhancer-type element with TTF-1 binding sites, which can upregulate the simian virus 40 (SV40) promoter in a human pulmonary adenocarcinoma cell line (H441) (13).

Although the expression of SP-B is known to be dependent, at least in part, on TTF-1 and HNF-3, other factors such as SP1 and SP3 are also required for its expression (11, 12). In transient transfection studies when TTF-1 was cotransfected with a reporter gene controlled by the SP-B promoter into H441 cells, the reporter gene expression was activated radically over background. However, when HNF-3 or - was cotransfected into the same cell, marked activation of SP-B gene expression was not observed (14). TTF-1 can also regulate SP-A (15, 16) and SP-C expression in mammals (17). Besides TTF-1, human upstream stimulatory factor-1, nuclear factor (NF)-1, and GT box binding protein also play roles in regulating SP-A and/or SP-C expression (18).

TTF-1 is a homeodomain-containing transcription factor that regulates some tissue-specific genes expression in thyroid, lung, and brain (21). In thyroid, TTF-1 can mediate the tissue-specific expression of thyroglobulin and thyroperoxidase genes. In lung, in addition to SPs (13, 16, 17, 22), TTF-1 also is required for the expression of Clara cell- specific protein (CCSP) (23) and type I pneumocyte-specific T1- gene (24). In embryo, the blockage of organ development in TTF-1 gene-targeted mice suggests a vital role of TTF-1 during the morphogenesis of these organs (21).

There are three primary domain structures in TTF-1: an N-terminal transactivation domain, a DNA-binding homeodomain in the middle, and a C-terminal activation domain (25). The N-terminal domain was shown to bear the main transactivation activity because the mutant without N-terminal domain completely abolished the transcriptional activity whereas mutant with the C-terminus deleted reduced the activity only partially (26). Structurally, it has been suggested that the N-terminal domain has functional properties similar to the typical transactivation acidic domain in VP16 (25). The characteristics of this domain imply that this may be the region where TTF-1 interacts with other potential factors/ cofactors to regulate responsive gene expression in vivo.

During the lung development, both TTF-1 and HNF-3 are expressed early at the onset of lung growth, but SP-B expression is not detected until a later gestational stage (22). This suggests that TTF-1 and/or HNF-3 alone cannot fully account for the differential regulation of SP-B gene expression during the development. An additional cofactor(s) or the modified form of TTF-1, therefore, might be required in regulating the SP-B gene expression in vivo.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmid Construction

Plasmid pCIH-ER was constructed by inserting the hormone binding domain (HBD) of estrogen receptor (ER) into pCIN-HA (a modified vector from pCI-neo [Promega, Madison, WI] containing the influenza virus hemagglutinin [HA] epitope [YPYDPDYA]). The ER-HBD fragment was amplified by Pyrococcus furiosus DNA polymerase using the primers CGTCGACTCTAGCGCTTTCGCGAAATGAAATGGGTGCTTCA and GGATCCTCTAGGTTTAAACCGGTGATATCGTGTTGGGGAAGCCCTC, and then was inserted into the Xba I-digested pCIN-HA vector by the exonuclease recession method as described (27). Plasmid pER-BR22 was constructed using the same method by subcloning a polymerase chain reaction (PCR)-amplified BR22 complementary DNA (cDNA) into pCIH-ER. Primers used for this amplification were ATATCACCGGGCCGCGTCGACCGGTCCGCG and GTTTAAACCGGTTTTTTGTACTGCCTTTGGGC. GAL4 activation domain (GAD)-BR22 plasmid DNA was used as the template for amplification. For GFP-NH construct, primers GGCCCGGGATCCATGTCGATGAGTCCAAAGCAC and TCCGGTGGATCAGGCCTGGCGCTTCATTTTGTA were used to amplify the NH fragment of human TTF-1 from the N-terminus to the end of homeodomain polypeptide at amino acid 220. PCR-amplified DNA was subcloned into the Bam HI site of pEGFP-C1 vector (Clontech, Palo Alto, CA). GBT-NH plasmid was constructed by inserting the same TTF-1 region into the Bam HI site of GAL4 DNA-binding domain vector (GBT-9) (Clontech). Primers used for this construct were CCCGGGGATCATGTCGATGAGTCCAAAGCAC and TCGACGGATCAGGCCTGGCGCTTCATTTTGTA. To construct GST-BR22 for protein expression, a PCR DNA fragment containing the same region as that in the GAD-BR22 was generated by the primers TCTGACGGATCCCGGTCCGCGAAGTGGCGGCCT and ATGGTGGGATCATTTTTTGTACTGCCTTTGGGC. The insert DNA was in-frame subcloned into the Bam HI site of pGEX-VH vector (28). Plasmid, pCIN-Flag, was constructed by ligating the Bam HI/Sal I Flag-tag containing fragment isolated from pVP-Flag7 (a gift from Dr. Richard Baer, Columbia University, New York, NY) and Eco RI/Sal I-digested pCI-neo vector. HA-BR22F and Flag- TTF-1 plasmids were generated by using exonuclease recessing method to subclone the PCR-amplified full length of BR22 and TTF-1 fragment into pCIN-HA and pCIN-Flag vectors at Sal I and Eco RI sites, respectively. Primers used for PCR amplification were CAGCTGGGAATTATGTCGATGAGTCCAAAGCAC and CCCGGGGAATTACCAGGTCCGACCGTATAGCAA for TTF-1 amplification; and GCCGCGTCGATGGCGCCGGTGAGGCGGT and TCTAGAGTCGACATTTTTCTTGTATTTTTTGAAGAA for BR22 amplification. All of those expression genes were sequenced to confirm their gene sequences as expected.

Yeast One-Hybrid Analysis

A DNA fragment of human SP-B promoter region from 118 to 64 including the TTF-1 and HNF-3 binding sites was synthesized and subcloned into the pHISi vector at the Sma I site. Two plasmids, pHISi-SPBP and pHISi-SPBR, were constructed with the promoter elements inserted in the same or reverse orientation as that in the human SP-B gene to activate the HIS3 reporter gene expression (Figure 1). Yeast strains BP and BR were generated through a homologous recombination by integrating pHISi-SPBP or pHISi-SPBR into the yeast strain YM4271 (Clontech). A human adult lung cDNA library was transformed into BP or BR individually and the transformed yeasts were plated on the medium lacking histidine to select for yeast clone containing interactive gene products. The transformation procedure was performed using the lithium acetate method. An inhibitor, 15 mM of 3-aminotriazole (3-AT) was added to medium to suppress the leaky HIS3 gene product. About 1 × 10⁶ colonies were screened in both yeast strains. Yeast colonies > 2-3 mm were selected and their plasmids were recovered for further analyses.

View larger version (9K): [in this window] [in a new window]   Figure 1.   Diagram of the SP-B promoter fragment insertion in BP and BR yeast. The DNA sequences inserted into the promoter region of HIS3 gene in the genome of YM4271 are shown. Both TTF-1 and HNF-3 binding sites are marked. Arrows indicate the orientation of DNA insertion in HIS3 promoter.

Yeast Two-Hybrid Analysis

The plasmid pGAD-BR22, which contained GAD and the inserted cDNA from the human lung library, was cotransformed with pGBT-NH into yeast strain HF7c. This strain has two reporter genes integrated in its genome, HIS3 and LacZ. Both genes are controlled by an inducible promoter with the GAL4 binding sites. Cotransformed cells were plated on the selection plate without tryptophan, leucine, and histidine. Inhibitor of 15 mM 3-AT was usually included in the medium to increase the selection specificity. To assay -galactosidase activity, yeast colonies were transferred to a nitrocellulose membrane, permeabilized by liquid nitrogen, and then incubated with X-Gal in Z buffer (100 mM phosphate buffer [pH 7.0], 10 mM KCl, and 1 mM MgSO₄). The incubation time was from 2 h to overnight at room temperature.

In Vitro Protein-Protein Interaction Analysis

GST-BR22 recombinant protein was produced in bacteria, BL21 (DE3)pLysS (Novagen, Madison, WI) and affinity-purified through glutathione-agarose beads. Equal amounts of purified recombinant protein, GST-BR22 or GST, were bound to beads and then incubated in vitro with either ³⁵S-methionine-labeled in vitro- translated TTF-1 or luciferase. After incubating at 4°C overnight, the beads were washed four times with binding buffer (20 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], pH 7.6; 100 mM KCl; 10% glycerol; 1 mM ethylenediaminetetraacetic acid [EDTA]; and 1 mM phenylmethylsulfonyl fluoride). The proteins retained on beads were eluted with 10 mM glutathione, fractionated in 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel, transferred to polyvinylidene difluoride membrane, and exposed to X-ray film.

Coimmunoprecipitation Analysis

Plasmids pCIH-ER, pER-BR22, and hTTF-1 were transcribed from the T7 RNA polymerase promoter for in vitro translation protocol. In vitro transcription/translation coupled reactions (Promega) were conducted in rabbit reticulocyte lysates in the presence or absence of [³⁵S]-methionine. The quantity and quality of the in vitro-translated polypeptides were analyzed by SDS-PAGE using the labeled reactions. Equal amounts of unlabeled ER and ER-BR22 were incubated with magnetic beads containing anti-HA monoclonal antibody (12CA5; Boehringer Mannheim/Roche, Indianapolis, IN) to the HA tag of ER-BR22 at 4°C for 1 h in binding buffer. The beads were washed and incubated with ³⁵S-labeled TTF-1 or luciferase at 4°C for another hour in the same buffer and then were extensively washed, boiled in the 1× SDS dye, and fractionated on 10% SDS-PAGE.

For the coimmunoprecipitation of protein complex from mammalian cells, cell lysates from transient transfected 293 cells (human embryonic kidney fibroblast) were used. Cells of 5 × 10⁵ were seeded on a 25-cm² flask 24 h before the transfection. Transfection procedure was carried out using Fugene 6 reagent (Boehringer Mannheim/Roche). Basically, 2 µg of total plasmid DNA was applied to each transfection. For cotransfection, 1.0 µg each of plasmids containing ER and GFP were applied to each sample. Transfected cells were incubated in RPMI with 10% horse serum plus 100 nM of 4-hydroxytamoxifen and harvested 48 h after transfection. Cell lysates were prepared in buffer containing 1 M NaCl, 10 mM Hepes (pH 7.6) 0.1% Nonidet P-40 (NP-40), 5 mM EDTA, and protease inhibitors (Boehringer Mannheim/Roche). For coimmunoprecipitation analysis, cell lysates from eight flasks (25 cm²) of cells were prepared by stirring cells in 1 ml of lysis buffer described earlier for 30 min at 4°C and spun at 4°C for 30 min. Supernatants (200 µl) were removed and incubated with 16 µg of 12CA5 monoclonal anti-HA antibody in total 1 ml of buffer containing 200 mM NaCl, 10 mM Hepes (pH 7.6), 0.1% NP-40, 5 mM EDTA, and protease inhibitors at 4°C overnight with end-to-end rotation. Protein A-sepharose beads (20 µl, 50% slurry) (Sigma, St. Louis, MO) were added and the mixture was incubated with rotation at 4°C for an additional 2 h. Supernatants were saved. The beads were washed six times in buffer with 250 mM NaCl and protease inhibitors. The protein complexes retained on the beads after washes were directly denatured in 1× SDS dye by boiling and then fractionated in 10% SDS-PAGE gel. Western blot analysis was conducted using anti-GFP antibody (1:3,000 dilution) (Clontech) followed by the addition of goat antirabbit antibody conjugated with alkaline phosphatase (AP) (1:5,000) (Bio-Rad, Hercules, CA) and AP substrate (CSPD from Boehringer Mannheim/Roche). Blots were subsequently exposed to X-ray film to record the signals.

Reverse Transcriptase-PCR Analysis

H441 cells were obtained from ATCC (Rockville, MD) and maintained in culture according to the suggestion by ATCC. Cells (5 × 10⁶) were harvested and lysed for total RNA preparation. SV total RNA isolation system (Promega) was used for RNA isolations. Human lung, placenta, and skeletal muscle total RNAs were purchased from Clontech. Superscript II Rnase H reverse transcriptase (RT) was used to synthesize the first-strand cDNA, which was then used as a template in the subsequent PCR amplification. A cDNA pool that could represent the message in original sample was generated using SMART PCR cDNA synthesis protocol (Clontech) (29). A total of 1 ng of synthesized cDNA was used for subsequent specific gene message detection. The PCR primers used for detecting various genes were: BR22RT5', CGGTCCGCGAAGTGGCGGCCT; BR22RT3', GTACTG CCTTTGGGCTTCTTC; GST-spb5', TTCCCGGGGATCTTCCCCATTCCTCTCCCCTAT; GST-spb3', GTCGACGGATCACATGGAGCACCGGAGGACGAG; glyceraldehyde-3-phosphate dehydrogenase (G3PDH)5', ACCACA- GTCCATGCCATCAC; and G3PDH3', TCCACCACCCTGTTGCTGTA.

Dual-Luciferase Analysis

Dual-luciferase assay was conducted to measure the human SP-B promoter transactivation activity. Plasmid, pGL3-SPBPF, was constructed by ligation of the human SP-B promoter fragment from 118 to +41 into pGL3-E vector (Promega) at the Bgl II/ Hind III site. The pGL3-E vector contains a promoterless firefly (Photinus pyralis) luciferase gene for promoter analysis. Two oligonucleotides containing the promoter region with a modification at +15 residue as described (30) were synthesized and annealed to generate the SP-B promoter fragment. Plasmid DNAs containing 450 ng each of HA- and Flag-plasmids, 100 ng of pGL3-SPBPF, and 5 ng of pRL-TK (a plasmid containing Renilla luciferase gene controlled by a thymidine kinase promoter) (Promega) were cotransfected into 293 cells. The Renilla luciferase controlled by a thymidine kinase promoter served as a control for the dual-luciferase cotransfection studies. Cell lysates were harvested 48 h after transfection and proceeded for dual- luciferase assay using TD-20/20 luminometer (Turner Designs, Sunnyvale, CA). The dual-luciferase assay procedure and the substrates used for measuring enzyme activity were carried out as described in the manufacturer's manual (Promega). Basically, the firefly luciferase activity was measured first, followed by adding Stop and Go mixture to quench the firefly luciferase activity and activate the Renilla luciferase activity. Triplicates were performed for each transfection study and three individual experiments were conducted to collect the data for analysis.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

A cDNA Clone, BR22, Was Identified Using the Human SP-B Promoter Proximal Region Containing TTF-1/HNF-3 Binding Sites in a Yeast One-Hybrid Library Screening Analysis

Two yeast strains, BR and BP, were generated by integrating a functional HIS3 reporter gene into yeast genome YM4271. In BP the HIS3 reporter gene is controlled by a DNA fragment from SP-B promoter 118 to 64, which includes both TTF-1 and HNF-3 DNA-binding sites. In BR, the SP-B promoter fragment is in reverse orientation with seven nucleotides shorter at the 5' end of the fragment (Figure 1). The seven-nucleotide truncation at the 5' end of the sequence may be due to the exonuclease degradation activity during the cloning procedure. However, the essential elements for TTF-1 and/or HNF-3 binding factors remain intact in BR. Both strains have the reporter HIS3 gene controlled by a single copy of SP-B promoter sequence. These two yeast strains were used as host to screen a human lung cDNA expression library for interacting protein(s). Several positive clones were isolated and their plasmids were recovered for DNA sequence analysis. One clone, GAD-BR22, has an open reading frame in-frame fused to the GAD domain. This clone contains a partial cDNA encoding a novel polypeptide with apparent bipartite nuclear localization signals. The growth of BR and BP yeast retransformed with GAD-BR22 on a selection medium plus 15 mM 3-AT confirmed the capability of BR22 binding to the promoter region of HIS3 gene in both BR and BP.

The cDNA sequence is shown in Figure 2. Both 5' and 3' end sequences were obtained by using a rapid amplification of cDNA ends (RACE) technique. The potential translation initiation codon, ATG, starts not far from the beginning of cDNA at +12 position. Two different reverse transcription enzymes and messenger RNA sources were used in this study to extend the 5' end sequence and both yielded similar results. It is possible that this specific message has a short 5' flanking untranslated region (UTR). At the 3' UTR, an AATAAA polyadenylation site is present at 18 nucleotides before the poly (A) track. This cDNA encodes a polypeptide with nuclear localization signals. There is no evidence of an N-terminal signal peptide sequence or membrane retention signals for cytoplasmic location. Three potential bipartite nuclear localization signals (NLS) were identified within the deduced amino acid sequence (Figure 2C). Several potential phosphorylation sites are observed in this polypeptide (Figure 2B).

View larger version (77K): [in this window] [in a new window]   Figure 2.   Sequence of BR22 cDNA. (A) The cDNA sequence of BR22. Isolated cDNA clone GAD-BR22 sequence is in upper case. DNA sequence obtained by the RACE method is shown in lower case. The potential translation initiation and termination codons, as well as the polyadenylation signal, are as marked. (B) Deduced polypeptide sequence of BR22. The potential phosphorylation sites are shaded. (C) Locations of bipartite nuclear localization signals.

Northern blot analysis indicated that a 1.3-kb message was present in all tissues surveyed; however, lower amounts of message were seen in some tissues, including the lung (Figure 3A). Because BR22 was isolated originally from a lung cDNA library, we assumed that message for BR22 should be in the lung as well. Using RT-PCR analysis, BR22 was detected in all of the tissues examined as well as in the lung carcinoma cells H441 (Figure 3B). As a control, G3PDH and SP-B messages were also evaluated by RT-PCR analysis. Only the SP-B message was detected in the lung and H441, while the rest of tissues examined were negative, as expected. For RT-PCR control, different sets of RT reaction were performed without adding RT, and the following PCR amplification reactions were carried out as described in the experiment. They were all negative. Therefore, the results clearly indicated that amplified messages were not from the genomic genes. Genebank searches for homologues of this gene revealed a matching sequence, but with no known function.

View larger version (32K): [in this window] [in a new window]   Figure 3.   Northern blot and RT-PCR analyses of BR22. (A) Northern blot of various human tissues RNA was probed with labeled BR22 (top) and -actin probes (bottom). Relative size markers for RNA are as labeled (in kb). (B) The message of BR22 in human lung, skeletal muscle, placenta, and H441 cells was detected by RT-PCR analysis. Primers used for PCR reaction are listed in MATERIALS AND METHODS. DNA fragments obtained after the amplification were analyzed in 1% agarose gel. The messages detected in each sample are labeled at the top. PCR amplification cycles used for detecting different message are also indicated.

BR22 Interacted with TTF-1 In Vivo in a Yeast Two-Hybrid System

Because TTF-1 plays an essential role in activating the SP-B promoter expression in the lung, we explored the possibility of forming a protein complex containing BR22 and TTF-1 using a yeast two-hybrid analysis. Because the carboxy-terminal portion of TTF-1 after the homeodomain was not required for the transactivation activity (25), a short form of TTF-1 without the carboxy-terminal domain (NH) was used. As shown in Figure 4, using a yeast two-hybrid analysis, colony #2-embedding plasmids with chimeric genes of GBT-NH and GAD-BR22 not only grew on the histidine-deficient plate, but also actively expressed -galactosidase (Figures 4A and 4C, respectively). While both HIS3 and LacZ gene promoters included GAL4 DNA-binding sites, the promoter sequences flanking the binding sites were different. Therefore, it is unlikely that activation of both genes is the result of a false positive owing to a nonspecific contact with the promoter. Colony #1 was a negative control showing that in the absence of TTF-1 the GAD-BR22 alone could not activate the reporter genes. This result also suggests that the binding of BR22 to the SP-B promoter has sequence specificity. Colony #3 was a clone testing the interactions between GAD-BR22 and an irrelevant nuclear protein. Both reporter genes were not activated. Thus, it is clear that not every GBT-fusion protein is able to interact with BR22 in yeast. Colonies #4 and #5 were positive controls using proteins previously shown to have direct interactions in yeast two-hybrid analysis (31, 32). They actively expressed both reporter genes.

View larger version (79K): [in this window] [in a new window]   Figure 4.   In vivo analysis of protein-protein interactions between BR22 and TTF-1. The cDNA clone GAD-BR22 was cotransformed with GBT-NH into yeast strain HF7C and their reporter gene activities were analyzed. Individual clones transformed with plasmid vectors are as labeled. (A) Test of HIS3 gene activity on a triple selection (Trp Leu His) plate. The growth of colonies on the selection medium requires the HIS3 reporter gene expression and suggests the occurrence of a protein-protein interaction in vivo. (B) Transformed yeast grew on a double selection (Trp Leu) plate. Colonies represent the presence of both DNA-binding domain and activation-domain vectors in yeast. (C) -galactosidase activity was detected by the colony-lift filter assay. Blue colors suggest the positive expression of -galactosidase reporter gene. Positive controls are the interaction between NonO and PSF (#4), and between p53 and SV40 large T antigen (#5).

It is important to show that GBT-NH, the bait, is not self-activating in vivo in the yeast two-hybrid system. This is essential inasmuch as the N-terminal domain of TTF-1 has been shown to possess transactivation capability (25). To clarify this issue, GBT-NH alone was transformed into HF7c to test its ability to stimulate reporter gene activity. Although yeast embracing GBT-NH was able to grow on the histidine selection plate without 3-AT, an inhibitor of histidine synthase, the colony-lift filter assay of -galactosidase activity was shown to be negative (data not shown). Further, 5 mM of 3-AT can completely suppress the growth of GBT-NH transformants (Figure 5). In our system, 15 mM 3-AT was routinely added to the medium to reduce the HIS3 background expression due to the "leaky" promoter. In summary, neither GBT-NH nor GAD-BR22 alone transactivated the promoters in our yeast two-hybrid system. Therefore, we conclude that BR22 and TTF-1 can form a protein complex in vivo in yeast.

View larger version (131K): [in this window] [in a new window]   Figure 5.   Test of the transactivation activity of GBT-NH in yeast. The plasmid GBT-NH was transformed into yeast strain HF7c and then streaked the yeast on the selection plate without tryptophan (-W) or histidine (-H) for its growth test. (A) Yeast grew on the tryptophan-deficient plate. Yeast colonies appeared after 2 d of incubation at 30°C. (B) Yeast grew on the histidine-deficient plate after 3 d of incubation. (C) and (D) are cells under the conditions of histidine deficiency plus various amount of 3-AT inhibitor to suppress the HIS3 leaky promoter expression.

BR22 and TTF-1 Formed a Protein Complex In Vitro

BR22 (amino acid 6-206) was constructed and expressed as a fusion protein with GST in Escherichia coli. Recombinant proteins expressed in bacteria were purified to homogeneity through affinity column chromatography (Figure 6A). Purified GST or GST-BR22 was bound to glutathione agarose beads and then incubated with in vitro-translated, ³⁵S-methionine-labeled TTF-1 or luciferase. As shown in Figure 6B, lane 4, TTF-1 formed a complex with GST-BR22 and was retained on agarose beads. However, it did not remain on beads containing GST alone (Figure 6B, lane 5). The interaction between BR22 and TTF-1 was specific because luciferase, an irrelevant protein control, did not remain on the GST-BR22 beads (Figure 6B, lane 6). In another experiment using anti-HA antibody to immunoprecipitate in vitro-translated ER-BR22 or ER, ³⁵S-labeled TTF-1 could be coimmunoprecipitated with ER-BR22 but not with ER (Figure 6C, lanes 10 and 9, respectively). A minor band seen with the ER-alone beads was caused by the background interactions because the signal intensity was not comparable to that with ER-BR22. These in vitro results confirm a direct protein-protein contact between BR22 and TTF-1.

View larger version (30K): [in this window] [in a new window]   Figure 6.   In vitro protein-protein interactions between BR22 and TTF-1. A total of 30 µg of affinity-purified recombinant GST-BR22 or GST was bound to glutathione beads and then incubated with 15 µl of in vitro-translated, 35S-methionine-labeled TTF-1 or luciferase overnight at 4°C. After washing the beads, retained proteins were eluted and analyzed by 10% SDS-PAGE. (A) Coomassie staining of eluted proteins fractionated in SDS-PAGE gel. The proteins bound to agarose beads are labeled at top. (B) Proteins obtained as in A were separated in SDS-PAGE gel, transferred to membrane, and exposed to an X-ray film to detect labeled TTF-1 (lanes 4 and 5) or luciferase (lane 6). The proteins that bound to agarose beads are labeled at top. Lane 7 is 1 µl of TTF-1 (filled arrow). Lane 8 is 1 µl of luciferase (open arrow). (C) In vitro-translated ER or ER-BR22 proteins were bound to the magnetic beads coupled with anti-HA tag antibody. The affinity beads were then incubated with 35S-methionine- labeled TTF-1 overnight at 4°C. Protein complexes retained on the beads were fractionated, transferred, and exposed to film as described earlier.

BR22 and TTF-1 Coimmunoprecipitated from Transfected Mammalian Cell Lysates

In addition to data collected in in vitro and in vivo in yeast studies, BR22 and TTF-1 protein complex also could be found in mammalian cell lysates. Mammalian expression plasmids containing chimeric gene constructs of HA-tagged ER-BR22 or GFP-NH were cotransfected into 293 cells. Cell lysates were harvested 48 h after the transfection for coimmunoprecipitation analysis. As shown in Figure 7, GFP-NH coimmunoprecipitated with HA-tagged ER-BR22 by 12CA5. Western blot analysis using a polyclonal antibody to GFP demonstrated the presence of GFP-NH in complex with ER-BR22 (Figure 7, lane 1), but not with GFP (Figure 7, lane 4). This result demonstrates the formation of BR22/TTF-1 complex inside the mammalian cells.

View larger version (52K): [in this window] [in a new window]   Figure 7.   Coimmunoprecipitation of ER-BR22 and GFP-NH from mammalian cells. Embryonic fibroblast 293 cells were cotransfected with HA-tagged ER-BR22 and GFP-NH (lanes 1-3) or ER-BR22 and GFP plasmids (lanes 4-6). After 48 h of transfection, the cells were harvested, lysed and prepared for whole-cell lysates. Aliquots of each lysate were immunoprecipitated with 12CA5. Immunocomplexes restrained on the protein A-agarose beads were separated from the supernant and washed, and the precipitant was fractionated by 10% SDS-PAGE. Protein compounds consisting of the complex were diagnosed by Western blot analysis using anti-GFP antibody to probe for the presence of GFP-NH or GFP. The full-length form of GFP-NH is denoted with an arrow and is present in both washed beads (lane 1) and the original lysates (lane 3). The supernatant left from the coimmunoprecipitation reaction does not have the full-length GFP-NH polypeptide. In control lysates cotransfected with ER-BR22 and GFP, GFP is not retained on the beads (lane 4), but left in the supernatant (lane 5) and the lysates (lane 6).

BR22 Acted with TTF-1 to Synergistically Transactivate the Human SP-B Promoter in Transfected Mammalian Cells

The capability of BR22 and/or BR22/TTF-1 complexes to transactivate the SP-B promoter was examined in vivo in transfected 293 cells using a dual-luciferase assay. Exponentially growing 293 cells were cotransfected with plasmids containing the human SP-B promoter reporter gene (pGL3-SPBPF), Renilla luciferase reporter gene (pRL-TK), HA-BR22F, and Flag-TTF-1. Cell lysates were harvested 48 h after transfection to measure both luciferase activities. As shown in Figure 8, overexpressing either BR22 or TTF-1 resulted in a mild increase in the SP-B promoter activity. However, when both plasmids were cotransfected, the SP-B promoter activity was markedly increased. This result suggests that the complex of BR22 and TTF-1 can synergistically increase the transactivation activity of the human SP-B promoter in vivo. In BR22 and TTF-1 transfected cells, a slight increase in Renilla luciferase activity was observed; however, the magnitude of increase was trivial compared with the SP-B promoter-driven firefly luciferase. Analysis by t test indicated that the increased firefly luciferase activity in the BR22 and TTF-1 cotransfected cells was significantly increased compared with the TTF-1 transfected cells. In contrast, there was no significant difference in the Renilla luciferase activity of the BR22 and TTF-1 cotransfected cells.

View larger version (10K): [in this window] [in a new window]   Figure 8.   Human SP-B promoter can be synergistically transactivated by BR22/TTF-1 complex in vivo. Plasmids containing HA-BR22F or Flag-TTF1, or empty vectors as indicated in the figure, were cotransfected with pGL3-SPBPF (P. pyralis luciferase) and pRL-TK (Renilla luciferase) plasmids into 293 cells. Whole-cell lysates were harvested 48 h after transfection for a dual-luciferase assay. Both firefly luciferase (open bar) and Renilla luciferase ( filled bar) activities in the cell lysate were measured. The values of triplicate firefly and Renilla luciferase activities were normalized first within each set and compared with the empty vector control (pCIN-HA + pCIN-Flag) respectively to calculate the folds of luciferase activity changes. The data shown are the averages of three separate transfections plotted with standard error, and Student's t test analyzed the results. The t test shows significant increase of the firefly luciferase activities in the HA-BR22 + Flag-TTF1 cells from the pCIN-HA + Flag-TTF1 cells, P < 0.002. No significant difference of Renilla luciferase activity can be concluded between these two transfected cells, P > 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, using both yeast one-hybrid and two-hybrid systems to explore proteins interacting with the SP-B promoter complex, we identified a novel polypeptide, BR22, that not only interacted with the SP-B promoter but also formed a complex with TTF-1 through direct protein-protein interactions. We individually identified two clones with various lengths of this gene, suggesting that the interaction between BR22 and the SP-B promoter is not a random nonspecific binding activity. Nuclear localization signal sequences in BR22 polypeptide imply that it is a nuclear protein. The detailed domain motifs of BR22 that are responsible for its cellular location and/or for the DNA-binding activity on the SP-B promoter remain to be elucidated. One important set of observations in this study is that the BR22/TTF-1 interaction can be demonstrated by using proteins synthesized in mammalian cells. Because the post-translational modification of proteins in yeast may be different from that in mammalian cells it is necessary to confirm the interaction in mammalian system. The coimmunoprecipitation of BR22 and TTF-1 from cotransfected 293 cell lysates indicates that BR22 and TTF-1 form a protein complex in mammalian cells.

TTF-1 is expressed in respiratory epithelial cells including type I, type II, and Clara cells. Many lung-cell type-specific gene expressions such as the type I specific protein , surfactant proteins, and CCSP are regulated by TTF-1. However, the expression of these genes cannot be explained solely by the presence of TTF-1. SP-B is expressed in both type II and Clara cells, but SP-C is expressed only in type II cells despite the presence of TTF-1 in Clara cells. Type I-specific protein and CCSP are expressed in type I cells and Clara cells, respectively, but not vice versa. Such a differential expression of lung-cell type-specific genes cannot be attributed to the presence of TTF-1 without other factors involved.

The phosphorylation status of TTF-1 has been thought to play a key role in regulating cell-specific gene expression in the lung. For example, cotransfecting protein kinase (PK) A catalytic subunit with TTF-1 into H441 can elevate the TTF-1 transcriptional activity by the phosphorylation of TTF-1 at the Thr⁹ residue (14), which is located in the N-terminus transactivation domain of TTF-1. In another study with cyclic adenosine monophosphate induction of SP-A gene expression in A549 lung adenocarcinoma cells, the stimulation of transcription activity could be mediated by the phosphorylation of TTF-1 in vivo. DNA-binding activity increased while TTF-1 was phosphorylated (15). However, the phosphorylated TTF-1 cannot account for the activation of all the affected genes. When PKA catalytic subunit was cotransfected into Ha-ras- transformed FRTL-5 cells, the TTF-1 transactivation activity was only partially restored. The hypophosphorylated TTF-1 in Ha-ras oncogene-transformed FRTL-5 cells is still able to bind its consensus elements within the thyroglobulin promoter (33). Further, many serine phosphorylation sites clustering in the N- and C-termini of TTF-1 transactivation domains can be phosphorylated in FRTL-5 and HeLa cells without altering their dependent gene expression patterns. The discordance could be explained by the differential response to TTF-1 phosphorylation in different cell types. Other mechanisms, including the modulation of the redox state (34) or the alteration of the cellular location of TTF-1 (3), also have been suggested as contributing to the differential regulations of gene expression in different tissues.

Northern blot analysis of various human tissues for the expression of BR22 indicates that BR22 is ubiquitously expressed, though the expression level in lung and kidney is less than in other tissues. Using RT-PCR to investigate the message of BR22 unequivocally revealed its existing in the lung and H441. Many reports have shown that a ubiquitously expressed protein can participate in regulating the cell type-specific gene expression. For example, the regulation of immunoglobulin gene expression in mature B-lymphocytes by NF-B occurs through a mechanism of sequestering protein complexes in different cellular compartments. The functional NF-B protein complex is retained in cytoplasm by an inhibitor, IB, to prevent the gene activation (35). Similarly, whereas tissue-specific cofactors were required for cardiac gene expression, a ubiquitous protein, NF-AT3, was found necessary to synergistically activate responsive gene expression (36). The pharmaceutical agents that block NF-AT3 activity can alter cardiac function.

In a recent report, a cofactor of TTF-1 was uncovered using a modified yeast two-hybrid system. Calreticulin, a calcium-binding protein in cytoplasm, a ubiquitously expressed Ro/SS-A autoantigen, has been shown to enhance the transcriptional activity of TTF-1 through a protein-protein interaction with its homeodomain (37). Through the protein-protein contacts calreticulin can stabilize the homeodomain structure of TTF-1 and increase its steady-state level to promote DNA-binding activity in vitro. BR22, a novel protein bearing NLS and DNA-binding capability, also can interact with TTF-1. The protein complexes involving BR22 and TTF-1 can occur in mammalian cells and synergistically increase the transactivation activity of the human SP-B promoter. This induction is specific inasmuch as the thymidine kinase promoter cannot be stimulated by the same complex.

A slightly increased Renilla luciferase activity detected during the transfection studies might be due to the elevated SP-B promoter activity, because it also occurred inside the BR22 or TTF-1 transfected cells (Figure 8). The t test P value (P > 0.05) suggested that this increase was not statistically significant. Without the SP-B promoter-driven firefly reporter gene the Renilla luciferase activity remained at a level similar to background level in BR22, TTF-1, and both genes' cotransfected cells (data not shown). The increased Renilla luciferase activity could not be explained by higher transfection efficiency because independent experiments were done with triplicate transfections and they all yielded similar results. In conclusion, our results suggest that BR22 could be a member of the transcription complex on the SP-B gene promoter. Mechanisms by which this BR22/TTF-1 complex upregulates the SP-B gene expression could be mediated through altering TTF-1 cellular location, varying TTF-1 DNA-binding activity, or modifying the protein to activate its transactivation activity.