Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Thioredoxin (Trx), a ubiquitous 12-kD thiol protein, is found in all cells ranging from bacteria to mammals. To exert its regulatory functions in cells, Trx serves as a reducing agent for various target proteins. Several of these target proteins have been identified, including such transcription factors as nuclear factor (NF)-B (1), glucocorticoid receptor (5, 6), and Jun-Fos (7, 8), as well as other enzymes known to be involved in injury and repair, such as the peroxidases (9, 10) and DNA polymerase (11, 12).

Although Trx is common, very little is known about its gene expression regulation. Previously, it was demonstrated that Trx gene expression is regulated by heat shock factor (13), which implies that Trx is a member of the stress gene family. We have demonstrated in primary cultures of monkey tracheobronchial epithelial (TBE) cells that Trx mRNA levels were increased by vitamin A and its derivatives (retinoids) (14). This effect was both time- and dose-dependent. Furthermore, a preliminary experiment based on a nuclear run-on assay demonstrated that the regulation of Trx gene expression by retinoids may occur, at least in part, at the transcriptional level. More recently, we have demonstrated that such an elevation of TRX gene expression by retinoic acid (RA) may play an important role in the regulation of the transcription of IL-8 and TNF--induced gene expression in airway epithelium (3, 4).

The importance of retinoids in regulating the development of the respiratory system and in epithelial cell differentiation is well recognized (15). Under normal circumstances, the transcriptional regulation of genes by retinoids is mediated by the interactions between the cis-acting elements (RA response elements, RAREs) and trans-acting factors (RA receptors and retinoid X receptors). In cells, retinol can be metabolized into different kinds of RA (retinoids) (19), which can serve as ligands for their corresponding receptors. The retinoid receptors are members of the nuclear signal receptor family.

Two major groups of retinoic receptors have been identified. One group is the retinoic acid receptors (RARs), which have three subtypes designated as RAR-, RAR-, and RAR-. The second group is the retinoid X receptors (RXRs), which also have three subtypes, designated as RXR-, RXR-, and RXR-. RARs can interact with the all-trans and 9-cis retinoic acids, whereas the RXRs can only interact with 9-cis retinoic acid (20; for recent review, see Ref. 21). Furthermore, the RXRs have been reported to form heterodimers or heterotetramers with RARs, vitamin D receptor, or thyroid hormone receptor. The specific DNA sequences and structures of RAREs that can interact with those dimerized receptors have been characterized (21). The functional RAREs are composed of two repeats (also called half sites), with one, two, or five nucleotides forming a space between the two half sites. However, wider spacing (10 to 200 bp) has also been identified (22). The consensus sequences of the RARE half sites are characterized as 5'-PuG(G/T)(T/A)CA (e.g., AGGTCA or AGTTCA, in the case of Trx gene). Different nucleotide spacing between the two half sites can cause the DNA fragments to have different affinities for the dimers of nuclear receptors.

To determine the transcriptional mechanism of the effect of RA on Trx gene expression, we used an approach similar to that described previously for isolating the human Trx gene (23, 24) . We isolated a 2.7-kb DNA fragment corresponding to the 5'-flanking region of the human Trx gene. Surprisingly, DNA sequencing revealed that there are four RARE-like motifs located at the 5'-flanking region (426, 453, 507, and 626). Furthermore, these "putative" RARE sites are not spaced regularly, as previously believed. There are 22, 49, and 114 nucleotides, respectively, between the first and second, second and third, and third and fourth RARE-like half sites, respectively. It is unclear from their physical location which pairs of RARE sites are involved. In this study, we used in vivo and in vitro genomic footprinting techniques, as well as electrophoretic mobility shift assay (EMSA), to demonstrate the protein-DNA interactions at some of these sites. Using a promoter-reporter gene transfection approach, we observed the participation of these RARE-like half sites in the upregulation of Trx gene expression by RA in conducting airway epithelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation and Characterization of the 5'-Flanking Region of the Human Trx Gene

The human Trx gene was cloned and sequenced previously (23, 24). A probe corresponding to the first exon of the human Trx gene was used to screen a human genomic library constructed in EMBL3 (Clontech Laboratories, Inc., Palo Alto, CA). Five genomic clones were identified by hybridization. Based on the published restriction site map (23), two of these five clones, Hg1 and Hg5, containing a 7.1-kb HindIII-digested DNA fragment which hybridized to the first exon DNA probe, were further studied. The 7.1-kb HindIII DNA fragment was identical to the predicted DNA fragment of part of the human Trx gene located between the upstream 5'-flanking region and a portion of the second intron. Further restriction enzyme digestion with HindIII and BamH1 yielded a 2.7-kb DNA fragment, which also hybridized to the first exon DNA probe. This DNA fragment was cloned into the pGEM-4Z (Promega, Madison, WI) for further DNA sequencing.

DNA Sequence Analysis

Human genomic clones were initially sequenced based on the di-deoxynucleotide DNA sequencing technique (25), and later by automated DNA sequencing, using the ABI Prism Model 377 Automated DNA sequencer (Applied Biosystems, Foster City, CA). Various primers corresponding to different 5'-flanking regions of the Trx gene were constructed for DNA sequencing. The sequencing data was analyzed and compared with existing data in GenBank using the software of GeneWorks (IntelliGenetics, Mountain View, CA).

Sources of Cells and Culture Conditions

Primary human tracheobronchial epithelial (TBE) cultures were prepared from tissues obtained from the Medical Center of the University of California at Davis (Sacramento, CA) and the Anatomic Gift Foundation (Laurel, MD) with consent, as previously described (26). All procedures were reviewed and approved by a campus Committee on the Protection of the Rights of Human Subjects at the University of California at Davis. These primary TBE cultures were maintained in a serum-free F12 medium supplemented with insulin (5 µg/ml), transferrin (5 µg/ml), epidermal growth factor (10 ng/ml), dexamethasone (0.1 µM), cholera toxin (20 ng/ml), and bovine hypothalamus extract (15 µg/ml). For retinoid treatment, all-trans RA was added to the culture medium at a level of 30 nM unless specified otherwise in the experiment. Cultures at 60-80% confluence were used for various studies including transfection, nuclear extract preparation, and genomic footprinting.

An immortalized human bronchial epithelial cell line was also used in this study for various transfection assays. This cell line, BEAS-2B (S clone), was generated from a primary culture of human bronchial epithelial cells, which were immortalized with the SV40 T antigen (27). The cell line was further subcloned into either serum-sensitive (S clone) or -resistant (R clone) clones and only the S clone was used in this study. This cell line was maintained in the serum-free, hormone-supplemented medium, as described for primary human TBE cultures.

Preparation of Promoter-CAT Reporter Gene Chimeric Construct and Transfection

The 2.7-kb DNA fragment of the BamH1 and HindIII digested human Trx genomic clone (from Hg5 clone) was directly cloned into the cloning sites of a promoterless CAT reporter gene vector (pBL-CAT3). This construct was designated as 2700-CAT3. Using this DNA construct as a template, various chimeric constructs containing deletions and mutations in the Trx promoter region were polymerase chain reaction (PCR)-amplified using appropriate primers. The nature of these chimeric constructs in various deletions and mutations was confirmed by DNA sequencing.

For transient transfection studies, chimeric DNAs were purified with Qiagen columns (Qiagen, Valencia, CA) (28). Transfection with Lipofectin-mediated gene transfer (Life Technologies, Inc., Carlsbad, CA) was performed as previously described (28). Five micrograms of DNA were used per transfection on each 60-mm dish with confluence at 60-80%. Each transfection included a control, pSV--gal reporter gene construct (1 µg/transfection), for normalizing the transfection efficiency. Three days after transfection, cell cultures were harvested for assaying the activities of reporter genes (29).

RA nuclear receptor expression clones were used in the cotransfection study. The RAR- expression clone was generously provided by R. Evans (The Salk Institute, La Jolla, CA), and the RAR-, RAR-, and RXR- clones were generously provided by P. Chambon (CNRS/INSERM/ULP, Reims Cedex, France). These clones were purified with Qiagen columns and used at 1 µg per transfection. A vector DNA, pSG5, without any inserted gene, was used to balance the total DNA per transfection.

Reporter Gene Assays

Cell extracts were prepared in three cycles of freeze-thaw in 0.25 M Tris-Cl buffer at pH 8.0. CAT reporter gene activity was determined by an ELISA method as previously described (29). -Galactosidase reporter gene activity was assayed according to instructions provided with the Promega kit. For each transfection, relative CAT activity was expressed after normalization with -galactosidase activity. The results were averaged from at least three dishes from two separate cultures. Significance was determined relative to the corresponding control by Student's t test (P < 0.05).

In Vivo Dimethyl Sulfate Genomic Footprinting

In vivo genomic footprinting was based on a protocol described by Mueller and Wold (30) with a slight modification as described previously (29). Briefly, cultures at 70-80% confluence were incubated with 0.1% dimethyl sulfate (DMS) in a serum-free culture medium at room temperature for 2 min. DMS selectively modifies guanine (G) residues, unless they are in contact (protected) with trans-acting factors. After rinsing with regular medium, cells were lysed and DNA pellets were prepared as previously described (29). DNA pellets were treated with 10% piperidine at 90-92°C for 30 min to cleave the methylated G residue. These modified G residues were mapped using a ligation-mediated PCR genomic DNA sequencing method, with primers near the TRX promoter region (listed in Figure 7A), as described previously (29). Briefly, oligonucleotide primers near these putative RARE-like half sites of the TRX gene were used to extend to the end of all modified G-terminating genomic fragments. A common linker was then ligated to the double-stranded fragments, after which a linker-specific primer and a primer nested within the original primers (e.g., P2, or R2) were used for PCR amplification. Finally, the third nested primer (e.g., P3, or R3), prelabeled with ³²P by polynucleotide kinase, was used to identify these PCR-amplified DNA fragments. The labeled DNA fragments were fractionated on a sequencing gel. A similar approach was used to methylate purified naked DNA in vitro with DMS, and this was processed as described above. The pattern of the in vitro methylated naked DNA was also included in the above DNA sequencing gel as a control.

View larger version (27K): [in this window] [in a new window]   View larger version (48K): [in this window] [in a new window]   Figure 7.   In vivo DMS genomic footprinting analysis of putative RARE-like half sites involved in RA-stimulation. Primers used in the footprinting are listed in (A) and their locations in the 5'-flanking region of Trx gene are indicated. (B) Genomic footprinting analysis was performed with DMS as described in the text. Filled circles indicate hyper-sensitive sites, whereas open circles indicate the protected area. The nt numbers correspond to the sequence of the Trx promoter region. For footprinting on the noncoding strand, primers R1, R2, and R3 were used. For footprinting on coding strand, another set of primers P1, P2, and P3 was used. N or lane 1: protein-free human DNA extracted from primary human airway epithelial cultures; -A or lane 2: primary human airway epithelial cells grown in vitamin A-depleted culture condition; +A or lane 3: primary human airway epithelial cells grown in RA-supplemented medium.

Preparation of Nuclear Extracts

Nuclear extracts were prepared as previously described (3). Briefly, cultured cells were collected and suspended in a hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.2 mM phenylmethyl sulfonyl fluoride [PMSF], 0.5 mM dithiothreitol [DTT]). The swollen cells were homogenized and nuclei were precipitated by centrifugation. Nuclei were gently treated with a high salt solution (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl₂, 1.2 M KCl, 0.2 mM ethylenediamine tetraacetic acid [EDTA], 0.2 mM PMSF, and 0.5 mM DTT) to release proteins from nuclei without lysing them. The nuclear protein extracts in this high salt solution were then dialyzed against a moderate salt solution (100 mM KCl). Precipitate was removed by centrifugation (10,000 × g for 30 min). Protein concentration of the extract was determined by the Bradford method using bovine serum albumin (BSA) as standard.

DNase I Footprinting

With a slight modification, the procedure described in "Current Protocols in Molecular Biology" (31) was used. Briefly, 5' end-labeled ³²P-DNA (40,000 cpm per reaction) was incubated with nuclear extract proteins in a final volume of 200 µl, containing 10 mM Tris-Cl (pH 8.0), 5 mM MgCl₂, 1 mM CaCl₂, 2 mM DTT, 50 µg/ml BSA, 2 µg/ml salmon sperm DNA, and 100 mM KCl. After incubation, 5 µl diluted DNase I solution (1:500 of 10 mg/ml DNase I in DNase I dilution buffer) was added. The DNase I treatment condition varied depending not only on the supplier but also on the storage and assay conditions. A 700-µl volume of "DNase I stopping solution" was added and DNA was phenol-extracted and ethanol-precipitated. The "DNase I stopping solution" contained a mixture of the following three components: 100% ethanol, tRNA stock solution (1 mg/ml), and saturated ammonium acetate in a ratio of 645:5:50. Precipitated DNA was gently suspended in 5 µl formamide loading dye buffer and applied to a sequencing gel for electrophoretic separation. After electrophoresis, the gel was exposed to X-ray film, and the autoradiograph was developed by a standard X-ray film developing procedure.

EMSA

The DNA-protein binding was performed in a 20-µl reaction volume containing 25 mM HEPES, pH 7.9, 10% (vol/vol) glycerol, 30 mM NaCl, 0.1 mg/ml BSA, 1 mM DTT, 5 mM MgCl₂, 0.5 mM EDTA, 0.5 µg of salmon sperm DNA, and 50-100 ng of poly(dI-dC). After incubation on ice for 10 min with 3-5 µg nuclear extract, 0.1-0.5 ng of ³²P-end-labeled probe (~ 20,000 cpm) was added and the result was incubated at room temperature for 15-20 min. The DNA-protein complexes were resolved in a native 4% acrylamide gel (30:1 ratio of acrylamide to bis-acrylamide) in 0.5× Tris borate-EDTA buffer. For antibody supershift analysis, nuclear extracts were mixed with 1 µg of anti-RAR- antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and incubated on ice before being added to the labeled DNA probe. For competition, cold consensus -RARE (Santa Cruz Biotechnology) of the human RAR- gene was added to the reaction mixture before adding the labeled probe.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of RARE-like Half Sites in the 5'-Flanking Region of the Human Trx Gene

Based on a nuclear run-on study, we previously demonstrated that retinoids stimulate Trx gene expression at the transcriptional level (14). To elucidate the nature of this stimulation, a 2.7-kb DNA fragment corresponding to the 5'-flanking region of human Trx gene was isolated and sequenced. DNA sequencing revealed that the precise size of the 2.7-kb DNA fragment is 2,659 bp and that it contains the nucleotide sequence of the 5'-end of the human Trx gene between +28 nt and 2631 nt (GenBank Accession No. AF084048). Sequence analysis revealed the presence of a TATA box at 28 and several "putative" transcriptional regulation sites, including four RARE-like half sites at 426, 453, 507, and 626. The number of nucleotides between those RARE-like half sites varies from 22 to 49 to 114 bp.

This genomic 2.7-kb DNA fragment was then cloned into the promoterless reporter construct (pBL-CAT3) and this clone, 2700-CAT3, was first used in a transient transfection study to determine whether the region contains the cis-element responsible for RA-stimulated reporter gene expression. As shown in Figure 1, 2700-CAT3 contained the basal CAT activity, which was 6-fold greater than the level of the activity in the promoterless pBL-CAT3 transfected cells. RA treatment further increased CAT activity in the 2700-CAT3 transfected cells up to 16-fold. These results suggest that the 2,700-bp DNA fragment may contain elements that are responsible for the basal and RA-stimulated reporter gene expression.

View larger version (28K): [in this window] [in a new window]   Figure 1.   Effects of RA on the expression of reporter gene CAT activity in 2700-CAT3 transfected cells. BEAS-2B cells at 60% confluence were transfected with 5 µg/dish of 2700-CAT3 or CAT3 plasmid DNA by a liposome-mediated procedure as described in MATERIALS AND METHODS. To normalize the transfection efficiency, pSV--gal plasmid DNA (1 µg/dish) was included in the transfection. One day later, RA (30 nM) was added and cultures were further maintained for two more days before harvest. Both CAT and -gal activities were assayed as described in the text. Relative CAT activity, representing an average and standard deviation of data derived from triplicate dishes, was expressed as CAT/-gal using the value in the control CAT3 vector-transfected culture as one. Significance was determined relative to the corresponding control by Student's t test (*P < 0.05). Solid bars, untransfected; open bars, pBL-CAT3; darkly shaded bars, 2700-CAT3 (RA); lightly shaded bars, 2700-CAT3 (+RA).

Transient CAT Gene Expression Assay

To elucidate which cis-element is involved in the RA-stimulated transcription of the Trx gene, various chimeric constructs were prepared containing different deleted DNA fragments of the Trx 5'-flanking region and the reporter gene, CAT. Because the four RARE half sites are within the 671 nt region, which has the SacI restriction site for cloning, we constructed a 671-CAT3 chimeric construct. As shown in Figure 2, RA stimulated CAT activity in the 671-CAT3 transfected cells in a dose-dependent manner (0-1 µM). A similar dose response to all-trans retinol (ROH, 0-1 µM) was observed (data not shown). However, RA was more effective than ROH. RA at 0.1 µM stimulated CAT activity 4-fold, whereas it required 1 µM of ROH to achieve this level of activity. RA at 1 µM stimulated CAT activity approximately 8-fold.

View larger version (22K): [in this window] [in a new window]   Figure 2.   Dose effects of retinoic acid (RA) on the expression of CAT reporter gene in 671-CAT3 transfected cells. BEAS-2B cells were transfected with 671-CAT3 and pSV--gal plasmid DNA, as described in Figure 1. One day after transfection, various amounts of RA were added as indicated. Two days later, cells were harvested for CAT and -gal assays as described in the text. Relative CAT activity was expressed as described in Figure 1. Significance was determined relative to the corresponding control by the Student's t test (*P < 0.05).

To further elaborate the response of the Trx promoter to RA, a cotransfection approach with various nuclear receptors was performed. As shown in Figure 3, cotransfection with RAR- or RAR- generated a 2-fold increase in basal CAT reporter gene activity in the absence of RA, and this increase was not seen in cultures cotransfected with vector (pSG5), or RAR- and RXR-. RA (30 nM) addition could further enhance the reporter gene CAT activity in some of these cultures, such as a 9-fold increase in cultures cotransfected with RAR- and a 5-fold increase in cultures cotransfected with RAR- and RXR-. However, RA did not further increase the CAT activity in cultures cotransfected with RAR-. These results suggest that specific RA nuclear receptors are required to further maximize the promoter activity regulated by RA.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Effects of RA nuclear receptors on CAT reporter gene expression in 671-CAT3 transfected cells. Plasmid DNAs (1 µg/ dish each) of RA nuclear receptor expression clones, RAR-, RAR-, RAR-, RXR-, and control vector pSG5, were included in the initial transfection with 671-CAT3 DNA. RA was added at 30 nM level (open bars). The control cultures were treated with no RA (solid bars). Experiments were performed as described in Figure 1. Significance was determined relative to corresponding control by Student's t test (*P < 0.05).

We used a deletion approach (Figure 4A) to determine which RARE-like half site was involved in the transcriptional regulation of the Trx gene by RA. These deletion constructs, 671-CAT3, 514-CAT3, 464-CAT3, and 430-CAT3, containing 4, 3, 2, and 1 RARE-like half sites, respectively, were used in the transfection study. As shown in Figure 4B, the relative CAT activity increases proportionately with the number of RARE-like half sites in these chimeric constructs. This phenomenon was most pronounced when cultures were cotransfected with the RAR- expression vector. After RA treatment, there was a 17-fold increase of the relative CAT activity in cultures cotransfected with 671-CAT3 and RAR-, compared with only 7- and 3-fold increases for cultures transfected with two other deleted chimeric constructs, 514-CAT3 and 464-CAT3, respectively. For 430-CAT3, which contained only the RARE-I half site at the proximal end, RA had no stimulating effect with or without RAR- cotransfection. A similar result was obtained when RAR- instead of RAR- was included in cotransfection (data not shown).

View larger version (28K): [in this window] [in a new window]   Figure 4.   Deletional analysis of Trx promoter activity. (A) A schematic presentation of various deletion chimeric constructs. PCR DNA amplification was performed to generate serial deletions of the 5'-flanking region of the human Trx gene from +23 BamH1 cloning site to the designated nucleotide site as indicated, using 671-CAT3 as a template. These DNA fragments, from -514 to +23, -464 to +23, and -430 to +23, were cloned into pBL-CAT3 under Sac1 and BamH1 cloning sites and the constructs were named 514-CAT3, 464-CAT3, and 430-CAT3, respectively. Putative RARE-like half sites are marked by a filled box, and the TATA box is marked by an open box at -28. (B) Relative CAT activities in cultures after transfection. BEAS-2B cells were cotransfected with various deleted Trx-CAT3 constructs and control vector pSG5 DNA or RAR- as described in Figure 1. RA was added at 30 nM level to some of these cultures as indicated (lighter bars). For data comparison, the normalized CAT activity in cells transfected with 430-CAT3 control (with pSG5, RA, filled bars) was used as "1.0" in this figure. Significance was determined relative to corresponding control by Student's t test (#P < 0.05 as compared with the control transfection with 430-CAT3 [with pSG5, RA]; *P < 0.05 as compared with the control transfection with 514-CAT3 [with pSG5, RA] and 671-CAT3 [with pSG5, RA]). Open bars, pSG5, +A; darkly shaded bars, A-g, A; lightly shaded bars, A-g, +A.

Enhancer Activity on a Heterologous Promoter

To demonstrate the nature of the enhancer activity in the four RARE-like half sites, the DNA fragment (671 to 357 nt) was cloned into the minimal promoter CAT-expression vector tk-CAT2. As shown in Figure 5, when the 671/357 DNA fragment was placed upstream of the 5'-end of the tk promoter, the relative CAT reporter gene activity increased 4-fold compared with the control transfection with tk-CAT2. RA treatment further increased CAT activity 3-fold. RA had no effect on the control vector-transfected cultures. These results suggested that the DNA fragment from 671 to 357 nt promotes both basal and RA-stimulated enhancer activities. However, if the 671/357 DNA fragment was placed downstream at the 3' end of the CAT reporter gene in a reverse direction, neither basal nor RA-enhanced CAT gene expression was observed.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Enhancer activity of Trx DNA fragment on a heterologous tk promoter. (A) A schematic illustration of Trx-CAT2 constructs used. A PCR amplified DNA fragment between -671 and -357 nt of Trx 5'-flanking region was cloned into tk-CAT2 vector (a) either upstream of the tk promoter (b: 671/357-tk-CAT2) or downstream from the CAT reporter gene in the reverse orientation (c: tk-CAT2-671/357). (B) Relative CAT activities in cultures transfected with the CAT2 chimeric constructs. Cultures treated with RA (30 nM) are shown in open bars, and untreated cultures are shown in filled bars. Significance was determined relative to corresponding control by Student's t test (*P < 0.05).

To further elucidate which of the RARE-like half sites were involved in basal and stimulated enhancer activities, we performed site-directed mutagenesis on each of the four RARE-like half sites. As shown in Figure 6A, different mutations were introduced into each half site, and these DNA fragments were introduced into the tk-CAT2 construct. As compared with wild type (wt), mutations in the RARE-I or RARE-II site at the proximal end (M1 and M2) severely reduced both basal and RA-stimulated enhancer activities. Mutations in the RARE-III site (M3) had less effect, whereas mutations in RARE-IV of the most distal site (M4) had no effect.

View larger version (30K): [in this window] [in a new window]   Figure 6.   Effects of site-directed mutations in the RARE-like half sites of the Trx promoter region on the reporter gene CAT expression activity. (A) A schematic presentation of the mutations on the four RARE-like half sites. M1, M2, M3, and M4 constructs represent mutations on RARE-I, -II, -III, and -IV half sites, respectively. Nucleotide mutations are indicated in lower cases, the wild-type (wt) nucleotide sequences in capital letters. The mutated DNA fragments were cloned into tk-CAT2 vector as described in the text. (B) Relative CAT activities in cultures transfected with the CAT2 chimeric constructs. Open bars, cultures treated with RA (30 nM); filled bars, untreated control cultures. Significance was determined relative to corresponding control by Student's t test (*P < 0.05).

In Vivo and In Vitro Genomic Footprinting

To further understand the regulation of Trx gene expression in vivo, the interactions between the upstream promoter DNA sequence containing the RARE sites and transcriptional proteins were studied by the genomic footprinting method at the nucleotide level. Using appropriate primer sets as shown in Figure 7A, the DNA-protein interaction sites on the coding (primer set P1, P2, and P3) and noncoding (primer set R1, R2, and R3) strands of the Trx promoter region were mapped by the ligation-mediated (LM)-PCR method (29, 30). Genomic footprinting data of the promoter region between -633 and -386, encompassing the four RARE sites, is displayed in Figure 7B. A strong footprint at the RARE-I- like site and its flanking region was observed. The -423 G residue of the noncoding strand of the RARE-I site was partially protected, whereas -427 G residue at the coding strand exhibited hyper-reactivity in untreated TBE cells. Strong footprints were also detected at the flanking region of the RARE-I site and the region flanking between putative RARE-I and RARE-II sites. These G residues at -437 of noncoding strand and 420 and -409 of coding strand were partially protected, whereas G residues at -406 and 504 of the noncoding region and -438, -427, 418, and -417 of coding strand displayed hyperreactivity to piperidine cleavage. However, a treatment of TBE cells with RA further enhanced the protection of several G residues of the noncoding strand. These changes are at -423 G of the RARE-I site, 437 and -440 between putative RARE-I and RARE-II sites, and range from a hyperreactive to a protective footprint at -504 near the putative RARE-III site. No footprint activity was detected at or near the putative RARE-IV site of both coding (Figure 7B) and noncoding strands (data not shown). These results strongly suggest that there are interactions between trans-activation proteins and the putative RARE-I site and its flanking region.

In vitro DNase I footprinting was used to further support the physical interactions between these putative RARE half sites and nuclear RA receptors. A radioactive, labeled Trx-RARE DNA fragment (from -671 to -357 nt) was incubated with nuclear extracts obtained from primary human TBE cultures both with and without RA treatment. Diluted DNase I solution was used to cleave unprotected, sensitive sites, which were presumably not bound by nuclear proteins. It was shown that the RARE-1/II regions were protected from DNase I cleavage after incubation with nuclear extracts (Figure 8), and that RA treatment slightly enhanced the protection. RA treatment also enhanced the protection of the region corresponding to the RARE-III site, which was not seen in cultures without RA treatment. The protection was not seen for the region corresponding to the fourth RARE-like half site, regardless of RA treatment.

View larger version (52K): [in this window] [in a new window]   Figure 8.   In vitro DNase I footprinting analysis of the Trx promoter region. 32P-labeled Trx DNA from -671 to -357 was incubated with nuclear extracts prepared from cultures treated with RA (+A) or left untreated (A), and then with diluted DNase I as described in the text. C: control DNA without incubation with nuclear extract, and treated with diluted DNase I to produce the DNA ladders. The labeled Trx probe DNA (protein-free) was also treated with DMS as described in Figure 7 to produce the DNA ladder of "G" nucleotide as shown in the "DMS" column. Regions in RARE-I, -II, and -III, protected by nuclear extracts, are shown by a thick solid line on the left side. The RARE-IV region was not sensitive to DNase I digestion and is indicated as a dotted line. The molecular weight markers are indicated on the right side with arrows.

Characterization of Interactions with RA Nuclear Receptors by Gel EMSA

To better understand these DNA-protein interactions, a DNA fragment of 25 or 26 bp, corresponding to the four RARE-like half sites, was synthesized and labeled for EMSA. As shown in Figure 9, there was specific binding of these oligonucleotide probes by nuclear extracts of primary cultures of human TBE cells. The specificity of the binding was confirmed by competition with a consensus RARE DNA fragment of the human RAR- gene. It appeared that the specificity and the intensity of binding in these RARE-like half sites was in the order of I > II > III > IV. To determine whether RAR is involved in the DNA- protein interaction at these putative RARE sites, anti- RAR- antibody was used in EMSA. As shown in Figure 10, treatment with anti-RAR- antibody caused a small supershift of DNA-protein complex of ³²P-labeled trx-RAREI/II (from nt 417 to 457) probe. Treatment with nonspecific, preimmune serum did not reveal a supershift.

View larger version (117K): [in this window] [in a new window]   Figure 9.   EMSA analysis of various Trx RARE-like half sites. An equal amount of 32P-labeled DNA probes (25-26 bp as shown) for each reaction was incubated with nuclear extract proteins. The DNA-protein complex was then separated on an acrylamide gel as described in the text. The specificity of the DNA-protein complexes (upper arrow) was demonstrated by a competition with increasing amounts (from 1× to 100×) of cold DNA fragment corresponding to the RARE consensus sequence in the human RAR- gene (competitor -RARE). Free probe migrating position is shown by the lower arrow.

View larger version (50K): [in this window] [in a new window]   Figure 10.   Characterization of EMSA DNA-protein complex by anti-RAR- antibody. 32P-labeled trx-RARE I-II (from -417 to -457 nt) was incubated with the nuclear extract from RA-treated primary TBE cultures and analyzed by EMSA. Anti-RAR- antibody was added at 4 and 8 µl (columns 2 and 4, respectively) to the reaction mixture. These treatments produced an extra super-shift band with a molecular weight larger than the original DNA- protein complex. The nonspecific rabbit serum, used at the same amount of protein (columns 1 and 3), did not produce the extra band, as compared with the untreated control (column 5).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Our genomic cloning data on the 5'-flanking region of the human Trx gene is consistent with earlier findings (23, 24). In the transcription factor database, we observed the presence of four RARE-like half sites (RARE-I to RARE-IV) in the 5'-flanking region of the Trx gene which had not been previously identified. In this communication, we demonstrate a potential involvement of multiple RARE half sites, with an unusual nucleotide spacing from 22 to 114 bp, in the regulation of Trx gene expression by RA. Normally, only one pair of RARE half sites with a nucleotide spacing from 1 to 5 bp is needed for retinoid-dependent gene expression and the binding of the homo- or heterodimers of RAR/RXR nuclear receptors (for recent review, see Ref. 21). Wider spacing has been reported before (22), except in a few examples. Our results were supported by the promoter-reporter gene expression analyses (Figures 4 and 6), which demonstrated a decrease of the basal and RA-enhanced reporter gene activities by a deletion or by a mutation of one of these four putative RARE-like half sites. The EMSA study (Figure 9) also demonstrated supporting evidence that these four RARE-like half site sequences were able to interact with nuclear proteins from cultured airway epithelial cells.

However, there are differences in the effectiveness of these four RARE-like half sites in the regulation of Trx promoter activity by RA. The deletion analysis (Figure 4) demonstrated dose-dependence on the presence of these four RARE-like half sites in the regulation of Trx promoter activity by RA. The deletion of RARE-IV caused a 2-fold reduction of the promoter activity. In addition, a deletion of RARE-III and RARE-IV ablated most RA-enhanced promoter activity, suggesting a greater role of RARE-III than RARE IV in the regulation of Trx promoter activity. In the mutagenesis analysis (Figure 6), which utilized a heterologous tk promoter, we demonstrated a significant reduction in the basal and enhanced promoter activity when one of these RARE-like half sites, except the RARE-IV, was mutated. The extent of this mutational effect was RARE-I/RARE-II > RARE-III > RARE-IV.

The results of these functional analyses at the promoter-reporter gene expression level are consistent with the results of the EMSA study, and with both in vivo and in vitro footprinting data. The EMSA study (Figure 9) demonstrated that all these RARE-like half sites could interact with nuclear protein extracts in vitro. However, the intensity of the interaction was again RARE-I/RARE-II > RARE-III > RARE-IV, which is similar to the promoter-reporter gene expression analysis results. Both in vivo (Figure 7) and in vitro (Figure 8) genomic footprintings demonstrated a persistent protein-DNA interaction in RARE-I/RARE-II and their flanking region in cultured cells without RA treatment. These interactions perhaps are important for the basally enhanced promoter activity of Trx gene expression. RA treatment not only further enhanced the interaction in this RARE-I/RARE-II region, but also caused a new protein-DNA interaction on the RARE-III site. The in vivo DMS footprinting study demonstrated a change at the G residue at -504 of the noncoding strand of RARE-III from a hypersensitive site to a protective one in cells after RA treatment. A similar increase of the protection of RARE-III from in vitro DNase I digestion was observed in the culture after RA treatment. These results suggest that the basal enhancer activity involves a persistent protein-DNA interaction at RARE-I/RARE-II and their flanking region, whereas the RA-stimulated enhancer activity requires the participation of RARE-III in addition to the RARE-I/RARE-II region.

How these interactions occur in vivo is not clear. For the basal enhancer activity, the DMS footprinting data (Figure 7) showed a persistent protein-DNA interaction at the RARE-I site and its flanking region. In the RARE-I site, the G residue at -423 of the noncoding strand was protected. In the flanking region, the G residues at -409 and -420 of the coding strand and -437 of the noncoding strand were also protected. The nature of the protein interacting at these sites is not clear. Cotransfection with RAR- and RAR- could further enhance the basal reporter gene activity in transfected cells (Figure 3). In a heterologous tk promoter, RAR- cotransfection had a stimulating effect on the basal promoter activity. These results suggest a classic nuclear receptor-RARE interaction in the basal enhancer activity. In support of this concept, an EMSA study demonstrated a supershift phenomenon with anti-RAR- antibody (Figure 10). However, the level of the supershift was quite low, suggesting other transcriptional factors may bind to the region. Interestingly, there is an AP-1 site at -446 upstream of the RARE-I site and its G residue at -437 of the noncoding strand was also protected in cells without RA treatment. This result suggests a possible involvement of the AP-1 transcriptional factor in the regulation of the basal enhancer activity. It is interesting to note that Trx is involved in the activation of the AP-1 transcriptional factor (7, 8). RA nuclear receptors are capable of interacting with the AP-1 transcription factor (31). Whether any of the AP-1 gene family members are involved in this interaction requires further study.

For RA-stimulated enhancer activity, the situation is more complex, because most of the basally protected sites were further preserved and new protein-DNA interactions occurred. One noticeable change was the involvement of RARE-III in the interaction, as shown by both in vivo DMS and in vitro DNase I footprinting data. How this works in vivo is not clear at this moment. One possible explanation is that the presence of RARE-III may further enhance the DNA-protein interactions in the region of RARE-I and RARE-II half sites. This enhancement may be related to the sharing of RARs/RXRs between these two pairs, in which RARs/RXRs may transiently bind to the distal RARE site and then move on to the proximal pair for stronger binding. Alternatively, there is a new RA nuclear receptor interaction at the RARE-III site that further enhances the existing protein-DNA interaction at the RARE-I/RARE-II site. Further studies are definitely needed to elucidate the nature of these interactions.

In summary, the current study supports the findings of our previous one (13) that RA stimulates Trx gene expression at the transcriptional level. Both the footprinting data and the promoter-reporter gene expression study suggest that three of these four RARE-like half sites in the 5'-flanking region of the Trx gene participate in transcriptional regulation. Both the RA nuclear receptor cotransfection experiments and the EMSA assays demonstrate the specificity of the nuclear receptors involved in the DNA-protein interactions at these putative RARE sites. Trx is critical to a variety of biologic functions, including modulation of the activity of nuclear factor-B (1), activator protein-1 (7, 8) and several other transcriptional factors, which have been implicated in the pathogenesis of various inflammatory diseases. It is therefore possible that RA's enhancement of Trx gene expression plays a significant physiologic role in cellular response to injury and inflammation. Our recent findings that Trx participates in RA-enhanced, interleukin-8 gene expression, (3) and the tumor necrosis factor- induced NF-B activation (4), are consistent with this conclusion.