HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 51, 39981-39990, December 22, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/51/39981    most recent M007529200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ludwig, M.-G. Articles by Anglard, P. Search for Related Content PubMed PubMed Citation Articles by Ludwig, M.-G. Articles by Anglard, P. Social Bookmarking                      What's this?

Multiple Regulatory Elements in the Murine Stromelysin-3 Promoter

EVIDENCE FOR DIRECT CONTROL BY CCAAT/ENHANCER-BINDING PROTEIN  AND THYROID AND RETINOID RECEPTORS^*

Marie-Gabrielle Ludwig, Paul Basset, and Patrick Anglard^§

From the Institut de Génétique et de Biologie Moléculaire et Cellulaire, INSERM, CNRS, Université Louis Pasteur, BP 163, 67404 Illkirch Cedex, C.U. de Strasbourg, France

Received for publication, August 18, 2000, and in revised form, September 15, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Stromelysin-3 (ST3) belongs to the matrix metalloproteinase (MMPs) family, a protease family involved in tissue remodeling. Although this family of enzymes is regulated by nuclear receptors, few hormone-responsive elements have been demonstrated in MMP promoters. In order to identify regulatory elements and/or factors that control the expression of the mouse st3 gene, we have analyzed genomic sequences encompassing 5 kilobase pairs of the ST3 promoter. Analysis of these sequences revealed several CCAAT/enhancer-binding proteins (C/EBP) and retinoic acid-responsive elements (RAREs), as well as one thyroid-responsive element. However, in contrast to most MMP promoters, no AP-1-binding sites were identified. Specific binding activities were demonstrated for all elements. Consistent with previous reports, retinoid X receptor is required for maximal binding to the ST3 RAREs and the TRE. The ST3-C/EBP element was shown to mediate dose-dependent promoter activation by C/EBP. Among the RAREs, the proximal DR1-RARE was shown to be sufficient for ST3 promoter activation by ligand-bound retinoid receptors, whereas the two distal DR2-RAREs appear to be involved more in the control of base-line promoter activity. Accordingly, ST3 expression was induced by retinoic acid and was reduced in cells where specific retinoic acid receptors had been inactivated. The involvement of these conserved regulatory elements is discussed in the context of physiological or pathological situations associated with st3 expression. Our findings therefore assign to C/EBP, retinoids, and thyroid hormone important roles in the regulation of ST3 gene expression.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Matrix metalloproteinases (MMPs)¹ are a family of structurally related proteinases that play a central role in the degradation of the extracellular matrix during both normal and pathological tissue remodeling processes. ST3, a member of this family, was originally identified in the stromal compartment of a range of human invasive carcinomas of the breast (1), where its fibroblastic expression levels correlate with tumor aggressiveness and poor clinical outcome (2, 3). Since then, stromal expression of ST3 has been identified during the progression of most types of human carcinomas, including basal cell carcinomas (4), colorectal carcinomas (5), non-small cell lung carcinomas (6), and head and neck carcinomas (7) suggesting that it contributes to tumor progression. Consistent with these observations, experimental models of tumorigenesis have shown that st3 promotes both tumor uptake in nude mice (8) and homing of malignant cells, the latter being a key process in both primary tumors and metastasis (9). Furthermore, ST3 expression has been observed in atherosclerotic vascular smooth muscle cells, endothelial cells, and macrophages present in human atherosclerotic lesions, whereas no expression is detected in normal aortic specimens (10). Like other MMP family members, st3 is expressed during normal physiological processes such as wound healing (11), trophoblast implantation (12), mammary gland involution (13), cycling endometrium (14), and development (15) all of which are associated with intense tissue remodeling.

Although certain physiological processes associated with MMP expression are controlled by hormones, few hormone-responsive elements have been characterized in MMP promoters (16). In the case of st3, several experimental observations suggest that it may be regulated by hormones in vivo. For example, in Xenopus laevis, tissues such as the intestine and the tail express high levels of st3 during metamorphosis, a process that is controlled by thyroid hormone (17). In human endometrium, which undergoes monthly cycles of rapid growth and remodeling in response to ovarian steroid hormones, ST3 is transiently expressed during the proliferative stage, and there is evidence for its repression by progesterone (18).

The expression of st3 is also observed during post-lactating involution of the mammary gland, an apoptotic process associated with intense extracellular matrix remodeling (13). In addition, during development, st3 expression has been detected in the fetal interdigital mesoderm of limb bud in human (1) and mouse (15) embryos, an area associated with programmed cell death. Interestingly, the formation of interdigital necrotic zones during digit separation has been shown to be controlled by RA (19), but the mechanism by which retinoic acid regulates the mouse ST3 gene in this process has not been addressed.

The expression pattern of mouse st3 suggests that it may be under either the direct or indirect control of hormones and/or vitamin A derivatives. In order to identify regulatory elements and/or factors controlling the expression of st3 in both physiological and pathological conditions, we first isolated and characterized its promoter. We have identified a number of responsive elements and provide evidence for direct control of st3 by C/EBP, thyroid hormone, and retinoid receptors. Taken together, our data provide new insights into the mechanism by which st3 can be regulated in response to stimuli from hormones or vitamin A derivatives in physiological and/or pathological situations associated with tissue remodeling processes.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Culture-- COS-1 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum. All F9 cell lines were cultured on gelatin-coated flasks in Dulbecco's modified Eagle's medium containing 4.5% glucose/Ham's F-12 (1/1), supplemented with 5% fetal calf serum. F9 cells were treated for 2 days with 10⁶ M 9-cis- or all-trans-RA. F9 cell lines harboring null mutations of RA receptors were kindly provided by Dr. D. Metzger (IGBMC).

Southern Blot Analysis-- Southern blot analysis was performed using a 6.3-kb BamHI fragment (BB6, Fig. 1) and total genomic DNA extracted from mouse tail biopsies. Plasmid DNA (0.5 µg) and total mouse genomic DNA (10 µg) were digested with appropriate restriction enzymes, separated by agarose gel electrophoresis, and transferred to Hybond N⁺ nylon membranes (Amersham Pharmacia Biotech). A probe encompassing the region from 3.8 to 3.3 kb was generated by PCR amplification (5' primer, 5'-GGCGGCCGCTGAGGGTGGTGGGTGCGCAAGAGGCGG-3'; 3' primer: 5'-CGTAGCTCCGCCCTGCACTGGGATAGCTCCGGAGCC-3'). Hybridization was performed in 1% SDS, 1 M NaCl, 50% formamide, 10% dextran sulfate, 0.2 mg/ml single-stranded DNA at 42 °C for 8 h. Membranes were washed twice for 30 min in 2× SSC, 0.1× SDS at room temperature and then twice for 30 min in 0.2× SSC, 0.1× SDS at 60 °C.

DNA Sequencing of the Mouse st3 Gene Promoter-- A  genomic library generated from mouse DNA was screened with a 5' st3 cDNA probe (9, 13). Several positive overlapping EcoRI and BamHI genomic fragments were subcloned in order to sequence exon/intron junctions and 3.8 kb of the mouse st3 gene promoter. Sequencing was performed by automated cycle sequencing, and nucleic acid sequence homology searches were performed using the FASTA programs of the GCG sequence analysis package and the combined GenBank^TM/EMBL data base.

Primer Extension Analysis-- Two [-³²P]dCTP -end-labeled antisense primers, PE1 (5'-CTTTCTTCACAGGGTGATGACG-3') and PE2 (5'-GTGGCCTGGCCCGGGCCATCAGCGGCGA-3'), were annealed at 37 °C to 10 µg of mouse involuting uterus RNA or tRNA as a negative control. The hybridization buffer was 40 mM PIPES, pH 6.4, 1 mM EDTA, 0.4 M NaCl, 80% formamide. Following ethanol precipitation, hybridized primers were extended at 42 °C for 1.5 h using 40 units of avian myeloblastosis virus-reverse transcriptase in 50 mM Tris-HCl, pH 8, 8 mM MgCl₂, 30 mM KCl, 0.6 mM of each dNTP, and 1 mM dithiothreitol. Samples were then incubated with 1 µg of RNase A for 30 min at 37 °C. Extension products were analyzed on a denaturing polyacrylamide gel along with dideoxy sequencing reactions of the 6.3-kb BamHI genomic fragment, using the SEQ primer (5'-GACCTTCAGAAGAGCAGTCGGCACTCT-3').

Polymerase Chain Reaction-- PCR mixtures contained 200 µM of each dNTP, 50 pmol of each primer, 10 ng of template DNA, 5 units of Taq polymerase (PerkinElmer Life Sciences), 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, in a total volume of 50 µl overlaid with 30 µl of mineral oil. The PCRs were performed with a Thermal Cycler from PerkinElmer Life Sciences.

Construction of Mouse st3 Promoter-CAT Reporter Plasmids-- st3 promoter sequences were inserted into either the pBLCAT6 (promoterless) or pBLCAT5 vectors (20). A 0.1-kb CAT construct, containing 121 bp of mouse st3 5'-flanking sequences, was generated by subcloning a SpeI-SmaI fragment into pBluescript II SK+ (pBS) digested with XbaI and HincII. A PstI-XhoI fragment was then inserted into pBLCAT6. The 0.1-kb -CAT construct was generated by PCR from the 0.1-kb CAT template using a 3' primer located on the pBLCAT6 vector and containing an XhoI site (5'-AATCTCGCCAAGCTCCTCGAG-3'), and a 5' primer containing a PstI site and generating a 14-bp deletion (5'-GGATCTCGAAGCTTGCATGCCTGCAGGCCGGTTGAGTTGACGAGGCGGGGCGGTGGGGGGTG-3'). The PCR product was digested with PstI and XhoI and cloned into pBLCAT6. The 1-kb CAT construct was made by first subcloning a 1-kb EcoRI-SmaI promoter fragment into the EcoRI-HincII sites of pBS. The resulting plasmid was then digested with BamHI and XhoI allowing insertion into pBLCAT6 digested with the same restriction enzymes. The 1.5-kb CAT construct was generated by replacing the SphI-SpeI fragment of the 1-kb CAT construct by the 1361-bp-long SphI-SpeI promoter fragment (1492 to 176) isolated from the BB6 plasmid. The 2.4-kb CAT construct was generated by in vivo recombination between a PCR product and the 1.5-kb CAT construct. This PCR product was obtained from the 6.3-kb BamHI plasmid (BB6) using an upstream primer (5'-CATGTCTGGATCTCGAAGCTTCAGACTCAGGCTAGAGAGGT-3'), which is homologous to the pBLCAT6 vector in the first half and to a fragment of the mouse st3 promoter (nucleotides 2436 to 2417) in the second half. The 3' primer corresponding to 5'-CCTTGAACTCAGAGATCTGCC-3' was complementary to the st3 promoter region from position 1476 to position 1456. 100 ng of the PCR-amplified product was co-transfected into competent bacteria with equimolar quantities of the 1.5-kb CAT construct digested with SphI. The 3.8-kb CAT construct was made by subcloning the 2.8-kb EagI-EcoRI promoter fragment together with the 1-kb EcoRI-XhoI fragment (isolated from 1 kb CAT) into pBS. The resulting 3.8-kb EagI-XhoI fragment was then inserted into modified pBLCAT6 in which an EagI site was added to the polylinker. The 5-kb CAT construct was generated as follows; the 6.3-kb BamHI fragment in pBS was digested with EcoRI and XhoI, and the released fragment was replaced by the EcoRI-XhoI fragment from the 1-kb CAT construct. The resulting insert in pBS including 5 kb of promoter sequences was then cloned into pBLCAT6 as a BamHI-XhoI fragment.

Isolated elements were cloned as double-stranded oligonucleotides (with 5' protruding compatible ends) into pBLCAT5 digested with HindIII and XbaI. Oligonucleotides from the coding strand were as follows: 5'-AGCTTCCTGAGGTCACATTTCAATAGTA-3' for DR1-tkCAT, 5'-AGCTTCTGAAGGTCCCGAGTTCAAATCCA-3' for 3'-DR2-tkCAT, 5'-AGCTTCCTGACATACCACTTACATAGTA-3' for DR1 µ-tkCAT, 5'-AGCTTCTGAACATACCGACTTACAATCCA-3' for DR2 µ-tkCAT, 5'-AGCTTAAGTGTGGTAATGTGGGGTAATGGGA-3' for C/EBP-tkCAT, 5'-AGCTTAAGCGTGGTAGTGTGGGATAGTGGGA-3' for C/EBPµ-tkCAT; the DR1G-tkCAT, consisting of two consensus motifs separated by one G, and the hC/EBP-tkCAT were described previously(21, 22).

Cell Transfection and CAT Assays-- The calcium phosphate procedure was used to transfect cells 4 h after plating in 10-cm diameter culture dishes. CAT reporter plasmids (5 µg of empty vector and equimolar amounts of promoter constructs) were co-transfected with 0.25 µg of RAR and 0.25 µg of RXR or with 0.1 to 1 µg of C/EBP expression vectors. 0.5 µg of pCH110 (Amersham Pharmacia Biotech) was used as an internal control to normalize for variations in transfection efficiency. Bluescribe plasmid (Stratagene) was used to adjust the total amount of transfected DNA to 15 µg. The culture medium was changed 16 h after transfection, and cells were further incubated for 24 h in 5% fetal calf serum or in medium supplemented with 10⁶ M RA. The preparation of cell extracts containing 30 units of -galactosidase activity and CAT assays were performed as described previously (23).

Preparation of Cell Extracts-- Whole cell extracts from COS-1 cells transiently transfected with either the RA receptor, the thyroid hormone receptor  (TR, provided by D. Metzger, IGBMC), or the C/EBP expression vectors were prepared as follows. Cells were washed with phosphate-buffered saline and harvested in extraction buffer (20 mM Tris-HCl, pH 7.5, 2 mM dithiothreitol, 20% glycerol, 0.4 M KCl, 2 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor mixture from Roche Molecular Biochemicals). Cell membranes were then disrupted by 3 freeze-thaw cycles and pelleted. Supernatants were aliquoted and stored at 80 °C. To prepare F9 cell nuclear extracts, cells were washed with phosphate-buffered saline and harvested in ice-cold low salt buffer (10 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol, 1× protease inhibitor mixture, 0.2 mM phenylmethylsulfonyl fluoride). Cells were incubated for 10 min on ice, pelleted, resuspended in low salt buffer, and incubated for 3 min on ice. After centrifugation, the cell pellet was resuspended in high salt buffer (20 mM HEPES-KOH, pH 7.9, 26% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor) to reach a final NaCl concentration of 0.3 M. After a 30-min incubation on ice and centrifugation, the nuclear extracts were aliquoted and stored at 80 °C.

Gel Shift Assays-- Double-stranded oligonucleotides harboring 5' overhangs were generated and labeled with [-³²P]dCTP using Klenow enzyme. DR2- and DR1-RAREs, C/EBP, C/EBPµ, and hC/EBP probes are described under "Construction of Mouse ST3 Promoter-CAT Reporter Plasmids"; oligonucleotides used for TRE and TREµ were 5'-AGCTTCCAAGTGTCAGTAGAGGTCAGAAGA-3' and 5'-AGCTTCCAAGTATCAGTAGAAATCAGAAGA-3', respectively. The DR5 and the mutated DR5µ (24) were gifts from Dr. C. Rochette-Egly, and the xTRE corresponds to the TRE found in the X. laevis TR gene promoter (25). 10 µg (unless otherwise specified) of protein extracts were incubated with 50,000 cpm of labeled probe and 1.5 µg of poly(dI-dC), in 10 mM HEPES, pH 7.5, 0.5 mM EDTA, 1 mM dithiothreitol, 80 mM KCl, 0.1% Nonidet P-40, 7.5% glycerol. Antibodies against RAR (99A6) (26), RAR (810B2) (27), and RXRs (4RX1D12) (28) were provided by Dr. C. Rochette-Egly. The samples (20 µl) were incubated for 20 min at room temperature, prior to loading on a 6% polyacrylamide gel. Gels were dried and exposed for autoradiography.

RNA Extraction and Northern Blot Analysis-- Cultured cells were washed with phosphate-buffered saline, and total RNA was prepared by the guanidinium thiocyanate/phenol/chloroform extraction protocol (29). Northern blot analysis was performed with 10 µg of RNA as previously reported (22).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Isolation and Characterization of the Mouse st3 Gene Promoter-- st3 genomic clones were isolated from a  genomic library screened with a mouse cDNA probe (9). Overlapping clones including the whole gene with its 5'-flanking sequences were used to characterize exon-intron boundaries (Table I) and promoter sequences. The physical map of the mouse st3 gene and promoter is illustrated in Fig. 1A and shows a similar genomic organization to that of the human gene (23), including 8 exons and a large first intron. A 6.3-kb BamHI restriction fragment (BB6) including 5 kb of 5'-flanking sequences was subcloned into pBluescript II SK+ in order to compare its Southern blot hybridization pattern to that of total mouse genomic DNA. Both DNAs were digested with various enzymes and hybridized to a 5' probe, as indicated (Fig. 1B). As the size of the detected restriction fragments (4.8, 3.2, and 3.6 kb, respectively) were identical for both DNAs, the 6.3-kb BamHI fragment was next used for sequence analysis and generation of promoter constructs.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   A, restriction map of the genomic region including both the promoter and the gene sequences of mouse ST3. Boxes indicate exons, and restriction sites are BamHI (B), SacI (S), EcoRI (E), EagI (Eg), and NheI (N); B* is a BamHI site created during the construction of the genomic library. The 6.3-kb BamHI fragment (BB6) was used to generate promoter constructs. The bold line indicates the position of the 0.5-kb PCR probe used for the Southern blot hybridization. B, Southern blot analysis of the ST3 5'-flanking region. Ten µg of mouse genomic DNA and 0.5 µg of BB6 plasmid DNA were digested with SacI, SacI and NheI, or SacI and EagI, and hybridized with the 0.5-kb PCR probe, as indicated. Autoradiographs were exposed for 4 days for mouse genomic DNA (left panel) and 1.5 h for BB6 plasmid DNA (right panel).

Identification of the Transcription Initiation Site-- The mouse st3 gene transcription start site was determined by primer extension with RNA recovered from mouse involuting uterus, a tissue known to express high levels of st3 (9). Primer extension was carried out using antisense primers PE1 and PE2 that are located in the second (+173 to +152) and the first exon (+140 to +113), respectively (Fig. 3). The transcription start site was demonstrated by the presence of a single band of 173 bp for the PE1 primer and of 140 bp for PE2 (Fig. 2). The same extended fragments were observed by using RNA from C26 cells, a murine tumor cell line expressing st3 (data not shown). This site was found to be located 27 bp downstream of a TATA sequence and 22 bp upstream of the translation start codon (Fig. 3). Moreover, alignment of the mouse, human (23), and X. laevis (30) stromelysin-3 sequences shows that the relative position of the TATA sequence and that of the transcription start site are well conserved (Fig. 4A). However, while the mouse and human sequences are well conserved between the TATA box and the ATG, no homology was observed with the X. laevis, as this sequence harbors an extended 5'-untranslated sequence (30).

View larger version (60K): [in this window] [in a new window]   Fig. 2.   Identification of the transcriptional start site of the ST3 gene by primer extension analysis. Two end-labeled antisense primers, complementary to the PE1 and PE2 sequences illustrated in Fig. 3, were hybridized with 10 µg of mouse involuting uterus RNA (2) or tRNA (1), as a negative control. Primer-extended products were analyzed on a denaturing polyacrylamide gel along with the sequencing reactions of the 6.3-kb BamHI restriction fragment (BB6, Fig. 1) using the primer complementary to the SEQ sequence underlined in Fig. 3. The primer-extended products observed after 3 days of exposure are of 173 bp for PE1 and 140 bp for PE2 and correspond to the same transcription initiation site.

View larger version (75K): [in this window] [in a new window]   Fig. 3.   Nucleotide sequence of 2.45 kb of 5'-flanking region of the mouse ST3 gene. Numbers on the right indicate positions relative to the transcription start site +1, as determined by primer extension analysis. The translation initiation site (ATG), at position +23, is indicated in bold. Sequences complementary to primers (PE1 and PE2) used for mapping of the transcription initiation start site or in sequencing reactions (SEQ) (Fig. 2) are indicated by arrows. Relevant restriction sites and putative responsive elements are underlined; the RAREs direct repeats are indicated by two arrows. TRE indicates thyroid hormone-responsive element. The boxed sequence is conserved between mouse, human, and X. laevis ST3 gene promoters. Sequences in brackets are two long repeated regions with 84% homology.

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Homology between mouse, human and X. laevis 5'-flanking sequences. A, the proximal mouse ST3 promoter sequence is compared with that of the human and X. laevis promoter. The transcription start site (+1) is indicated by an arrow and the ATG translation start codon is shown in bold. Vertical lines indicate homologous nucleotides. Putative transcription factor binding sites are boxed. B, the sequences and positions of the mouse TRE, C/EBP. and RAREs are shown in comparison to those present in the human ST3 promoter. Nucleotides differing from previously defined consensus sequences are underlined.

Putative Responsive Elements in the 5'-Flanking Gene Sequence and Inter-species Homologies-- We next searched for putative responsive elements within the first 3.8 kb of 5'-flanking sequences of the mouse st3 gene (Fig. 3). A region of 80% homology between the mouse and the human sequences was found to be limited to the first 100 bp of the promoter (Fig. 4A). This region includes the TATA box, several GC boxes, and a 14-bp sequence (5'-AGAACAAAAGGCAG-3') that is also conserved in the X. laevis sequence. No consensus AP-1-binding site was identified in the proximal promoter region, indicating that the mouse st3 promoter, like its human homologue, differs in this respect from most MMP proximal promoters (16). Although upstream sequences of the mouse promoter did not exhibit highly conserved regions, several putative regulatory elements were identified at relatively well conserved positions (Fig. 4B). These putative elements include a thyroid hormone-responsive element (TRE), binding sites for the CCAAT/enhancer-binding proteins (C/EBP), and several retinoic acid response elements (RAREs) (Fig. 3). The TRE-like sequence at position 558 of the mouse st3 promoter (Fig. 4B) only differs by 1 bp from the TRE consensus sequence of the DR4 type, as does its human homologue present at position 740 (23). It is noteworthy that the expression of the X. laevis st3 gene has been associated with thyroid hormone-induced tadpole metamorphosis (31). We also identified two consensus C/EBP elements T(T/G)NNGNAA(T/G) (32), at positions 1816 and 1826 of the murine st3 promoter, whereas only a single copy of this element is present in the human ST3 gene promoter at position 2165 (22). In contrast to other MMPs, both the mouse and human st3 promoters are characterized by the presence of RAREs consisting of a direct repeat of the (A/G)G(G/T)TCA motif spaced by 1 (DR1) or 2 bp (DR2). However, although the position of these elements is conserved, only the human DR1- and DR2-RAREs are consensus binding sites, whereas the mouse DR1 at position 367 and the two DR2 at positions 2340 and 2141 differ by 1 base pair from the consensus sequence (Fig. 4B). In addition, these two DR2-RAREs are included in a 165-bp repeated region of 84% identity, as indicated by brackets in Fig. 3. Although the sequence of the X. laevis st3 promoter differs significantly from that of the murine promoter, we identified a conserved GAGA repeated sequence at 1191 bp on the mouse promoter (Fig. 3) which was previously identified at 72 bp in the X. laevis sequence (30).

Functional Analysis of the Mouse st3 Gene Promoter-- To determine the contribution of 5'-flanking sequences to ST3 promoter activity, we performed transient transfection assays in COS-1 cells with a CAT reporter gene driven by various lengths of the murine st3 5'-flanking region. The activities of these sequences, which range in size from 0.1 to 5 kb but with all sharing a common 3'-end corresponding to the +15 position, were then evaluated by CAT assays (Fig. 5). The largest construct, which contains 5 kb of 5'-flanking sequences, is able to drive CAT expression from the reporter construct. Sequential deletion of this construct revealed the presence of several positive and negative regulatory regions. Although both the 5-kb CAT and 3.8-kb CAT constructs exhibit only a weak promoter activity, deletion of their 5'-end which generates the 2.4-kb CAT construct is associated with a 4-fold increase in promoter activity. This is indicative of the presence of a silencing region between positions 3.8 and 2.4 kb relative to the transcription initiation site. In contrast, further deletions of this 2.4-kb promoter construct lead to a progressive decrease in CAT activity, demonstrating a progressive deletion of positive regulatory regions. The first region, from position 2.4 to 1.5 kb, includes both of the DR2-RAREs and the C/EBP elements, and the second, from position 1 to 0.1 kb, contains the TRE and an additional DR1-type RARE, indicating that these elements could participate in the control of st3 promoter activity. In contrast, no significant difference in CAT activity was observed when a region containing the GAGA repeats was deleted, although such repeats have been shown to be important for basal activity of the X. laevis st3 promoter (30). Finally, the 14-bp box which is conserved in the human, mouse, and X. laevis proximal promoters was shown to participate in minimal promoter activity, since its deletion resulted in 50% decrease in CAT expression. The relative activity of all the constructs was found to be similar in embryonal carcinoma F9 cells (data not shown), which express st3.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Analysis of mouse ST3 gene promoter activity. Various lengths of the mouse ST3 promoter were cloned into the promoterless vector pBLCAT6 (CAT), and equimolar amounts of each construct were transfected into COS-1 cells together with a -galactosidase expression vector. CAT and -galactosidase activities were determined, as described under "Experimental Procedures." The values are the means of duplicates and are representative of three independent experiments.

Binding of Thyroid Hormone Receptors and C/EBP to Their Cognate Elements in the st3 Promoter-- Given that the putative the st3 TRE- and C/EBP-binding sites are conserved and included in sequences controlling promoter activity, we next investigated by EMSA whether they could bind to their corresponding factors (Fig. 6). Gel shift assays were initially carried out with double-stranded oligonucleotides containing the putative st3-TRE and nuclear extracts from COS-1 cells transfected with TR and/or RXR (Fig. 6A). TR was found to bind to the st3 TRE-like element (lane 2), as indicated by the presence of a complex of similar size to that observed for a consensus TRE (lane 9). The binding affinity of this complex appeared to be stronger when RXR was added to TR (lane 3), indicating a higher affinity for TR/RXR heterodimers than for TR alone. This complex was shown to be specific, as it was only competed with a 10-50-fold excess of unlabeled st3-TRE (lanes 4-6) but not by a 50-fold excess of a mutated TRE (lane 7). Consistent with this observation, no binding was observed with the mutated TRE (lane 8). These results demonstrate that TR binds efficiently to the st3-TRE, and furthermore that RXRs (also called TRAP, thyroid receptor-associated proteins) appear to increase the binding activity of the TR to their cognate elements, an effect previously reported for other T3-responsive gene promoters (33).

View larger version (48K): [in this window] [in a new window]   Fig. 6.   Binding activity of the ST3-TRE and the C/EBP element. A, end-labeled oligonucleotides containing the mouse ST3-TRE (lanes 1-7), the mutated ST3-TRE (TREµ, lane 8) or a consensus TRE (xTRE, lane 9) were incubated with extracts from COS-1 cells transfected with RXR and/or TR expression plasmids, as indicated. Competition experiments were performed with a 10-, 30-, and 50-fold excess of the "cold" wild type ST3-TRE (wt, lanes 4-6) or a 50-fold excess of a mutated TRE (µ, lane 7). B, the binding activity of the ST3 C/EBP element was tested in the presence of protein extracts from untransfected COS-1 cells (lane 1) or from cells transfected with a C/EBP expression vector (lanes 2-6). Competition assays were performed with a 50-fold molar excess of cold C/EBP-binding site (wt, lane 3) or mutated (µ, lane 4). Labeled mutated C/EBP (lane 5) and human ST3-C/EBP (lane 6) oligonucleotides have been used as negative and positive controls, respectively.

The two identified C/EBP elements in the mouse st3 promoter were analyzed by using a double-stranded oligonucleotide encompassing both sites. Although no complex was observed with an extract from untransfected COS-1 cells (lane 1), complexes of the same size were observed for both the mouse (lane 2) and the human ST3-C/EBP elements (lane 6) in the presence of extracts from cells transfected with a C/EBP expression vector. This complex was shown to be specific, since it was only competed by a 50-fold excess of unlabeled mouse C/EBP (wt, lane 3) and not by the same excess of a mutated C/EBP-binding site (µ, lane 4), for which no binding was observed (lane 5). We therefore concluded that C/EBP efficiently binds to the st3-C/EBP elements, suggesting that it may regulate the expression of the mouse st3 gene expression in a manner similar to that of the human ST3 (22).

Both of the Distal DR2-RAREs and the Proximal DR1-RARE Efficiently Bind to Endogenous RAR/RXR Heterodimers-- EMSA experiments using the two distal 5'- and 3'-DR2 probes were performed (Fig. 7A) initially with whole cell extracts derived from untransfected COS-1 cells or cells transfected with RAR (, , and ) and RXR (, , and ) expression vectors. Although no complex was observed with extracts derived from untransfected COS-1 cells (lanes 2 and 5), complexes were formed with both 5'- and 3'-DR2-RAREs in the presence of all six retinoid receptors (lanes 3 and 6). The 3'-DR2 complex (Fig. 7B, lane 2) had similar electrophoretic migration properties to those observed with a consensus DR5-RARE (Fig. 7B, lane 1) and was competed by a 10-500-fold excess of either unlabeled st3 3'-DR2 (lanes 3-5) or the DR5-RARE (lanes 6-8) but not by a mutated DR5 (lanes 9-11) that has been shown previously not to bind the RA receptors (24). Hence, we concluded that the 3'-DR2-RARE complex that forms in the presence of the retinoid receptors is specific and that the binding affinity of this element is similar to that observed for the consensus DR5-RARE. Similar results were observed with the 5'-DR2-RARE (data not shown).

View larger version (46K): [in this window] [in a new window]   Fig. 7.   RAR/RXR heterodimers bind specifically to the mouse ST3 promoter RAREs. A, gel shift assays were performed using the 5'- (lanes 1-3) and 3'- (lanes 4-6) ST3 DR2-RARE probes with either whole cell extracts from untransfected COS-1 cells (lanes 2 and 5) or a pool of extracts from COS-1 cells containing the six types of RA receptors (lanes 3 and 6). B, the 3'-DR2-RARE complex observed in the presence of all 6 RA receptors (lane 2) is competed with a 10-500 molar excess of unlabeled 3'-DR2-RARE (lanes 3-5) and consensus DR5-RARE (lanes 6-8) but not with a similar excess of a mutated DR5-RARE (DR5µ, lanes 9-11). The labeled consensus DR5-RARE (lane 1) was used as a positive control. C, binding of RAR/RXR heterodimers on the DR2. The binding activity of the 3'-DR2 probe was evaluated in the presence extract from COS-1 cells transfected either with RXR (lane 1), or RAR (lane 2), or both RXR and RAR (lane 3) expression vectors. The RAR·RXR complex observed in lane 3 was not observed in the presence of antibodies raised against either RXRs (lane 4) or RAR (lane 5). D, binding of endogenous receptors on the DR2. The 3'-DR2 probe was incubated with F9 cell nuclear extracts (lane 1) in the presence of a 100-fold molar excess of unlabeled 3'-DR2 oligonucleotide (lane 2), or anti-RXRs (lane 3), or anti-RAR (lane 4) antibodies. E, binding of recombinant and endogenous RA receptors to the DR1. The DR1 probe was incubated with 10 and 20 µg of cell extracts derived from either transfected COS-1 cells (lanes 1 and 2) or from F9 cell nuclear extracts (lanes 3 and 4). The specific shifted or supershifted complexes and the unspecific bands are indicated by S, SS, and NS, respectively. X-ray films were exposed overnight (A-C) or for 2 (D) or 3 (E) days.

As RXR and RAR have previously been shown to mediate efficiently the activation of the human ST3 gene by RA (21), we investigated whether these specific receptors could also bind to the mouse 3'-DR2-RARE (Fig. 7C). We found that the binding of RXR or RAR homodimers to this element was barely detectable (lane 1 and 2, respectively), whereas a strong complex was formed in the presence of RAR/RXR heterodimers (lane 3). These data confirm that RAR/RXR heterodimers are the major functional forms regulating retinoid-responsive elements (34). As expected, this complex was not observed in the presence of antibodies raised against either RXRs (lane 4) or RAR (lane 5) and was partially supershifted.

The binding of endogenous retinoid receptors was then addressed by using nuclear extracts derived from F9 cells expressing st3 (Fig. 9). As shown in Fig. 7D, incubation of the 3'-DR2-RARE with F9 nuclear extracts resulted in a strong and specific complex (lane 1), which was competed with a 100-fold excess of unlabeled probe (lane 2). This complex was partially supershifted, or at least destabilized, in the presence of either RXRs (lane 3) or RAR (lane 4) antibodies, indicating the presence of these receptors in the complex. These observations demonstrate that endogenous RARs and RXRs present in F9 cells also bind to the st3 DR2-RARE as heterodimers. An EMSA analysis was also performed using the proximal DR1-RARE. As shown in Fig. 7E, recombinant RARs and RXRs overexpressed in COS-1 cells (lanes 1 and 2) as well as F9 cell endogenous receptors (lanes 3 and 4) bind efficiently to this proximal element in a dose-dependent manner.

Activation of the Mouse st3 Promoter by C/EBP and RA-- By having demonstrated that both C/EBP and RA receptors bind to the st3 promoter, we tested their ability to mediate promoter activation in transient transfection assays (Fig. 8, A and B). The same wild type or mutated C/EBP containing oligonucleotide sequences as those used in gel shift assay were inserted into the pBLCAT5 vector (tkCAT). We found that the activity of the tk promoter driven by the mouse st3 C/EBP was specifically induced in the presence of the C/EBP expression plasmid (Fig. 8A). Indeed, a 3-5-fold activation was observed for the C/EBP-tkCAT construct in response to increasing amounts of C/EBP expression plasmid ranging from 0.1 to 1 µg. A similar level of activation was observed when the human ST3 C/EBP element was used as a positive control. A mutation in the mouse C/EBP-binding site abolished promoter activation by C/EBP.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Transcriptional activation by C/EBP (A) and by ligand-bound RA receptors (B). ST3 promoter fragments or isolated elements were transfected in COS-1 cells together with either C/EBP (A) or with 0.25 µg of RAR and 0.25 µg of RXR (B) expression plasmids. Following transfection, cells were cultured for 24 h with or without 9-cis-RA, and CAT activity was determined. The CAT activity is expressed as fold induction by C/EBP (A) or RA (B) over control cells. The values are the means derived from three independent experiments.

The DR1- and DR2-RAREs were then studied, either as isolated elements or in the context of the st3 promoter, for their capacity to mediate promoter response to RA. As shown in Fig. 8B, a 2-fold activation in response to RA was observed for both DR1- and DR2-tkCAT constructs in the presence of RAR and RXR expression vectors, whereas a 3-fold induction was observed for the consensus DR1G-RARE (35). Similar results were obtained with RAR and RXR receptors (data not shown). In contrast, the activity of the tkCAT and the DR1µ- and DR2µ-tkCAT constructs was not affected by RA, demonstrating that mutations within these RAREs that prevent binding (data not shown) also abolish their capacity to mediate transactivation. Various lengths of the st3 promoter were then analyzed for their response to RA (Fig. 8B). Whereas the presence of RA did not affect the 0.1-kb tkCAT construct, which does not include any RARE, all promoter fragments that included at least one of the identified RAREs (1-, 2.4-, and 5-kb CAT) were activated when cells were treated with RA. The 1-kb CAT construct, containing the DR1-RARE, exhibits a 2-fold induction. Addition of the two DR2-RAREs (in 2.4- and 5-kb CAT) does not further increase the response to RA. This suggests that the DR1-RARE is necessary and sufficient to mediate induction by RA, whereas the role of the DR2-RAREs would be restricted to the control of basal promoter activity. It is worth noting that similar RA-induced activities were observed with endogenous RA receptors when transfections were performed in F9 cells (data not shown).

st3 Gene Expression Is Induced by RA and Is Altered in Cell Mutants for RARs and/or RXRs-- Consistent with the transcriptional activation of the st3 promoter, we showed that the expression of st3 in F9 cells was induced approximately 10-fold by both 9-cis-RA and all-trans-RA, as evaluated by Northern blot analysis (Fig. 9). Since the regulation of several RA-responsive genes has been shown to be mediated preferentially by distinct RA receptors in F9 cells (36), we subsequently looked for st3 expression in mutant F9 cells in which one or two of the six receptors have been inactivated by homologous recombination (37). As shown in Fig. 10A, st3 induction by 9-cis-RA was still observed in all mutant cell lines but to a lesser extent as this induction only reached 2-3-fold. Although this weaker induction clearly indicates that the presence of RA receptors is required in order to mediate an optimal induction by RA, this observation also indicates that these receptors appear to be redundant for st3 induction and may in fact substitute for each other. In addition, we noted that the base-line expression of st3 was significantly reduced in F9 mutant cells, since longer Northern blot exposure was required to detect a signal comparable to that observed in wild type cells. This expression was actually reduced by around 50% in cell lines where a single receptor was inactivated and was further decreased in double knock-out F9 cells. This was particularly apparent in RAR/ RXR/ cells in which a 4-fold decrease in st3 expression was observed (Fig. 10B). However, because this decrease results from the lack of both of the receptors and is suggestive of a qualitative requirement, it is important to note that RAR and RXR are the predominant RA receptors expressed in F9 cells (38). In this respect, it is possible that the strong decrease in st3 expression associated with their inactivation may correspond simply to the lower levels of the other receptors that are not present in sufficient quantities to substitute for RAR and RXR (34).

View larger version (26K): [in this window] [in a new window]   Fig. 9.   ST3 RNA expression in response to 9-cis-RA or all-trans-RA. Total RNA (10 µg) was extracted from F9 cells exposed to 1 µM 9-cis-RA or all-trans-RA for 2 days. ST3 and 36B4 expression were then analyzed by Northern blot, as described under "Experimental Procedures." Autoradiographs were exposed for either 4 h (ST3) or 7 h (36B4).

View larger version (42K): [in this window] [in a new window]   Fig. 10.   Analysis of ST3 gene expression in RA receptor knock-out F9 cell lines. A, ST3 response to 1 µM 9-cis-RA was evaluated by Northern blot in mutant F9 cells. Genotypes are indicated on the top of each lane. The same RNA samples were hybridized with mouse ST3 and 36B4 probes, and the transcripts were visualized by autoradiography, after 14 h of exposure for 36B4 and ST3. B, quantification of the basal ST3 RNA expression in wild type (wt) and mutant F9 cells. Relative levels of ST3 transcripts were evaluated after normalization with 36B4 RNA levels.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The mouse st3 gene has been shown to be induced in various tissue remodeling processes that are characterized by extracellular matrix turnover and/or apoptosis and that are controlled by hormones (18) and/or retinoic acid (19). However, the mechanisms by which its expression is regulated have not been addressed. In the present study, the mouse st3 promoter was isolated and analyzed to investigate this question. The transcription initiation site was mapped, and minimal promoter sequences driving transcription were identified between positions 121 and +15. This minimal promoter activity was shown to be modulated through elements contained within various upstream regions. The analysis of 5 kb of st3 promoter led to the identification of a number of regulatory elements including TRE, RAREs, and C/EBPs that are conserved between human, mouse, and/or X. laevis. We have shown that these elements are functional in binding assays and/or promoter studies and are likely to participate in the in vivo control of mouse st3 expression.

In normal physiological conditions, the X. laevis st3 has been shown to be tightly regulated by the thyroid hormone during tadpole metamorphosis. However, although a TRE of the DR4 type has been recently identified in both the X. laevis² and the human ST3 gene (Fig. 4), this element has not yet been characterized. With respect to the mouse TRE, we have shown here that it binds thyroid hormone receptors in the presence of RXR. This finding supports previous studies that have described RXRs as being members of the thyroid hormone receptor-associated proteins and that the binding affinity of TR/RXR heterodimers to TRE is higher than that of TR alone (33). However, although the conservation of this element suggests a physiological relevance, no target mouse or human cells that express st3 in response to T3 have been reported thus far. Expression of the X. laevis st3 in response to T3 during tadpole metamorphosis therefore remains the only example of the physiological expression of st3 in T3-controlled apoptotic processes (31).

In pathological situations, a number of clinical and experimental studies suggest an involvement for T3 and its receptors in carcinogenesis. Amplification or deletion of erbA genes has been described in breast cancer (39-41), and there is experimental evidence for an increased number of thyroid receptors in spontaneous rat mammary tumors (42). In this respect, it is interesting to note that the expression and secretion of stromelysin-1, which promotes mammary carcinogenesis (43), is actually regulated by thyroid hormone in cells derived from the mammary gland (44). As st3 was originally isolated from breast cancer tissue, the possible role for thyroid hormone in the regulation of this gene in vivo cannot be excluded.

As with most MMP genes, both the human and mouse st3 genes are induced by TPA (1, 9). Such TPA activation is normally mediated through TPA-responsive elements or AP-1-binding sites. However, while the position of this element is very well conserved in most MMP promoters, no AP-1-binding site was found within the mouse st3 gene 5'-flanking sequences studied here, suggesting either that this element is located elsewhere in the promoter or that an alternative mechanism is involved in st3 transcriptional activation by TPA. A recent report has demonstrated that st3 can still be induced by TPA in vivo, in transgenic mice in which a c-jun transactivation mutant dramatically represses TPA-induced AP-1 activation (45). This observation implies that alternative regulatory element(s) exist that could mediate st3 transcriptional activation by TPA. In this respect, the conserved C/EBP element that we have identified in the mouse st3 promoter appears to be a good candidate, as other studies have shown that C/EBP-binding sites mediate the activation of a number of TPA-responsive genes (46, 47), including the human ST3 gene (22). Indeed, we have shown that this element can bind efficiently to the C/EBP transcription factor, which in turn activates transcription. Interestingly, among the various physiological processes in which C/EBP has been reported to play a role (48, 49), the involution of the mammary gland for one (50) has been associated with high levels of st3 (13). Furthermore, C/EBP has been shown to control macrophage differentiation (51), a process associated with the expression of several MMPs (52, 53) including ST3 (10). Therefore, in conjunction with data from these studies, our data suggest that the requirement of C/EBP in the control of stromelysin-3 expression may be physiologically relevant.

Another aspect of st3 regulation that differs from other MMPs is its positive regulation by RA. RA has been shown to repress the expression of AP-1-regulated genes and, consistent with this observation, the expression of various MMP genes is repressed by RA (16). The present study supports further our previous findings that demonstrated that expression of the human ST3 gene is induced by RA (21). We found indeed that the human DR1- and DR2-RAREs are highly conserved in the mouse promoter. Furthermore, we have shown that all RAREs present in the mouse st3 promoter are capable of binding RA receptors. This binding occurred preferentially with RAR/RXR heterodimers and is consistent with what we and others (36) have observed for the human ST3 gene and for several RA-responsive genes, as these heterodimers are considered to be the functional units controlling RARE-dependent promoter transactivation. However, although multiple RAREs present in several gene promoters have been shown to be required for promoter activation and to cooperate for an optimal response to RA (54), we found that the proximal DR1-RARE of ST3 alone was sufficient to mediate promoter activation by RA and that this response was not further induced in promoter constructs including the two additional DR2-RAREs. When we evaluated st3 expression in cells exposed to RA, we found that it was indeed induced by both 9-cis-RA and all-trans-RA. In addition, the base-line expression of st3 was significantly reduced in cells in which RA receptors were inactivated by homologous recombination. Together, these observations support that RA receptors contribute to the regulation of st3 base-line expression and suggest furthermore that whereas the DR1 is required for the st3 response to RA, the DR2-RAREs contribute to the base-line expression of ST3.

Certain observations in vivo are supportive of physiological relevance in the regulation of st3 expression by RA. For example, the process of interdigital cell death in digit separation during embryogenesis is controlled by RA (19) and is associated with the expression of the human (1) and mouse st3 (15) and other genes such as bone morphogenic proteins and tissue transglutaminase that have previously been shown to play a role in apoptosis (13, 55). This process does not occur in mice in which RA receptors such as RAR and RAR have been inactivated (19). Interestingly, since the expression level of both the st3 and bmp-7 genes is almost undetectable in these knock-out mice, certain authors (56) have suggested that RA might increase cell death in interdigital necrotic zones through a modulation of both of these genes, thereby contributing directly to this tissue remodeling process. With respect to st3, by showing for the first time that RA receptors bind to st3-RAREs and that their inactivation alters st3 expression and response to RA, our data provide evidence for the direct transcriptional control of st3 by RA in vivo. Furthermore, one should note that certain physiological processes such as tail resorption controlled by T3 during frog metamorphosis, and mouse interdigital separation controlled by RA, are associated with cell death suggesting a role for st3 in apoptosis that has been conserved across evolution.

In summary, we have isolated the mouse st3 promoter in order to provide further insight into the regulation of the expression of this gene. As discussed here, although st3 expression has been detected in a variety of tissues (see Introduction), very little was known about the molecular mechanisms controlling its induction. Although homology between the human and the mouse st3 promoters is restricted to a limited proximal region, we have identified and characterized distal regulatory elements such as a TRE, C/EBPs, and RAREs that are well conserved and whose corresponding binding factors are known to be involved in the regulation tissue remodeling processes associated with st3 expression. Therefore, we can speculate that the direct control of st3 expression by these factors may also be of importance in vivo. Determining how the st3 gene is regulated in vivo in a cell-specific manner represents a future challenge that will be not only of basic interest but may also have interesting applications in the context of gene therapy directed against the tumor stroma.

	ACKNOWLEDGEMENTS

We thank J. C. Young, W. Mohan, and B. Wasylyk for critical reading of the manuscript and E. Guérin and D. Luo for stimulating discussions. We are grateful to O. Lefebvre (IGBMC) for kindly providing mouse ST3 genomic clones; P. F. Johnson (National Institutes of Health, Bethesda) for the C/EBP expression plasmid; D. Metzger (IGBMC) for the thyroid hormone expression vector and the F9 cell mutants; C. Rochette Egly (IGBMC) for retinoic acid antibodies; and P. Kastner (IGBMC) for RARs and RXRs expression plasmids.

	FOOTNOTES

^* This work was supported by funds from INSERM, the CNRS, the Hôpital Universitaire de Strasbourg, the Bristol-Myers Squibb Pharmaceutical Research Institute, the Association pour la Recherche sur le Cancer Project 5421, the Ligue Nationale Française Contre le Cancer, and the Comité du Bas Rhin et du Haut-Rhin, and the Fondation pour la Recherche Médicale Française.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

This work is dedicated to the memory of Professor Paul Basset.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF297645.

Deceased.

Supported by a fellowship from the Association pour la Recherche sur le Cancer.

^§ To whom correspondence should be addressed: Institut de Génétique et de Biologie Moléculaire et Cellulaire, BP 163, 67404 Illkirch Cedex, France. Tel.: 33-388-65-3420; Fax: 33-388-65-3201; E-mail: anglard@igbmc.u-strasbg.fr.

Published, JBC Papers in Press, September 18, 2000, DOI 10.1074/jbc.M007529200

² Y. B. Shi, personal communication.

	ABBREVIATIONS

The abbreviations used are: MMP, matrix metalloproteinases; bp, base pair(s); CAT, chloramphenicol acetyltransferase; C/EBP, CCAAT/enhancer-binding protein; DR1 to DR5, direct repeats of the (A/G)G(G/T)TCA motif separated by 1-5 nucleotides; EMSA, electromobility shift assay; IGBMC, Institut de Génétique et de Biologie Moléculaire et Cellulaire; kb, kilobase pair(s); PCR, polymerase chain reaction; RA, retinoic acid; RAR, retinoic acid receptor; RARE, retinoic acid responsive element; RXR, retinoid X receptor; ST3, stromelysin-3; T3, thyroid hormone; tk, thymidine kinase; TR, T3 receptor; TRE, T3-responsive element; PIPES, 1,4-piperazinediethanesulfonic acid; TPA, 12-O-tetradecanoylphorbol-13-acetate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES