Isolation and Functional Analysis of the Mouse RXR (cid:1) 1 Gene Promoter in Anterior Pituitary Cells*

The retinoid X receptor (RXR) isoform RXR (cid:1) has limited tissue expression, including brain, skeletal muscle, and anterior pituitary gland. Within the anterior pituitary gland, RXR (cid:1) expression is limited primarily to the thyrotropes. In this report, we have isolated (cid:1) 3 kb of 5 (cid:1) -flanking DNA of the mouse RXR (cid:1) 1 gene. We have identified the major transcription start site in the thy-rotrope-derived TtT-97 cells. Transient transfection studies show that a 1.4-kb promoter fragment has full promoter activity in TtT-97 cells. This promoter has much less activity in thyrotrope-derived (cid:2) TSH cells, pi-tuitary-derived GH3 somatomammotropes, and non-pi-tuitary CV-1 cells. None of these cell lines has detectable RXR (cid:1) 1 mRNA. A previous report has identified a nonconsensus direct repeat (DR-1) element in the RXR (cid:1) 2 gene promoter region that mediates stimulation of promoter activity by 9- cis -retinoic acid (9- cis mouse RXR 1. To obtain a large fragment of the 5 (cid:2) -flanking region of the RXR (cid:3) 1 gene, a (cid:4) EMBL-3 SP6/T7 mouse genomic library was screened with a nick-translated 32 P-labeled PCR product of the 5 (cid:2) -untranslated region of mouse RXR (cid:3) 1. Rapid Amplification of cDNA Ends (RACE)— 5 (cid:2) -RACE was carried out using the CLONTECH Marathon cDNA Amplification kit (CLONTECH, Laboratories). TtT-97 poly(A) (cid:3) RNA was treated with reverse transcriptase using a 52-bp oligonucleotide with a poly(dT) end.

Tissue and cell-specific expression of many genes occurs at the level of gene transcription. Within the anterior pituitary gland, transcription factors including thyroid hormone receptor (TR␤2), Pit-1, GATA-2, ETS-1, and steroidogenic factor (SF-1) have expression limited to certain cell types, leading to specific cell phenotypes and hormone production (1)(2)(3)(4)(5)(6)(7). The retinoid X receptor (RXR) 1 has three distinct isoforms encoded on separate genes. The RXR␣ and RXR␤ isoforms are widely distributed throughout the developing embryo and in the adult, while the RXR␥ isoform has restricted tissue distribution, which includes a high level of expression in the pituitary gland early in development (8). RXR␥ has two isotypes, RXR␥1 and RXR␥2, which are generated from the same gene and differ at the N terminus (9). RXR␥1 and RXR␥2 gene expression is believed to be controlled by two separate regulatory regions. The RXR␥2 coding region is entirely contained within RXR␥1 which has an additional 123 amino acids on the N-terminal end (9). Our group and others (9,10) have shown that the RXR␥1 subtype is expressed in the anterior pituitary and has expression that is restricted to thyrotropes. RXR␥1 is not required for thyrotrope development or basal expression of the TSH␤ subunit promoter within the thyrotropes, but does appear to be required for vitamin A-mediated suppression of TSH␤ promoter activity within thyrotropes seen both in vitro and in vivo (10,11). The human RXR␥2 gene 5Ј-flanking region has been identified and contains a retinoid-responsive element, which confers positive regulation of promoter activity by retinoic acid (12). In this report, we have isolated the RXR␥1 promoter and begun to characterize its regulation in thyrotropes. An understanding of RXR␥1 promoter function in thyrotropes and other cell types will provide insight into mechanisms governing thyrotropespecific gene expression and phenotype, as well as the generalized mechanism of cell-specific gene expression.

MATERIALS AND METHODS
cDNA and Genomic Library Screening-A cDNA phage library was constructed from mouse thyrotrope tumor (TtT-97) poly(A) ϩ RNA as described previously (13). The library was screened with a nick-translated 32 P-labeled NotI cDNA fragment of the mouse RXR␥1 coding region (plasmid kindly provided by R. Evans). This was performed to obtain 5Ј-untranslated sequence of mouse RXR␥1.
To obtain a large fragment of the 5Ј-flanking region of the RXR␥1 gene, a EMBL-3 SP6/T7 mouse genomic library (CLONTECH Laboratories, Palo Alto, CA) was screened with a nick-translated 32 P-labeled PCR product of the 5Ј-untranslated region of mouse RXR␥1.
Rapid Amplification of cDNA Ends (RACE)-5Ј-RACE was carried out using the CLONTECH Marathon cDNA Amplification kit (CLONTECH, Laboratories). TtT-97 poly(A) ϩ RNA was treated with reverse transcriptase using a 52-bp oligonucleotide with a poly(dT) end.
Following second strand synthesis, double-stranded cDNA was bluntended with T4 DNA polymerase I and ligated to an adapter primer supplied with the Marathon kit. Adapter ligated cDNAs were amplified by PCR using an antisense oligonucleotide corresponding to the translational start (complement of ATG in bold) of RXR␥1 (5Ј-CCATACAT-GTTGGCTGCTCAGTT-3Ј) and a sense oligonucleotide within the 5Ј adapter (AP-1). PCR was carried out as follows: 94°C for 1 min, 30 cycles of denaturation at 94°C for 30 s, annealing at 65°C for 30 s, and extension at 68°C for 3 min, followed by extension at 72°C for 7 min using TaKara LA Taq polymerase (Roche Molecular Biochemicals). Amplified products were size-separated on a 1.5% agarose gel, transferred to a Nytran membrane, and hybridized with a 20-bp oligonucleotide corresponding to a region just upstream of the translational start of RXR␥1 to verify the size and specificity of the amplification. PCR products were excised the appropriate section of the gel, purified, and subcloned into the pCR2.1 vector (Invitrogen) for sequencing.
RNase Protection-A 431-bp NsiI/HindIII fragment of the RXR␥1 5Ј-flanking region, purified from the 3.1-kb XhoI/NotI 5Ј-flanking region in pGL2-basic, was subcloned into pGEM3zfϩ (digested with PstI/ HindIII). The resultant plasmid was linearized with BamHI, and a radiolabeled antisense RNA probe was generated using SP6 RNA polymerase and [␣-32 P]dUTP (800 Ci/mmol). A positive control RNA (unlabeled sense strand) was generated using a product from 5Ј-RACE subcloned into pGEM7zfϩ and linearized with NcoI. RNA was generated using SP6 RNA polymerase, and the product was treated with RQ DNase I to eliminate contaminating DNA. The resultant product, when hybridized with the 431-bp radiolabeled antisense probe and digested with RNase, should generate a 177-bp protected product. The protection assay was carried out as outlined in the Ribonuclease Protection Assay Kit (Ambion). Briefly, 5 g of TtT-97, ␣TSH, or GH3 poly(A) ϩ RNA was precipitated with 3 ϫ 10 6 cpm of antisense riboprobe. The pellets were resuspended, and hybridization of the probe and RNA was carried out at 42°C overnight. The annealed products were treated with a 1:5 dilution of RNase (A ϩ T1) for 30 min at 37°C, precipitated, and the digested products were size-separated on a 5% denaturing polyacrylamide gel and exposed to radiographic film.
Reverse Transcription-PCR of RXR␥1 5Ј-Untranslated Region-5 g of TtT-97 poly(A) ϩ RNA was treated with avian myeloblastosis virus reverse transcriptase (Promega) and random hexamer oligonucleotides. Each DNA product from 0.5 g of poly(A) ϩ RNA was subsequently used for each PCR reaction with a common 3Ј antisense oligonucleotide (5Ј-GCGAATTCCATGGCAGAAGTGATGACT-3Ј) corresponding to sequence in exon 2 of the RXR␥1 coding region and different 5Ј sense oligonucleotides corresponding to regions upstream and downstream of the transcription start site identified by RNase protection analysis Fig. 3). After 35 cycles (94°C for 1 min, 50°C for 1 min, 72°C for 1 min), products were size separated on an agarose gel, transferred to nitrocellulose, and hybridized with a radiolabeled oligonucleotide corresponding to a common region in all potential products (ϩ64/ϩ84). The 3.1-kb XhoI/NotI DNA fragment of the RXR␥1 gene ( Fig. 1A) was used as a positive control for each of the 5Ј oligonucleotides with the common 3Ј oligonucleotide (ϩ64/ϩ84, Fig. 1B).
Transfection Plasmids-The 3.1-kb XhoI/NotI fragment of the RXR␥1 gene (Ϫ2900 to ϩ207) was cloned into the pGL2-basic plasmid (Promega) at the SmaI site by blunt ligation. A pGL1.4 reporter plasmid was generated by blunt ligation of the 1.4-kb SacI/NotI fragment of the RXR␥1 gene (Ϫ1139 to ϩ207) into the pGL2-basic plasmid at the SmaI site. A SacI/PvuII fragment of the RXR␥1 gene (Ϫ1142 to ϩ194) was cloned into the SmaI site of the pA3luciferase (pA3luc) plasmid in both orientations by blunt ligation. A pGL556 reporter plasmid containing the Ϫ483 to ϩ73 region of the RXR␥1 gene was generated by PCR and subcloned by blunt ligation into the pGL2-basic plasmid at the SmaI site. A pGL556DR1mut reporter plasmid was generated in a similar fashion, except the DR-1 element was replaced with an XhoI site by ligation of two separate PCR products. An RXR␣ expression plasmid was generated by ligation of the coding region of mRXR␣ into the pW1 human ␤-actin promoter plasmid, a modified pGEM9zf(Ϫ) (Promega) vector with an actin promoter (14). All plasmid constructs were verified for ligation fidelity and orientation by sequencing.
Transient Transfection Studies-Transient transfection assays have been previously described (10). Briefly, 20 g of the different RXR␥1 promoter fragments in pA3luc or pGL2 and 1 g pCMV-␤-galactosidase DNA as an internal control were transfected by electroporation into 7-10 million TtT-97 cells, 3 million ␣TSH cells or 5 million GH3 cells. Similar amounts of plasmid DNA were transfected into 800,000 CV-1 cells by calcium phosphate precipitation. Cells were incubated at 37°C for 16 h in Dulbecco's modified Eagle's medium with 10% charcoalstripped fetal calf serum. After harvest, cells were subjected to freezethaw extraction and assayed for both luciferase and ␤-galactosidase activity as described previously (15).
P19 mouse embryonal carcinoma cells were passed at a density of 2 ϫ 10 5 /10-cm plate. The following day, 10 g of luciferase reporter plasmid, 0 -1 g of RXR␣ expression plasmid, and 0.5 g of a ␤-galactosidase reporter plasmid were transfected by calcium phosphate precipitation (16). The media was changed the following day, and cells were incubated at 37°C for 24 h with or without 500 nM 9-cis-retinoic acid (9-cis-RA) from a 10 mM stock solution in 100% ethanol. All experiments were performed in triplicate, and luciferase activity was normalized to ␤-galactosidase activity.
Quantitative RT-PCR-Tissue were collected from 5 male wild-type SvJ129 mice, and total RNA was prepared from each tissue separately (TriReagent, Sigma). Quantitative mRNA analysis was carried out for mRXR␥1 and mRXR␥2 using the ABI 7700 system (PerkinElmer Life Sciences). Primers and probes (containing fluorochrome and quencher) were generated against each specific RNA using a Primer Express program (PerkinElmer Life Sciences). Primer and probe sequences are available upon request. Primer and probe concentrations were optimized against total TtT-97 RNA (17). TtT-97 total RNA was subsequently used to generate quantitative standard curves for sample analysis. Standard curves were linear between 0.1 and 200 ng of total RNA for RXR␥1 (r ϭ 0.992) and RXR␥2 (r ϭ 0.994). Amplification reactions were performed in MicroAmp optical tubes (PerkinElmer Life Sciences) in a 50-l volume containing 8% glycerol, 1ϫ TaqMan  Samples were run in duplicate with a control-lacking reverse transcriptase (no RT). The no RT signal was consistently Ͻ1% of the RT-PCR reaction. Input total RNA was first determined against an 18 S ribosomal RNA (rRNA) control (PerkinElmer Life Sciences, catalog number 4308310), which correlated well with RNA amount determined by optical density. Input total RNA was 300 -700 ng for the RXR␥ isotype measurements. Individual target RNA concentrations were corrected for input RNA based on rRNA meaurements. Reverse transcription was performed at 48°C for 30 min followed by activation of TaqGold at 95°C for 10 min. Subsequently 40 cycles of amplification were performed at 95°C for 15 s and 60°C for 1 min. The detection threshold was set above the mean base-line fluorescence determined during the first 15 cycles. Threshold cycle was determined when fluorescence intensity first increased above detection threshold, and sample values were generated from the standard curve.

RESULTS
Cloning of the 5Ј-Flanking Region of RXR␥1-A TtT-97 thyrotrope cDNA library was screened with a radiolabeled probe corresponding to the coding region of RXR␥1 to obtain additional 5Ј-untranslated sequence present in thyrotropes. A clone was obtained containing the entire coding sequence as well as both 5Ј-and 3Ј-untranslated sequence. The 5Ј-untranslated sequence corresponded to a sequence previously reported from a murine skeletal muscle cDNA library (9). To define the 5Ј extent of the RXR␥1 mRNA in thyrotropes, 5Ј-RACE was performed on TtT-97 poly(A) ϩ RNA. PCR was performed on the cDNA population with a specific adapter on the 5Ј end using an antisense oligonucleotide corresponding to sequence in exon 2 of RXR␥1 and a sense oligonucleotide corresponding to se- Separated products were hybridized with a radiolabeled oligonucleotide corresponding to a common region in all products. B, the 3.1-kb XhoI/ NotI fragment of the RXR␥1 gene (Fig. 1A) was used as a positive control for each of the 5Ј oligonucleotides with the common 3Ј oligonucleotide (ϩ64/ϩ84, Fig. 1B), generating smaller products than the RT-PCR products. quence in the adapter at the 5Ј end of the insert. The largest fragments generated were subcloned and sequenced. The two largest clones were ϳ570 and 600 bp corresponding to 431 and 455 bp of 5Ј-untranslated sequence, respectively.
A EMBL-3 SP6/T7 mouse genomic library was screened by using a cDNA fragment corresponding to the 5Ј-untranslated region of RXR␥1. A 3.1-kb XhoI to NotI fragment was identified, subcloned, and partially sequenced (Fig. 1). The 3Ј end of the genomic fragment was identical to 205 bp of 5Ј-untranslated sequence identified by 5Ј-RACE.
Transcriptional Start Site Identification in Thyrotropes-RNase protection analysis was carried out by subcloning an NsiI/HindIII fragment of RXR␥1 (Fig. 1A) into pGEM3 and generating a 431-bp radiolabeled antisense RNA by linearization with BamHI. The radiolabeled fragment was hybridized with 5 g of TtT-97, ␣TSH, and GH3 poly(A) ϩ RNA and treated with RNase. TtT-97 thyrotropes express RXR␥1 mRNA, while the thyrotrope-derived ␣TSH cells and somatotrope-derived GH3 cells do not (10). The protected fragments are shown in Fig. 2. The largest major protected band was 208 bp using TtT-97 RNA (arrow, lane 1). No protected bands were detected using ␣TSH or GH3 RNA (lanes 2 and 3). Lane 4 shows the positive control RNA, which generated the predicted 177-bp protected product. Lane 5 shows the untreated 431-bp riboprobe. The transcription start site (ϩ1) was mapped to a position 3 bp upstream of the largest 5Ј-RACE product, which results in 458 bp of 5Ј-untranslated sequence. A TAATA element is located at Ϫ36 bp relative to the transcription start site (Fig. 1B).
To confirm that no RXR␥ RNA transcripts were derived from sequences upstream of this start site in thyrotropes, RT-PCR was performed using randomly primed TtT-97 mRNA. A common antisense oligonucleotide corresponding to sequence in exon 2 was used to generate PCR products across an exon-intron boundary to avoid genomic DNA contamination. Sense oligonucleotides were synthesized corresponding to ϩ33/ϩ50 (A), Ϫ146/Ϫ130 (B), Ϫ241/Ϫ227 (C), and Ϫ437/Ϫ420 (D). PCR products were sizeseparated and transferred to nitrocellulose for Southern blot analysis. Fig. 3A shows a strong signal using oligonucleotide A, while a faint signal was generated using oligonucleotide B. Oligonucleotides C and D generated no products. PCR of a DNA plasmid control revealed relatively similar amounts of PCR products using each of the sense oligonucleotides (Fig. 3B), suggesting that the major transcription start is between Ϫ146 and ϩ33, but a minority of larger transcripts are generated, as can also be seen in the RNase protection analysis (Fig. 2).
Functional Analysis of the 5Ј-Flanking Region of RXR␥1-Gene transfer of the XhoI/NotI (3.1 kb) and SacI/NotI (1.4 kb) 5Ј-flanking fragments of RXR␥1 in pGL2basic was carried out by electroporation in TtT-97 thyrotropes. Both fragments had 350 -450-fold higher luciferase activity compared with the promoterless pGL2basic vector (data not shown), suggesting that the SacI/NotI fragment contained all of the sequences necessary for full promoter activity.
To verify the 1.4-kb RXR␥1 fragment as a promoter region in thyrotropes, a Ϫ1133 to ϩ207 (SacI/PvuII) fragment of the RXR␥1 5Ј-flanking region was subcloned into pA3luc in both orientations. Gene transfer by transient transfection was carried out in TtT-97 thyrotropes, which showed promoter activity at 390-fold over the promoterless vector (Fig. 4). Reverse orientation of the plasmid displayed 4% of the activity of the forward orientation. The same experiments were carried out in ␣TSH, GH3, and CV-1 cells, which either lack or have very low levels of RXR␥1 mRNA. Promoter activity was higher than the promoterless vector in all cell types tested; however, relative luciferase activity (forward compared with reverse) was greater in the TtT-97 thyrotropes compared with these other cell types (Fig. 5), suggesting cell-specific activity of the RXR␥1 promoter.
Expression of RXR␥1 and RXR␥2 mRNA in Different Murine Tissues and TtT-97 Thyrotropes-RXR␥1 promoter activity and mRNA expression are low in GH3 and ␣TSH cells, but RXR␥2 mRNA is detectable in these tissues (10). One hypothesis is that RXR␥2 down-regulates RXR␥1 expression in tissues and cell lines. To test this, we measured RXR␥1 and RXR␥2 mRNA expression in TtT-97 thyrotropes and a variety of mouse tissues. Fig. 6 shows that RXR␥1 and RXR␥2 are both highly expressed in the TtT-97 thyrotropes, suggesting that endogenous RXR␥2 does not inhibit the expression of RXR␥1. RXR␥1 is also expressed in skeletal muscle and brain, which confirms previous data using Northern blot analysis (9). These data also show that RXR␥1 is expressed in mouse pituitary, but not adipose or thyroid tissue. RXR␥2 is expressed in skeletal muscle, heart, and liver, which also confirms previous studies (9).
Regulation of RXR␥1 Promoter Activity-Activity of the RXR␥2 promoter is stimulated by 9-cis-RA through an imperfect DR-1 element (12). Sequence examination of the RXR␥1 promoter revealed a putative DR-1 element at position Ϫ232 (Fig. 1B). We therefore examined the effects of 9-cis-RA on the RXR␥1 promoter in thyrotropes. Fig. 7 shows that 9-cis-RA suppressed promoter activity in a dose-dependent manner in thyrotropes, suggesting opposite regulation of the two RXR␥ gene promoters by retinoic acid. To examine the effect of retinoic acid on the RXR␥1 promoter in non-pituitary cells, similar experiments were carried out in mouse embryonal carcinoma cells (P19). 9-cis-RA also suppressed RXR␥1 promoter activity in these cells in an RXR-dependent manner as shown by the addition of increasing amounts of an RXR␣ expression vector (Fig. 8A). An RXR␥1 promoter fragment with a mutation in the DR-1 element was generated by PCR as described under "Materials and Methods." A promoter fragment with the wild-type DR-1 element retained suppression of activity with 9-cis-RA (Fig. 8B), while mutagenesis of DR-1 in this promoter context lost suppression of activity with 9-cis-RA. DISCUSSION In this report, we have isolated the mouse RXR␥1 5Ј-flanking region and begun characterization of promoter activity in anterior pituitary cells. RXR␣ and RXR␤ are widely expressed throughout development and in the adult, while RXR␥ has a more limited tissue distribution. In development, RXR␥ is expressed primarily in the brain, developing skeletal muscle, and anterior pituitary gland (8). Within the anterior pituitary, RXR␥ mRNA and protein are limited to the thyrotropes in mice, rats, and humans (9, 10, 18 -20). Studies from our group and others (10,17,21) have shown that vitamin A and retinoid derivatives suppress production of TSH and activity of the TSH␤ promoter both in vitro and in vivo. The in vitro activity of retinoids appears to require RXR␥.
Thyrotropes express a number of genes with limited tissue distribution. The TSH␤ subunit is expressed only in the thyro- FIG. 8. The 9-cis-retinoic acid effect on RXR␥1 promoter activity is mediated through an imperfect DR-1 site. Transient transfections were carried out as described under "Materials and Methods." A, P19 cells were transfected in triplicate with 10 g of luciferase reporter plasmid containing the SacI/NotI fragment of the RXR␥1 promoter. Cells were co-transfected with either empty vector or increasing amounts of an RXR␣ expression plasmid, then subjected to treatment with either vehicle (open bars) or 0.5 M 9-cis-RA (black bars). Results are expressed as -fold promoter activity versus promoterless vector. B, P19 cells were transfected in triplicate with 10 g of luciferase reporter plasmid containing the Ϫ483 to ϩ73 region of the RXR␥1 promoter containing a mutation of the putative DR-1 element (Fig. 1B) generated by PCR mutagenesis (pGL556DR1mut) or without the mutation (pGL556). Cells were co-transfected with 1 g of an RXR␣ expression plasmid, then subjected to treatment with either vehicle (open bars) or 0.5 M 9-cis-RA (black bars). tropes, with the exception of a few reports showing limited expression in the enterocytes and pars intermedia (22). Mechanisms governing this limited expression appear to include gene activation by two transcription factors, Pit-1 and GATA-2, both of which also have limited tissue distribution (4). Interestingly, two other pituitary-restricted hormone genes, growth hormone and prolactin, also require Pit-1 for gene expression, but appear to use different partners interacting with a composite element for unique cell-type expression. The growth hormone promoter contains a binding site for the transcription factor Zn-15, which is flanked by two Pit-1 binding sites (23,24). In contrast, the prolactin promoter contains a composite Pit-1/ETS-1 binding site believed to direct unique expression of this gene in lactotropes (5). Inspection of the RXR␥1 promoter region does not reveal any consensus Pit-1/GATA-2 elements. This would suggest that expression of this gene in thyrotropes is through nonconsensus composite Pit-1/GATA-2 elements, or expression of RXR␥1 in thyrotropes is through a different mechanism than TSH␤ subunit expression.
Two subtypes of RXR␥ (␥1 and ␥2) are generated from the RXR␥ gene locus. These two subtypes appear to be under control of separate regulatory regions. The RXR␥2 protein is 123 amino acids shorter than RXR␥1 on the N-terminal end, but the remaining 340 amino acids, including the DNA binding and ligand binding domains, are identical. RXR␥2 appears to be generated from a separate promoter and contains a unique 5Ј exon and 5Ј-untranslated region (9). The human RXR␥2 promoter has been isolated, but cell-specific expression of this promoter has not been examined (12). We have previously shown that RXR␥1 mRNA is highly expressed in mouse pituitaries, while RXR␥2 is expressed at low levels (11). We show here that the mouse RXR␥1 promoter exhibits high activity in thyrotrope-derived TtT-97 cells, which express RXR␥1 mRNA, but relatively low activity in pituitary-derived cell types (␣TSH, GH3), which lack detectable RXR␥1 mRNA (10). These data suggest that the regulation of RXR␥1 gene expression is influenced, at least in part, by tissue-specific regulation in the Ϫ1142 to ϩ194 region of the RXR␥1 promoter.
Analysis of the RXR␥2 human gene promoter reveals an imperfect retinoid X response element (GGTTGAAAGGTCA) immediately upstream of the transcription start site (12). The authors further show that this element is required for retinoic acid stimulation of promoter activity, which is believed to mediate RXR-dependent transcriptional autoactivation. This DR-1 element behaves like a classic RXRE in the RXR␥2 promoter. Examination of the mouse RXR␥1 promoter reveals a similar imperfect DR-1 element (CTTTCACAGATCA) at position Ϫ220 to Ϫ232. In contrast to the RXR␥2 promoter, our studies show that RXR␥1 promoter activity is down-regulated by retinoic acid in the thyrotrope-derived TtT-97 cells. One could argue that this suppression of promoter activity is thyrotrope-specific, but we also observed this effect in the nonpituitary P19 cells, suggesting a generalized mechanism of negative regulation of this promoter/gene by retinoic acid. Mutagenesis of the DR-1 element blocked negative regulation of RXR␥1 promoter activity by 9-cis-RA confirming this DR-1 element as necessary for negative regulation by retinoic acid.
The mouse RXR␤ gene has been isolated and the promoter region has been studied (25)(26)(27). Like RXR␥, there appears to be two different isotypes, RXR␤1 and RXR␤2, which have unique regulatory regions. Examination of a limited region of the RXR␤1 promoter did not reveal a classical DR-1 element. Transient transfection studies using the RXR␤1 promoter in GH3 cells showed no regulatory response of promoter activity by treatment with 9-cis-RA (26). Both RXR␥1 and RXR␥2 promoter regions contain a DR-1 element, and both promoter activities are regulated by 9-cis-RA, albeit in different directions. These data would suggest that RXR␥1 not only has unique expression in a limited number of cell types, but unique regulation by retinoic acid in comparison to other retinoid X receptors.