JBC Anatrace, Inc.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M704845200 on September 17, 2007

J. Biol. Chem., Vol. 282, Issue 46, 33421-33434, November 16, 2007
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
282/46/33421    most recent
M704845200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gillespie, R. F.
Right arrow Articles by Gudas, L. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gillespie, R. F.
Right arrow Articles by Gudas, L. J.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Retinoic Acid Receptor Isotype Specificity in F9 Teratocarcinoma Stem Cells Results from the Differential Recruitment of Coregulators to Retinoic Acid Response Elements*

Robert F. Gillespie{ddagger} and Lorraine J. Gudas{ddagger}§1

From the {ddagger}Molecular Biology Program, Weill Graduate School of Medical Sciences, Cornell University and §Pharmacology Department, Weill Medical College of Cornell University, New York, New York 10021

Received for publication, June 12, 2007 , and in revised form, September 12, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The retinoic acid receptor (RAR) {alpha}, beta2, and {gamma} isotypes each regulate specific subsets of target genes in F9 teratocarcinoma stem cells. We used chromatin immunoprecipitation assays to monitor the association of RAR{gamma}, retinoic X receptor (RXR) {alpha}, and coregulators with the RARbeta2, Hoxa1, and Cyp26A1 retinoic acid response elements (RAREs) in F9 wild type and RAR{alpha}, -beta2, and -{gamma} null cells. Additionally we quantitatively monitored expression of the corresponding mRNAs. We demonstrated that the association of RAR{gamma} and/or RXR{alpha} with a RARE was not sufficient for retinoic acid (RA)-mediated transcription of the corresponding target gene. However, the ability of RAR{gamma} and/or RXR{alpha} to recruit pCIP (AIB1/ACTR/RAC-3/TRAM-1/SRC-3) and p300 to a RARE did correlate with RA-associated transcription of target mRNAs. Therefore, the specific functions of the RAR isotypes do not manifest at the level of their DNA binding but rather from a differential ability to recruit specific components of the transcriptional machinery. We also demonstrated that RA-mediated displacement of the polycomb group protein SUZ12 from a RARE was inhibited in the absence of RAR{gamma}. Thus, transcriptional components of the RAR signaling pathway are specifically required for displacement of SUZ12 from RAREs during RA-mediated differentiation of F9 cells.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Retinoic acid (RA)2 is an important regulator of vertebrate development and homeostasis because of its role in essential processes such as apoptosis, cell differentiation, and proliferation (1, 2). The effects of RA are mediated through binding to the retinoic acid receptors (RARs) (3) and retinoic X receptors (RXRs) (4), which are members of the nuclear receptor superfamily (5). The RARs and RXRs each have three isotypes ({alpha}, beta, and {gamma}), which are encoded by distinct genes. In addition, for each RAR isotype, there are several isoforms generated by differential promoter usage and splicing (3, 4). The multiple RAR and RXR isotypes and isoforms are conserved in vertebrate evolution and display distinct spatiotemporal expression patterns in developing embryos and adult tissues, suggesting that each receptor performs some unique functions (1).

RXR-RAR heterodimers bind much more efficiently to retinoic acid response elements (RAREs) than their respective homodimers in vitro (6), and several lines of evidence support the idea that these heterodimers represent the functional units transducing the retinoid signal in vivo (3). RAR-RXR heterodimers are thought to be constitutively associated with RAREs and to actively repress transcription in the absence of ligand through association with the corepressors nuclear receptor corepressor or SMRT (silencing mediator of retinoic acid and thyroid hormone receptor) (7–9). Nuclear receptor corepressor and SMRT are found in repressor complexes containing the histone deacetylase HDAC3 (10, 11).

The F9 murine embryonal carcinoma cell line has been extensively used as a cell autonomous model system to study RA signaling. F9 cells resemble the pluripotent stem cells of the inner cell mass of the early embryo and differentiate into three distinct extraembryonic (primitive, parietal, and visceral) endodermal cell types upon treatment with RA, depending on the cell culture conditions (12). F9 cells express all known RA receptors, but RARbeta2 mRNA is only present in high amounts after RA addition (13).

Specific functions for each of the RARs have been demonstrated through the use of F9 cells in which each of the individual RARs has been knocked out by homologous recombination as well as through the use of synthetic isotype-selective ligands. F9 RAR null cells exhibit marked modulation of a variety of genes when compared with the F9 wild type (Wt) cells. For example, the RA-mediated transcription of the homeobox gene Hoxa1 is specifically regulated by RAR{gamma} because the RA-induced expression of Hoxa1 mRNA is abrogated in F9 RAR{gamma}–/– cells (14). Additionally a synthetic retinoid selective for RAR{gamma} can induce expression of Hoxa1 mRNA in F9 Wt cells but not in F9 RAR{gamma}–/– cells (15). Furthermore RAR{gamma}–/– cells fail to exhibit complete morphological differentiation in culture when treated with RA (14). Another example of isotype-specific function comes from studies utilizing a F9 RARbeta –/–2 cell line. In contrast to F9 Wt, RAR{alpha}–/–, and RAR{gamma}–/– cell lines, the F9 RARbeta –/–2 cell line exhibits no growth arrest in response to RA (16). Additionally a number of genes specifically regulated by the RARbeta2 isoform in F9 cells have been identified through use of subtractive hybridization and DNA array analysis (17). RAR{alpha} also specifically regulates RA target genes in F9 cells as the expression of both Hoxb1 and CRABP-II is reduced in F9 RAR{alpha}–/– cells compared with F9 Wt and F9 RAR{gamma}–/– cells (18). It should be noted that in the F9 RAR null cell lines the expression levels of the undisrupted RARs were similar to those found in F9 Wt cells (19). These results demonstrate that each of the RAR isotypes regulates a specific subset of target genes in F9 cells when the RAR isotypes are expressed at endogenous levels.

A degree of functional redundancy among the three RAR isotypes has also been demonstrated through use of the F9 RAR null cells and isotype-selective ligands. For example, ~10-fold overexpression of RAR{alpha} in F9 RAR{gamma}–/– cells could restore target gene activation of RAR{gamma} target genes such as Hoxa1 as well as the differentiation potential of F9 RAR{gamma}–/– cells (20). In the same set of experiments, however, overexpression of RARbeta2 could not restore Hoxa1 mRNA expression in F9 RAR{gamma}–/– cells. Additionally the expression of Hoxa1 mRNA in F9 RAR{gamma}–/– cells could be restored by exposure to an RAR{alpha}-selective ligand (15). However, this same RAR{alpha}-selective ligand inefficiently induced expression of Hoxa1 mRNA in F9 Wt cells (15), indicating that the presence of RAR{gamma} can hinder the ability of RAR{alpha}, bound to an RAR{alpha}-selective agonist, to induce RA target genes. Furthermore an RARbeta-specific agonist could induce expression of another RA target gene, RARbeta2 itself, in F9 RAR{gamma}–/– cells but not in F9 Wt or RAR{alpha}–/– cells (15). These results demonstrate that some of the functional redundancies observed among individual RARs in F9 RAR null cells do not exist in the context of wild type cells.

In this study we used ChIP assays to monitor how the association of RAR{gamma}, RXR{alpha}, and coregulators with the RAREs regulating expression of the Hoxa1 (21), RARbeta2 (22), and Cyp26A1 (23, 24) genes is affected in the various F9 RAR null cell lines as compared with F9 Wt cells. Additionally we quantitatively monitored the expression of these target genes in the aforementioned F9 cell lines by real time PCR. We demonstrated that RAR{gamma} was associated with the RAREs that regulate the transcription of the Hoxa1 and Cyp26A1 mRNAs. We also showed that RAR{gamma} was associated with the RARbeta2 RARE even though RAR{gamma} was not required for the RA-induced expression of RARbeta2 mRNA. Furthermore we demonstrated that the presence of RAR{gamma} and RXR{alpha} at an RARE was not sufficient for the recruitment of factors required for transcription, such as p300 and pCIP, to the corresponding target genes. We also showed that the levels of the polycomb repressive protein SUZ12 associated with the RAREs monitored in this study were significantly higher in RAR{gamma}–/– cells as compared with F9 Wt and F9 RARbeta –/–2 cells. Thus, we demonstrated that components of the RAR signaling apparatus were specifically required for displacement of SUZ12 from RAREs during the RA-mediated differentiation of F9 cells.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Cell Culture—F9 Wt, RAR{alpha}–/–, RARbeta2–/–, and RAR{gamma} –/– embryonal carcinoma cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (Invitrogen), 2 mM glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin. Cells were plated in gelatin-coated tissue culture plates ~48 h prior to RNA harvesting (5 x 105 cells/60-mm dish) or formaldehyde fixation (2.5 x 106 cells/20-cm dish). Cells were treated with 1 µM retinoic acid for 24 h.

Antibodies and Chemicals—All-trans-RA was obtained from Sigma-Aldrich and dissolved in ethanol. Anti-RAR{gamma} serum was generated by immunization of rabbits with a peptide corresponding to the F region of RAR{gamma} (NH2-PGPHPKASSEDEAPGGQGKRGQS-COOH). Polyclonal anti-RAR{gamma} IgG was purified from the crude serum through use of a DEAE Affi-Gel blue gel column (Bio-Rad). Anti-RXR{alpha} (D-20, sc-553), anti-pCIP (M-397, sc-9119), anti-p300 (N-15, sc-584), and anti-actin (I-19, sc-1616) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-SUZ12 (07-379) antibody was purchased from Upstate%20Biotechnology">Upstate Biotechnology (Lake Placid, NY). Anti-phospho-Ser-5 carboxyl-terminal domain (CTD) of RNA polymerase II (pCTDser5) was purchased from Covance Research Products (Richmond, CA).

RNA Preparation and RT-PCR—RNA was prepared and subjected to semiquantitative or real time RT-PCR analysis as described previously (25).

Characterization of F9 RAR{alpha}, -beta2, and -{gamma} Null Cell Lines—RNA was prepared and subjected to semiquantitative RT-PCR analysis. The primers used to detect RAR{alpha} spanned the exon in which the disruption cassette was inserted and were as follows: 5'-ATCGAGACCCAGAGCAGCAG-3' and 5'-CCTGGTGCGCTTTGCGAACC-3' (18). The primers used to detect RAR{gamma} also spanned the exon in which the disruption cassette was inserted (14) and were as follows: 5'-CAATAAGGAGAGACTCTTTGCG-3' and 5'-TTGCTGACCTTGGTGATGAGTT-3'. The primers used to detect RARbeta2 (16) were as follows: 5'-GATCCTGGATTTCTACACCG-3' and 5'-CACTGACGCCATAGTGGTA-3'.

ChIP Assays—ChIP assays were performed as described previously (25).

Semiquantitative and Real Time PCR—Semi-quantitative PCRs and real time PCRs were performed as described previously (25).

Western Blot Analysis—Whole cell extracts were prepared from COS cells that were either mock-transfected or transfected with a plasmid expressing RAR{alpha}, RARbeta, or RAR{gamma}. Five micrograms of each of the COS whole cell extracts were resolved by 12% SDS-PAGE followed by transfer to a nitrocellulose membrane (0.45-µm pore size; catalog number 162-0090, Bio-Rad). Primary antibody incubation was done overnight at 4 °C. The anti-RAR{gamma} blue eluate, as described above, was used at a 1:200 dilution to detect RAR{gamma}. After a 1-h incubation with an immunoglobulin G horseradish peroxidase-conjugated secondary antibody at room temperature (anti-rabbit, 1:40,000 dilution; sc-2030, Santa Cruz Biotechnology), the membranes were developed with SuperSignal Substrate (Pierce) for 5 min and exposed to BioMax film (Eastman Kodak Co.). Primary and secondary antibodies were diluted in phosphate-buffered saline containing 5% Blotto (Santa Cruz Biotechnology) and 0.1% Tween 20. Blots were stripped with Restore Plus Western Blot Stripping Buffer (Pierce, 46430) and then reprobed with an anti-actin antibody (1:400 dilution) followed by incubation with the appropriate horseradish peroxidase-conjugated secondary antibody (anti-goat, 1:2000 dilution; sc-2056, Santa Cruz Biotechnology).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
We have previously studied the association of RAR{gamma}, RXR{alpha}, and other proteins involved in transcriptional regulation with the Hoxa1, RARbeta2, and Cyp26A1 RAREs (R1 and R2) during the course of RA treatment in F9 Wt cells (25). The main purpose of this study was to determine how the association of RAR{gamma}, RXR{alpha}, and coregulators with the aforementioned RAREs is affected by the absence of each of the three RAR isotypes. Therefore, we utilized F9 cells lines in which both alleles of RAR{alpha} (18), RARbeta2 (the predominant isoform of RARbeta) (16), and RAR{gamma} (14) were individually knocked out by homologous recombination.

To confirm that the RAR isotypes were individually knocked out in our cell lines, we monitored expression of the RAR isotypes in the various F9 RAR null cell lines by semiquantitative RT-PCR. F9 Wt and F9 RAR{alpha}–/–, -beta –/–2, and -{gamma}–/– cell lines were either untreated or treated with 1 µM RA for 24 h, and RNA was harvested. RAR{alpha} mRNA was detected in RAR{gamma}–/– and RARbeta –/–2 cells at levels similar to those in F9 Wt cells irrespective of the presence of RA treatment (Fig. 1A). As expected, no RAR{alpha} mRNA could be detected in F9 RAR{alpha}–/– cells (Fig. 1A). The expression of RAR{gamma} mRNA could not be detected in RAR{gamma}–/– cells, although RAR{gamma} mRNA was expressed in RAR{alpha}–/– and RARbeta –/–2 cells at levels similar to those seen in F9 Wt cells irrespective of the presence of RA (Fig. 1A). In contrast to RAR{alpha} and RAR{gamma}, the expression of RARbeta2 was strongly induced by the presence of RA in F9 Wt cells (Fig. 1A). RA-induced expression of RARbeta2 mRNA was observed in RAR{alpha}–/– and RAR{gamma}–/– cells, and as expected, RARbeta2 mRNA could not be detected in RA-treated RARbeta –/–2 cells (Fig. 1A). To ensure that equivalent amounts of RNA were used in the RT-PCR assays, expression of the ribosomal phosphoprotein 36B4 "housekeeping gene" (26) was monitored. All samples expressed 36B4 mRNA to similar levels (Fig. 1A). Thus, we confirmed that the three different RAR null cell lines each specifically lacked expression of one RAR isotype.

RA Induced Expression of Hoxa1, RARbeta2, and Cyp26A1 mRNAs in F9 Wt and RAR Null Cell Lines—We next determined whether Hoxa1, RARbeta2, and Cyp26A1 mRNAs could be induced by RA in the F9 RAR{alpha}, -beta2, and -{gamma} null cell lines as these three genes are all strongly induced by RA in F9 Wt cells (27–29). F9 Wt and the RAR null cell lines were either untreated or treated with 1 µM RA for 24 h, and RNA was harvested for use in RT-PCR. Real time PCR was used to quantitate RT-PCR products corresponding to the three RA target genes. As expected, all three RA target genes were strongly induced (>22-fold) by RA treatment in F9 Wt cells (Fig. 1B). Again we also monitored the levels of 36B4 mRNA in the F9 Wt, RAR{alpha}–/–, RARbeta –/–2, and RAR{gamma}–/–cell lines. Similar levels of 36B4 mRNA were observed in all of the cell lines both in the presence and absence of RA (Fig. 1B).

Hoxa1 mRNA levels in RAR{alpha}–/– cells were similar to those seen in F9 Wt cells cultured in the presence or absence of RA (Fig. 1B) in agreement with previous reports (18). This result demonstrates that RAR{alpha} is not required for expression of the Hoxa1 gene. Additionally in the presence of RA, RARbeta2 mRNA levels were similar in F9 Wt and RAR{alpha}–/– cells, demonstrating that RAR{alpha} is not required for induction of RARbeta2 by RA. However, in untreated RAR{alpha}–/– cells, as compared with F9 Wt cells, ~2.9-fold higher levels of RARbeta2 transcripts (p value <0.0001) were expressed (Fig. 1B). This latter result indicates that in the absence of RA RAR{alpha} may have a role in the repression of RARbeta2 (30). Levels of Cyp26A1 mRNA after RA treatment of F9 RAR{alpha}–/– cells were ~57% of the levels observed in F9 Wt cells (Fig. 1C), indicating that RAR{alpha} has a role in the RA-induced expression of this gene.

Previous reports have demonstrated that RAR{gamma} is required for the expression of both the Hoxa1 (18) and Cyp26A1 (27) genes in F9 Wt cells. Our results are in agreement with these reports as we did not observe expression of either Hoxa1 or Cyp26A1 mRNA in RA-treated F9 RAR{gamma}–/– cells (Fig. 1, B and C), whereas a >22-fold induction of these mRNAs was observed in F9 Wt cells. RA-induced expression of RARbeta2 also was reduced in RAR{gamma}–/– cells (~43% of F9 Wt) relative to F9 Wt cells (Fig. 1, B and C). Thus, we concluded that RAR{gamma} has a prominent role in the regulation of Hoxa1 and Cyp26A1 and a lesser role in the induction of RARbeta2 transcripts in response to RA.

As expected, expression of RARbeta2 mRNA was abrogated in RARbeta –/–2 cells as RARbeta2 transcripts detected in RARbeta –/–2 cells were less than 1% of the levels observed in F9 Wt cells (Fig. 1, B and C). The low levels of RARbeta2 transcripts detected in RARbeta –/–2 cells may represent fusion transcripts generated through the RARbeta2 RARE that remains intact in these cells. Additionally the RA-induced expression of Hoxa1 (~28%) and Cyp26A1 (~34%) transcripts was lower in F9 RARbeta –/–2 cells relative to levels observed in F9 Wt cells (Fig. 1C), consistent with data from our laboratory (16).

The Association Patterns of RAR{gamma} and RXR{alpha} with the RAREs Regulating Expression of the Hoxa1 Gene in F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} Null Cell Lines—We previously demonstrated that RAR{gamma} and RXR{alpha}, both in the presence and absence of RA, are associated with the RAREs that regulate expression of the Cyp26A1 (R1 and R2), Hoxa1, and RARbeta2 genes in F9 Wt cells (25). These results are consistent with previous studies that have also demonstrated that RAR-RXR heterodimers are constitutively associated with RAREs (31, 32). In this study we wanted to determine whether the absence of each of the RAR isotypes in F9 cells would affect the association of RAR{gamma} and RXR{alpha} with these RAREs. Therefore, we monitored the association of RAR{gamma} and RXR{alpha} with the aforementioned RAREs through use of a two-step ChIP assay (33). F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell lines were cultured in the presence or absence of RA for 24 h and then subjected to the protein-protein cross-linking reagent disuccinimidyl glutarate. Cells were then formaldehyde-fixed as in conventional ChIP assays, and soluble chromatin was prepared as described under "Materials and Methods." Antibodies to RAR{gamma} and RXR{alpha} were used to immunoprecipitate protein-DNA complexes from soluble chromatin. Nonspecific rabbit IgG antibodies were also used as a negative control in the two-step ChIP assays.


Figure 1
View larger version (45K):
[in this window]
[in a new window]

  		FIGURE 1.
Expression of RAR isotype mRNAs and RA target gene mRNAs in F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell lines as measured by RT-PCR. A, confirmation of the identities of the three F9 RAR null cell lines by semiquantitative RT-PCR. F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell lines were cultured in the absence or presence of 1 µM RA for 24 h, and RNA was harvested. Harvested RNA was assayed by semiquantitative RT-PCR. Each experiment was repeated at least three times; data shown are from one experiment. Linearity of the PCR was demonstrated by serial dilution of the F9 Wt 24-h RA sample (data not shown). Products were visualized by standard gel electrophoresis. B, induction of the RA target gene mRNAs in F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell lines as measured by quantitative real time RT-PCR analysis. Error bars indicate standard error of three biological replicates with each quantitative PCR performed in triplicate. The y axis of each graph has a different scale. C, expression of the RA target mRNAs in the F9 RAR null lines relative to the F9 Wt cells. The data of B were normalized to the F9 Wt 24-h RA sample for each RA target mRNA.

 
The anti-RAR{gamma} IgG used in the two-step ChIP assays was generated in our laboratory (see "Materials and Methods"). To determine the specificity of the anti-RAR{gamma} IgG, we prepared individual COS cell extracts from cells overexpressing each of the RAR isotypes. COS cell extract was also prepared from cells that were mock-transfected. The COS cell extracts were resolved by SDS-PAGE and subsequently subjected to immunoblot analysis against the anti-RAR{gamma} IgG. The anti-RAR{gamma} IgG specifically recognized antigen from COS cells overexpressing RAR{gamma}, and the positive signal was at the expected molecular mass for RAR{gamma} (Fig. 2B). These results demonstrate that the anti-RAR{gamma} IgG specifically recognizes the RAR{gamma} isotype.

The Hoxa1 RARE is located ~2 kb downstream of the Hoxa1 gene, whereas the RARbeta2 RARE is located ~55 bp upstream of the transcription start site (Fig. 2A). Cyp26A1 contains a RARE ~70 bp upstream of the transcription start site denoted as R1 (23) as well as a more recently described RARE denoted as R2 (24) found ~1950 bp upstream of R1 (Fig. 2A). As a control for the nonspecific IP of DNA in ChIP assays, we also measured a gene-free region located–18 kb downstream of the Hoxb1 gene (Hoxb1–18 kb). Levels of these five DNA regions recovered in ChIP assays were quantitated by real time PCR assays. We define -fold enrichment as the percentage of input of a specific locus in an IP divided by the percentage of input of the Hoxb1–18 kb 3' negative control region in the same IP (Fig. 3). The data are presented as -fold enrichment to normalize for the higher levels of nonspecific DNA (Hoxb1–18 kb 3' locus) found in RXR{alpha} IPs as compared with RAR{gamma} IPs in the two-step ChIP assays.


Figure 2
View larger version (16K):
[in this window]
[in a new window]

  		FIGURE 2.
A, diagram of the loci monitored in ChIP assays in relation to respective transcription start sites. DNA is represented with a thin black line. Gray boxes denote introns; black boxes refer to exons. The location of RAREs are represented by arrows pointing downward, and the actual RARE sequences are shown directly below the arrows with bold letters denoting binding sites. Bent arrows indicate transcription start sites; hatch marks signify exon/intron gene architecture that is not detailed. Thick black lines indicate regions of DNA amplified during ChIP assays. Schematics were drawn approximately to scale. B, demonstration of the specificity of the anti-RAR{gamma} antibody. Whole cell extracts were prepared from COS cells that were either mock-transfected (m) or transfected with a plasmid expressing RAR{alpha} ({alpha}), RARbeta (beta), or RAR{gamma} ({gamma}). Five micrograms of each of the COS whole cell extracts were resolved by 12% SDS-PAGE followed by Western blot analysis. Anti-RAR{gamma} blue eluate was used at a 1:200 dilution to detect RAR{gamma}. The experiment was performed three times. PP, promoter proximal region.

 
The levels of RAR{gamma} associated with the Hoxa1 RARE in the presence and absence of RA in F9 Wt cells were similar (Fig. 3A, middle panel; ~10-fold enrichment). However, the levels of RAR{gamma} associated with the Hoxa1 RARE increased in F9 RAR{alpha}–/– (~3.5-fold, p value <0.05) and RARbeta –/–2 (~2.7-fold, p value <0.01) cells as a result of RA treatment (Fig. 3A, middle panel). Additionally higher levels of RAR{gamma} were associated with the Hoxa1 RARE in RAR{alpha}–/– cells (~2.7-fold, p value <0.05) and in RARbeta –/–2 cells (~1.7-fold, p value <0.06) treated with RA as compared with RA-treated F9 Wt cells (Fig. 3A, middle panel). The levels of Hoxa1 RARE DNA immunoprecipitated with RAR{gamma} when using soluble chromatin derived from RAR{gamma}–/– cells (Fig. 3B, middle and right panels) were comparable to the background levels of Hoxa1 RARE DNA immunoprecipitated with nonspecific rabbit IgG (Fig. 3A, right panel). This result further confirms the isotype specificity of the anti-RAR{gamma} IgG used in this study.

We also examined the association of RXR{alpha} with the Hoxa1 RARE in F9 Wt and the F9 RAR{alpha}, -beta2, and -{gamma} null cell lines (Fig. 3A, left panel). Previous studies have demonstrated that RXR{alpha} is required for the RA-induced expression of Hoxa1 in F9 cells because Hoxa1 is not expressed in F9 RXR{alpha}–/– cells treated with RA (34). Although Hoxa1 mRNA is not expressed in RAR{gamma}–/– cells (Fig. 1B), RXR{alpha} was associated with the Hoxa1 RARE in these F9 RAR{gamma}–/– cells presumably as a heterodimer with either RARbeta2 or RAR{alpha}. However, the levels of RXR{alpha} associated with the Hoxa1 RARE in untreated (~39% of Wt) and RA-treated (~75% of Wt) F9 RAR{gamma} null cells were lower (p value <0.05 for both comparisons) than those in F9 Wt cells (Fig. 3A, left panel), indicating that RAR-RXR{alpha} heterodimer association with the Hoxa1 RARE is reduced, although not eliminated, in F9 RAR{gamma}–/– cells as compared with F9 Wt cells.

The Association Patterns of RAR{gamma} and RXR{alpha} with the RAREs Regulating Transcription of Cyp26A1 and RARbeta2 mRNAs in F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} Null Cell Lines—Consistent with our previous report (25) high levels of RAR{gamma} were associated with the Cyp26A1 R2 RARE in the absence and presence of RA in F9 Wt cells (Fig. 3B, middle panel). The levels of RAR{gamma} associated with the R2 RARE in the F9 RAR{alpha} null and F9 RARbeta2 null cells were similar to the levels observed in F9 Wt cells (Fig. 3B, middle panel). Again the Cyp26A1 R2 RARE was not detected in RAR{gamma} IPs utilizing soluble chromatin derived from F9 RAR{gamma}–/– cells; only backgrounds levels were seen (Fig. 3B, middle panel).

The levels of RXR{alpha} associated with the Cyp26A1 R2 RARE in F9 RARbeta2 null and RAR{alpha} null cells were similar to those seen in F9 Wt cells (Fig. 3B, left panel). Therefore, although RAR{gamma} is required for expression of Cyp26A1 (Fig. 1B), RXR{alpha} can still associate with the R2 RARE in the absence of RAR{gamma} presumably as a heterodimer with RARbeta2 or RAR{alpha}. This result suggests that RARbeta2-RXR{alpha} or RAR{alpha}-RXR{alpha} heterodimers bound at the R2 RARE are not able to transduce a RA signal, culminating in transcription of the Cyp26A1 gene, in response to RA. Similar patterns of association of RXR{alpha} and RAR{gamma} with the Cyp26A1 R1 RARE compared with the R2 RARE were observed, although -fold enrichment levels were lower at the R1 RARE compared with the R2 RARE (Fig. 3, B and C).

We wanted to address whether RAR{gamma} associates with the RARbeta2 RARE in F9 RARbeta –/–2 cells even though these cells do not express RARbeta2 mRNA in response to RA (Fig. 1, B and C). In contrast to the Hoxa1 and Cyp26A1 genes, RAR{gamma} was not required for transcription of RARbeta2, although RA-induced expression levels of RARbeta2 were reduced in RAR{gamma}–/– cells as compared with F9 Wt cells (Fig. 1, B and C). RAR{gamma} was associated with the RARbeta2 RARE in both the presence and absence of RA (Fig. 3D, middle panel) in F9 Wt cells. Additionally the levels of RAR{gamma} associated with the RARbeta2 RARE in the F9 RARbeta2 null and F9 RAR{alpha} null cell lines were comparable to the levels seen in the F9 Wt cells (Fig. 3D, middle panel). Therefore, we have demonstrated that RAR{gamma} associated with the RARbeta2 RARE is incapable of transducing the RA signal required for expression of RARbeta2 transcripts in the absence of RARbeta2 protein.


Figure 3
View larger version (22K):
[in this window]
[in a new window]

  		FIGURE 3.
RAR{gamma} and RXR{alpha} associate with the target RAREs in F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cells. F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell cells were either untreated or treated with 1 µM RA for 24 h. Cells were then fixed with disuccinimidyl glutarate and formaldehyde and processed into soluble chromatin. Chromatin samples were immunoprecipitated with antibodies to RAR{gamma}, RXR{alpha}, or IgG, and bound DNA was quantitated by real time PCR. The data are presented as -fold enrichment (mean ± S.E.). -Fold enrichment is defined as the percentage of input of a specific locus in an IP divided by the percentage of input of the Hoxb1–18 kb 3' negative control region in the same IP. Error bars indicate standard error of three biological replicates with each quantitative PCR performed in triplicate. KO, knock-out.

 
We also determined whether the association level of RXR{alpha} with the RARbeta2 RARE is perturbed in F9 RARbeta –/–2 cells as compared with F9 Wt, F9 RAR{gamma}–/–, and F9 RAR{alpha}–/– cells. In contrast to RAR{gamma}, lower levels of RXR{alpha} were associated with the RARbeta2 RARE, both in the presence (~43% of Wt) and absence (~25% of Wt) of RA, in F9 RARbeta –/–2 cells as compared with F9 Wt cells (Fig. 3D, left panel; p value <0.05 for both comparisons). Additionally the levels of RXR{alpha} associated with the RARbeta2 RARE were lower in the F9 RARbeta –/–2 cells as compared with the F9 RAR{alpha} and {gamma} null cell lines (Fig. 3D, left panel). Therefore, the absence of RARbeta2 impedes the association of RXR{alpha} with the RARbeta2 RARE, consistent with previous reports that demonstrated an autoregulatory role for the RARbeta2 protein (15, 20, 35).

RA-mediated Recruitment of pCIP and p300 to the RAREs That Regulate Expression of the Hoxa1 and Cyp26A1 Genes Is Compromised in F9 RAR{gamma}–/– Cells as Compared with F9 Wt and RARbeta –/–2 Cells—We demonstrated that RAR{gamma} was required for the RA-induced expression of Hoxa1 and Cyp26A1 mRNAs (Fig. 1, B and C) in F9 cells. Additionally we showed that RARbeta2 transcripts were not expressed in F9 RARbeta –/–2 cells. Therefore, we wanted to examine whether the recruitment of coregulators to the RAREs regulating the aforementioned genes is perturbed in F9 RARbeta –/–2 and F9 RAR{gamma}–/– cell lines. We were able to utilize conventional one-step ChIP assays to monitor the association of the p160 coactivator pCIP (also referred to as AIB1/ACTR/RAC-3/TRAM-1/SRC-3) (36) as well as the histone acetyltransferase coactivator p300 (37) to the RAREs in the F9 RARbeta –/–2 and F9 RAR{gamma}–/– cell lines as well as in the F9 Wt cells.

Consistent with our previous results (25), the levels of pCIP associated with the Hoxa1 RARE rose ~4.3-fold in F9 Wt cells as a result of RA treatment (Fig. 4A, left panel). The levels of pCIP associated with the Hoxa1 RARE in untreated and RA-treated RARbeta –/–2 cells were similar to those seen in F9 Wt cells (Fig. 4A, left panel), although the RA-induced expression of Hoxa1 mRNA was reduced in F9 RARbeta –/–2 cells to ~25% of the level observed in F9 Wt cells (Fig. 1, B and C). However, in F9 RAR{gamma}–/– cells, which do not express Hoxa1, the levels of pCIP associated with the Hoxa1 RARE were reduced both in the absence (~3-fold) and presence (~4-fold) of RA relative to F9 Wt cells (Fig. 4A, left panel). Additionally the levels of pCIP associated with the Hoxa1 RARE in RA-treated F9 RAR{gamma}–/– cells were similar to the basal levels of pCIP associated with the Hoxa1 RARE in F9 Wt and F9 RARbeta –/–2 cells (Fig. 4A, left panel). Therefore, we concluded that coactivators are not recruited by RAR-RXR heterodimers bound at the Hoxa1 RARE in F9 RAR{gamma}–/– cells treated with RA.

We also monitored pCIP association with the Cyp26A1 R1 RARE located immediately upstream of the Cyp26A1 transcription start site (Fig. 2A). The recruitment patterns of pCIP to the Cyp26A1 R1 RARE in the three F9 cell lines were similar to those seen for the Hoxa1 RARE (Fig. 4A, compare left panel and middle panel). The levels of pCIP associated with the Cyp26A1 R1 RARE rose in the F9 Wt by ~3.9-fold and in F9 RARbeta –/–2 cells by ~5.4-fold as a result of RA treatment, and this RA-associated increase was not observed in the F9 RAR{gamma}–/– cell line (Fig. 4A, middle panel). Therefore, we concluded that recruitment of pCIP to the RAREs that regulate the transcription of Hoxa1 and Cyp26A1 requires RAR{gamma}. Furthermore the lack of RA-induced recruitment of pCIP to the Hoxa1 and Cyp26A1 RAREs likely explains why these genes are not expressed in RAR{gamma}–/– cells.

The levels of p300 increased at both the Hoxa1 (~6.9-fold) and Cyp26A1 R1 (~4-fold) RAREs in F9 Wt cells as a result of RA treatment (Fig. 4B, left and middle panels). Additionally comparable levels of p300 were associated with the Hoxa1 and Cyp26A1 R1 RAREs in F9 RARbeta –/–2 cells as compared with F9 Wt cells both in the absence and presence of RA (Fig. 4B, left and middle panels). In contrast, RA-associated p300 recruitment to the Hoxa1 RARE was reduced by ~4.5-fold in F9 RAR{gamma}–/– cells relative to F9 Wt cells. Moreover p300 levels did not increase at the Cyp26A1 R1 RARE as a result of RA treatment in F9 RAR{gamma}–/– cells. These results further demonstrate that RA-mediated coregulator recruitment to the RAREs that control transcription of Hoxa1 and Cyp26A1 is compromised in F9 RAR{gamma}–/– cells.

The Levels of pCIP and p300 Associated with the RARbeta2 RARE Do Not Increase in Response to RA in F9 RARbeta –/–2 Cells—RA-induced transcription of RARbeta2 is abrogated in F9 RARbeta –/–2 cells and reduced ~2-fold in F9 RAR{gamma}–/– cells relative to F9 Wt cells (Fig. 1, B and C). Thus, we monitored the association of pCIP and p300 with the RARbeta2 RARE in F9 Wt, RARbeta –/–2, and RAR{gamma}–/– cell lines. The levels of p300 associated with the RARbeta2 RARE increased ~2.3-fold (p value <0.08) as a result of RA treatment in F9 Wt cells (Fig. 4B, right panel). Levels of p300 also increased at the RARbeta2 RARE by ~1.8-fold as a result of RA treatment (p value <0.05) in the F9 RAR{gamma}–/– cell line. The high basal levels of pCIP and p300 observed at the RARbeta2 RARE are consistent with a previous study that demonstrated that much of the transcriptional machinery is associated with the RARbeta2 RARE in P19 embryonal carcinoma cells prior to RA treatment (32). In contrast, we did not observe an RA-induced increase of p300 at the RARbeta2 RARE in the F9 RARbeta –/–2 cell line (Fig. 4B, right panel), although the basal level of p300 associated with the RARbeta2 RARE was higher in RARbeta –/–2 cells than in the F9 Wt (p value <0.01) and the F9 RAR{gamma}–/– cell lines (p value <0.01). Recruitment patterns of pCIP to the RARbeta2 RARE mirrored those seen for p300 in the three cell lines (compare Fig. 4B, right panel, with Fig. 4A, right panel). The fact that p300 and pCIP levels did not increase as a result of RA treatment at the RARbeta2 RARE in the F9 RARbeta –/–2 cell line indicates that RARbeta2 protein itself is required to increase the levels of the coregulators necessary for RARbeta2 transcription.


Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		FIGURE 4.
pCIP and p300 are not recruited to the Hoxa1 RARE in response to RA in F9 RAR{gamma}–/– cells. F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cells were either untreated or treated with 1 µM RA for 24 h. Cells were then fixed with formaldehyde and processed into soluble chromatin. Chromatin samples were immunoprecipitated with anti-pCIP antibody (A) or anti-p300 antibody (B), and bound DNA was quantitated by real time PCR. Each experiment was repeated at least three times, and the quantitative PCR analyses were performed in triplicate for each sample. The data are presented as percentages of input DNA before immunoprecipitation (mean ± S.E.). Error bars indicate standard error of three biological replicates with each quantitative PCR performed in triplicate.

 
Levels of RNA Polymerase II Associated with Transcription Initiation Were Reduced at the RAREs Regulating Expression of the Hoxa1 and Cyp26A1 Genes in F9 RAR{gamma}–/– Cells as Compared with F9 Wt and RARbeta –/–2 Cells—We also determined how the association of the initiating form of RNA polymerase II (initiating pol II) with the RAREs monitored in this study is perturbed in the F9 RARbeta –/–2 and the F9 RAR{gamma}–/– cell lines as compared with F9 Wt cells. We monitored the association of pol II by using an antibody that recognizes phosphorylated serine 5 of the CTD of initiating pol II (38). In other systems, serine 5 phosphorylation of the CTD has been associated with transcriptional initiation (39). Similar basal levels of initiating pol II were associated with the Hoxa1 RARE in F9 Wt, F9 RARbeta –/–2, and F9 RAR{gamma}–/– cells (Fig. 5A, left panel, 0h). Treatment of F9 Wt cells with RA resulted in a ~9.3-fold (p value <0.02) increase in the levels of initiating pol II associated with the Hoxa1 RARE. In the F9 RARbeta –/–2 cells, RA treatment resulted in a similar increase (~8.7-fold) in the levels of initiating pol II associated with the Hoxa1 RARE as compared with F9 Wt cells (Fig. 5A, left panel). However, in the F9 RAR{gamma}–/– cell line, initiating pol II levels rose only ~1.7-fold at the Hoxa1 RARE after RA treatment. Additionally the levels of initiating pol II associated with the Hoxa1 RARE in the F9 RAR{gamma}–/– cells never rose above the basal levels observed at this RARE in F9 Wt cells (Fig. 5A, left panel). These results suggest that in the absence of RAR{gamma}, RAR ({alpha} + beta)-RXR{alpha} heterodimers (Fig. 3A) bound at the Hoxa1 RARE are unable to associate and/or increase the levels of initiating pol II in response to RA treatment.


Figure 5
View larger version (23K):
[in this window]
[in a new window]

  		FIGURE 5.
The association patterns of polymerase II and SUZ12 to RAREs in response to RA. F9 Wt and F9 RAR{alpha}, -beta2, and -{gamma} null cell were either untreated or treated with 1 µM RA for 24 h, and then cells were fixed with formaldehyde and processed into soluble chromatin. Chromatin samples were immunoprecipitated with a monoclonal antibody that recognizes phosphorylated serine 5 of the CTD of RNA polymerase II (A) or an anti-SUZ12 antibody (B), and bound DNA was quantitated by real time PCR. Each experiment was repeated at least three times, and quantitative PCR analyses were performed in triplicate. The data are presented as percentages of input DNA before immunoprecipitation (mean ± S.E.). Error bars indicate standard error of three biological replicates with each quantitative PCR performed in triplicate.

 
The recruitment patterns of initiating pol II to the Cyp26A1 R1 RARE in the F9 RARbeta2 –/– and F9 RAR{gamma} –/– cell lines, as compared with F9 Wt cells, were as follows. Initiating pol II levels rose much higher at the Cyp26A1 R1 RARE (~90-fold in F9 Wt cells) as compared with the Hoxa1 RARE (~9.3-fold in F9 Wt cells) in response to RA (Fig. 5A, compare middle and left panels). The increase of initiating pol II levels at the Cyp26A1 R1 RARE (~12.3-fold) was much lower (p value <0.02) in the F9 RAR{gamma}–/– cells as compared with the increase observed in F9 Wt cells (~90-fold) (Fig. 5A, middle panel). Therefore, in the absence of RAR{gamma}, the RAR ({alpha} + beta)-RXR heterodimers bound at the Cyp26A1 RAREs (Fig. 3, B and C) were also unable to recruit and/or increase the levels of initiating pol II efficiently in response to RA treatment.

Differences in the Association of Initiating pol II with the RARbeta2 RARE in F9 RARbeta –/–2 Cells as Compared with F9 Wt and F9 RAR{gamma}–/– Cells—The recruitment patterns of initiating pol II to the RARbeta2 RARE in the F9 Wt, RARbeta –/–2, and RAR{gamma}–/– cells were different from those seen for the Hoxa1 and Cyp26A1 RAREs. In contrast to the levels of initiating pol II recruitment observed for the Hoxa1 and Cyp26A1 RAREs, the increase in initiating pol II recruited to the RARbeta2 RARE in RA-treated F9 RAR{gamma}–/– cells was ~5.4-fold (p value <0.02), and this increase was similar to the ~4-fold increase (p value <0.02) in levels of initiating pol II recruited to the RARbeta2 RARE in F9 Wt cells (Fig. 5A, right panel). No RA-mediated increase in initiating pol II associated with the RARbeta2 RARE was observed in the F9 RARbeta –/–2 cells, and these data are consistent with the observation that RARbeta2 transcripts were not expressed in F9 RARbeta –/–2 cells (Fig. 5A, right panel). The high basal levels of initiating pol II associated with the RARbeta2 RARE in the three cell lines studied are consistent with previous reports (32, 40). We concluded that although RAR{gamma} was associated with the RARbeta2 RARE in F9 RARbeta –/–2 cells (Fig. 3D) the presence of RAR{gamma} is not sufficient for RA-mediated recruitment and/or increase in the levels of initiating pol II (Fig. 5A) or the RA-induced transcription of the RARbeta2 gene (Fig. 1, B and C).

The Polycomb Group Protein SUZ12 Is Not Efficiently Displaced from the Hoxa1 and Cyp26A1 R1 RAREs in F9 RAR{gamma}–/– Cells in Response to RA—We have previously shown that the polycomb group protein SUZ12 is associated with the Hoxa1, Cyp26A1 R1, and RARbeta2 RAREs in F9 Wt cells and that dissociation of SUZ12 occurs upon exposure to RA (25). Polycomb proteins have been shown to negatively regulate Hox genes (41–43), and the association of SUZ12 with numerous RAREs suggests that polycomb proteins may have a more global role in the repression of RA target genes. We therefore addressed the question of whether there were differences in the association patterns of SUZ12 with the Hoxa1, Cyp26A1, and RARbeta2 RAREs in the F9 RARbeta –/–2 and F9 RAR{gamma}–/– cell lines as compared with F9 Wt cells.

The levels of SUZ12 associated with the Hoxa1 RARE in F9 Wt cells decreased ~33-fold in response to RA treatment (Fig. 5B, left panel). An ~8.3-fold decrease in the level of SUZ12 associated with the Hoxa1 RARE was seen in F9 RARbeta –/–2 cells after RA treatment. Although the basal levels of SUZ12 associated with the Hoxa1 RARE were similar in F9 Wt and F9 RARbeta –/–2 cells (Fig. 5B, left panel, 0h), RA treatment caused a ~3.4-fold greater decrease in the levels of SUZ12 associated with the Hoxa1 RARE in F9 Wt cells as compared with F9 RARbeta –/–2 cells (Fig. 5B, left panel, 24 h; p value <0.01).

The basal levels of SUZ12 associated with the Hoxa1 RARE in F9 RAR{gamma}–/– cells were also statistically similar to the basal levels of SUZ12 associated with this RARE in F9 Wt cells (Fig. 5B, left panel, 0h; p value >0.10). However, ~26-fold lower levels of SUZ12 remain associated with the Hoxa1 RARE in F9 Wt cells as compared with F9 RAR{gamma}–/– cells after RA treatment (Fig. 5B, left panel, 24 h). In fact, the level of SUZ12 associated with the Hoxa1 RARE in RA-treated F9 RAR{gamma}–/– cells was similar to (<2-fold difference) the level seen in untreated F9 RAR{gamma}–/– cells (Fig. 5A, left panel; p value <0.05).

The association patterns of SUZ12 with the Cyp26A1 R1 RARE before and after RA treatment in the F9 Wt, RARbeta –/–2, and RAR{gamma}–/– cell lines were as follows. The basal levels of SUZ12 associated with the Cyp26A1 R1 were similar in the three F9 cell lines (Fig. 5B, middle panel, 0h). However, after RA treatment the levels of SUZ12 associated with the Cyp26A1 R1 RARE were ~18.3-fold lower in F9 Wt cells than in F9 RAR{gamma}–/– cells (p value <0.0001) and ~2.7-fold lower in F9 Wt cells as compared with F9 RARbeta –/–2 cells (Fig. 5B, middle panel, 24 h; p value <0.01).

These results demonstrate that the addition of RA does not result in the efficient displacement of SUZ12 from the Hoxa1 and Cyp26A1 RAREs in F9 RAR{gamma}–/– cells. Additionally the levels of SUZ12 displaced from the Hoxa1 RARE after RA treatment in the three cell lines were inversely related to the expression levels of Hoxa1 in these same lines. Whereas transcription of Hoxa1 was strongly induced by RA in F9 Wt cells, transcription of Hoxa1 was reduced by ~50% in F9 RARbeta –/–2 cells, and it was almost completely abrogated in the F9 RAR{gamma}–/– cell line (Fig. 1A). Conversely after RA treatment the highest levels of SUZ12 were associated with the Hoxa1 RARE in the RAR{gamma}–/– cells with lower levels in F9 RARbeta –/–2 cells and the lowest level in the F9 Wt cells (Fig. 5A, left panel). An inverse relationship between the expression of Cyp26A1 mRNA in the three cell lines and the level of SUZ12 associated with the Cyp26A1 R1 RARE in these three same cell lines was also observed (Fig. 1, B and C, and Fig. 5A, middle panel).

The Levels of SUZ12 Associated with the RARbeta2 RARE Correlate with Transcription and the Presence of RAR{gamma}—The levels of SUZ12 associated with the RARbeta2 RARE decreased ~3.4-fold (p value <0.01) in F9 Wt cells and ~7.9-fold (p value <0.06) in the F9 RAR{gamma}–/– cell line after exposure to RA (Fig. 5B, right panel). However, in the F9 RARbeta –/–2 cell line, which does not express RARbeta2 transcripts, an RA-dependent decrease in association of SUZ12 with the RARbeta2 RARE was not observed (Fig. 5A, right panel). Therefore, these results indicate that RA-associated transcription of the RARbeta2 gene is correlated with the displacement of SUZ12 from the RARbeta2 RARE.

Interestingly higher levels of SUZ12 were associated with the RARbeta2 RARE in untreated F9 RAR{gamma}–/– cells as compared with untreated F9 RARbeta –/–2 cells (p value <0.06) and untreated F9 Wt cells (p value <0.06). Additionally the levels of SUZ12 that were associated with the RARbeta2 RARE were similar in RA-treated RAR{gamma}–/– cells and RA-treated F9 RARbeta –/–2 cells (Fig. 5B, right panel, 24 h), although transcription of RARbeta2 mRNA was higher in RA-treated RAR{gamma}–/– cells as compared with RA-treated F9 RARbeta –/–2 cells (Fig. 1, B and C). Furthermore the levels of SUZ12 associated with the RARbeta2 RARE were ~3.5-fold higher in RA-treated F9 RAR{gamma}–/– cells as compared with RA-treated F9 Wt cells (Fig. 5B, 24 h; p value <0.01). Therefore, irrespective of the transcriptional status of a gene, the levels of SUZ12 associated with a RARE may be influenced by the presence of RAR{gamma}. The binding of RAR{gamma} to a RARE may block the association of SUZ12 with the aforementioned RARE. In the absence of RAR{gamma}, higher levels of SUZ12 may be able to associate with a RARE.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Previous studies have demonstrated that the RAR{alpha}, -beta2, and -{gamma} isotypes each regulate a specific subset of target genes (44). In this study, we used chromatin immunoprecipitation assays to monitor the association of RAR{gamma}, RXR{alpha}, and other proteins involved in transcription with the RARbeta2, Hoxa1, and Cyp26A1 RAREs in F9 Wt cells. We then monitored the association patterns of these factors to the aforementioned RAREs in F9 RAR{alpha}, -beta2, and -{gamma} null cell lines and compared these results to what was observed in the F9 Wt cells. Additionally we determined the expression of the corresponding target genes in the F9 Wt and the F9 RAR{alpha}, -beta2, and -{gamma} null cell lines via quantitative real time RT-PCR. By using this approach we were able to demonstrate that the association of RAR{gamma} and/or RXR{alpha} with a RARE does not suffice for the RA-mediated transcription of the corresponding target gene. However, the ability of RAR{gamma} and/or RXR{alpha} to recruit the coactivators pCIP and p300 to an RARE did correlate with the RA-mediated transcription of the corresponding target mRNAs. Additionally the ability of RAR{gamma} and/or RXR{alpha} to recruit and/or increase the levels of initiating RNA polymerase II to a RARE in response to RA mirrored the ability to recruit pCIP and p300. From these results we conclude that the specific functions of the RAR isotypes do not manifest at the level of DNA binding but rather from a differential ability to recruit specific components of the transcriptional machinery. We also demonstrated that the RA-mediated displacement of SUZ12 from a RARE was greatly inhibited in the absence of RAR{gamma}. Thus we demonstrated that components of the RAR signaling machinery are specifically required for the displacement of SUZ12 from RAREs during the RA-mediated differentiation of F9 cells.

RAR Isotype-specific Function Does Not Manifest at the Level of DNA Binding—RAR-RXR heterodimers bind to RAREs, which are composed typically of two direct repeats of a core hexameric motif, PuG(G/T) TCA where Pu is a purine, spaced by 5 bp (referred to as DR5) or less commonly 2 bp (DR2) (6). The DNA binding domain is the most highly conserved region among the RAR isotypes (6). In vitro studies have demonstrated that RAR isotypes can bind to the same sequences as each RAR isotype could bind to the Hoxa1/RARbeta2 RARE in gel shift mobility assays (14, 20). Additionally although the RA-induced expression of Hoxa1 and RARbeta2 mRNAs is regulated by the same DR5 element (Fig. 2A), expression of Hoxa1 in F9 cells requires RAR{gamma}, whereas RARbeta2 is required for its own expression (Fig. 1, B and C). These results suggest that each of the RAR isotypes has the ability to associate with a given RARE and that the specific functions of a RAR do not manifest at the level of DNA binding. Consistent with this conclusion, we demonstrated that RXR{alpha}, presumably bound as a heterodimer with RAR{alpha} or RARbeta, was associated with the Hoxa1 and Cyp26A1 RAREs in F9 RAR{gamma}–/– cells (Fig. 3, A, B, and C, left panels) even though these two genes are not induced by RA in this cell line (Fig. 1, B and C). Additionally we showed that RAR{gamma} was associated with the RARbeta2 RARE in F9 Wt, F9 RAR{alpha}–/–, and F9 RARbeta –/–2 cells (Fig. 3D, middle panel) even though RAR{gamma} was not required for the expression of RAR