The retinoic acid receptor-2 contains two separate cell specific transactivation domains, in the A-region and the hormone binding domain

Gert E. Folkers, Bas-Jan M. van der Leede, and Paul T. van der Saag

Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Utrecht

Molecular Endocrinology 7:616-627 (1993)
 

Abstract

In contrast to other members of the steroid/thyroid hormone superfamily not much is known about the regions involved in transactivation of the receptors for retinoic acid. To determine the transactivation function of RARs, fusion-proteins between the DNA binding domain of the yeast transcription factor GAL4 and RAR or RAR were made. Transfection of these constructs resulted in RA-induced activation of a GAL4 responsive element- containing promoter. Deletion analysis revealed that RAR-2 has two transcription activation functions (TAFs). TAF-1 activates transcription constitutively and was mapped to the first thirty-two amino acids of the A-region. TAF-2 is located in the ligand-binding domain between amino acid 137 and 410 and activated transcription only in the presence of retinoic acid. The presence of two TAFs was confirmed by co-transfection of RAR deletion constructs with the hRAR-2 promoter as reporter, showing that the absence of RAR TAF-1 causes a decrease in transactivation, whereas truncation of TAF-2 completely blocks this function. Internal deletions in the ligand-binding domain in both GAL-RAR and RAR expression constructs resulted in a non-functional receptor, indicating that the complete ligand-binding domain is required for its transactivation function. Furthermore we have shown that the contribution of the two TAFs in transcription activation varies between different cell lines, suggesting that they act in a cell-specific manner.

Back

Introduction

Retinoic acid (RA) has been shown to be a regulator of cell growth and differentiation (1); moreover an important role has been suggested for RA during embryonic development (2). Cloning of the receptors for RA, hRAR (3,4), hRAR (5) and hRAR (6) has made it possible to study the mechanism by which RA exerts its regulatory function. RA has been shown to activate transcription via retinoic acid response elements (RAREs), as found in the hRAR-2 promoter; these consist of the direct repeat ^A/_GGTTCA with 5 bp spacing between the two half sites (7,8). Recently members of a closely related family of retinoid receptors (RXRs) have been identified (9) that were shown to bind with high affinity to the retinoic acid metabolite 9-cis RA (10,11). These proteins also appeared to be required for high affinity binding of RARs to RAREs (12,13,14,15).

Both RARs and RXRs are members of the steroid/thyroid hormone receptor superfamily and function as ligand-dependent transcription factors. These proteins share a similar domain structure and, based on their amino acid similarity, they can be divided in 6 sub-domains, namely A-F (for reviews see 16,17). The DNA binding domain (DBD or C-region) is the most conserved region and contains the two zinc finger structures necessary for binding to a hormone response element (HRE). The E-region mediates ligand binding, dimerization and transactivation. Forman and Samuels have put forward a model that couples these events structurally and functionally (18). Members of this receptor family are believed to bind as homo- or heterodimers to their response elements. This dimerization is achieved by 9 heptad repeats, consisting of hydrophobic amino acids at position 1 and 8 of every repeat (18). In the absence of hormone (ligand) the receptor is inactive; this is possibly caused by a region designated _i. The repression is released upon ligand binding to two regions at the N- and C-terminus of the ligand-binding domain, designated ligand 1 and ligand 2. The model proposes that through binding of the ligand a conformational change in the protein structure leads to inactivation of _i, thereby allowing activation.

It is known that transcriptional activation of steroid hormone receptors is achieved through two transcriptional activation functions (TAFs, for review see 19) within the molecule. TAF-1 is located in the AB region and fusion of this region of the estrogen receptor (ER) to the DBD of the yeast transcription factor GAL4 results in a hormone-independent activation of a GAL-response element-containing promoter. Deletion of this region in ER results in reduced activation of ERE-containing reporter-plasmids (20). The second TAF is located in the hormone-binding domain (HBD) and is ligand-inducible (20,21). Fusion of all the individual exons of ER-HBD to GAL-DBD revealed that this transactivation function is not located in one exon (22), indicating that the tertiary structure of the HBD is required for its activation function. In the glucocorticoid receptor (GR) a region located between the DBD and the HBD named tau-2, consisting of thirty predominantly acidic amino acids (aa), appeared to possess a transactivation function (23). Moreover, is has been found that GR also contains a hormone-induced activation function (21).

It has been shown that these TAFs act in a promoter- and cell-specific fashion. TAF-1 contributes mainly to the transactivation of minimal promoters, whereas on more complex promoters transcription is enhanced by TAF-2 in the presence of its hormone (24). Also the relative contribution of the two TAFs to the activity of GR, ER and PR (progesterone receptor) is variable in different cell lines (25).

The mechanisms by which transcription activation of RARs is achieved are unknown. In order to elucidate this we have studied the transactivation function of RAR. In this paper we show that RAR-2 has two TAFs; TAF-1 is constitutively active and is located in the first thirty-two aa of the A-region, while TAF-2 co-localizes with the hormone-binding domain, between aa 137 and 410 and activates transcription in a ligand-dependent fashion. Deletions in this region completely abolish the transactivation function, indicating that the entire ligand-binding domain is required for this function. Furthermore we show that both TAFs can activate transcription in a cell-specific manner.

back

Material and Methods
 

Cell culture and transient transfections

Cells were cultured as monolayers on gelatin-coated dishes in a 1:1 mixture of Dulbecco's modified Eagle medium and Ham's F12 (DF), supplemented by 7.5% (v/v) fetal calf serum (FCS) at 37^oC and 7.5% CO₂. For transient transfection assays 10⁵, 8x10⁴, 2.5x10⁵, 2x10⁵ P19-EC, RAC65, 293 and COS-1 cells respectively, were cultured in a tissue culture dish (diameter 3.5 cm). 24 hrs after plating the cells were transfected by the calcium phosphate co-precipitation method (26) with 10µg reporter plasmid, 1µg expression-vector (PSG5, GAL-RAR fusion-proteins) and 1.5 µg SV2-Lac-Z or PDM-Lac-Z as an internal control for transfection efficiency. After 18 hrs the precipitate was removed and fresh medium, containing 1 µM RA was added as indicated. After an additional 24 hrs cells were harvested in phosphate-buffered saline (PBS) without calcium or magnesium salts and lysed by three subsequent cycles of freeze/thaw in 250 mM Tris pH 7.5 and 25 mM EDTA. CAT assays were performed as described by Gorman et al (27). After separation of the reaction products by thin layer chromatography quantification was carried out with a Phospho-Imager using Image Quant software (Molecular Dynamics).
 

Construction of expesssion plasmids

GAL-RAR was constructed by the cloning of a Kpn1-Stu1 fragment into pBluescript SK^- (Stratagene); from this plasmid a Kpn1-Xba1 fragment was cloned in pSG424 (GAL-DBD). GAL-RAR was generated by cloning the Sau3A-Sal1 and Sal1-BamH1 fragments of pSG5-RAR into the BamH1 site of pSG424.

GAL-RAR -F and -EF were made by BamH1 and partial EcoR1 digestion, the restriction-sites were made blunt with the Klenow fragment and re-ligated. GAL-RAR -DF and -CF were constructed by cloning the Sau3A-Sal1 and Sau3A-Xho1 fragments respectively in the BamH1-Sal1 digested GAL-DBD vector. Digestion of GAL-RAR with Xho1 and Sal1 resulted after religation in GAL-RAR -C. GAL-RAR -AC and -AB were generated by cloning the Xho1-BamH1 and Sal1-BamH1 fragments of pSG5-RAR in the Sal1-BamH1 sites of pSG424.

1-76 is the construct GAL-RAR -CF; 1-61, 1-32, 1-10 were constructed by cloning the blunt Sma1-Dde1, Sma1-Alu1, Sma1-HinF1 fragments respectively of GAL-RAR -CF in the Sma1 site of pSG424. Cloning of the blunt ended HinF1-Xba1 and HinF1-HinF1 fragments of GAL-RAR- CF in the blunt BamH1 site of pSG424 resulted in the constructs 11-76, 22-76 and 11-22. Cloning of the blunt Alu1-Xho1 fragment in the Sma1 site of pSG424 resulted in the construct 60-76. 11-22 was made by cloning the two GAL-RAR -CF fragments Sma1-HinF1 and HinF1-Xba1 in the Sma1-Xba1 site of pSG424.

1-410, 1-204, 1-137, 137-448 are respectively GAL-RAR -F, GAL-RAR -EF, GAL-RAR -DF, GAL-RAR -AC. 1-370, 1-320, 1-224 were constructed by digesting GAL-RAR with BamH1 and partial digestion with Bgl2, Stu1 or Msc1 respectively, filling in the sites followed by ligation. 232-448, 321-448, 370-448 were made by ligation of the following GAL-RAR fragments: partial Sau3A-Xba1, Stu1-Xba1, partial Bgl2-Xba1 in pSG424 digested with BamH1-Xba1, Sma1-Xba1 and BamH1-Xba1 respectively. 226-270 was made by digesting GAL-RAR with Msc1 and Xba1, and cloning the partially digested and filled in Acc1-Xba1 fragment of GAL-RAR in this vector. 296-320 was constructed by partial Taq1 digestion of GAL-RAR, filling in this site, digesting with Xba1 and ligating the fragment Stu1-Xba1 in this vector. 321-365 was generated by cloning the blunt ended Afl3-Xba1 fragment in GAL-RAR, which was digested with Stu1-Xba1.

The pSG5-RAR construct RAR -F was made by partial digestion with EcoR1 and BamH1 which were filled in before ligation. RAR -E was made by partial EcoR1 digestion and subsequent ligation. E1 and E2 were made by cloning the Cla1-BamH1 fragments of 226-270 and 321-365 respectively in the Cla1-BamH1 sites of pSG5-RAR. RAR-AB was made by cloning the Nde1-Cla1 fragment of RAR-DBD in the blunt Sst1 and the Cla1 site of pSG5-RAR.

All constructs were sequenced to confirm the reading frame using the T7 sequencing kit (Pharmacia).
 

Preparation of whole cell extracts and gel mobility shift assay

Ten µg of various GAL-RAR-expression-plasmids were transfected together with 2 ug of SV2-Lac-Z as described above in either P19-EC or COS-cells. Cells were harvested in cold PBSø and after a short spin resauspended in 50 µl 20mM Tris pH 7.5, 20% glycerol, 400mM KCl, 2mM DTT 1mM PMSF and 0,1µl aprotinin (30 TIU/ml) and lysed by three subsequent rounds of freeze/thaw. Gel mobility shift assays were performed in 20 ul containing 4 µl whole cell extract (5 µg) and 6 µl Dignam D in 2mM MgCl_2, 50mM NaCl, 10mM Tris pH 7.5, 1mM EDTA, 5% glycerol (w/v) 2 µg poly dIdC and 1mM DTT. After addition of 0.1 ng end labeled double stranded consensus GAL-4 binding site (agcttCGGAGGACAGTCCTCCGc, 10⁸cpm/ug) the mixture was incubated for 30 minutes on ice and thereafter loaded on a 4%PAA gel (0.5x TBE). The specifity of the complexes was confirmed by competition with a 100 fold molar excess of unlabeled GAL-4 oligo. No competition was observed with an unrelated oligo.

back

Results

Fusion proteins between GAL-DBD and RAR activate transcription in a ligand-dependent manner

In order to identify receptor domains involved in transcriptional activation two different types of experiments can be performed. Firstly, deletions or mutations can be introduced in the receptor resulting in a protein with altered transactivation properties. In this study we used the hRAR-2 promoter fused to the bacterial CAT gene and cotransfected different RAR expression plasmids (Fig.1). However this has the disadvantage that mutations in regions not directly involved in transactivation may affect this function by disrupting for instance the DNA binding or dimerization domain, thereby disabling the transactivation function of the receptor indirectly. To overcome this problem the DBD of the yeast transcription factor GAL4 has already been very useful and therefore we have chosen for this option. In addition to DNA-binding activity to the GAL4 binding site (UAS) this DBD also contains a dimerization (28) and a nuclear localization domain (29), which makes it possible to study activation functions of a protein upon fusion. Thus we co-transfected fusion constructs of RAR and GAL-DBD together with a GAL-responsive reporter construct containing an E1b TATA-box with five GAL-response elements fused to the CAT gene (Fig 1). Without co-transfection of an expression plasmid no activity could be detected on the GAL-responsive promoter in embryonal carcinoma cells (P19-EC), while the empty expression vector GAL-DBD showed only very low activity and was not affected by RA (not shown). Fusion proteins of GAL-DBD and RAR or RAR however showed at least a 200 times higher activity upon addition of RA than the empty expression vector (Fig. 2). As a positive control GAL-VP16, which is known to contain one of the most potent activators (30), was transfected. Promoter-activation by GAL-RAR constructs was approximately 10 times lower than by GAL-VP16 (data not shown) In the absence of RA only GAL-RAR activated the GAL responsive promoter to a limited extent, which can be caused by the presence of a constitutive activation function.

RAR has two activation functions

As described in the preceding part the GAL-RAR fusion proteins showed a respectable induction by RA, while RAR activated transcription also in the absence of RA. To investigate how activation occurs deletion constructs of GAL-RAR as shown in figure 3A were transfected in P19-EC cells. To confirm the expression of the different fusion-proteins, a gel mobility shift assay on a consensus GAL-binding site was performed with whole cell extracts of both P19-EC and COS-1 cells transfected with these constructs. Figure 3B shows a representative experiment with extracts from COS-1 cells. All proteins were expressed albeit with some variations both between different experiments and with the various constucts. Quantification of gels from two different series of transfections with the Phospho-imager revealed that expression levels of the different fusion-proteins varied 2 to 3 fold maximally (data not shown). In the transfection experiments the activity of GAL-RAR in the presence of RA is set at 100%. As can be seen in Fig. 3B, deletion of the F-region resulted in a slight decrease in activity in the presence of RA (85% of GAL-RAR), showing that this region is probably not of great importance for transactivation. To date no specific function has been assigned to the F-region; it could contribute to determine the structure or stability of the protein. Further deletion of almost the entire E-region resulted in complete loss of activation potential, comparable to levels found with the empty expression vector GAL-DBD; this indicates the presence of an activation function in this region (TAF-2). Deletion of the complete ligand- and DNA-binding domain (GAL-RAR-CF) however resulted in RA-independent activation of the GAL-responsive promoter. Moreover a construct lacking this region (GAL-RAR-AB) activated the reporter also to a lesser extent than the full length GAL-RAR protein confirming the existence of a TAF in the AB-region. Surprisingly, deletion of only the C-region elevated the RA-induced activation by more than two-fold, while the construct lacking both AB and C also gave a higher activity than GAL-RAR -AB in the presence of RA. This could possibly be explained by the presence of two DNA-binding domains that cause a change in the protein structure leading to a less efficient DNA binding or transactivation function.

Transfection of GAL-RAR in the absence of RA showed some transactivation of a GAL-responsive promoter, probably caused by the presence of a constitutive activator. Furthermore we observed that GAL-RAR -CF is more active than GAL-RAR in the absence of RA, which indicates that TAF-1 is responsible for the constitutive activation function and that the ligand-binding domain affects this activity, since it represses it in the absence of RA (see Fig. 5 and discussion). However, the construct lacking TAF-1 (GAL-RAR -AB) also has activity in the absence of RA, albeit less than constructs containing TAF-1. An explanation for this RA independent activity of the ligand binding domain could be that protein structure has been altered by the fusion with GAL-DBD, thus causing the repression of TAF-2 to be incomplete in the absence of ligand. Another possible explanation for this activity could be the presence of residual RA in the culture medium. This hypothesis is supported by our observations that cells grown in charcoal-stripped serum resulted in a 3 to 4 fold lower activity of GAL-RAR -AC, whereas the activity in the presence of RA was the same as in the normal medium (data not shown). From these data we can not exclude the presence of a weak constitutive activation function although we were unable to etect any sequences whithin the ligand binding domain causing this activity in the absence ligand (see Fig. 5). From these data however, it is clear that RAR has two domains involved in transactivation: TAF-1 located in the AB-region is constitutively active and TAF-2 in the ligand-binding domain activates transcription in the presence of RA.
 

TAF-1 is located at the N-terminus of RAR between aa 1 and 32

We have shown that the TAF-1 containing construct (GAL-RAR- CF) activated transcription equally well in the presence and absence of RA (Fig. 3B) and that it is located between aa 1 and 76. We wanted to specify the activity further by using deletion constructs of this region fused to GAL-DBD (Fig. 4). These constructs were transfected into P19-EC cells; Table 1 shows the results of these transfections in the absence of RA. Similar results were obtained when RA was added to the medium, the presence of all the different GAL-RAR TAF-1 fusion-proteins was confirmed by gel mobility shift assays (data not shown).

A construct lacking only fifteen aa at the C-terminal part (B-region) was not as active as GAL-RAR 1-76 (=GAL-RAR -CF) while further truncation in the AB-region of the next twenty-nine C-terminal amino acids (GAL-RAR 1-32) permitted higher transcriptional activation than GAL-RAR 1-76. This could be explained by the fact that the protein structure of this relatively small protein is more accessible for protein-protein interactions which are known to be required for transactivation. A GAL-RAR fusion protein containing only the ten most N-terminal amino acids activated transcription only to a limited extent, suggesting that the transactivation function is located between aa 1 and 32. This was confirmed by a truncation of only ten aa at the N-terminus resulting in near background levels of transactivation, while upon subsequent deletion of the next eleven aa at the N-terminal region (=GAL-RAR 22-76) this residual activity was also lost, indicating that the A-region of RAR is effectively needed for transactivation. To prove that the first thirty-two aa are involved, a construct with an internal deletion of aa 11 to 22 was transfected in P19-EC cells resulting in a significant reduction in activation; this 12 amino acids containing region on its own is not sufficientfor full activation. When fused to GAL-DBD (GAL-RAR 11-22), this fusion-protein showed only a limited activation of a GAL-responsive promoter. From these experiments we conclude that RAR TAF-1 is located between amino acid 1 and 32.

TAF-2 requires the complete ligand-binding domain

To further characterize TAF-2 located in the ligand-binding domain and which functions as a RA-inducible activator, a series of deletion constructs comprising this region were prepared and transfected into P19-EC cells. Figure 5 shows the CAT activities of a representative experiment using the different TAF-2 -containing constructs. It is clear that any deletion in the ligand-binding domain, between aa 137 and 410, resulted in a complete loss of the activation function. Deletion of only forty aa at the C- terminal region (1-370), of the smallest construct still showing activity (GAL-RAR -F, 1-410) or deletion of ninety-five aa at the N-terminus (232-448), resulting in a lack of the D-region and part of the E-region, completely blocks the RA-induced activation function. All further C- and N- terminal truncations in this region showed no detectable RA-induced activation of transcription. Subsequently internal deletions in this region were made to investigate whether the complete ligand-binding domain is required for this function. Transfection into P19-EC cells revealed that successive deletions of 45, 25 and 45 aa in the ligand-binding domain also completely inhibited the activation function (Fig. 5). Only the constructs 1-224 and 1-204 showed a 5-10 fold higher activity than the empty expression-vector GAL-DBD; this is most likely caused by TAF-1, the activity of which is not affected by RA. Interestingly the activity of all C-terminal truncations and internal deletions between aa 137 and 410 showed a much lower activity than GAL-RAR 1-76 (maximally 5% of this construct, Fig. 3B and data not shown) despite the fact that they all contain TAF-1. An explanation for this observation could be that one or more repression functions are present between aa 76 and 410. Recently it has been shown that such silencing domains are present in hRAR (31) which could contribute to the lack of RA-independent activity of the TAF-1 -containing C-terminal deletion constructs. However we cannot rule out the possibility that the presence of a truncated ligand-binding domain results in a tertiary protein structure which prevents activation by TAF-1.

Since any deletion made in the ligand binding leads to a complete loss of the transactivation function of this region we conclude that probably the whole ligand binding domain is required for its transactivation function.
 
 
 
 

Different roles for RAR TAF-1 and TAF-2 in transactivation

To confirm that TAFs identified with the RAR-GAL fusion approach as described above, contribute to the transactivation properties of this receptor, deletion constructs of the RAR cDNA were prepared (Fig. 6A) and transfected into RAC-65 cells, together with a hRAR-2 promoter construct (Fig. 1). This cell line is derived from P19-EC cells and is resistant to the growth-inhibitory and differentiation-inducing activity of RA (32,33). Without cotransfection only a limited RA-induced activation was observed (fig 6B), which is caused by the presence of a truncated RAR protein, thought to be at least partially the reason for its resistance (34,35). Co-transfection of hRAR-2 expression plasmid resulted in a large induction of the hRAR-2 promoter (100 fold). A construct lacking the F region showed a somewhat higher basal level of activity than hRAR, but the transactivation in the presence of RA is almost the same, again indicating that this region is not directly involved in transactivation (see Fig. 3). The higher basal level found here and the lower activities of the GAL-RAR construct lacking the F-region (in the presence of RA) indicate that this region somehow, by an unknown mechanism, contributes to the receptor stability or activity. Deletion of the complete E region however resulted in a basal level of activity, while receptors with internal deletions in this region also did not activate the hRAR promoter. According to the model of Forman and Samuels (18) deletions in the E-region destroy either the ligand-binding domain, the dimerization domain, and _i or the transactivation region. Mutations in any of these regions can cause a disruption of secondary or tertiary protein structure, thereby creating a receptor unable to activate a RARE-containing promoter. It should be mentioned here that the same mutations when fused to the DBD of GAL4 were also not capable of activating transcription from a GAL-responsive promoter (Fig. 5). Therefore these data taken together demonstrate that the entire ligand-binding domain is required for its function as a transactivator.

We finally examined the influence of TAF-1 on receptor activity by creating a construct which lacks the AB-region. Upon transfection of this construct into RAC-65 cells in the presence of RA, activation of the hRAR-promoter was observed for only 40% of the activity of full length RAR. To exclude the possibility that the results with these TAF-1-lacking constructs are influenced by the presence of the truncated receptor in RAC-65 cells (34,35), co-transfection was also performed in COS-1 cells. In this cell line a lower level of transactivation of the hRAR promoter was observed also compared to that with the full length receptor (50% reduction, data not shown), confirming that there is a TAF present in this region.

RAR-2 TAF-1 and TAF-2 activate transcription in a cell-specific manner

For a limited number of members of the steroid hormone family (GR, ER and PR), cell specificity for functioning of TAFs has been demonstrated (25): in one cell line TAF-1 is more active whereas in another cell line TAF-2 has the highest contribution to overall receptor activity, in the presence of its corresponding hormone. These findings indicate that different activation mechanisms are involved in the functioning of the different TAFs. To analyze whether RAR-TAFs can activate transcription in a cell-specific way, we first transfected GAL-RAR and GAL-RAR constructs in P19-EC, the foetal kidney cell line 293, NIH-3T3 and COS-1 cells. Only small but significant differences were observed between the two constructs in the cell lines tested, both in the presence and absence of RA (data not shown).

This observation left open the possibility that the relative contributions of both TAFs in receptor activity differ between cell lines. To test this hypothesis constructs containg only TAF-1 (GAL-RAR -CF), TAF-2 (GAL-RAR -AC), or both TAFs (GAL-RAR) were transfected in P19-EC, 293, NIH-3T3 and COS-1 cells. Figure 7 shows a representative experiment of these transfections and quantification of three independent experiments is presented in Table 2. Significant differences were observed : the construct lacking TAF-1 showed a lower activity than GAL-RAR in P19-EC and 293 cells in the presence of RA, whereas in 3T3 and COS cells a higher activity of TAF-2 (in the presence of RA) was observed when compared to the construct containing both TAFs (Table 2). More pronounced differences in activation function between the various cell lines were seen when the construct containing only TAF-1 was transfected. Both in the presence and absence of RA, TAF-1 transactivation was less than that of full length receptor in P19-EC and 293 cells, whereas in 3T3 and COS cells TAF-1 activated transcription more than GAL-RAR containing both TAFs (Table 2, Fig. 7). In P19-EC and 293 cells the sum of the activity of the two TAFs separately is approximately equal to that of the construct containing both TAFs, while in 3T3 and COS cells the level of transcriptional activation by the two TAFs separately is 3-4 times higher than the activity of GAL-RAR. Since the latter construct contains the DBD of both RAR and GAL4, it is possible that the presence of two DBDs affects this activity. It is unlikely that the differences in transactivation by the various TAFs, observed between the cell-lines are caused by variation in the expression-levels of the GAL-RAR proteins since in both COS-1 and P19-EC cells no significant differences in expression-levels between the fusion-proteins were observed. We observed however differences in the expression-levels between both cell lines, COS having a 10-20 fold higher expression than P19-EC cells, as determined by gel mobility shift assays (Fig. 3B, data not shown). At present we can not exclude that the differences are the consequence of these elevated expression-levels. In conclusion, it is clear from these data that the relative contribution of TAF-1 and TAF-2 to overall receptor activity is different in different cell lines and therefor we conclude that the TAFs present in RAR-2 can activate transcription in a cell-specific manner.

back

Discussion

We have shown here that RAR-2 contains two transactivation functions, the first of which (TAF-1) is located in the N-terminal region and is constitutively active, and the other if which (TAF-2) is located within the ligand-binding domain and functions in a ligand-dependent manner. We identified these activities on the basis of two criteria: firstly by fusion of potential TAFs to the heterologous DBD domain of the yeast transcription factor GAL4, which resulted in an activation of a GAL-responsive promoter. Secondly, deletion of such a region decreases the activity of the receptor, both when fused to GAL-DBD and also in a RAR-cDNA construct lacking the TAF.

Deletion of the AB-region of RAR resulted in a reduced RA-dependent hRAR-2 promoter activity in both RAC-65 and COS cells. Deletion analysis of this region when fused to GAL-DBD showed that the N-terminal thirty-two aa contain this activation function. Upon screening of this region against the EMBL-Databank (version 30) no homologies with other known activators were found. Sequence analysis revealed that this region has an overall negative charge but it does not contain many acidic amino acids; furthermore no glutamine- or proline-rich stretches of amino acids are present, which are known from other transcriptional activators to be involved in transactivation (for a review see 36). However, the presence of serine residues allows phosphorylation to be involved in the transactivation process. Interestingly phosphorylation has been observed in the N-terminus of the PR (37), GR (38) and the thyroid hormone receptor (c-ERBA-, 39). The latter receptor was shown to be phosphorylated by casein kinase 2 (CK-2). In the N-terminal thirty-two aa of RAR some potential casein kinase sites can be found. Recently it has also been shown that the N-terminus of RAR can be phosporylated (40), indicating that phosphorylation is potentially an important feature of this receptor as well. Functionality of phosphorylation was confirmed by a mutation of Ser 51 in the N-terminus of the vitamin D3 receptor, thus preventing phosphorylation by PKC in vitro and resulting in a markedly impaired transactivation by vitamin D3 through this mutated receptor upon transfection. (41). Further research is presently underway to investigate whether phosphorylation of the A-region is involved in transactivation by RAR.

The transactivation function assigned to the hormone binding domain could not be limited to a region of thirty aa as in the case of GR (tau2, 18); however we observed that the entire ligand-binding domain is required for its function as is the case with AR (42) and T3R (31). Any deletion in the D-E region in both GAL-RAR and pSG5-RAR expression constructs completely blocks transactivation by RA. A truncation of only forty aa at the C-terminus of the ligand-binding domain thus deleting heptad repeat 9 and ligand 2, caused inactivity. In fact a recent study by Danielian et al (43) confirmed the importance of this region by showing that a conserved region between aa 538 and 552 of the mouse ER (RAR: 398-413) is involved in transactivation. Despite the fact that mutations in this region did not influence hormone- or DNA-binding a significant reduction of ligand-dependent transactivation was observed. Fusion of the corresponding RAR region to GAL-DBD is not sufficient for transactivation since the construct GAL-RAR 370-448 is inactive. The RAR mutant E2 lacking heptad repeat 5-8 also showed no activity; this is most likely caused by its inability to dimerize. The link between dimerization, proposed to be achieved by the heptad repeats, and transactivation has been shown by many groups. Mutations in T3R-1 between aa 286 and 305 (RAR: 235-254) did not affect hormone binding and nuclear localization but both transactivation and interaction with a nuclear protein (TRAP) (44,45) was prevented; heterodimerization between RAR and T3R also required this region (46). Recently it has been shown that RAR, T3R, VD3 heterodimerize with RXR and also in this case the heptad repeats seem to be involved (12). Together these data confirm the idea that the heptad repeat region is important for dimerization and that this dimerization is required for transactivation. However, deletions in this region of GAL-RAR fusion constructs caused also inactivation (GAL-RAR 296-320 and 321-365), despite the fact that these fusion products do not require the dimerization function of the receptor since this function is already contained by GAL-DBD (28). The fact that the constructs RAR E1 and GAL-RAR 226-270 lacking _i as well as GAL-RAR 232-448 missing the N-terminal part of the ligand binding domain, were all inactive, confirms that besides the dimerization domain other regions of the ligand-binding domain are also required for its activation function. It is tempting to speculate that upon ligand binding a conformational change takes place that allows the receptor directly or through protein-protein interactions to interact with the basal transcription machinery. Deletions in this region either prevent binding of the ligand or change the protein structure, which destroys this function and thereby cause the protein to be transcriptionally inactive. In both v-ERBA and RAR the ability to repress transcription appeared also to require parts of the ligand-binding domain (31) explaining the loss of activity of TAF 1 when fused to the whole ligand binding domain or parts of it. To obtain a better idea of how the ligand binding domain functions in both activation and repression it will be necessary to make more subtle deletions or point mutations in this region.

We have shown that in different cell lines the contribution of the two TAFs to receptor activity varies depending on the cell line tested. In P19-EC and 293 cells TAF-2 has the major contribution to total activity, while in 3T3 and COS cells the activity by both TAFs is about equal. Studies with GAL-ER construct showed similar differences: in chicken embryo fibroblasts TAF-1 was more active than TAF-2 whereas in Hela cells the reverse has been found (24). Also with GR-, ER- and PR- containing expression constructs transfected into Hela or CV1 cells differences were observed between both constructs containing TAF-1, TAF-2 or both, depending on the cell line used (25). The ability of different TAFs to squelch, synergize or hetero-synergize with one another revealed that TAFs of different receptors can activate through different mechanisms (20,47). The results obtained with GAL-RAR constructs containing TAF-1 or TAF-2 in different cell lines indicate that both TAFs function by different mechanisms. This could be explained by the presence/absence and/or the amount of intermediary proteins needed for transactivation or repression in these cell lines. In all cell-lines tested we observed the presence of the two TAFs but we did not find a no correlation between the activity of GAL-RAR in the absence of RA and the GAL-RAR-CF (TAF 1) activity in the cell-lines tested, we think that the activity observed in the absence of added ligand is caused by residual RA in the serum. Transfection of GAL-RAR in P19-EC or 293 cells grown in the presence of charcoal-depleted serum leads to a 3 to 4 fold decrease in transactivation in the absence of RA, without altering the RA-induced activation (data not shown).

The presence of two activating functions in this receptor which activate transcription in a cell-specific manner permits it to regulate transcription in a tissue-specific and developmentally programmed manner. Since it is known that expression of most RARs is spatially and temporally regulated (48-51) and that RARs have isoforms differing in their A-regions (52,53,54) it is likely that these isoforms have a specific regulatory function during development and in homeostasis. One mechanism through which RA may exert its complex regulatory functions could be that different isoforms contain cell-specific activators. From co-transfection studies with RAR1 and RAR2 it became clear that RAR1 repressed, while RAR2 activated transcription of a RARE-containing promoter; on a palindromic TRE however both receptors activated transcription equally well (55). Also with RAR differences have been observed between full length receptor and a receptor lacking the A-region (hRARA) in a both promoter- and cell-specific fashion (56).

During the preparation of this manuscript Nagpal et al (57) reported the presence of an activation function in all RARs and RXRs tested in the ligand binding domain. These authors also observed differences caused by deletion of the N-terminus of the various receptors depending on the reporter tested. Further functional analysis of both RAR-TAFs, understanding the specific role of TAF-1 in different RAR-isoforms and the molecular dissection of TAF-2 will require a combination of structural, mutational and functional in vitro and in vivo studies.
 

Acknowledgements

We thank P. Chambon (Strasbourg) for providing us with hRAR-2 cDNA. We are grateful to S.W. de Laat for continuous support and interest. We thank C.L. Mummery, F.A.E. Kruyt and A. van Puijenbroek for critically reading the manuscript. Finally we thank F.J.M. Vervoordeldonk and J. Heinen for photography.

G.E.F. and B.M. vd L. are supported by the Netherlands Cancer Society.

back

References

1. Roberts AB, Sporn MB 1984. In Sporn MB, Roberts AB, Goodman DS, eds. The Retinoids vol 2:209-286

2. Eichele G 1989 Retinoids in vertebrate and limb pattern formation. TIG 5:246-251

3. Petkovich M, Brand NJ, Krust A, Chambon P 1987 A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 330: 444-450

4. Giguere V, Ong ES, Segui P, Evans RM 1987 Identification of a receptor for the morphogen retinoic acid. Nature 330: 624-629

5. Brand N, Petkovich M, Krust A, Chambon P, Thé H de, Marchio A, Tiollais P, Dejean A 1988 Identification of a second human retinoic acid receptor. Nature 332: 850-853

6. Krust A, Kastner , Petkovich M, Zelent A, Chambon P 1989 A third human retinoic acid receptor, hRAR-. Proc Natl Acad Sci USA 86: 5310-5314

7. Thé H de, Mar Vivanco-Ruiz M del, Tiollais P, Stunnenberg H, Dejean A 1990 Identification of a retinoic acid responsive element in the retinoic acid receptor gene. Nature 343: 177-180

8. Sucov HM, Murakami KK, Evans RM 1990 Characterization of an autoregulated response element in the mouse retinoic acid receptor type gene. Proc Natl Acad Sci USA 87: 5392-5396

9. Mangelsdorf DJ, Ong ES, Dyck JA, Evans RM 1990 Nuclear receptor that identifies a novel retinoic acid pathway. Nature 345:224-229

10. Heyman RA, Mangelsdorf DJ, Dyck JA, Stein RB, Eichele G, Evans RM, Thaller C 1992 9-Cis retinoic acid is a high affinity ligand for the retinoic acid X receptor. Cell 68: 397-406

11. Levin AA, Sturzenbecker LJ, Kazmer S, Bosakowski T, Huselton C, Allenby G, Speck J, Kratzeisen C, Rosenberger M, Lovey A, Grippo JF 1992 9-Cis retinoic acid sterioisomer binds and activates the nuclear receptor RXR. Nature 355: 359-361

12. Leid M, Kastner P, Lyons R, Nakshatri H, Saunders M, Zacharewski T, Chen J-Y, Staub A, Garnier J-M, Mader S, Chambon P 1992 Purification, cloning, and RXR identity of the HeLa cell factor with which RAR or TR heterodimerizes to bind target sequences efficiently. Cell 68: 377-395

13. Yu VC, Delsert C, Andrsen B, Holloway JM, Devary OV, Näär AM, Yun Kim S, Boutin J-M, Glass CK, Rosenfeld MG 1991 RXR: a coregulator that enhances binding of retinoic acid, thyroid hormone, and vitamin D receptors to their cognate response elements. Cell 67: 1251-1266

14. Zhang X-K, Hoffmann B, Tran P B-V, Graupner G, Pfahl M 1992 Retinoid X receptor is an auxiliary protein for thyroid hormone and retinoic acid receptors. Nature 355: 441-446

15 Kliewer SA, Umesono K, Mangelsdorf DJ, Evans RM 1992 Retinoic X receptor interacts with nuclear receptors in retinoic acid, thyroid hormone and vitamin D₃ signalling. Nature 355: 446-449

16. Green S, Chambon P 1988 Nuclear receptors enhance our understanding of transcription regulation. TIG 4: 11, 309-314

17. Evans RM 1988 The steroid and thyroid hormone receptor superfamily. Nature:889-895

18. Forman BM, Samuels HH 1990 Interactions among a subfamily of nuclear hormone receptors: the regulatory zipper model. Mol Endocrinol 4: 1293-1301

19. Gronemeyer H 1991 Transcription activation by estrogen and progesterone receptors. Annu Rev Genet 25: 89-123

20. Tora L, White J, Brou C, Tasset D, Webster N, Scheer E, Chambon P 1989 The human estrogen receptor has two independent nonacidic transcriptional activation functions. Cell 59: 477-487

21. Webster NJG, Green S, Rui Jin J, Chambon P 1988 The hormone-binding domains of the estrogen and glucocorticoid receptors contain an inducible transcription activation function. Cell 54: 199-207

22. Webster NJG, Green S, Tasset D, Ponglikitmongkol M, Chambon P 1989 The transcriptional activation function located in the hormone-binding domain of the human oestrogen receptor is not encoded in a single exon. EMBO J 8: 1441-1446

23. Hollenberg SM, Evans RM 1988 Multiple and cooperative trans-activation domains of the human glycocorticoid receptor. Cell 55: 899-906

24. Berry M, Metzger D, Chambon P 1990 Role of the two activating domains of the oestrogen receptor in the cell-type and promoter context dependent agonistic activity of the anti-oestrogen 4-hydroxytamoxifen. EMBO J 9: 2811-2818

25. Bocquel MT, Kumar V, Stricker C, Chambon P 1989 The contribution of the N- and C-terminal regions of steroid receptors to activation of transcription is both receptor and cell-specific. Nucleic Acids Res 17: 7, 2581-2595

26. Banerji J, Rusconi S, Schaffner W 1981 Expression of a -globin gene is enhanced by remote SV40 DNA sequences. Cell 27:299-308

27. Gorman CM, Moffat LF, Howard BH 1982 Recombinant genomes which express chloramphenicol acetyl transferase in mammalian cells. Mol Cell Biol 2:1044-1051

28. Carey M, Lin Y-S, Green MR, Ptashne M 1990 A mechanism for synergistic activation of a mammalian gene by GAL4 derivatives. Nature 345: 361-364

29. Silver PA, Chaing A, Sadler I 1988 Mutations that alter both localization and production of a yeast nuclear protein. Genes Dev 2: 707-717

30. Sadowski I, Ma J, Triezenberg S, Ptashne M 1988 GAL4-VP16 is an unusually potent transcriptional activator. Nature 335: 563-564

31. Baniahmad A, Köhne AC, Renkawitz R 1992 A transferable silencing domain is present in the thyroid hormone receptor, in the v-erbA oncogene product and in the retinoic acid receptor. EMBO J 11: 1015-1023

32. Campione-Piccardo J, Sun J-J, Craig J, McBurney MW, 1985 Cell-cell interactions can influence drug induced differentiation of murine embryonal carcinoma cells. Dev Biol 109:25-31

33. Jones-Villeneuve EMV, Rudnicki MA, Harris JF, McBurney MW 1983 Retinoic acid induced neural differentiation of embryonal carcinoma cells. Mol Cell Biol 3:2271-2279

34. Pratt MAC, Kralova J, McBurney MW 1990 A dominant negative mutation of the alpha retinoic acid receptor gene in a retinoic acid-nonresponsive embryonal carcinoma cell. Mol Cell Biol 10: 6445-6453

35. Kruyt FAE, Veer LJ van der, Mader S, Brink CE, Feijen A, Jonk LJC, Kruijer W, Saag PT van der 1992 Retinoic acid resistance of the variant embryonal carcinoma cell line RAC65 is caused by expression of a truncated RAR. Differentiation 49: 27-37

36. Mitchell PJ Tijan R 1989 Transcriptional regulation in mammalian cells by sequence specific DNA binding proteins. Science 245:371-378

37. Sullivan WP, Madden BJ, McCormick AJ, Toft DO 1988 Hormone-dependent phosphorylation of the avian progesterone receptor. J Biol Chem 263: 14717-14723

38. Hoeck W, Groner B 1990 Hormone-dependent phosphorylation of the glucocorticoid receptor occurs mainly in the amino-terminal transactivation domain. J Biol Chem 265: 5403-5408

39. Glineur C, Bailly M, Ghysdael J 1989 The c-erb -encoded thyroid hormone receptor is phosphorylated in its amino terminal domain by casein kinase II. Oncogene 4: 1247-1254

40. Rochette-Egly C, Lutz Y, Saunders M, Scheuer I, Gaub M-P, Chambon P 1991 Retinoic acid receptor : specific immunodetection and phosphorylation. J Cell Biol 115: 535-545

41. Hsieh J-C, Jurutka PW , Calligan MA, Terpening CM, Haussler CA, Samuels DS, Shimizu Y, Shimizy N 1991 Human vitamin D receptor is selectively phosphorylated by protein kinase C on serine 51, a residue crucial to its trans-activation function. Proc Natl Acad Sci USA 88: 9315-9319

42. Jenster G, Korput AGM van der, Vroonhoven C van, Kwast TH van der, Trapman J, Brinkmann AO 1991 Domains of the human androgen receptor involved in steroid binding, transcriptional activation, and subcellular localization. Mol Endocrinol 5: 1396-1404

43. Danielian PS, White R, Lees JA, Parker MG 1992 Identification of a conserved region required for hormone dependent transcriptional activation by steroid hormone receptors. EMBO J 11: 1025-1033

44. O'Donnell AL, Koenig RJ 1990 Mutational analysis identifies a new functional domain of the thyroid hormone receptor. Molecular Endocrinology 4: 715-720

45. O'Donnell AL, Rosen ED, Darling DS, Koenig R J 1991 Thyroid hormone receptor mutations that interfere with transcriptional activation also interfere with receptor interaction with a nuclear protein. Molecular Endocrinology 5: 94-99

46. Glass KC, Lipkin SM, Devary OV, Rosenfeld MG, 1989 Positive and negative regulation of gene transcription by retinoic acid -thyroid hormone receptor heterodimer. Cell 59: 697-708

47. Tasset D, Tora L, Fromental C, Scheer E, Chambon P 1990 Distinct classes of transcriptional activating domains function by different mechanisms. Cell 62: 1177-1187

48. Dollé P, Ruberte E, Kastner P, Petkovich M, Stoner C, Gudas L, Chambon P 1989 Differential expression of genes encoding , and retinoic acid receptors and CRABP in the developing limbs of the mouse. Nature 342: 702-705

49. Dollé P, Ruberte E, Leroy P, Morriss-Kay G, Chambon P 1990 Retinoic acid receptors and cellular retinoid binding proteins. A systematic study of their differential pattern of transcription during mouse organogenesis. Development 110: 1133-1151

50. Ruberte E, Dollé P, Krust A, Zelent A, Morriss-Kay G, Chambon P 1990 Specific spatial and temporal distribution of retinoic acid receptor gamma transcripts during mouse embryogenesis. Development 108: 213-222

51. Ruberte E, Dollé P, Chambon P, Morriss-Kay G 1991 Retinoic acid receptors and cellular retinoid binding proteins II. Their differential pattern of transcription during early morphogenesis in mouse embryos. Development 111: 45-60

52. Kastner P, Krust A, Mendelsohn C, Garnier J-M, Zelent A, Leroy P, Staub A, Chambon P 1990 Murine isoforms of retinoic acid receptor with specific patterns of expression. Proc Natl Acad Sci USA 87: 2700-2704

53. Leroy P, Krust A, Zelent A, Mendelsohn C, Garnier J-M, Kastner P, Dierich A, Chambon P 1991 Multiple isoforms of the mouse retinoic acid receptor are generated by alternative splicing and differential induction by retinoic acid. EMBO J 10: 59-69

54. Zelent A, Mendelsohn C, Kastner P, Krust A, Garnier J-M, Ruffenach F, Leroy P, Chambon P 1991 Differentially expressed isoforms of the mouse retinoic acid receptor are generated by usage of two promoters and alternative splicing. EMBO J 10: 71-81

55. Husmann M, Lehmann J, Hoffmann B, Hermann T, Tzukerman M, Pfahl M 1991 Antagonism between retinoic acid receptors. Mol Cell Biol 11:4097-4103

56. Kastner P, Perez A, Lutz Y, Rochette-Egly C, Gaub M-P, Durand B, Lanotte M, Berger R, Chambon P 1992 Structure, localization and transcriptional properties of two classes of retinoic acid receptor fusion proteins in acute promyelocytic leukemia (APL): structural similarities with a new family of oncoproteins. EMBO J 11: 629-642

57. Nagpal S, Saunders M, Kastner P, Durand B, Nakshatri H,Chambon P 1992 Promoter context- and response element-dependent specificity of the transcriptional activation andmodulation function of retinoic acid. Cell 70: 1007-1019
 
index
homepage