A role for cofactors in synergistic and cell-specific activation by retinoic acid receptors and retinoid X receptor.
 
Gert E. Folkers, Bart van der Burg and Paul T. van der Saag
 
Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Utrecht, the Netherlands
 
The Journal of Steroid Biochemistry and Molecular biology 56:119-129 (1996)
 

Abstract

Transcriptional activation is thought to be mediated by DNA-bound activators by interaction with a basal transcription factor thereby stabilizing the pre-initiation complex. For such interaction cofactors such as TAFs, bridging proteins, mediators or intermediary proteins, are required by binding simultaneously to the activator and the target. We have investigated the activation functions (AFs) of both RAR and RXR and show that both activators contain two homologous AFs. By comparing the capacity to activate transcription by these AFs on several promoters, both as full-length receptors and as fusion-proteins of AFs with the DNA-binding domain of the yeast transcription factor GAL-4, we were able to show that these AFs function by different mechanisms. We found that the activity of these AFs is cell-type specific, as they are more active in certain cell lines than in others. Furthermore we observed that the AFs of RAR and RXR can activate transcription synergistically both as GAL-fusion protein and with full-length receptors. For AF-2 of RAR we observed cell-type dependent differences in synergistic activation and we show that the E1A protein, which functions as a cofactor for RAR, permits synergistic activation in cell lines in which in the absence of this cofactor transcription occurs non-synergistically. We propose a model in which several non cell type specific cofactors and cell-specific cofactors act together to form a more stable pre-initiation complex explaining the observed cell-specific synergistic activation.

Back

Introduction

Retinoic acid (RA) is an important regulator of cell growth, differentiation and homeostasis. Furthermore an important role has been suggested for this molecule during vertebrate development [1]. Identification of receptors for retinoic acid (RARs and RXRs) made it possible to study the molecular mechanism underlying the role of RA in these processes [1]. RAR and RXR both belong to the steroid/thyroid hormone receptor superfamily which all share a similar structure [2]. The C-terminal part of the protein contains the ligand binding domain (E-region), which also contains a dimerization domain and a transactivation domain (AF-2, formerly called TAF-2), the N-terminal part of the protein (AB-region) contains a ligand independent transactivation function (AF-1) [3]. The DNA-binding domain (C-region) is located in between and is the most conserved region among the various members of this family [2]. These receptors exert their effect on promoters containing hormone response elements, recognized by the DNA-binding domain. Retinoic acid response elements (RAREs) were found in many promoters [4,5] and consist of direct repeats with the sequence a/gGg/tTCA generally separated by five base-pairs, although RA can also activate promoters with direct repeats with one or two base-pairs spacing. It has been shown that only a heterodimer of RAR and RXR can bind to such element and both receptors are required for RA-induced activation of a RARE [3].

Recently we [6] and others [7,8] have shown that RARs and RXRs contain two regions involved in transactivation (AF-1 and AF-2) similarly as reported for the estrogen receptor [9] and the glucocorticoid receptor [10]. These experiments further indicated that these regions of RAR and RXR function both in a promoter- and cell-specific manner [6,7]. Similar findings have also been reported for the estrogen receptor [9,11,12]. Transfection experiments with different activators from several members of this family have shown that some activators can interfere with the activity of other AFs (squelching) whereas other activators enhance transcription together in a more than additive fashion (hetero-synergism). Finally synergistic activation by AFs can also occur when multiple binding sites are present in a target promoter (homo-synergism) [9-14]. Transcriptional activation by DNA-bound activators is believed to be caused by an acceleration of the rate limiting step in the formation of the pre-initiation complex, possibly by a stabilization of the complex through protein-protein interactions of activators with basal transcription factors [15,16]. In vitro transcription experiments have shown that for activated transcription by AFs coactivators/cofactors are required (for review see [17]). Several classes of proteins have been shown to fullfill such functions. The tata-binding protein (TBP)-associated factors (TAFs) have been shown to functionally interact with several activators, thereby acting as a coactivator [18]. Recent data suggest that various activators require different subsets of TAFs for transcriptional enhancement [19]. Furthermore a class of cofactors like USA are required for activated transcription but are not associated with TBP [20]. A third group of proteins, identified in squelching experiments which is acting as cofactors in activated transcription is designated as adaptors, mediators, or transcription intermediary proteins (TIFs), functioning by interacting simultaneously with activator and basal transcription factor [13,14,21,22].

Whether the differences in mechanism of activation of various activators leading to cell- and promoter-specific activation, as well as the capacity to squelch or synergize the activity of an activation function, is caused by the presence and/or levels of expression of different cofactors/coactivators are questions which remain to be answered.

In this paper we have analyzed the mechanism underlying activation of the individual AFs of RAR and RXR in RA-induced activation. The requirement of both RAR and RXR for high affinity binding to a RARE (3), the presence of multiple activation functions within the RARE bound complex [6,7,8] and the presence of several ligands which can also function in a receptor-dependent way (23,24) makes the study of the role of specific AFs in RA-induced activation difficult. To overcome these difficulties we fused RAR and RXR sequences containing AFs, to the DNA-binding domain of the yeast transcription factor GAL-4. Transfection of these fusion proteins established that both receptors contain two autonomous activation domains: AF-1 is functioning independently of ligand when present as GAL-fusion protein, while AF-2 can activated transcription only in the presence of ligand. Comparing the activity of these activation domains in various cell lines and on promoters containing one or five GAL-binding sites suggested that each activation domain functions by different mechanisms. Furthermore the activity of a specific activator is cell type-dependent, as it is more active in certain cell lines than in others. We present evidence that the presence of cell-specific cofactors contribute to this cell specificity.

back

Material and Methods

Plasmids

GAL-RAR -AC (AF-2) and GAL-RAR -CF (AF-1) have been described before [6], GAL-RXR -AB (AF-2) was made by cloning a Stu1-Sma1 fragment of pSG5 mRXR into the Sma1 site of pSG424. GAL-RXR -CF (AF-1) was generated by cloning the EcoR1-Stu1 fragment after filling in with Klenow into the Sma1 site of pSG424, resulting in a fusion protein with 7 additional amino acids at the N-terminus. The reporter E1b 5xGAL-CAT have been described before [25]. By cloning of the HindIII-XbaI fragment from this construct, containing the 5 GAL sites, into pBLCAT2 we made the construct 5xGAL-tk-CAT. 1xGAL-E1b-CAT and 1xGAL-tk-CAT were constructed by cloning an oligo containing a dubble strand GAL-binding site agcttCGGAGGACAGTCCTCCGc into the HindIII site of E1b-CAT and pBLCAT2 respectively. The 5*GAL-e1b-Luc reporter has been described before [34] and the 1 and 3*RARE-tk-CAT has been made by cloning one or 3 copies (head to tail) of an oligo containing the RARE as found in the RAR promoter in the BamHI site of pBLCAT2.
 

Transfection and CAT assay.

Cells were cultured and transfected as described before [6]. In short, the cells were seeded in Dulbecco's Modified Eagle Medium/HAM's F12 1:1 (DF), supplemented with 7.5% (v/v) fetal calf serum, transfected by the calcium phosphate coprecipitation method using 5-10µg reporter (CAT or LUC), 2 µg SV2-Lac-Z and 0.05-5.0 µg expression plasmids. After 14-18 h the precipitate was removed and fresh medium containing 1 µM RA was added. After an additional 24 h cell extracts for CAT-assay were prepared by freeze-thaw lysis in 0.25 M Tris pH 7.5 and 25 mM EDTA. Chlorampenicol acetyl transferae assay and quantification was performed as described previously [6], and normalized for transfection efficiency by measuring -Galactosidase activity.

Luciferase assay. Luciferase assays were performed as described by Brasier et al [26] with minor modifications. Extracts for this assay were made by lysing the cells on plates in 400µl Triton-lysis buffer (1% Triton X-100, 25 mM glycylglycin (pH 7.8), 15 mM MgSO4, 4 mM EGTA and 1 mM DTT). Lysates were transferred to Eppendorf tubes and pelleted by centrifugation (10 min, 12000 rpm at 4^o C) and directly used for assaying luciferase activity. 75µl cell extract was added to 265µl reaction mixture consisting of 111µl H₂O, 10µl 100 mM ATP, 90µl 100 mM glycylglycin (pH 7.8) and 54µl 100 mM MgSO₄. Reactions were started upon addition of 100µl 0.2 mM luciferin, light emmission was integrated over 10 sec in a Lumac/3M Biocounter, and normalized for transfection efficiency by measuring -galactosidase activity.

For the -galactosidase assay 5µl CAT extract or 50µl Luc extract, 145 or 100µl 100 mM phosphate buffer (pH 7.8) and 90µl Lac Z buffer (30 mM NAPO₄ (pH7.0), 3 mM KCl, 0.3 mM MgSO₄, 15 mM -mercaptoethanol and 20µg ortho-nitrophenyl--galactopyranoside were added together in a microtiter plate well. After 5-60 min -galactosidase activity was calculated by measuring absorbance at 417 nM in an Elisa plate reader.
 

Gel retardation assay.

Extracts and probes were made and the binding assay on the GAL-binding site was performed as described before [6]. Supershifts with antibodies against GAL-DBD, RAR or RAR, kindly provided by P. Chambon, were performed by addition of an antibody 10 min after addition of the labelled GAL-binding site, incubation was continued for 20 min and then loaded on a 4% (w/v)polyacrylamide gel, containing 0.25 x TBE running buffer.

back

Results
 

Experimental design

To study the mechanism underlying transcriptional activation by the individual AFs of RAR and RXR we fused sequences of both RAR and RXR that contain the transactivation domains of these receptors to the DNA-binding domain of the yeast transcription factor GAL-4. We have shown previously that RAR contains two domains involved in RA-induced transactivation [6] and similar constructs for RXR as shown in Figure 1a were made. These constructs were transfected in COS cells and extracts from these cells were used in a gel mobility shift assay on a GAL-binding site. Figure 1B shows that all constructs were expressed to similar levels (the expression of RXR AF-2 is two fold higher) and migrated according to their molecular weight as judged by Western blot analysis with

RXR (as RAR) also contains two autonomous activation functions

We first transfected the GAL-RXR constructs in COS cells with a GAL-responsive promoter containing five binding sites in front of an E1b tata-box coupled to the CAT gene (5xGAL-e1b-CAT). GAL-RXR AF-2 activated transcription only upon addition of 10^-6M RA, whereas the construct containing the AB region of RXR (AF-1) activated this promoter also in the absence of RA, while upon addition of RA only a slight increase was observed (Figure 2). As it has been shown previously that RXR only binds 9-cis RA [23,24] we transfected RXR AF-2 with both all-trans RA and 9-cis RA and observed that 9-cis RA reaches half maximal activity at 3x10^-8M and all-trans RA requires 10^-6M for half maximal response ([27], data not shown). Although we have used 10^-6M all-trans RA in our transfections, which is not enough for maximal response this is apparently sufficient to activate AF-2 of RXR, probably by isomerization of all-trans RA to 9-cis RA. By comparing the activity of transactivation functions of both RAR and RXR in various cells and on different promoters we wanted to analyze the mechanism underlying the RA-induced, cell- and promoter-specific activity of the receptors.

Synergistic activation by GAL-fusion protein

The activity of GAL AF-1 fusion proteins on a GAL responsive promoter is in marked contrast with the results reported by others, where no activity of GAL AF-1 fusion proteins has been observed [7]. In our experiments we used a reporter containing five GAL binding sites in front of a simple promoter (E1b-tata), whereas Nagpal et al. used more complex promoters containing only one or two GAL binding sites, a difference which could possibly explain the differences. To investigate the mechanism underlying this discrepancy we transfected the various GAL fusion constructs in COS cells together with GAL reporters containing one or five GAL-binding sites in front of either the E1b-tata box or the HSV thymidine kinase (tk) promoter. Figure 3 gives the result of a representative experiment showing as expected that AF-2 of both receptors activated the one GAL site containing promoters in a ligand- dependent manner. The autonomous AF-1s showed only weak, if any activity on both promoters containing one GAL site, whereas the same constructs activated transcription significantly on a promoter containing five binding sites (Figure 2, Table 1, data not shown). The most obvious explanation for this observation would be that the autonomous AF-1 of RAR and RXR activate transcription in a cooperative fashion on a promoter with multiple binding sites, as has been observed for the viral activation domain of VP16 [28]. In order to test this hypothesis we compared the inductions by these activators on the simple promoter (e1b-tata) or a more complex promoter (tk) containing either one or five GAL binding sites in COS cells. Table 1 shows the fold induction of the GAL-fusion constructs relatively to the empty expression vector GAL-DBD in the presence of RA. As expected GAL-VP16 activated transcription cooperatively: the activity of the five GAL sites-containing promoter is higher than five times the activity of a single binding site containing promoter. The ligand-dependent AFs, which already have activity on a single GAL site containing promoter showed only a limited increase when tested on a promoter containing five binding sites in COS cells. AF-1 of RAR however, showed cooperative enhancement on a multiple binding site containing promoter in comparison with the single binding site containing promoter. The corresponding GAL-RXR fusion showed only a five-fold increase (additive).

The same experiment was also performed on a more complex promoter, the thymidine kinase promoter. On this promoter we did not see any cooperativity of any of the GAL fusions; only the GAL-RAR AF-1 construct showed a significant transcriptional enhancement upon multimerization of GAL-binding sites, while also the GAL-VP16 fusion protein showed no synergistic activation on this promoter. A possible explanation for this observation could be that the binding of other transcription factors on the tk-promoter such as SP1, required for promoter activity, are preventing these activators to interact with the basal transcription factors or TAFs while on the E1b-tata promoter this is not the case. To further investigate the homosynergistic properties of these activators we transfected increasing amounts of the GAL-fusion constructs on a Luciferase reporter with five GAL binding sites in front of an e1b tata box in P19 EC cells. Figure 4 shows that AF-1 of RAR and RXR activated transcription similarly, the activity increased when more activator was transfected. At the highest concentrations we observed an exponential enhancement in activity (RAR AF-1) indicative for the presence of synergism by these activators. The ligand-dependent activation functions showed a concentration-dependent activity until reaching its maximal activity. Transfection of more activator resulted in a decrease in activity for AF-2 of RXR, whereas the activity of GAL-RAR AF-2 remained constant. In COS cells we observed similar paterns of activation, but in this case only the ligand-dependent constructs activated transcription almost maximally at the lowest concentration tested (data not shown). The GAL constructs containing AF-1 activated similarly as in P19-EC, showing an exponential activation at high concentrations of activators. These results indicate that the equivalent transactivation domains of RAR and RXR activate transcription by a similar mechanisms. In short these data show that in COS cells AF-2 cannot but that AF-1 can synergistically enhance transcription on a promoter containing multiple binding sites, while this enhancement is depending on the amount of activator present and the type of promoter used.
 

Synergistic activation of the RAR/RXR heterodimer on a multiple RARE-containing promoter.

Having shown that GAL AF-1 constructs can activate transcription synergistically on multimerized binding sites we next wanted to know whether the RAR/RXR heterodimer can also activate transcription synergistically. Transfection of RAR and RXR expression- constructs in COS cells with a reporter containing one or three RAREs in front of the Herpes simplex virus tk promoter revealed that the heterodimer could synergize since a more than three-fold enhancement was observed (up to ten-fold) when the activity, in the presence of RA, of a 3x RARE-containing tk-promoter was compared with a 1x RARE-containing tk promoter (Figure 5). The activity of the reporter containing one or three copies of the RARE transfected together with an AF-1-lacking RAR and full length RXR in COS cells was slightly less active than the full length receptor. Because of the high activity in the absence of cotransfected receptor and the moderate increase upon cotransfection of RARs we were unable to investigate the role of AF-1 in this synergism. In P19 EC cells we also observed synergistic activation, even in the absence of cotransfected receptors (data not shown).

No hetero-synergism with the homologous AFs fused to GAL-DBD.

Previously it has been shown that RAR and RXR when present together can activate transcription highly cooperatively on a RARE containing promoter in the presence of RA likely as the consequence of a much stronger binding to such element [see 3]. We showed above that the heterodimer could synergistically activate transcription upon binding-site multimerization (Figure 5). To test whether this synergism is caused by the formation of a more potent activator when the two corresponding AFs are present together on one promoter, we transfected the two corresponding GAL-RAR and GAL-RXR AFs in both COS cells and P19-EC cells. As shown in Figure 6 no synergistic activation was observed when the two homologous AFs were present together, but instead a slight decrease in activity was observed in both COS and P19-EC cells (data not shown). Although other reasons might also explain these results, we propose that the homologous AFs have a common target for transcriptional activation, thus preventing synergistic activation.
 

The activity of the AFs of RAR and RXR is cell type dependent.

We have shown that the homologous activation functions of RAR and RXR function by a similar mechanism showing similar activation characteristics on various promoters (Figure 4 and 6). However the contribution of the various AFs in RA-induced promoter activity on different promoters is not equal [7] and furthermore we have reported that the activity of the two GAL-RAR activation functions is cell-specific [6]. This suggests that the activators function by different mechanisms. If each activator functions through a different mechanism, we expect that activation by each AF would differ between various cell-lines. To test this hypothesis we transfected the RAR and RXR GAL-fusion proteins together with the five GAL sites containing-e1b reporter in five different cell-lines : COS cells, P19 EC cells, the human embryonic kidney cell line 293, NIH 3T3 cells and the human breast tumor cell line T47-D. The results of these experiments, presented as fold induction, in the presence of 10^-6 M RA, relative to the activity of the empty expression vector GAL-DBD are shown in Table 2. In all cases the ligand inducible-activator of RAR is the most potent one, although in some cell lines (COS, 293 and 3T3) GAL-RXR AF-2 is almost as active. Also the two ligand-independent AFs of RAR and RXR behave differently in this panel of cell lines: the GAL-RAR AF-1 construct is always more active than the corresponding RXR fusion and in 3T3 and T47-D cells hardly any activity was detected with the latter activator while RAR AF-1 was active in these cells. In P19 EC and 293 cells both activators function almost equally well. We cannot exclude that part of these differences are the consequence of different accumulation levels of the various activators in each cell although we observed in COS cells comparable accumulations of all GAL fusions (Figure 1b). These data together show that all four activators activated transcription differently in the cell lines tested and suggest that the activators function by different mechanisms.
 

Cell-specific synergistic activation is dependent on cofactors.

It has been shown that for activation of transcription several classes of cofactors are required and activators are thought to interact with these cofactors [17,18]. In vitro transcription experiments and in vitro binding studies indicate that different activators are binding different cofactors which in turn contact the basal transcription factors [29-31]. It has been proposed that several classes of cofactors are cell-specific [12,14] and thus differences in activation between various cell-lines might be the result of differential expression of cofactor. Furthermore, synergistic activation has been shown to rely on multiple interactions between more than one activator with more than one target [32,33]. If this model for synergistic activation and cell-specific cofactor requirement is correct synergistic activation may be cell-specific. To test this we compared the activity of the GAL-RAR and GAL-RXR fusion proteins on promoters containing one or five GAL-binding sites in front of E1b-CAT, and only found significant differences in synergistic activation by RAR AF-2 between various cell lines (Figure 7, data not shown). In 293, P19 EC and T47D synergism was observed, whereas in COS and 3T3 cells no or only minor enhancement was observed on a reporter containing 5 GAL sites (Figure 7).

Recently we have shown that the adenoviral E1A is acting as a cofactor for RAR by contacting simultaneously AF-2 and TBP [34]. We showed that as a consequence of this interaction an enhancement in transcription by AF-2 of RAR, but not by the other activators of RAR and RXR, was observed. Therefore we now wanted to investigate whether addition of this cofactor could enable synergistic activation in cell lines that were unable to activate transcription synergistically upon multimerization of binding sites. Cotransfection of E1A 13S (but not 12S) caused an 4-6 fold enhancement in activity only on the five GAL-binding sites-containing reporter but not on the single binding site-containing reporter (Figure 7). Consequently synergistic activation could be observed in all cell lines and the levels of synergy as well as the fold induction (relative to the activity in the absence of cotransfected activator) also were comparable between the various cell lines (Figure 7, data not shown). Together these data suggest that the activators of RAR and RXR function by different mechanisms, probably by relying on (a) specific as well as common cofactor(s). The levels of expression of such cofactors contribute to both the activity and the capacity to activate transcription synergistically and in a cell-specific fashion.

back

Discussion

All members of the steroid/thyroid hormone superfamily analyzed so far have a similar modular organization with two regions involved in activation [2]. The activation function AF-2, located in the hormone-binding domain was recently characterized [35-39] and reported to be conserved among all members of this family. We have recently characterized the activation function present at the N-terminal part of RAR, and presented evidence that this activator belongs to the class of acidic activators [40]. No obvious sequence homology among the family members are present although some similarities were observed, among which regularly spaced hydrophobic residues surrounding negatively charged residues [40]. In agreement with these homologies some functional similarities between the homologous AFs of RAR and RXR were observed. The activation characteristics of homologous AFs at different activator concentrations are comparable (Figure 4) and the inability to synergize with each other suggest that both activators have a common target for transactivation (Figure 6). This could mean that some features of these AFs have been conserved during evolution, although at the same time clear differences were observed between the corresponding AFs of RAR and RXR: (1) Different promoter specificity by the individual AFs both on natural promoters [7] and when analyzed as individual activators on simple (E1b-tata) or more complex promoters (TK) (Figure 2, Table 1). (2) Cell-specific differences were found (Table 2). (3) Differences in the capacity to activate transcription synergistically (Table 1, 3) on reporters containing multiple binding sites. (4) Differences in responsiveness to E1A, a protein which we have shown to function as a cofactor for AF-2 of RAR [34]. We hypothesize that each activator functions by a different mechanism having on or more common targets as well as (cell)-specific targets explaining the observed cell specificity and synergism. The presence of multiple isoforms in these receptors [41,42,43] differing at their N-termini (containing AF-1), and therefore possibly contributing differentially to the expression of target genes further strengthens the hypothesis that each activator functions by a different mechanism.

Both AF-1 and AF-2 of RAR and RXR could, depending on the cell line, activate transcription synergistically upon multimerization of binding sites, both when present as GAL-fusion with a GAL-responsive reporter (Figure 2, Table 2) and as heterodimer on a RARE-containing reporter (Figure 4).

It is thought that synergism is caused by interactions of the activator with more than one target of the pre-initiation complex [32,34,44]. The best characterized activator so far VP16 has been shown to activate transcription synergistically upon multimerization of binding sites both in vitro [45] and in vivo [28]. VP16 is capable of interacting with TBP, TFIIB and also TAF40 in vitro and these interactions were reported to be required for activation by this activator [29]. Together these data suggest that this activator can as a consequence of simultaneous or subsequent interactions with the basal transription factors stabilize the pre-initiation complex either directly or through different cofactors enabling this activator to function synergistically. Whether this is also the mechanism underlying the synergistic activation for the activators of RAR and RXR remains to be seen but our cotransfection data with E1A (a cofactor for RAR [34]), are supporting this model. Synergistic activation by AF-2 of RAR was observed in cell lines expressing E1A (293) or E1A-like activity (P19 EC), but not in cell lines lacking such activity (COS, 3T3). Cotransfection of E1A in the latter cell type resulted in a more than additive enhancement in transcription upon multimerization of binding sites (Figure 7). We did not observe hetero-synergism in both P19 EC and COS cells when the corresponding AFs of RAR and RXR were added together, possibly because the AFs of RAR and RXR have one or more common targets so they are competing for this limiting target leading to the observed decrease rather than to an increase. Furthermore the two AF-2s can form heterodimers which could impaire the binding to a GAL-binding site.

The ability to activate transcription by the activation domains of RAR and RXR was varying among the different cell lines tested both on single and multiple binding sites containing promoters (Table 2, Figure 7; data not shown). In general the ligand-dependent AF-2 is more active than AF-1 and the variations between the different cell lines are more pronounced for AF-1 than for AF-2. Similar results were also reported for ER [11,12]. The RAR/RXR heterodimers also activate transcription cell-type specifically on a RARE containing promoter (data not shown).

Although we hypothesize here that these differences are the consequence of differential cofactor/coactivator expression (see later), other mechanisms like cell-specific phosphorylation may also be important in explaining the observed differences in activation (for review see [46]). We have observed that RAR is phosphorylated at multiple sites including the regions required for activation (unpublished results) making it possible that phosphorylation has modulatory role in the activity of RAR AFs. Furthermore the presence of activator specific repressors might also contribute to the observed differences. Such repressors have recently been identified for the Hela-cell activator TEF-1 [47]. Finally, we cannot exclude that endogenously present RARs and RXRs contribute to the observed cell-specificity of AF-2 by forming heterodimers with the GAL-fusion proteins as reported before by Nagpal et al [8] although such complexes were not observed in gel shift assays (Figure 1c).

Data from the recent literature are providing us some possible explanations concerning the mechanism underlying specificity in transcriptional activation by different activators which might account for the observed cell specificity and synergistic activation. In vitro transcription experiments indicate that each type of activator needs only a subset of TAFs to activate transcription, while this interaction with one or more of these TAFs is a requirement for activated transcription [19]. Furtermore Brou et al have chromatographically separated TFIID fractions that contain different combinations of TAFs, each exhibiting different functional properties with respect to activation by different activators [48,49]. Cloning of hTAF_II30 which is associated with a subset of TFIID complexes revealed that this TAF is required for AF-2 activation of ER [50], indicating that the presence and levels of expression of such TAFs possibly contribute to cell-specific activation of nuclear receptors. Furthermore recently several cofactors other than TAFs have been identified, functioning by bringing activators into proximity of the basal transcription machinery through interactions with both the activator and a target in the basal transcription machinery (CBP [51,52], MBF [53] Trip1 [54], ada2, ada3 and gcn5 [55] sug1 [56]). We have recently reported that E1A (for review see [57,58]) acts as a cofactor for RAR through an interaction with AF-2 and TBP [34]. Here we found that the presence of such activity contributed to the cell-specific and synergistic activition of this activator. In this paper we observed that the activity of AF-2 varied between different cell-lines, being more active in 293 and P19 EC cells than in COS or 3T3 cells; the first two cell lines contain E1A or E1A-like activity respectively, whereas the latter two do not. Cotransfection of E1A in cells lacking E1A or E1A-like activity caused transcriptional activation [59,60], especially on a promoter containing multiple binding sites (Figure 7), enabling this activator to activate transcription synergistically. On the other hand in cells containing such cofactor, cotransfection of E1A only marginally enhanced the activity of AF-2; synergistic activation was already observed without cotransfection of E1A and no further activation was observed upon cotransfection of E1A 13S. However some upregulation in AF-2 activity was observed in P19 EC cells but not as strong as in 3T3 or COS cells.

On the basis of these results we propose a model for synergistic and cell specific activation whereby AF-2 can interact with the basal transcription machinery through a non-cell-specific cofactor, possibly hTAF_II30, the TAF involved in AF-2 activation of ER [50]. The presence of a cell-specific cofactor (E1A or E1A-like activity) adds on another opportunity for AF-2 to interact through this cofactor with TBP [30,31,61,62]. Together these interactions create a stronger binding of the activator to the basal transcription factors than when only one type of interaction would be present. When multiple binding sites are present it is more likely that both these interactions can take place simultaneously since on single sites possibly only one interaction can take place at a time, causing the formation of a less stable pre-initiation complex. Occurrence of such multiple interactions has been proposed as a criterium for synergistic activation [32,33,45].

The presence of multiple activators within the RAR/RXR heterodimer which may also work cell-type specifically adds another possibility for RA-regulated target gene expression. Besides the opportunity to regulate promoters by different activators it is also possible that promoters are only transcribed in certain cell lines where the activator regulating the activity of a given promoter is active. Therefore the presence of cell-specific activators adds another level of regulation for RA-induced promoter activation.

back

Acknowledgements

We are grateful to Dr P. Chambon for providing an RAR and RXR expression constructs and antisera against GAL-DBD, RAR and RAR. Furthermore we thank Patricia Swanink, Ferdinand Vervoordeldonk and Jaap Heinen for excellent technical assistance. G.E.F and B.v.B are supported by the Dutch Cancer Society.
 

References

1. Sporn M.B., Roberts A.B. and Goodman D.S.: The retinoids, biology, chemistry and medicine. Raven Press, New York (1994).

2. Evans R.M.: The steroid and thyroid hormone receptor superfamily. Science 240 (1988) 889-895.

3. Leid M., Kastner P. and Chambon P. Multiplicity generates diversity in retinoic acid signalling pathways. Trends Biochem. Sci. 17 (1992) 427-433.

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

5. Sucov H.M., Murakami K.K. and Evans R.M.: Characterization of an autoregulated response element in the mouse retinoic acid receptor type gene. Proc. Natl. Acad. Sci. USA 87 (1990) 5392-5396.

6. Folkers G.E., van der Leede B.-J. M. and van der Saag P.T.: The retinoic acid receptor-2 contains two separate cell-specific transactivation domains, at the N-terminus and in the ligand binding domain. Mol. Endocrin. 7 (1992) 616-627.

7. Nagpal S., Saunders M., Kastner P., Durand B., Nakshatri H. and Chambon P.: Promoter context- and response element-dependent specificity of the transcriptional activation modulating functions of retinoic acid receptors. Cell 70 (1992) 1007-1019.

8. Nagpal S., Friand S., Nakshatri H. and Chambon P.: RARs and RXRs: evidence for two autonomous transcativation functions (AF-1 and AF-2) and heterodimerization in vivo. EMBO J. 12 (1993) 2349-2360.

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

10. Hollenberg S.M. and Evans R.M.: Multiple and cooperative trans-activation domains of the human glycocorticoid receptor. Cell 55 (1988) 899-906.

11. Berry M., Metzger D. and Chambon P.: 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 (1990) 2811-2818.

12. Bocquel M.T. Kumar V., Stricker C. and Chambon, P.: 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 (1989) 2581-2595.

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

14. Meyer M.-E., Gronemeyer H., Turcotte B. Bocquel M.-T., Tasset D. and Chambon P.: Steroid hormone receptors compete for factors that mediate their enhancer function. Cell 57 (1989) 433-442.

15. Ptashne, M.: How eukaryotic transcriptional activators work. Nature 335 (1988) 683-689.

16. Choy B. and Green M.R.: Eukaryotic activators function during multiple steps of preinitiation complex assembly. Nature 366 (1993) 531-536.

17. Conaway R.C. and Conaway J.W.: General initiation factors for RNA polymerase II. Annu. Rev. Biochem. 62 (1993) 161-190.

18. Goodrich J.A. and Tjian R.: TBB-TAF complexes: selectivity factors for eukaryotic transcription. Curr. Opin. Cell. 6 (1994) 403-409.

19. Chen J.-L., Donatella Attardi L., Verrijzen C.P., Yokimori K. and Tjian R.: Assembly of recombinant TFIID reveals differential coactivator requirements for distinct transcriptional activators. Cell 79 (1994) 93-105.

20. Meisterernst M., Roy A.L., Lieu H.M. and Roeder R.G.: Activation of class II gene transcription by regulatory factors is potentiated by a novel activity. Cell 66 (1991) 981-993.

21. Martin K.J., Lillie J.W. and Green M.R.: Evidence for interaction of different eukaryotic transcriptional activators with distinct cellular targets. Nature 346 (1990) 147-152.

22. Ptashne M., and Gann A.F.F.: Activators and targets. Nature 346 (1990) 329-331.

23. Heyman R.A., Mangelsdorf D.J., Dyck J.A., Stein R.B., Eichele G., Evans R.M. and Thaller C.: 9-Cis retinoic acid is a high affinity ligand for the retinoic acid X receptor. Cell 68 (1992) 397-406.

24. Levin A.A., Sturzenbecker L.J., Kazmer S., Bosakowski T., Huselton C., Allenby G., Speck J., Kratzeisen C., Rosenberger M., Lovey A. and Grippo J.F.: 9-Cis retinoic acid stereoisomer binds and activates the nuclear receptor RXR. Nature 355 (1992) 359-361.

25. Lillie J.W. and Green M.R.: Transcription activation by the adenovirus E1a protein. Nature 338 (1989) 39-44.

26. Brasier, A.R., Tate J.E. and Habener, J.F: Optimized use of the firefly luciferase assay as a reporter in mammalian cell lines. Biotechniques 7 (1989) 1116-1122.

27. Pijnappel W.W.M., Hendriks H.F.J., Folkers G.E., van den Brink C.E., Dekker E.J., Edelenbosch C., van der Saag, P.T. and Durston A.J.: The retinoid ligand 4-oxo-retinoic acid is a highly active modulator of positional specification. Nature 366 (1993) 340-344.

28. Emami K., and Carey M.: A synergistic increase in potency of a multimerized VP16 transcriptional activation domain EMBO J. 11 (1992) 5005-5012.

29. Goodrich J.A., T. Hoey T., Thut C.J., Admon A. and R. Tjian.: Drosophila TA_II40 interacts with both a VP16 activation domain and the basal transcription factor TFIIB. Cell 75 (1993) 519-530.

30. Horikoshi, N., Maguire K., Kralli A, Maldonado E, Reinberg D. and Weinmann R.: Direct interaction between adenovirus E1A protein and the TATA box binding transcription factor IID. Proc. Natl. Acad. Sci. USA 88 (1991) 5124-5128.

31. Lee W.S., Kao C.C., Bryant G.O., Liu X. and Berk A.J.: Adenovirus E1A activation domain binds the basic repeat in the TATA box transcription factor. Cell 67 (1991) 365-376.

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

33. Lin Y.-S., Carey, M., Ptashne M. and Green M.R.: How different eukaryotic transcriptional activators can cooperate promiscuously. Nature 345 (1990) 359-361.

34. Folkers G.E. and van der Saag P.T.: Adenoviral E1A functions as cofactor for retinoic acid receptor through direct interaction with TBP and RAR. submitted for publication.

35. Danielian P.S., White R., Lees J.A. and Parker M.G.: Identification of a conserved region required for hormone dependent transcriptional activation by steroid hormone receptors. EMBO J. 11 (1992) 1025-1033.

36. Saatcioglou, F., Bartunek P., Deng T., Zenke M., and M. Karin.: A conserved C-terminal sequence that is deleted in v-ErbA (the thyroid hormone receptor). Mol. Cell. Biol. 13 (1993) 3675-3685.

37. Barettino D., del Mar Vivanco-Ruiz M. and Stunnenberg H.G.: Characterization of the ligand-dependent transactivation domain of the thyroid hormone receptor. EMBO J. 13 (1994) 3039-3049.

38. Durand B., Saunders M., Gaudon C., Roy B., Losson R. and Chambon P.: Activation function 2 (AF-2) of retinoic acid receptor and 9-cis retinoic acid receptor: presence of a conserved autonomous constitutive activating domain and influence of the nature of the response element on AF-2 activity. EMBO J. 13 (1994) 5370-5382.

39. Leng X., Blanco J., Tsai S.Y., Ozato K., O'Malley B.W. and Tsai M.-J.: Mouse retinoid X receptor contains a separable ligand binding and transactivation domain in its E region. Mol. Cell. Biol. 15 (1995) 255-263.

40. Folkers G.E., van Heerde E.C. and van der Saag P.T.: Activation function-1 (AF-1) of retinoic acid receptor -2 is an acidic activator resembling VP16. Submitted for publication.

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

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

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

44. Herschlag D. and Johnson F.B.: Synergism in transcriptional activation: a kinetic view. Genes and Dev. 7 (1993) 173-179.

45. Ohashi Y., Brickman, J.M., Furman E., Middleton B. and Carey M.: Modulating the potency of an activator in yeast in vitro transcription system. Mol. Cell. Biol. 14 (1994) 2731-2739.

46. Hunter, T. and Karin, M.: The regulation of transcription by phosphorylation. Cell 70 (1992) 375-387.

47. Chaudhary S., Brou C., Burton M.E., Tora L., Chambon P. and Davidson I.: A cell-specific factor represses stimulation of transcription in vitro by transcriptional enhancer factor-1. Mol.Cell. Biol. 14 (1994) 5290-5299.

48. Brou C., Chaudhary S., Davidson I., Lutz Y., Wu J., Egly J.-M., Tora. L. and Chambon P.: Distinct TFIID complexes mediate the effect of different transcriptional activators. EMBO J. 12 (1993) 489-499.

49. Brou C., Wu J., Ali S., Scheer E., Lang C., Davidson I., Chambon P. and Tora, L.: Different TBP-associated factors are required for mediating the stimulation of transcription in vitro by the acidic transactivator GAL-VP16 and the two nonacidic activation functions of the estrogen receptor. Nucleic Acid Res. 21 (1993) 5-12.

50. Jacq X., Brou C., Lutz Y., Davidson I., Chambon, P. and Tora L.: Human TAFII30 is present in a distinct TFIID complex and is required for transcriptional activation by the estrogen receptor. Cell 79 (1994) 107-117.

51. Arias J., Alberts A.S., Brindle P., Claret F.X., Smeal T., Karin M., Feramisco J. and Montminy M. Activation of cAMP and mitogen responsive genes relies on a common nuclear factor. Nature 370 (1994) 226-229.

52. Kwok R.P.S., Lundblad J.R., Chrivia J.C., Richards J.P., Bächinger, H.P., Brennan R.G., Roberts S.G.E., Green M.R. and Goodman, R.H.: Nuclear protein CBP is a coactivator for the transcription factor CREB. Nature 370 (1994) 223-226.

53. Li F.-Q., Ueda H. and Hirose S.: Mediators of activation of fushi tarazu gene transcription by BmFTZ-F1. Mol. Cell. Biol. 14 (1994) 3013-3021.

54. Lee J.W., Ryan F., Swaffield J.C., Johnston, S.A., Moore D.D.: Interaction of the thyroid-hormone receptor with a conserved mediator. Nature 374 (1995) 91-94.

55. Horiuchi J., Silverman N., Marcus G.A. and Guarente L.: Ada3, a putative transcriptional adaptor, consists of two separable domains and interacts with ada2 and gcn5 in a trimeric complex. Mol. Cell. Biol. 15 (1995) 1203-1209.

56. Swaffield J.C., Melcher K. and Johnston S.A.: A highly conserved ATPase as a mediator between acidic activation domains and the tata-binding protein. Nature 374 (1995) 88-91.

57. Flint, J., and Schenk T.: Adenovirus E1a protein paradigm viral activator. Annu. Rev. Genet. 23 (1989) 141-161.

58. Boulanger P.A. and Blair G.E.: Expression and interaction of human adenovirusn oncoproteins. Bioch. J. 275 (1991) 281-299.

59. Berkenstam A., del Mar Vivanco-Ruiz M, Barettino D., Horikoshi M. and Stunnenberg H.G.: Cooperativity in transactivation between retinoic acid receptors and TFIID requires an activity analogous to E1A. Cell 69 (1992) 401-412.

60. Kruyt F.A.E., Folkers G.E., Walhout A.J.M., van der Leede B.-J. M. and van der Saag P.T.: E1A functions as a coactivator of retinoic acid-dependent retinoic acid receptor-2 promoter activation. Mol. Endocrin. 7 (1993) 604-615.

61. Keaveney M., Berkenstam A., Feigenbutz M., Vriend G. and Stunnenberg H.G.: Residues in the TATA-binding protein required to mediate a transcriptional response to retinoic acid in EC cells. Nature 365 (1993) 562-566.

62. Boyer T.G. and Berk A.J.: Functional interaction of adenovirus E1A with holo-TFIID. Genes Dev. 7 (1993): 1810-1823.

index
homepage