Abstract Introduction Material and Methods Results Discussion References

Adenoviral E1A functions as cofactor for retinoic acid receptor through direct interaction with RAR
 

Gert E. Folkers and Paul T. van der Saag
 

Hubrecht Laboratory, Netherlands Institute for Developmental Biology,Utrecht, the Netherlands
 

Molecular and Cellular Biology 15:5868-5878 (1995)
 

Abstract

Transcription regulation by DNA-bound activators is thought to be mediated by a direct interaction between these proteins and TBP, TFIIB or TAFs, although occasionally cofactors or adaptors are required. For ligand-induced activation by the RAR/RXR heterodimer, the RAR2 promoter is dependent on the presence of E1A or E1A-like activity since this promoter is only activated by RA in cells expressing such proteins. The mechanism underlying this E1A requirement is largely unknown. We now show that direct interaction between RAR and E1A is a requirement for RA-induced RAR2 activation. The activity of the hormone-dependent activation function AF-2 of RAR is upregulated by E1A and an interaction between this region and E1A was observed, but not with AF-1 or AF-2 of RXR. This interaction is dependent on conserved region III (CRIII), the 13S-specific region of E1A. Deletion analysis within this region indicated that the complete 13S specific region is needed for activation. The putative zinc finger region is crucial, probably as a consequence of interaction with TBP. Furthermore, the region surrounding aa 178, partially overlapping with the TBP-binding region is involved in both binding to and activation by AF-2. We propose that E1A functions as cofactor by interacting with both TBP and RAR, thereby stabilizing the pre-initiation complex.

back

Introduction

Transcription of RNA polymerase II (Pol II) promoters requires the assembly of the preinitiation complex consisting of basal transcription factors. This process begins with the binding of TFIID, consisting of the TATA-binding protein (TBP) and TBP-associated factors (TAFs), to the TATA-box, followed by ordered binding of the other transcription factors (TFIIA,-B,-E,-F,-H) and RNA polymerase forming the initiation-complex (28). Transcription factors bound to promoter or enhancer sequences modulate the activity of Pol II promoters. These transcriptional activators are thought to function by directly or indirectly interacting with a component of the basal transcription machinery, thereby stabilizing the preinitiation complex (14,47). Several activators have been shown to interact with TBP (33,51) TFIIB (3,37) or TAFs (27,30) However, there is evidence that some activators require additional proteins designated as adaptors, cofactors or bridging proteins(28,64). The presence of such factors was first hypothesized from experiments where overexpression of a given activator can inhibit its own activity as well as the activity of activators belonging to the same class of activators, thus functioning by the same mechanism (29,55,56). This repression, referred to as squelching, is thought to be caused by the titration of a limiting component required for activation by promoter-bound activators (29).

The adenovirus E1A protein has been shown to activate a variety of both viral and cellular promoters (24), while this protein is unable to bind DNA (23). This has led to the suggestion that E1A could activate transcription through an interaction with various transcription factors and could thus behave as a cofactor-like protein. Comparison of E1A between various adenovirus strains revealed the presence of three conserved regions (CR) which fulfill specific functions (48,55). From the E1A gene two alternatively spliced mRNAs are transcribed, a 12S and a 13S form encoding proteins of 243 and 289 amino acids, respectively (6). The two proteins differ by the presence of CRIII in the longer form being the most important region for transcriptional activation (24) which contains a zinc-binding region (15). A direct interaction has been reported between E1A and TBP (34,43) or TFIID (7), which was found to be dependent on the presence of the 13S-specific region. Furthermore it was shown that this region is responsible for activation by promoter-bound transcription factors (50).

The activity of retinoic acid is mediated by two types of receptors, the RARs and RXRs. Both receptors belong to the steroid/thyroid hormone receptor superfamily and can activate transcription in the presence of their ligands, all-trans RA or 9-cis RA, the last being specific for RXR whereas RARs can be activated by both isomers. RAR and RXR, as other members of this family of transcription factors have a similar domain structure: an autonomous activation domain (AF-1) in the N-terminal part of the receptor (AB-region), a highly conserved DNA-binding domain (C-region), and a hormone-binding domain (E-region) containing a region involved in dimerization as well as an autonomous activation function (AF-2) (22,44).

Retinoic acid has profound effects on cell growth and differentiation, and furthermore an important role has been suggested for RA in embryonal development (20,21,31). For the study of the mechanism of retinoids in differentiation, embryonal carcinoma cell lines have been very helpful as they can differentiate into a variety of cell types upon administration of RA (54), causing extensive changes in gene expression (31). One of the first events observed upon RA-induced differentiation of EC-cells is the induction of RAR2. Cloning of the promoter (17) and subsequent analysis of other RA-responsive promoters revealed that an element consisting of two repeats of the sequence ^A/_GG^G/_TTCA separated by 2 or 5 basepairs is required for RA-induced activation of such promoters (44, and references therein). Binding to such element is only possible by RAR/RXR heterodimers of which RAR has been shown to bind the most 3` repeat, while RXR binds preferentially to the most 5` repeat (40,52,62).

We have observed that the RAR2 promoter is highly active in P19-EC cells in the presence of RA, but not in its differentiated derivative, END2 cells (39). In EC cells E1A-like activity has been identified by the ability of specific adenovirus mutants, unable to produce functional E1A protein, to grow in these cells, while in differentiated EC-cells this was not observed (35,42). Furthermore several promoters which require E1A, are active in EC cells but not in their differentiated counterparts (24). This E1A-like activity is also present in mouse oocytes and pre-implantation embryos but is lost during subsequent embryonal development (70). Introduction of E1A proteins in EC cells can result in differentiation (57,69), whereas RA-differentiated cell lines can be reverted to a more undifferentiated form (16,74), suggesting that regulation of E1A-like activity expression is important for EC cell differentiation. This led us to hypothesize that E1A or E1A like activity may be involved in RAR promoter activation. Earlier we have observed that many cell lines do not express the RAR2 isoform, both in the presence or absence of RA (72). Transfection experiments in some of these cell lines have revealed that the RAR2 promoter cannot be activated by RA, while upon cotransfection of E1A 13S RA-dependent activation could be observed (39). We identified two promoter regions to be involved in this E1A-dependent upregulation: a CRE-like element (-99/-92) and the region from -63 to +156. The first element has been demonstrated to bind members of the CREB/ATF family (38). Some members of this family have been shown to interact with E1A causing transcriptional activation, possibly by bringing the upstream activator in close proximity of the basal transcription machinery (12,49) . The second region contains the transcription start site, the TATA-box and the RARE, of which the latter has been shown to be the most important regulatory element for this promoter (17).

In this paper we focus on the mechanism underlying the E1A-dependent upregulation of the RARE-containing promoter element and investigated whether the activity of the RAR/RXR heterodimer is upregulated by E1A. We show here that the activity of AF-2 of RAR is upregulated by E1A 13S but not by the shorter 12S form. Transfection experiments using high amounts AF-2 containing activators indicate that cotransfection of E1A 13S could prevent squelching caused by this overexpression, suggesting that E1A functions as a bridging protein/cofactor. For this upregulation by E1A a direct interaction between the hormone binding domain of RAR and the C-terminal part of the 13S-specific region of E1A is required since all mutants that lost the ability to bind, were also impaired in E1A-dependent activation. We suggest that E1A forms a bridge between AF-2 of RAR and TBP through interactions of these proteins with E1A.

back

Material and Methods.

Plasmids

GST-RAR (aa 138-448) was produced by cloning the blunt SalI-XbaI fragment of pSG5 RAR in the SmaI site of pGEX2T. GST RXR (aa 1-466) was made by cloning the EcoRI fragment of pSG5 mRXR in pGEX2T. GST-E1A, GST-TBP and GST-TFIIB were kindly provided by M. Timmers. The expression plasmids E1A 12S and E1A 13S containing genomic sequences of serotype 5 E1A genes have been described before (38). The RSV driven E1A expression constructs E1A 12S, 13S, 5/3, NC, CX, 3/2 and C3/MX, were provided by N. Jones (67). The GST-E1A mutants were constructed by amplifying parts of the E1A gene from these plasmids by PCR using forward primer (P1) containing a BamHI site linked to the first codon of E1A in frame with pGEX 2T and a reverse primer (P2) ending after the stop codon, extended with an EcoRI site. The BamHI/EcoRI digested PCR products were subsequently cloned in the corresponding sites of pGEX2T. An EcoRV site at aa 139/140 and a ScaI site at 177/178 (without changing the codons), were introduced in E1A 13S by site-directed mutagenesis. Internal deletions were made using these introduced sites together with SmaI (150/151) or the end-filled DraII (164/165) sites. N1 was produced by digesting GST-E1A with BamH1 and EcoRV (139-140) and ligated back, after Klenow treatment. The SmaI-EcoR1 and ScaI-EcoRI fragments were cloned in SmaI-EcoRI digested pGEX 3X to generate N2 and N3 respectively. C1 and C4 were created by cloning the BamHI-XbaI (made blunt ended with Klenow) and BamHI-ScaI fragments, respectively, in BamHI-SmaI digested pGEX 2T. For C3 a PCR fragment was generated using P1 and P3 (primer ending at codon 185 extended with an EcoRI site) and after digestion with BamHI and EcoRI and cloned in the corresponding sites of pGEX 2T. Digestion of this clone with BamHI and DraIII, treatment with Klenow and ligation, resulted in 132-185. C2 was made by cloning the BamHI-DdeI (made blunt) digested fragment in SmaI-BamHI digested pGEX 2T. C157S, C174S, C179S, S185R and V187L were made by PCR using primers that alter the codons concerned and P1 or P2. These PCR products were digested with AccI-SmaI or AccI-XbaI and cloned in the corresponding sites of pGEX-E1A. The deletion constructs 178-179, 179-185 and 185-187 were made using the mutants described above in which we introduced the restriction sites: ScaI (177/178), NaeI (179/180), StuI (184/185) and AflII (186-188). The mutant 185-187 was made by cloning the EcoRI-AflII (blunt) fragment of mutant V187A mutant in the StuI-EcoRI digested mutant S185R. Expression constructs are made by cloning the corresponding StyI-XbaI fragments from the GST-E1A mutants in the E1A containing expression vector digested with the same enzymes.

Gal-Luc, the reporter containing five GAL-binding sites in front of the E1b tata-box, and the hRAR promoter construct -1470+156 Luc where made by cloning the promoter sequences from the reporters GAL-CAT (48) and -1470+156 CAT (38) to the promoterless Luciferase reporter construct pLUC.

Cell culture and transient transfection

Cells were cultured as described before (25) and transient transfections were performed by calcium phoshate precipitation as described (25) using 5-8µg of the indicated reporter together with 1µg of the indicated activator and 1µg of E1A 12S or 13S when indicated. As internal control for transfection efficiency 1µg SV2-LacZ (not affected by RA treatment or E1A cotransfection) was added. After 16 hrs the precipitate was removed and cells were cultured for another 24 hrs in the presence of 1.0µM RA. CAT activity was determined and normalized for transfection efficiency using the -galactosidase assay (25). Results are the mean (+/- SEM ) of at least five independent or three duplicate experiments.

Luciferase assays were performed as described by Brasier et al (8) with minor modifications. Extracts for this assay were made by lysing the cells on plates in 400µl Triton-lysis buffer (1% Triton X-100, 25mM glycylglycin (pH 7.8), 15mM MgSO4, 4mM EGTA and 1mM 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 (3.7mM ATP, 34mM glycylglycin (pH 7.8) and 20mM MgSO₄). Reactions were started upon addition of 100µl 0.2mM 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 (30mM NAPO₄ (pH7.0), 3mM KCl, 0.3mM MgSO₄, 15mM -mercaptoethanol and 20µg ortho-nitrophenyl--galactopyranoside was added together in a microtiter plate well. After 5-60 min -galactosidase activity was determined by measuring absorbance at 417 nM in an Elisa plate reader.
 

GST-pull down assay

Linearized plasmids (pSG5-RAR, pSG5-RXR, provided by P. Chambon) were transcribed using T7 RNA polymerase and translated in vitro using rabbit reticulocyte lysate in the presence of [³⁵S]-methionine according to the manufacturer (Gibco-Life). GST-fusion proteins were purified as described by Hateboer et al (33). Binding was performed in 450µl binding buffer (250mM NaCl, 50mM Hepes-KOH pH 7.5, 0.5mM EDTA, 0.1% (v/v) NP40, 0.2mM PMSF, 1mM DTT and 100µg BSA/ml) using one to five µl in vitro translated protein together with 1-2µg GST-fusion protein bound to glutathion-Sepharose for two hours at 4^oC. After four washings in binding buffer (4^oC, 1ml), bound proteins were eluted by boiling for 5 min in sample buffer and separated on 10-12.5% (w/v) SDS-PAGE. To compare the binding strength 1/10 of the total input of in vitro translated protein was run next to the GST- purified proteins. Quantification of GST-bound complexes was performed using a Phosphor-imager (Molecular Dynamics).
 

Western blotting

Whole cell extract (WCE) from transiently transfected COS cells was prepared by three subsequent freeze/thaw cycles (-80^oC/4^oC) in 50-100µl lysis buffer (20mM Tris pH 7.5, 20% (v/v) glycerol and 400mM KCl) together with 0.2mM PMSF and protease inhibitors (aprotinin, leupeptin, pepstatin and chymostatin, final concentration 1.0µg/ml of each). Equivalent amounts of extract were loaded and separated on 8-12.5% (w/v) SDS-PAGE and transferred to nitrocellulose using a semi-dry blot apparatus. Membranes were blocked in 4% (w/v) nonfat dry milk in PBST (150 mMNaCl, 16mM Na₂HPO₄ and 4mM NaH₂PO₄ and 1% (v/v) Tween 20) for 1 hr. Blots were incubated using either a monoclonal against E1A or HA-tag in PBST containing 2% (w/v) nonfat dry milk for 2 hrs. After 7 washes in PBST, blots were incubated with peroxidase-conjugated second antibodies in PBST containing 2% (w/v) nonfat dry milk. After 7 washes with PBST blots were developed using the ECL kit (Amersham).
 

Immunoprecipitations

A monoclonal antibody to the hemagglutinin tag (12CA5) was chemically crosslinked to protein A-Sepharose using dimethylpimelidate (DMP). The antibody was bound to protein A-Sepharose beads, washed three times with 0.1M borate buffer (pH 8.0), three times with 0.2M triethanolamine (pH 8.5) and then cross-linked for 1 hour at room temperature with 40mM DMP in triethanolamine (pH 8.5). Subsequently three washes with 40mM triethanolamine (pH 8.5) and and three washes with 0.1M borate buffer (pH 8.0) were performed. Immunoprecipitations were performed at 4^oC, using 50µl of chemically crosslinked E1A/protein A-Sepharose and 25-100µg whole cell extract of E1A 13S, HA-TBP or HA-RAR transfected COS cells in 750µl binding buffer (see GST-pull down assay). After two hours beads were washed four times with binding buffer. Antigen-antibody complexes were eluted by incubation at 100^oC in sample buffer, subsequently loaded on SDS-PAGE and transferred to nitrocellulose and immunoprobed as described above.

back

RESULTS

AF-2 of RAR is upregulated by E1A

We and others have previously shown that both RAR and RXR contain two autonomous cell- and promoter-specific activation functions (AFs) (25,59,60). To test whether any AF of the RAR/RXR heterodimer is involved in E1A-induced RAR promoter activation, AFs of both RAR and RXR were coupled to the DNA-binding domain of the yeast transcription factor GAL4 (Figure 1A) and transfected into COS cells together with GAL-CAT, a GAL-responsive reporter (48). As shown in Figure 1B the activity of RAR AF-2 is significantly upregulated (7-fold) by E1A 13S and not by the 12S form, whereas RXR AF-2 is not affected by cotransfection of E1A. Similar results where obtained with a Rous sarcoma virus (RSV) driven E1A expression construct (data not shown). No induction by E1A was observed in the absence of RA or activator (data not shown). The ligand-independent activation functions (AF-1) of both RAR and RXR were both slightly (two-fold) upregulated by E1A 13S in the presence or absence of RA (Figure 1B, data not shown). Cotransfection with increasing amounts of E1A 12S or 13S, in the presence of GAL-RAR AF-2 resulted in a slight decrease by the 12S form and a concentration-dependent increase in activity by E1A 13S (Figure 1C).
 

Cotransfection of E1A can prevent squelching by overexpressed AF-2

Previous experiments suggested that E1A 13S or E1A-like activity functions as cofactor in RA-dependent activation by RAR (5,39). If E1A can function as cofactor for RAR, addition of high amounts of RAR AF-2 titrates this cofactor out, while addition of E1A should then release squelching. In order to test this hypothesis we transfected increasing amounts of hRAR2 in P19 EC cells (containing E1A-like activity) using the hRAR2 promoter as reporter. We observed a concentration-dependent decrease in RAR promoter activity (Figure 2A) while addition of E1A prevented this decrease, but instead a small increase was found. Only at the highest concentration of hRAR2 a decrease in activity was observed, possibly because at this RAR concentration RXR, the dimerization partner for RAR, is limiting. In order to confirm that this squelching and the

Direct interaction between RAR and E1A

Subsequently various deletion constructs of RAR (Figure 3B) translated in vitro were used to map the region interacting with GST-E1A. As shown in Figure 3C both the F-region as well as the AB-region (containing AF-1) could be omitted without affecting the interaction with both E1A and RXR. Unexpectedly, deletion of the DNA binding domain (C-region) reduced binding of RAR to both RXR and E1A, while the interaction of RAR and RXR has been shown in similar assays to rely on the hormone-binding domain (53). Possibly the folding of the ligand-binding domain in vitro in the absence of the DNA binding domain is altered resulting in a weaker interaction with both RXR and E1A. As expected, deletion of the hormone-binding domain resulted in complete loss of binding between RAR and RXR. Similarly, the interaction between RAR and E1A was almost completely lost, in agreement with the transfection data (Figure 1B) indicating that direct interaction between RAR AF-2 and E1A is a requirement for the E1A dependent activation of AF-2.
 

Activation and binding of AF-2 by E1A is dependent on CRIII

To identify the region(s) of E1A required for activation we transfected various deletion constructs of E1A (67) (schematically presented in Figure 4A) into COS cells together with GAL-RAR AF-2 and a GAL-responsive reporter. Figure 4B shows that deletion of all regions except the conserved region III (CRIII) permits activation by these E1A constructs. These transfection experiments show that the CRI and CRII which are known to bind p300, RB and p107 (61) are not required for the upregulation of AF-2. Binding of RAR to GST-E1A deletion constructs with intact CRIII, is comparable with full length E1A or slightly reduced and, in agreement with the activation capacity, completely lost when the 13S specific region is deleted (Figure 4C). Subsequently N- and C-terminal deletions, as well as internal deletions affecting CRIII were tested for their ability to bind RAR. The region between amino acids (aa) 151 and 191 were sufficient for optimal binding (N1, C1, C2), demonstrating that the 13S-specific region is important for RAR binding (Figure 4C).

Behavior of mutants N1 and N2 indicates that possibly the N-terminal part of the protein (1-150) stabilizes the interaction between RAR and E1A since the binding of these N-terminal deletion constructs is impaired, although alternatively deletion of this region could disrupt the structure of the remaining protein. Deletion of the complete putative zinc finger region resulted in complete loss of binding (12S, N3, m3). The region C-terminal from the putative zinc finger region (aa 175-188), which has been shown to be required for E1A/ATF-2 interaction (49), also contributed to binding of RAR to E1A since the deletion constructs N3, C4, m3 and 132-185, containing only parts of this region, bound only weakly to RAR. Finally lack of effect of the internal deletion mutants (m1-m3) in binding to RAR indicates that the N-terminal part of CRIII is not required for binding to RAR (compare m1 with m2 and m3). Differences in binding of RAR to these GST-E1a mutants were not caused by variations in the amount of protein added in the GST-pulldown assays as is shown by Coomassie blue staining of such an assay (Figure 4C).

These findings strongly suggest that the region C-terminal part of CRIII is involved in RAR/E1A interaction. Transfection of these mutants on the RAR promoter or with GAL-RAR AF-2 on a GAL-responsive reporter in COS cells revealed that all deletions within the 13S specific region are deleterious for E1A-dependent upregulation of RAR AF-2 activity. C3/MX lacking only three aa of CRIII is impaired in activation, while binding to RAR was only marginally reduced. Transfection of these E1A mutants with the RAR2-promoter show that the same regions are required for upregulation of this promoter by E1A (Table 1).

The zinc finger region of E1A 13S is required for transcriptional activation

In general we observed a good correlation between the capacity of E1A to interact with RAR and the ability to enhance transcription by this receptor (Figure 4, Table I). However few exceptions were observed: the internal deletion mutants C3/MX, m1 and m2 could bind RAR, but their E1A-induced activation was lost or diminished. This can be explained by the requirement of both the activator and TBP to bind to E1A for transcriptional activation. The importance for an interaction between E1A and TBP for RA-dependent upregulation in activity of the RARs is supported by mutational analyses of TBP, showing that the core 1 region, close to the basic region of TBP, is required for activation by E1A-like activity (36). Recently a detailed mutational analysis revealed that the region between aa 147 and 177, containing the putative finger structure, is important for TBP/E1A interaction (26). This implies that m1 and m2, although still capable of binding to RAR, are unable to activate transcription, because they can no longer bind to TBP.

back

Discussion

Based on transfection experiments and binding studies we propose that the E1A-dependent upregulation of the RAR2 promoter in cells that are otherwise unable to activate the RAR2 promoter in response to RA is caused by an interaction between RAR AF-2 and TBP mediated by a direct interaction of these proteins with E1A 13S. On the basis of these data we propose that E1A-like activity is functioning like E1A, as a cofactor for RARs by simultaneously interacting with RAR and TBP, thereby enhancing the activity of the hormone-dependent activation funcion (AF-2).
 

E1A is functioning as a cofactor for RAR

The presence of intermediary factors, cofactors or adaptors in transcriptional activation has been proposed initially after the observation that overexpression of activators leads to squelching (64). These cofactors are thought to contact both a component of the basal transcription machinery as well as the promoter-bound activator, thereby stabilizing the pre-initiation complex. Finally this model predicts that addition of extra cofactors could prevent squelching caused by overexpression of the activator.

Previous experiments by us and others (5,39) have shown that E1A can enhance transcription of the RAR/RXR complex bound to the RARE. Here we present evidence that the viral activator E1A enhances RA-dependent activation of the RAR gene, functioning as a cofactor for RARs. This is supported by the following observations: (1) The RAR2 promoter is inactive in cell lines lacking E1A or E1A-like activity, whereas it is highly upregulated by RA in cells having such proteins like P19-EC, F9-EC, HEPG2 and 293 cells. Introduction of E1A 13S in the former cell type enables RARs to further activate this promoter (5,39). (2) Transfection experiments using this promoter have shown that the hormone-dependent AF (AF-2) of RARs is most important for activation of this promoter (75). We show here that the activity of the ligand-dependent activation function (AF 2) of RAR (but not of RXR) is upregulated by E1A, while the ligand-independent activation functions (AF-1) of these receptors are not affected (Figure 1). This further strengthens the importance of E1A in the RA-dependent upregulation of the RAR2 promoter since the AF that is upregulated by E1A is also the critical AF for RAR2 promoter activation. (3) Squelching by overexpression of AF-2 of RAR could be prevented by co-transfection of E1A 13S (Figure 2). (4) The activity of RAR AF-2 is cell-specific (25) and a good correlation was observed between the presence of E1A-like activity/E1A and the activity of this activator in the presence of RA (25). Cotransfection of E1A 13S in cells not expressing E1A or E1A-like activity caused an upregulation in activity, whereas in cells already expressing these proteins no or only marginal upregulation in activity was observed (Folkers et al, manuscript in preparation), indicating that there is a correlation between the strength of this activator and the presence of E1A or E1A-like activity. (5) Using in vitro binding assays we have shown that RAR but not RXR can interact with E1A and that this interaction is dependent on the presence of the 13S-specific region and the ligand-binding domain of RAR, containing AF-2, the region shown to be required for the E1A-dependent upregulation in activity (Figure 3). (6) Deletion and point mutants within the 13S specific region show that this activation is dependent on the TBP-binding region, on a distinct region required for RAR binding and finally on a region not involved in binding of RAR or TBP. This indicates that interaction with both RAR and TBP and a third (hypothetical) protein are all important for E1A-dependent upregulation of AF-2 activity (Figure 4,5, Table 1). Altogether these data indicate that E1A fulfills all characteristics for being a cofactor for RARs.

The requirement for cofactors has been shown to be important for members of the steroid/ thyroid hormone receptor superfamily both in vivo (56) and in vitro (9,63), but so far however no cofactor for a member of this family has been cloned. Recently however two groups have purified proteins that interact with AF-2 of the estrogen receptor and may fulfill a cofactor like function (11,32). Furthermore transcriptional enhancement by the SWI genes as well as their human homologs has been reported to function in the glucocorticoid receptor-dependent transcriptional activation, probably at the level of chromatin disruption (13,58,76).
 

Mechanisms underlying cofactor action of E1A

In recent literature there is compelling evidence that certain activators are brought into proximity of the basal transcription machinery by cofactors interacting with both the activator and a target in the basal transcription machinery. The nuclear protein CBP, is contacting simultaneously phosphorylated CREB and TFIIB, causing transactivation by the AF of CBP (2,41). Transactivation by BmFTZ-F1 is mediated by the presence of two proteins, MBF1 and MBF2, forming a dimer which is necessary for complex formation between TBP and BmFTZ-F1 (46). A similar model has been proposed for E1A acting as cofactor in Oct-4 dependent transcription. In gel mobility shift assays a direct interaction between Oct4 and E1A was observed; in this case the 13S-specific region was required for both activation and binding to Oct4, suggesting that E1A functions as a bridging factor (68). For E1A and the CCAAT-Binding factor (CBF) an interaction between the CRIII region and CBF was required to mediate transcriptional activation (1).

The interaction of E1A with TBP has been studied in detail. This interaction can both take place with recombinant TBP (34,43) and holo-TFIID (7) and is dependent on the zinc-binding region of E1A (CR3) (26). Furthermore the residues C-terminal from this region are also important for transcriptional activation (49,73). Based on transactivation, trans-repression and binding experiments with E1A mutants (49,73) it was suggested that two discrete functions are present in this region: binding of the promoter-bound activator and a different binding site for a limiting factor. For RAR the results are similar: all mutants that were reported to prevent TBP binding, are unable to activate transcription (C157S and C174S) although these mutants can still bind to RAR. Also the mutant unable to bind RAR (178-179) is impaired in transactivation. Furthermore binding experiments indicate that the N-terminal part of E1A (aa 1-150) contributes to binding of E1A to RAR. This region has been shown to interact with various proteins including RB, p107, p300 (61) indicating that possibly multiple interactions between RAR and E1A are required for transcriptional activation by E1A. A third type of E1A mutants not directly involved in RAR or TBP binding (C3/MX, S185R, V187L, and 185-187) created proteins unable to function as cofactor for RAR. We suggest that these mutants are unable to bind a third partner within this complex as a requirement for further formation of the preinintiation-complex. This protein could be an additional cofactor, a TAF or otherwise a basal transcription factor. It is unclear whether binding of all proteins is required simultaneously or subsequently for the functioning of E1A as a cofactor for RAR.

Taken together we therefore suggest that three distinct regions of the 13S-specific region are required for transcriptional activation which each bind different proteins: the TBP binding region (147-174), the RAR binding region arround aa 178 possibly requiring the zinc-finger structure since deletion mutants deleting part of the zinc finger are defective or impaired in binding to RAR; point mutants however, proposed to destroy the zinc-finger structure are still capable of binding to RAR (C157S, C174S). And finally a region that possibly interacts with an additional protein, binding in a region partially overlapping with the RAR-binding region and extending to aa 188. Mutational analysis of ATF2 has shown that deletion of aa 179-193 destroys both binding to and activation by E1A (49). Based on transactivation and transrepression experiments using E1A mutants, it was concluded that the extreme C-terminal part of the 13S specific region (183-188) is involved in ATF binding and the region N-terminal from this region, partially overlapping with the TBP binding region, was proposed to bind an unknown limiting factor (26,49). Detailed mutational analyses using more point mutants will be required to prove that RAR and ATF2 bind to different regions of E1A and to establish whether different putative limiting factors are bound to the C-terminal part of E1A when acting as cofactor for RAR or ATF.

Recently several groups have identified an autonomous transactivation function within the hormone-binding domain of members of this superfamily (4,19,45,66,71), which is proposed to form an -helix with hydrophobic residues on one side and negatively charged glutamic acid on the other side of the helix. Since the region interacting with E1A co-localizes with the region required for transactivation we propose that the activation domain is involved in this interaction. The region of E1A 13S important for RAR binding contains several hydrophobic residues and one positively charged arginine suggesting that a stable interaction could be formed by hydrogen bonds or ionic interactions between these proteins . Another possible explanation could be that the interaction is not mediated by AF-2 of RAR directly but rather the consequence of the E1A/RAR interaction bringing the region required for transcriptional activation in proximity with a target molecule of the basal transcription machinery, which cannot or more difficult take place in the absence of E1A.
 

Functional importance of E1A

The presence of an E1A like activity in cells where the activity of the RAR promoter can be activated by RA (P19-EC, F9, HepG2), and the inability of RAR2 promoter activation in cells that have lost their E1A-like activity (differentiated EC cells), makes it likely that this protein(s) function by a similar mechanism as its viral counterpart. This was further confirmed by squelching experiments by overexpression of AF-2 in P19-EC cells indicating that E1A can functionally replace E1A-like activity in its ability to activate through AF-2 (Figure 2). These results suggest that E1A-like activity is also acting as a cofactor mediating activation by an interaction with both RAR and TBP. The importance for an interaction of E1A-like activity with TBP for RA-dependent upregulation of the RAR-promoter became apparent from mutational analyses of TBP showing that a region in core 1 is required for activation (36), although it remains to be seen whether E1A or E1A-like activity indeed contacts the same region (43). This could open the way for cloning E1A-like activity by virtue of its ability to bind to both RAR and TBP. The identification of E1A-like activity and the subsequent study of the role in RAR-promoter activation and embryonal development can ultimately establish the role of this protein as cofactor. The availability of the cofactor required for RA-dependent transactivation by RARs also could make it possible to investigate whether the restricted expression pattern of RAR during embryonal development (18,65) is caused by the absence of these cofactors for RARs.
 

Acknowledgements

We are grateful to Dr. P. Chambon for the RAR and RXR expression constructs; to N.C. Jones for E1A constructs and M.Timmers for TBP expression constructs and pGEX constructs, and to A. Zantema for the monoclonal antibody M73. We further thank P.J. Coffer, A.J.M. Walhout and C.L. Mummery for critically reading the manuscript. This work was supported by the Netherlands Cancer Society.

back

References

1. Agoff, S.N., and B. Wu. 1994. CBF mediates adenovirus E1a trans-activation by interaction at the C-terminal promoter targeting domain of conserved region 3. Oncogene 9:3707-3711.

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

3. Baniahmad, A., I. Ha, D. Reinberg, S. Tsai, M.-J. Tsai, and B.W. O'Malley. 1993. Interaction of human thyroid hormone receptor beta with transcription factor TFIIB may mediate target gene derepression and activation by thyroid hormone. Proc. natl. Acad. Sci. USA 90:8832-8836.

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

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

6. Boulanger, P.A., and G.E. Blair. 1991. Expression and interaction of human adenovirus oncoproteins. Biochem. J. 275:281-299.

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

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

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

10. Byers Kraus, V., J.A. Inostroza, K. Yeung, D. Reinberg, and J.R. Nevins. 1994. Interaction of the Dr1 inhibitory factor with the TATA binding protein is disrupted by adenovirus E1A. Proc. Natl. Acad. Sci. USA 91:6279-6282.

11. Cavaillès, V., S. Dauvois, P.S. Danielian, and M.G. Parker. 1994. Interaction of proteins with transcriptionally active estrogen receptors. Proc. Natl. Acad. Sci. USA 91:10009-10013.

12. . Chatton, B., J.L. Bocco, M. Gaire, C. Hauss, B. Reimund, J. Goetz, and C. Kedinger. 1993. Transcriptional activation by the adenovirus larger E1a product is mediated by members of the cellular transcription factor ATF family which can directly associate with E1a. Mol. Cell. Biol. 13:561-570.

13. Chiba, H., M. Muramatsu, A. Nomoto, and H. Kato. 1994. Two human homologues of Sacchromyces cerevisiae SWI2/SNF2 and Drosophila brahma are transcriptional coactivators cooperating with the estrogen receptor and the retinoic acid receptor Nucl. Acid Res. 22:1815-1820.

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

15. Culp, J.S., L.C. Webster, D.J. Friedman, C.L. Smith, W.-J. Huang,

F.Y.-H. Wu, M. Rosenberg, and R.P. Ricciardi. 1988. The 289-amino acid E1A protein of adenovirus binds zinc in a region that is important for trans-activation. Proc. natl. Acad. Sci. USA 85:6650-6454.

16. de Groot, R.P., N. Foulkes, M. Mulder, W. Kruyer and P. Sassone-Corsi. 1991. Positive regulation of jun/AP1 by E1A. Mol. Cell. Biol. 11:192-201.

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

18. Dollé, P., E. Ruberte, P. Leroy, G. Morris-Kay, and P. Chambon. 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.

19. Durand B., M. Saunders, C. Gaudon, B. Roy, R. Losson, and P. Chambon. 1994. 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. 13:5370-5382.

20. Durston, A.J., J.P.M. Timmermans, W.J. Hage, H.F.J. Hendriks, N.J. de Vries, M. Heideveld, and P.D. Nieuwkoop. 1989. Retinoic acid causes an anteroposterior transformation in the developing central nervous system. Nature 340:140-144.

21 Eichele, G. 1989. Retinoids in vertebrate and limb pattern formation. Trends Genet. 5:246-251.

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

23. Ferguson,B., B. Krippl, O. Andrisani, N. Jones, H. Westphal, and M. Rosenberg. 1985. E1A 13S and 12S mRNA products made in Escherichaia coli both function as nucleus-localized transcription activators but do not directly bind DNA. Mol. Cell. Biol. 5:2653-2661.

24. Flint, J., and T., Schenk. 1989. Adenovirus E1a protein paradigm viral activator. Annu rev. Genet. 23:141-161.

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

26. Geisberg, J.V., W.S. Lee, A.J. Berk, and R.P. Ricciardi. 1994. The zinc finger region of the adenovirus E1A transactivating domain complexes with the TATA box binding protein. 1994. Proc. natl. Acad. Sci. USA 91:2488-2492.

27. Gill, G., E. Pascal, Z.H. Tseng, and R. Tjian. 1994. A glutamine-rich hydrophobic patch in transcription factor Sp1 contacts the dTAF_II110 component of the Drosophila TFIID complex and mediates transcriptional activation. Proc. natl. Acad. Sci. USA 91:192-196.

28. Gill, G., and R. Tjian. 1992. Eukaryotic coactivators associated with the TATA box binding protein Curr. Op. Gen. Dev. 2:236-242.

29. Gill, G., and M. Ptashne. 1988. Negative effect of the transcriptional activator GAL4 Nature 334:721-724.

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

31. Gudas, L.J. 1992. Retinoids, retinoid-responsive genes, cell differentiation, and cancer. Cell Growth Diff. 3:655-662.

32. Halachmi, S., E. Marden, G. Martin, H. MacKay, C. Abbondaza and M. Brown. 1994. Estrogen receptor-associated proteins: possible mediators of hormone-induced transcription. Science 264:1455-1458.

33. Hateboer, G., H.T.M. Timmers, A.K. Rustgi, M. Billaud, L.J. van 't Veer, and R. Bernards. 1993. TATA-binding protein and the retinoblastoma gene product bind to overlapping epitopes on c-Myc and adenovirus E1A protein. Proc. natl. Acad. Sci. USA 90:8489-8493.

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

35. Imperiale, M.J., H.T. Kao, L.T. Feldman, J.R. Nevins, and S. Strickland. 1984. Common control of the heat shock gene and early adenovirus genes: evidence for a cellular E1a-like activity. Mol. Cell. Biol. 4:867-874.

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

37. Kook Kim, T., and R.G. Roeder. 1994. Proline-rich activator CTF1 targets the TFIIB assembly step during transcriptional activation. Proc. natl. Acad. Sci. USA 91:4170-4174.

38. Kruyt, F.A.E., G.E. Folkers, C.E. van den Brink, and P.T. van der Saag. 1992. A cyclic AMP response element is involved in retinoic acid-dependent RAR2 promoter activation. Nucleic Acid Res. 20:6393-6399.

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

40. Kurokawa, R., V. Yu, A. Näär, S. Kyakumoto, Z. Han, S. Silverman, M.G. Rosenfeld, and C.K. Glass. 1993. Differential orrientations of the DNA-binding domain and C-terminal dimerization interface regulate the binding site selection of nuclear receptor heterodimers. Genes Dev. 7:1423-1435

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

42. La Thanque, N.B., and P.W.J. Rigby. 1987. An adenovirus E1A-like transcription factor is regulated during the differentiation of murine embryonal carcinoma stem cells. Cell 49:507-513.

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

44. Leid, M., Kastner, P. & Chambon, P. 1992. Multiplicity generates diversity in retinoic acid signalling pathways. Trends Biochem. Sci. 17:427-433.

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

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

47. Lieberman, P.M., and A.J. Berk. 1994. A mechanism for TAFs in transcriptional activation: activation domain enhancement of TFIID-TFIIA-promoter DNA complex formation. Genes Dev. 8:995-1006.

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

49. Lui, F., and M.R. Green. 1994. Promoter targeting by adenovirus E1a through interaction with different cellular DNA-binding domains. Nature 368:520-525.

50. Lui, F., and M.R. Green. 1990. A specific member of the ATF transcription factor family can mediate transcription activation by adenovirus E1A protein. Cell 61:1217-1224.

51. Liu, X., C.W. Miller, P.H. Koeffler, and A.J. Berk. 1993. The p53 activation domain binds the TATA box binding polypeptide in holo-TFIID, and a neighboring p53 domain inhibits transcription. Molec. Cell. Biol. 13:3291-3300.

52. Mader, S., J.-Y. Chen, Z. Chen, J. White, P. Chambon, and H. Gronemeyer. 1993. The pattern of binding of RAR, RXR and TR homo- and heterodimers to direct repeats are dictated by the binding specificities of the DNA binding domains. EMBO J. 12:5029-5041.

53. Marks, M.S., P.L. Hallenbeck, T. Nagata, J.H. Segars, E. Appella, V.M. Nikodem, and K. Ozato. 1992. H-2RIIBP (RXR) heterodimerization provides a mechanism for combinatorial diversity in the regulation of the retinoic acid and thyroid hormone responsive genes EMBO J. 11:1419-1435.

54. Martin, G.R. 1975. Teratocarcinomas as a model system for the study of embryogenesis and neoplasia. Cell 5:229-243.

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

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

57. Montano, X., and D.P. Lane. 1987. The adenovirus E1a gene induces differentiation of F9 teratocarcinoma cells. Mol. Cell. Biol. 7:1782-1790.

58. Muchardt, C., and M. Yaniv. 1993. A human homologue of Sacchromyces cerevisiae SNF2/SWI2 and Drosophila brm genes potentiates transcriptional activation by the glucocorticoid receptor. EMBO J. 12:4279-4290.

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

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

61. Nevins, J.R. 1993. Transcriptional activation by the adenovirus E1A proteins. Sem. Vir. 4:25-31.

62. Perlmann,T., P.N. Rangarajan, K. Umensone, and R.M. Evans. 1993 Determinants for selective RAR and TR recognition of direct repeat HREs. Genes Dev. 7:1411-1422.

63. Pfitzner, E., A. Sak, V. Ulber, G.U. Ryffel, and L. Klein-Hitpass. 1993. Recombinant activation domains of virion protein 16 and human estrogen receptor generate transcriptional interference in vitro by distinct mechanisms. Mol. Endocrin. 7:1061-1071.

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

65. Ruberte, E., P. Dollé, P. Chambon, and G. Morriss-Kay. 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.

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

67. Schneider, J.F., F. Fisher, C.R. Goding and N.C. Jones. 1987. Mutational analysis of the adenovirus E1a gene: the role of transcriptional regulation in transformation. EMBO J. 6:2053-2060.

68. Schöler, H.R., T. Ciesiolka, and P. Gruss. 1991. A nexus between Oct-4 and E1A: implications for gene regulation in embryonic stem cells. Cell 66:291-304.

69. Slack, R.S., J. Craig, S. Costa and M.W. McBurney. 1995. Adenovirus 5 E1A induced differentiation of P19 embryonal carcinoma cells requires binding to p300. Oncogene 10:19-25.

70. Suemori, H., S. Hashimoto, and N. Nakatsuji. 1988. Presence of the adenovirus E1A-like activity in preimplantation stage mouse embryos. Mol. Cell. Biol. 8:3553-3555.

71. Tone Y, T.N. Collingwood, M. Adams, and V.K. Chatterjee. 1994. Functional analyses of a transactivation domain in the thyroid hormone receptor. J. Biol. Chem. 269:31157-31161.

72. Van der Leede, B.-J.M., C.E. van den Brink, P.T. van der Saag. 1993. Retinoic acid receptor and retinoid X receptor expression in retinoic acid-resistant human tumor cell lines. Mol. Carcinog. 8:112-122.

73. Webster, L.C., and R.P. Ricciardi. 1991. trans-dominant mutants of E1A provide genetic evidence that the zinc finger of the trans-activating domain binds a transcription factor. Mol. Cell. Biol. 11:4287-4296.

74. Weigel R.J., and J.R. Nevins. 1990. Adenovirus infection of differentiated F9 cells results in a global shut-off of differentiation-induced gene expression. Nucleic Acid Res. 18:6107-6112.

75. Xiao, J.-H, B. Durand, P. Chambon, and J.J. Voorhees. 1995. Endogenous retinoic acid receptor (RAR)-retinoid X receptor (RXR) heterodimers are the major functional forms regulating retinoid-responsive elements in adult human keratinocytes. J. Biol. Chem. 270:3001-3011.

76. Yoshinaga, S.K., C.L. Peterson, I. Herskowitz and K.R. Yamamoto. 1992. Roles of SWI1, SWI2 and SWI3 proteins for transcriptional enhancement by steroid receptors. Science 258:1598-1604.
 
index
homepage