INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

TFIID is a general transcription factor composed of a small TATA-binding polypeptide and a large number of TATA-binding protein (TBP)-associated factors (TAFs), all of which are highly conserved in evolution (reviewed in references 7 and 17). TFIID is involved, along with other general initiation factors (TFIIA, TFIIB, TFIIE, TFIIF, and TFIIH), in both activator-independent (basal) and activator-dependent transcription. Furthermore, and in contrast to basal transcription, activator-dependent transcription in mammalian cell-free systems reconstituted with purified factors generally requires cofactor activities that include both USA (upstream factor stimulatory activity)-derived factors and TBP-associated factors (TAF_IIs) within TFIID (for reviews, see references 28, 35, 37, and 38).

In general, the efficiency of preinitiation complex (PIC) assembly or function is controlled by the presence of transcription factors that are usually bound to specific distal sequences. Some models of how transcription regulatory factors influence PIC assembly invoke interactions with TFIID that, through qualitative and/or quantitative effects on TFIID binding, enhance the recruitment of downstream factors (reviewed in reference 7). Whereas TFIID from metazoans was found to mediate both basal and activator-dependent transcription in cell-free systems reconstituted with partially purified components, TBP elicited mainly basal transcription. This led to the hypothesis that TAF_IIs within TFIID interact directly with activators to promote PIC assembly. Conversely, and using reconstituted TFIID complexes, a seemingly good correlation was drawn between the activity of a specific activator and the ability of its activation domain to selectively bind a specific given TAF_II (for reviews, see references 7 and 43). In addition, in vitro studies have shown an important role for TAF_IIs within TFIID in core promoter recognition and transcriptional strength, especially on TATA-less promoters that contain other core promoter elements such as the initiator (Inr) and/or downstream promoter elements (for a review, see reference 40). In this regard, early in vivo studies in Saccharomyces cerevisiae suggested that individual TAF_IIs are dispensable for activated transcription of most genes (31, 44) and that core promoter elements, rather than upstream binding sites, confer TAF_II-dependence on some genes (39).

Consistent with these latter observations is the finding that the S. cerevisiae Mediator complex can support activated transcription in vitro with TBP alone (22, 25). In addition, it was also reported that TAF_IIs are not required either for activated transcription by GAL4-VP16 in unfractionated HeLa nuclear extracts (34) or for activation by thyroid hormone receptor in association with the human TRAP/Mediator in a partially purified system (15). However, in a reexamination of the TAF_II requirement for activator function in S. cerevisiae, more extensive genetic analyses have suggested that at least some TAF_IIs (notably the histone-related TAF_IIs) are broadly required for transcription and that the TAF_II dependency, in some cases, may require upstream activators (reviewed in reference 17). Finally, and further complicating the interpretation of the in vivo assays, is the discovery that a subset of TAF_IIs found in the TFIID complex are also integral components of histone acetyltransferase complexes in S. cerevisiae and humans (for review, see reference 6). Thus, TAF_IIs have been shown to serve as conventional coactivators acting at the DNA level, as core promoter-selective factors, and within coactivators implicated in chromatin modifications, and at least one TAF_II has several catalytic activities that are potentially involved in transcription (reviewed in reference 17). It thus seems likely that different TAF_IIs may function by distinct mechanisms that depend on the specific regulatory elements and chromosomal architecture of a given promoter.

An understanding of the various TFIID functions is based on a resolution of the overall architecture of TFIID that requires knowledge of both the primary sequences and structures of individual subunits and their interactions and topological organization within TFIID (1, 5). Studies of S. cerevisiae, Drosophila melanogaster, and human TAF_IIs have provided valuable information on a number of protein-protein interactions and, of special interest, the potential for a histone octamer-like structure within TFIID that would comprise, in the human system, the H4-like (hTAF_II80), the H3-like (hTAF_II31), and the H2B-like (hTAF_II20) subunits (21).

Here we describe interactions of the H2A-related hTAF_II135 with the H2B-related hTAF_II20 through histone-like folds (16, 21) and the importance of this interaction for human TFIID assembly. We also report important new insights into TFIID function based on the demonstration of synergistic hTAF_II135-TFIIA interactions that relieve hTAF_II250-mediated inhibition of TBP binding and function, as well as a functional synergism between TAF_IIs and the TRAP/Mediator complex.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of cDNA clone for hTAF_II135. TFIID was affinity purified from either HeLa nuclear extract (NE) or a phosphocellulose (Whatman P11) fraction (0.85 KCl) of HeLa NE by use of an antibody against the N-terminal portion of hTAF_II100 or anti-TBP antibodies, respectively (42). The 135-kDa polypeptide was excised and digested with endoproteinase Lys-C (13). Earlier sequence analysis of the purified peptides yielded several sequences that were used to design degenerate oligonucleotides and to isolate the cDNA corresponding to the truncated hTAF_II135 sequence (amino acids 239 to 1083). Subsequent hTAF_II135 peptide sequence analysis revealed two additional peptides. The sequence of one of these two peptides matched the translation of a genomic DNA clone (accession number Z22478). The sequences of this genomic clone were used to obtain the missing 5' region of hTAF_II135 DNA coding region. The full-length coding DNA sequence (amino acids 1 to 1083) that was obtained is identical to the one reported by Mengus et al. (30).

Expression and purification of TFIID subunits, activators, and GST derivatives. The pVL derivatives for the expression of hemagglutinin (HA)-hTAF_II250, Flag-hTAF_II100, Flag-hTAF_II80, Flag-hTAF_II55, Flag-hTBP, Flag-hTAF_II31, and Flag-hTAF_II28 have been described (11, 18, 19, 42). Expression pVL plasmids for Flag-hTAF_II150, Flag-hTAF_II135, Flag-hTAF_II20, and HA-hTAF_II20 were constructed by PCR. In each case, an NdeI site at the N-terminal end and an appropriate restriction enzyme site at the C-terminal end following the natural stop codon were created. The large number of primers used in the PCRs has precluded description of their exact sequences, but the information is available upon request. The PCR-generated fragments were then inserted into adapter pFlag(S)-7 and pFlag (AS)-7 plasmids carrying the appropriate tag epitope (10) and subsequently subcloned into either pVL-1392 or pVL-1393 (41). Each construct was verified by sequencing.

Generation of antibodies against hTAF_II135. A cDNA corresponding to the C-terminal portion of hTAF_II135 was amplified by PCR with the appropriate restriction enzymes. The PCR-generated fragment was then inserted into bacterial expression construct pET11d (Novagen), which carries the six-histidine tag. The recombinant protein was then expressed in and purified from bacteria and subjected to preparative SDS-PAGE. Gel slices containing the corresponding protein were crushed and used for immunization of two rabbits.

Generation of hTAF_II135 cell line. A HeLa cell line constitutively expressing Flag-hTAF_II135 (f:135) was made using the pCIN4 expression vector (36).

Far Western blotting. The baculovirus-expressed and purified hTAF_II135 was labeled with ³²P at the heart muscle kinase site present in the Flag-tagged sequence (10) by incubation with heart muscle kinase and [-³²P]ATP for 30 min at 30°C. Labeled protein was then purified through a nick column (Pharmacia) and used for protein-blot interactions as described (4).

Gel filtration. Purified TFIID preparations from either f:135 or f:TBP cell lines were fractionated on Superose 6 (Smart System; Pharmacia) in buffer BC200 containing 0.05% NP-40. Fractionated proteins were resolved by SDS-PAGE and visualized by silver staining.

DNase I footprinting. Plasmid pML4, containing the major late promoter, was used for DNase I footprinting as described (10). Briefly, the EcoRI-HindIII DNA fragment from pML4 was isolated and end labeled with ³²P by T4 polynucleotide kinase. Cleavage of the DNA fragment with XbaI generated a specific labeled transcribed strand. DNase footprinting reactions and the processing of the labeled products were performed essentially as described previously (10).

In vitro RNA polymerase II transcription assays. Nuclear extracts were prepared as previously described (12). TFIID, TFIIH, and the Flag-thyroid hormone receptor alpha (TR)-TRAP complex were purified from cell lines expressing Flag-TBP, Flag-hTAF_II135, Flag-ERCC3, and Flag-TR, respectively, using affinity purification on anti-Flag antibody columns as previously described (10, 15, 18). Flag-TR and Flag-retinoid X receptor alpha (RXR) were purified from Sf9 cells as previously described (14). Native TFIIA was purified as previously described (18). For TFIIA (p55 and p12), TFIIB, TFIIE, and TFIIE recombinant Flag-tagged proteins were expressed in and purified from E. coli using an anti-Flag antibody column (M2 agarose; Kodak). Histidine-tagged TBP and TFIIF subunits (RAP30 and RAP74) were prepared as described previously (18). TFIIA and TFIIF were then reconstituted from individually purified components following denaturation and renaturation (45). RNA polymerase II was purified essentially as described previously (3).

Partial TFIID reconstitution. Following the method that we described earlier (18), human hTAF_II250 or hTAF_II20 containing a fused N-terminal HA epitope tag was immobilized on protein A-Sepharose containing covalently linked monoclonal antibodies directed against the HA epitope. After extensive washing (BC1000 with 0.1% NP-40), the beads were incubated sequentially (at 4°C for 4 h) with molar excesses of additional TFIID subunits. After each incubation, unbound materials were removed by several washes with 100 volumes of BC150 (with 0.1% NP-40). Finally, the resulting complex was eluted with HA peptide (1 mg/ml) in BC100 (with 0.1% NP-40).

In vitro protein-protein interaction assays. Sp1 activation domains A and B, TFIIA subunits p12 and p55, hTAF_II20, and hTBP were expressed in and purified from bacteria as GST fusion proteins. Histidine-tagged full-length hTAF_II135 and hTAF_II135 deletion mutant constructs were expressed in the TNT reticulocyte lysate system (Promega) and labeled with [³⁵S]methionine according to the manufacturer's instructions. Equivalent inputs of radioactive material of full-length hTAF_II135 and hTAF_II135 deletion mutants were used for binding studies with different GST derivatives. In each reaction, 1 µg of purified GST or GST derivative was immobilized on glutathione, and the appropriate input material was added to each reaction (in 300-µl total volume in BC300 plus 0.1% NP-40). After incubation at 4°C for 2 h, the beads were washed four times with 300 µl of incubation buffer, and bound proteins were eluted with SDS loading buffer and analyzed by SDS-PAGE and autoradiography.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Strong and specific interaction of hTAF_II135 with hTAF_II20. The multisubunit nature of TFIID (TBP and TAF_IIs) and the overall stability of human TFIID suggest a multiplicity of protein-protein interactions, with any single TAF_II expected to interact with a number of other subunits. To test for direct and comparative interactions of purified hTAF_II135 with affinity-purified TBP and other human TAF_IIs (hTAF_II250, hTAF_II150, hTAF_II135, hTAF_II100, hTAF_II80, hTAF_II55, hTAF_II43, hTAF_II31, hTAF_II28, hTAF_II20, and hTAF_II18), purified proteins (Fig. 1A) were probed with a Flag-tagged, ³²P-labeled hTAF_II135 in a far Western blot. This analysis (Fig. 1B) revealed a strong interaction of hTAF_II135 with hTAF_II20, which was unexpected on the basis of Drosophila studies (9), as well as a moderate interaction with hTAF_II150 and weak interactions with hTAF_II250 and hTAF_II43. Since the interaction of hTAF_II135 with hTAF_II20 was by far the strongest among all TFIID subunits, we speculated that it might have important consequences for de novo assembly of TFIID and, in particular, the recruitment of hTAF_II135. Furthermore, although one essential feature of the original octamer-like model was the presence of two hTAF_II20 molecules in TFIID, there was no apparent H2A-like partner for hTAF_II20 (which was assumed to heterodimerize).

View larger version (52K): [in this window] [in a new window]   FIG. 1.   Comparative far Western assays showing strong interactions between TAFII135 and hTAFII20. (A) Individually expressed (Sf9 cells via baculovirus vectors) and purified TFIID subunits were resolved on SDS-PAGE and visualized by Coomassie staining, M, protein markers, with molecular sizes indicated in kilodaltons on the left. (B) A set of gels equivalent to those in panel A were transferred onto nitrocellulose, denatured-renatured, and used for binding with 32P-labeled and purified hTAFII135. After extensive washing, bound material was analyzed by autoradiography.

View larger version (27K): [in this window] [in a new window]   FIG. 2.   The extreme C-terminal portion of hTAFII135 interacts with hTAFII20. The indicated deletion mutants of hTAFII135 were [35S]methionine-labeled and incubated with GST-hTAFII20. Input samples (I) contained 10% of the amount used for binding (B). The arrow indicates the band corresponding to the appropriate mutant hTAFII135 deletion.

Functional in vivo relevance of interaction between TAF_II135 and TAF_II20. To assess the physiological relevance of the interaction of TAF_II135 and TAF_II20 and its potential involvement in a histone octamer-like structure, we generated a cell line that stably expressed Flag-tagged TAF_II135 (f:135). TFIID complexes were purified from this cell line and from a cell line expressing a Flag-tagged TBP (f:TBP). A Western blot of nuclear extract from f:135 cells revealed a level of f:135 protein expression at least fivefold higher than that of endogenous TAF_II135 protein (Fig. 3A). Similar Western blots using antibodies against TBP and various TAFs (hTAF_II250, hTAF_II150, hTAF_II135, hTAF_II100, hTAF_II80, hTAF_II55, hTAF_II31, hTAF_II20, and hTAF_II15) revealed the presence of these components both in TFIID purified from the f:135 cell line and in TFIID prepared from the f:TBP cell line (Fig. 3B). Significantly, however, there was a clear enrichment of hTAF_II20 (and the hTAF_II15 isoform) in the TFIID preparation from the f:135 cell line compared to that from the f:TBP cell line when normalized to the content of TBP and other TAFs (Fig. 3B). These data are consistent with the observed in vitro interaction between hTAF_II20 and hTAF_II135. The data further show copurification of natural endogenous hTAF_II135 and exogenous Flag-hTAF_II135 (Fig. 3A) in the TFIID that was affinity purified (via f:135) on anti-Flag antibody columns and in the same ratio that they are expressed in the f:135 cell line (Fig. 3C). Furthermore, a Western blot with anti-Flag antibody revealed only one band, corresponding to the exogenous Flag-tagged TAF_II135, in the TFIID purified from the f:135 cell line and no reactive bands in TFIID purified from the f:TBP cell line (data not shown). Since there is no apparent self-association of hTAF_II135 (Fig. 1), this clearly indicates the presence of at least two molecules of hTAF_II135 within the TFIID complex.

View larger version (43K): [in this window] [in a new window]   FIG. 3.   At least two molecules of TAFII135 coexist in TFIID. (A) Western blot analysis was performed using anti-hTAFII135 (lanes 1 and 2) and anti-Flag antibodies (lanes 3 and 4) to probe SDS-PAGE-resolved proteins in nuclear extracts. Both the endogenous hTAFII135 (open arrow) and Flag-TAFII135 (solid) are present in nuclear extract (NE) prepared from the Flag-TAFII135 (f:135) cell line (lane 1), whereas only the endogenous hTAFII135 band is present (lane 2) in NE from the Flag-TBP (f:TBP) cell line. (B) Comparison of TBP and TAFII expression in the TFIID purified from f:TBP and f:135 cell lines. As indicated, increasing amounts (microliters) of purified TFIID from either cell line were analyzed by Western blotting with antibodies generated against each indicated TFIID subunit. (C) Western blot analysis was performed using anti-hTAFII135 antibody. The endogenous hTAFII135 (open arrow) and Flag-hTAFII135 (solid arrow) proteins coexist in the purified TFIID prepared from the f:135 cell line (lane 2). In the TFIID purified from the f:TBP cell line, only the endogenous hTAFII135 (open arrow) is detected (lanes 1 and 3). (D and E) Indicated fractions, obtained from gel filtration on Superose 6, of purified TFIID from f:TBP and f:135 cell lines, respectively, were resolved by SDS-PAGE and silver stained.

Role of TAF_II20 and TAF_II135 interaction in human TFIID assembly. The interaction of hTAF_II135, through its histone fold, with the H2B-related hTAF_II20 constitutes a novel histone-like pair that parallels the well-characterized H3-related hTAF_II31 and H4-related hTAF_II80 pair. This is consistent with the possibility of a histone octamer-like substructure within TFIID and a central role for TAF-TAF interactions through histone-like folds in TFIID assembly. In this regard, we now show that the interaction between hTAF_II20 and hTAF_II135 is critical for TFIID assembly in vitro. By use of an earlier-described method for the assembly of human TFIID (18), several combinations of human TAFs and TBP were assembled on an immobilized HA epitope-tagged hTAF_II250 or hTAF_II20 and eluted with HA peptides. Although hTAF_II135 does assemble with the TAF_II250-TBP complex in a stoichiometric fashion (Fig. 4, lanes 2 and 3), the subsequent addition of other TAF_IIs (hTAF_II150, hTAF_II100, and hTAF_II80) resulted in the dissociation of previously bound hTAF_II135 (data not shown). In addition, hTAF_II135 failed to stably associate with a preformed complex comprised of immobilized hTAF_II250, TBP, hTAF_II80, and hTAF_II31 (Fig. 4, lane 4), indicating that such an association requires an additional TAF(s). In this regard, addition to the TFIID assembly mixture of hTAF_II20, along with the histone H3- and H4-like TAF_IIs (hTAF_II31 and hTAF_II80), greatly facilitated the incorporation of hTAF_II135 (Fig. 4, lane 6 versus lane 5) and, subsequently, hTAF_II100 (Fig. 4, lane 7). It is also possible that hTAF_II135 gets incorporated into de novo-assembled TFIID as a complex with hTAF_II20. These results show a requirement for hTAF_II20, along with other histone fold-containing TAFs (hTAF_II31 and hTAF_II80), for the stable incorporation of hTAF_II135 into assembling TFIID, which is consistent with the in vivo data obtained with the f:135 cell line (Fig. 3). Our data clearly demonstrate a new pathway for the assembly of human TFIID that stresses the importance of hTAF_II135-hTAF_II20 interactions and is thus distinct from the pathway proposed for Drosophila TFIID. In the latter case, earlier assembly studies reported a simple association of dTAF_II110 (Drosophila homologue of hTAF_II135) with a TBP-dTAF_II250 complex, followed by sequential association of dTAF_II150 and the smaller Drosophila TAFs (9).

View larger version (20K): [in this window] [in a new window]   FIG. 4.   Role of TAFII135-hTAFII20 interactions in human TFIID assembly. Partially reconstituted TFIID species were resolved by SDS-PAGE and visualized by silver staining. The complexes were assembled with purified subunits that were individually expressed in Sf9 cells via baculovirus vectors (see text). Lane 1 contained TBP alone. Complexes analyzed in other lanes contained hTBP-hTAFII250 (lane 2), hTBP-TAFII250-TAFII135 (lane 3), hTBP-TAFII250-TAFII80-TAFII31, to which added TAFII135 failed to bind in a stoichiometric manner (lane 4), hTBP-TAFII250-TAFII80-TAFII31-hTAFII20 (lane 5), hTBP-TAFII250-hTAFII135-hTAFII80-hTAFII31-hTAFII20 (lane 6), and hTBP-TAFII250-hTAFII135-hTAFII100-hTAFII80-hTAFII31-hTAFII20 (lane 7).

TAF_II135 interacts with TFIIA subunits. In light of previously described interactions of the Drosophila homologue of hTAF_II135, dTAF_II110, with general factor TFIIA and with activation domains of Sp1 (20, 46), we examined protein-protein interactions of hTAF_II135 with individual TFIIA subunits and with Sp1 activation domains A and B. The individual p35, p19, and p12 subunits of TFIIA, the p55 precursor of p35 and p19, and Sp1 domains A and B were expressed as GST fusion proteins and used for solution interaction studies with [³⁵S]methionine-radiolabeled full-length hTAF_II135. The results show that hTAF_II135 interacts with both the unprocessed TFIIA p55 and the derived p35 subunit, but not with the derived p19 subunit, and with the p12 subunit (Fig. 5A). Further studies with several deletion mutants of hTAF_II135 (Fig. 5B) revealed interactions of both TFIIA p55 (Fig. 5C) and TFIIA p12 (Fig. 5D) with the extreme C-terminal region (897 to 1083) of hTAF_II135.

View larger version (34K): [in this window] [in a new window]   FIG. 5.   Strong in vitro interactions between hTAFII135 and TFIIA subunits. (A) Equivalent amounts of full-length [35S]methionine-labeled hTAFII135 were incubated with the indicated GST derivatives (1 µg of each). After extensive washing, bound proteins were resolved by SDS-PAGE and analyzed by autoradiography. Input samples contained 10% of the amount used for binding. The arrow indicates the band corresponding to full length hTAFII135. (B) Deletion mutants of hTAFII135 that were generated, [35S]methionine labeled, and subsequently used for interaction studies (C and D). Equivalent amounts of [35S]methionine-labeled full length hTAFII135 and different deletion mutants were incubated with either GST-IIAp55 subunit (C) or GST-IIAp12 subunit (D). After washing, bound material was processed as in A.

Functional relevance of TAF_II135 interactions with TFIIA. To assess the functional significance of the observed hTAF_II135-TFIIA interactions, and because TFIIA can potentiate the binding of TBP to DNA (for reviews, see references 35 and 37), we asked whether hTAF_II135-TFIIA interactions could affect the binding activity of TBP within a minimal complex containing only TBP, hTAF_II250, and hTAF_II135. To this end we employed DNase footprinting to compare binding of TBP-hTAF_II250 and TBP-hTAF_II250-hTAF_II135 in both the absence and the presence of purified natural TFIIA. We also used TBP alone for binding experiments as a control. First, in the absence of TFIIA, limiting amounts of TBP alone bound weakly but specifically to the TATA box region of the adenovirus major late promoter (Fig. 6A, lane 3 versus lane 2). As expected, the addition of TFIIA greatly potentiated TBP binding (Fig. 6A, lane 4 versus lane 3). Second, in the absence of TFIIA, the TBP-hTAF_II250 and TBP-hTAF_II250-hTAF_II135 complexes both failed to show significant binding to either the TATA box region or the downstream region of the adenovirus major late promoter (Fig. 6A, lane 5 versus lane 2, and Fig. 6B, lanes 3 and 5 versus lane 2). This reflects the well-documented inability of TBP, when complexed to hTAF_II250 in the absence of a complete set of TAF_IIs, to efficiently bind DNA (8, 23, 27, 33). Whereas addition of TFIIA to the TBP-TAF_II250 complex only marginally increased binding to DNA (Fig. 6A, lane 6 versus lane 5, and Fig. 6B, lane 4 versus lane 3), addition of TFIIA to the TBP-hTAF_II250-hTAF_II135 complex, which alone showed no binding (Fig. 6B, lane 5), led to a strong binding of TBP to the TATA region (Fig. 6B, lane 6 versus lane 5). This clearly shows a synergy between TFIIA and hTAF_II135 that specifically relieves the inhibitory function of hTAF_II250 on TBP binding. These results are relevant to the natural TFIID, since the binding of natural TFIID on the adenovirus major late promoter is potentiated by TFIIA (data not shown). Furthermore, we have shown that a complex composed of TAF_II250, TAF_II80, and TBP failed to cooperate with TFIIA to relieve TAF_II250-mediated inhibition of TBP binding, thus demonstrating that the action of TAF_II135 and TFIIA is specific (data not shown).

View larger version (86K): [in this window] [in a new window]   FIG. 6.   hTAFII135 and TFIIA synergize to specifically relieve the hTAFII250-mediated inhibition of TBP binding. DNase I footprinting of hTBP, hTBP-hTAFII250, and hTBP-TAFII250-TAFII135. The transcribed DNA template spanning positions 91 to +85 of the adenovirus major late promoter was prepared. The Maxam-Gilbert sequencing method was used to prepare the G/A footprinting marker (lanes A1 and B1). No protein (lanes A2 and B2), hTBP (2.5 ng for lanes A2 and A3), hTBP-hTAFII250 (10 ng of TBP content for lanes A5, A6, B3, and B4), hTBP-TAFII250-TAFII135 (10 ng of TBP content for lanes B5 and B6). Purified TFIIA was added to reactions corresponding to lanes A4, A6, A7, B4, and B6.

Effects of TFIIA on basal and activated transcription in the presence of TFIID and partial TFIID complexes. To study the effects of TAF_IIs on basal and activated transcription, we compared the effects of equimolar amounts (based on TBP content) of f-TFIID and partial TFIID complexes. Partial in vitro-assembled TFIID species that contained hTAF_II135 in addition to TBP-TAF_II250 were designed to test the potential core promoter and coactivator functions of hTAF_II135. The transcription system consisted of recombinant and highly purified transcription factors from HeLa cells. The ability of this system to support both basal and activator-dependent transcription was tested simultaneously by using two templates whose correctly initiated G-less transcripts could be differentiated by their size. The activator-responsive template contained three thyroid hormone-responsive elements (TRE) upstream of the adenovirus major late promoter TATA box and natural initiator regions, and activation was mediated by TR, isolated in association with TRAP complex (14), and RXR. First, there was a higher basal activity with the TBP-TAF_II250-TAF_II135 complex compared to the TBP-TAF_II250 complex (Fig. 7A, lane 3 versus lane 1). This could be due, at least in part, to the above-described (Fig. 6) ameliorative effect of TAF_II135 on the TAF_II250-mediated repression of TBP binding to DNA, especially since these effects are more apparent in the presence of TFIIA (below). Second, basal transcription in the presence of the complete TFIID was totally repressed as a result of the strong repressive effect of the TAFs as a group on TBP function (Fig. 7A, lane 5 versus lanes 1 and 3). Thus, the complete complement of TAFs in TFIID appear to conditionally constrain the ability of TAF_II135 to reverse the TAF_II250-mediated repression of TBP function in basal transcription, but it is possible that this potential is reactivated in the presence of transcriptional activators. Third, and significantly, the absolute levels of TR-TRAP-activated transcription were slightly higher with TFIID than with the two partial complexes (Fig. 7A, lane 6 versus lanes 4 and 2). In addition, the fold activation (activated/basal transcription ratio) was much higher with TFIID than with the partial TFIID complexes, due mainly to the potent inhibitory effect of the TAFs as a group, in the complete TFIID, on TBP function in basal transcription (for quantitation, see legend to Fig. 7).

View larger version (41K): [in this window] [in a new window]   FIG. 7.   Effect of TFIIA on basal and TR-TRAP-activated transcription in the presence of TFIID and partial TFIID complexes. (A) The TFIID-dependent transcription system was supplemented with hTBP-TAFII250 (lanes 1 to 2 and 7 to 8), hTBP-TAFII250-TAFII135 (lanes 3, 4, 9, and 10), and f-TFIID (lanes 5, 6, 11, and 12). Added amounts of TFIID and partial TFIID complexes contained the same amount of TBP (4 ng). Transcription was tested in the absence (lanes 1, 3, 5, 7, 9, and 11) or presence (lanes 2, 4, 6, 8, 10, and 12) of 80 ng of TR-TRAP complex (12 ng of TR). Recombinant TFIIA was added to lanes 7 to 12. The TR-responsive template was TRE3ML53 (50 ng), and the control template consisted of 20 ng of pML200, which contains the major late core promoter. Relative transcriptional activities (quantitated by phosphorimager analysis) from the TR-responsive template were 1 (lanes 1, 5, and 11), 2 (lane 3), 2.4 (lane 7), 4.6 (lane 9), 17 (lane 2), 24 (lane 4), 28 (lanes 8 and 10), 30 (lane 6), and 56 (lane 12). (B) Transcription was performed as in A except that f-TFIID was added to lanes 1, 2, 5, and 6 and TBP was added to lanes 3, 4, 7, and 8. Recombinant TFIIA was added to lanes 5 to 8. Relative transcriptional activities from the TR-responsive template were 1 (lane 3), 1.3 (lanes 1 and 5), 3 (lane 7), 11 (lane 8), 14 (lane 4), 42 (lane 2), and 56 (lane 6). The lower band, below the arrow indicating pML200, is a transcript generated from the activator-responsive template.

Functional synergism between the TRAP complex and TAF_IIs. Since the above experiments concerned activated transcription in the presence of the activator TR and the interacting TRAP complex, we next asked whether the TRAP complex, which has been shown to mediate the function of many activators in concert with other positive cofactors such as PC4 (reviewed in reference 28), would exhibit synergistic or redundant functions with the TAFs in the presence of a different activator, fGAL4p65. fGAL4p65 consists of the DNA-binding domain (amino acids 1 to 94) of S. cerevisiae GAL4 fused to the potent C-terminal acidic activation domain of the NF-B p65 subunit. First, to study the effects of TAF_IIs on basal and activated transcription, we compared the effects of equimolar amounts (based on TBP content) of TBP and f-TFIID on both basal and fGAL4p65-activated transcription in the assay system described above. The test template contained five GAL4 binding sites upstream of the adenovirus E1b TATA and natural Inr regions. Basal transcription was assayed on a template containing only the adenovirus major late core promoter sequence and in a system reconstituted with general transcription factors and the general coactivator. In this assay, and in the absence of TRAPs, fGAL4-p65 strongly activated transcription in the presence of either natural TFIID or TBP (Fig. 8, lane 2 versus lane 1 and lane 6 versus lane 5, respectively; for quantitation, see figure legend). While absolute levels of activated transcription on the E1b promoter were threefold higher with TBP than with TFIID (Fig. 8, lane 6 versus lane 2), basal activity was significantly less with TFIID than with TBP alone (Fig. 8, lane 5 versus lane 1; see figure legend). This leads to a fold activation (activation/basal transcription ratio) that is actually higher for TFIID than for TBP alone. It is important to note that this high level of induction in the presence of TFIID is due to the potent inhibitory effect of TAFs as a group, within the complete TFIID, on TBP basal activity and the partial reversal of these effects by the activator (18).

View larger version (31K): [in this window] [in a new window]   FIG. 8.   Synergy of TAFIIs with the TRAP complex. The TFIID-dependent transcription system was supplemented with either f-TFIID (lanes 1 to 4) or hTBP (lanes 5 to 8). TRAP complex was added to lanes 3, 4, 7, and 8. Transcription was tested in the absence (lanes 1, 3, 5, and 7) or presence (lanes 2, 4, 6, and 8) of 20 ng of purified fGAL-p65. Added amounts of f-TFIID and TBP alone contained the same amount of TBP (4 ng). The test template contains the adenovirus E1b TATA and initiator regions and consisted of 50 ng of pG5E1b template. The control template consisted of 20 ng of pML53 and contains the major late core promoter sequence. Relative transcriptional activities from the fGAL4p65-responsive template were: 1 (lanes 5 and 7), 3 (lane 2), 10 (lanes 6 and 8), and 14 (lane 4). Basal levels of transcription with TFIID, either with or without TRAPs (lanes 3 and 1, respectively), were too low to be determined, in contrast to the low but measurable levels of basal transcription with TBP (lanes 5 and 7). The lower band, below the arrow indicating pML53, is a transcript generated from the activator-responsive template.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The TAF_II subunits of TFIID have been implicated both as targets for gene-specific activators (reviewed in references 7 and 43) and as modulators (negative and positive) of TFIID binding to diverse core promoter elements (for a review, see reference 38). Although not generally essential for basal transcription directed by TATA elements in core promoters, TAFs are essential for the function of other core promoter elements (initiator and downstream promoter element) either alone or in conjunction with the TATA elements (reviewed in references 17 and 38). As part of our ongoing effort to understand the assembly, structure, and function of human TFIID, we have focused, subsequent to cognate cDNA cloning and expression, on structure-function studies of hTAF_II135. We report both extended and novel functions of hTAF_II135 that include a critical role, dependent upon histone fold interactions, in TFIID assembly; a synergistic interaction with TFIIA that relieves the well-documented hTAF_II250-mediated inhibition of TBP binding and function (8, 23, 27, 33); and contributions to a functional synergy between TAFs and the human TRAP/Mediator complex in transcriptional activation. These findings support an increasing appreciation of TAFs as multifunctional components and TFIID as a dynamic complex subject to a variety of internal and external controls.

hTAF_II135-hTAF_II20 interactions, through the histone folds, are critical for human TFIID assembly. Taken together, the interaction and mapping data obtained with the baculovirus-expressed proteins and the biochemical characterization of the immunopurified TFIID from the Flag-tagged hTAF_II135 cell line show that hTAF_II135 and hTAF_II20 interact, through histone-like folds (16), to form a complex that coexists in TFIID. The multimeric nature of this complex, presumably a heterotetramer, within TFIID is suggested by the presence of at least two molecules each of hTAF_II20 (21) and hTAF_II135 (this report) in a single TFIID species. This hTAF_II135-hTAF_II20 complex may be involved, together with the H4-like hTAF_II80 and the H3-like hTAF_II31, in the formation of an octamer-like substructure, as previously proposed for the human TFIID (21). Other human TAFs, such as TAF_II28 (H3-like) and TAF_II18 (H4-like), have also been shown to contain histone folds (2) and to interact strongly with one another (29).

hTAF_II135 interacts with TFIIA to specifically relieve hTAF_II250-mediated inhibition of TBP binding and function. We have shown that hTAF_II135 interacts with two (the largest and the smallest) of the three subunits of TFIIA. These data confirm and extend an earlier observation of an interaction of dTAF_II110, the Drosophila homologue of hTAF_II135, with the large subunit of Drosophila TFIIA (46). A more detailed analysis with hTAF_II135 deletion mutants further shows that the human TFIIA subunit interactions are mediated through the C-terminal portion of hTAF_II135. Most importantly, from a functional standpoint, an analysis with highly purified TFIIA and in vitro-assembled TFIID subspecies has shown that the interaction of TFIIA with hTAF_II135 has a critical role in relieving hTAF_II250-mediated repression. Thus, these studies with partial TFIID species have revealed an internal mechanism involving hTAF_II135 that could be used in a structurally dynamic natural TFIID to facilitate the transition from TAF_II-mediated repression to activation. This leads to speculation that hTAF_II135 may be a direct or indirect (e.g., via TFIIA) target for factors that use it to relieve repression during activated transcription. Although TFIIA was shown to have a role in countering inhibitory interactions of the amino terminus of the S. cerevisiae TAF_II145 (the S. cerevisiae homologue of the human hTAF_II250) with TBP (24), our observations are the first to show a synergism between TFIIA and a specific TAF_II subunit in relieving hTAF_II250-mediated repression. They further demonstrate new core promoter functions for both TFIIA and hTAF_II135.

TAF_II135 core promoter function. In the context of TFIID, TAF_IIs appear to have coactivator functions mainly on the basis of in vitro studies with metazoan factors (for reviews, see references 38 and 43), whereas core promoter-selective functions are evident from in vivo studies in S. cerevisiae (for a review, see reference 17) and from both in vivo and in vitro studies in metazoans (for review, see reference 38). The transcriptional requirement of TAF_IIs was assessed originally in purified (reconstituted) cell-free systems in which TBP alone efficiently supported basal but not activator-mediated transcription for several activators. Since TFIID supported both basal and activated transcription in vitro, one or more TAF_IIs appeared to have a critical coactivator function under the conditions employed. These results, the demonstrations of in vitro interactions between activation domains and isolated TAF_IIs, and studies with partial (reconstituted) TFIID complexes led to the proposal, consistent with earlier demonstrations of qualitative and quantitative effects of activators on TFIID binding, that TAF_IIs are direct targets for activators (reviewed in references 7 and 43). The prototype for this paradigm was the activator Sp1 with its proposed "obligate" direct target, dTAF_II110, the Drosophila homologue of hTAF_II135 (9, 20).

Synergy of TAF_IIs with TRAP/Mediator complex. The Mediator complex has emerged as a major conduit for communication between gene-specific regulatory factors and the general transcription machinery in both S. cerevisiae (reviewed in references 26 and 32) and human (reviewed in reference 28). Given the observation that individual TAF_IIs are not universally required for transcription activation in either S. cerevisiae or metazoans, as well as the function of some activators in the absence of TFIID-specific TAF_IIs in certain cell-free systems from both metazoans and S. cerevisiae, a question of increasing importance is whether the TAF_IIs may have either redundant or synergistic effects with other prominent coactivators, such as the Mediator.