c-Jun Binds the N Terminus of Human TAFII250 to Derepress RNA Polymerase II Transcription in Vitro *

c-Jun is an oncoprotein that activates transcription of many genes involved in cell growth and proliferation. We studied the mechanism of transcriptional activation by human c-Jun in a human RNA polymerase II transcription system composed of highly purified recombinant and native transcription factors. Transcriptional activation by c-Jun depends on the TATA-binding protein (TBP)-associated factor (TAF) subunits of transcription factor IID (TFIID). Protein-protein interaction assays revealed that c-Jun binds with high specificity to the largest subunit of human TFIID, TAFII250. The region of TAFII250 bound by c-Jun lies in the N-terminal 163 amino acids. This same region of TAFII250 binds to TBP and represses its interaction with TATA boxes, thereby decreasing DNA binding by TFIID. We hypothesized that c-Jun is capable of derepressing the effect of the TAFII250 N terminus on TFIID-driven transcription. In support of this hypothesis, we found that c-Jun increased levels of TFIID-driven transcription in vitro when added at high concentrations to a DNA template lacking activator protein 1 (AP-1) sites. Moreover, c-Jun blocked the repression of TBP DNA binding caused by the N terminus of TAFII250. In addition to revealing a mechanism by which c-Jun activates transcription, our studies provide the first evidence that an activator can bind directly to the N terminus of TAFII250 to derepress RNA polymerase II transcription in vitro.

mediator complexes (for review, see Refs. 1 and 2) as well as by recruiting chromatin modifying complexes (for review, see Ref. 3).
Whereas unique combinations of activators and coactivators function at different promoters, RNA polymerase II and the general transcription factors (TFIIA, -B, -D, -E, -F, and -H) are thought to function at most mRNA-encoding genes (for review, see Ref. 4). The TFIID complex plays a central role in core promoter recognition and the assembly of preinitiation complexes regardless of the manner by which the remaining basal transcriptional components are recruited, either via a stepwise pathway or as a holoenzyme (5). TFIID is composed of the TATA-binding protein (TBP) and multiple TBP-associated factors (TAFs) whose sizes range from 18 to 250 kDa (for review, see Ref. 6). Subunits of TFIID contact core promoter elements including the TATA box, initiator, and downstream promoter element (7)(8)(9)(10).
At some promoters, the recruitment of TFIID limits the overall level of transcription in vitro. Efficient recruitment of TFIID can be achieved by upstream activators that interact with TFIID components (11)(12)(13). In addition, activators can elicit a conformational change in the TFIID-TFIIA-promoter complex, as observed by DNase I footprinting (14,15). Although direct interactions between TBP and activators have been noted, the primary functional targets of activators within TFIID are the TAF subunits. TAFs were coined coactivators in recognition of their requirement for activated transcription in some in vitro systems (16,17). Recently, the ability of TAFs to adopt multifunctional roles as components of distinct complexes including TFIID, Spt-Ada-Gen5-acetyltransferase (SAGA), p300/CBP-associated factor (PCAF), and TBP-free TAF II complex (TFTC) has been demonstrated (for review, see Ref. 1). The validity of the coactivator function of TAFs has been firmly established from examples in yeast, Drosophila, and human systems.
Although TAFs can function as coactivators, regulated transcription of some promoters is not dependent on TAFs in vitro and in vivo (18 -22). The mediator and other complexes are also targets of transcriptional activators and serve to coactivate transcription (for review, see Ref. 2). In the context of chromatin, nucleosome-remodeling complexes and histone acetyltransferases such as cAMP-response element-binding protein (CREB)-binding protein (CBP) function to coactivate transcription (3). It is likely that activators can stimulate transcription by many different mechanisms depending upon the conditions and the context of the promoter at which the activator binds.
One mechanism by which activators can function is by re-lieving the effects of factors that repress basal transcription. Although this is certainly true of activators in the context of chromatin, mechanisms of transcriptional repression and derepression have also been observed in the absence of chromatin. For example, the binding of TFIID to core promoters is internally repressed by the largest subunit of TFIID, TAF II 250 (23). Most of the work on the mechanism of this repression has been carried out with dTAF II 230 and yTAF II 145, the Drosophila and yeast homologs of human TAF II 250. Repression is mediated by a small domain (ϳ150 amino acids) in the N termini of these proteins that binds directly to TBP and inhibits its ability to bind TATA boxes (23,24). Furthermore, the N-terminal domain of dTAF II 230 can repress TBP-driven transcription when added in excess of TBP in an in vitro transcription system (23). Mutational analysis identified the concave DNA binding surface of TBP as a binding site of the N terminus of dTAF II 230 (25). In NMR studies, amino acids 11-77 of dTAF II 230 appeared to mimic DNA by binding to the DNA binding cleft of TBP (26). A second region of dTAF II 230 (amino acids 118 -143) interacts with the convex surface of TBP and further stabilizes the dTAF II 230-TBP interaction (27). The importance of TAF II 250 repression to transcriptional regulation in cells and the potential for targeting the N terminus of TAF II 250 as a means to derepress transcription are currently under study. TFIIA, which aids TFIID in binding core promoters, can prevent TAF II 250 from inhibiting TBP-DNA binding in vitro (28).
In addition, yTAF II 145 and yTFIIA bind competitively to TBP (24). The acidic activation domain of VP16 also binds the underside of TBP and can block the interaction of dTAF II 230 with TBP (25,29). Activators have not, however, been shown to bind directly to the N terminus of TAF II 250 to remove its repressive effects on TBP-DNA binding. AP-1 transcription factors of the Jun and Fos families activate many genes including those involved in cell growth and proliferation (for review, see Ref. 30). AP-1 proteins each contain an extended ␣-helical dimerization/DNA binding domain of ϳ62 amino acids that consists of a leucine zipper found C-terminal of a basic region (31). Dimers bind DNA through sequence-specific contacts between amino acids in the basic regions and the major groove of the DNA helix (32). Members of the Jun family can form homo-and heterodimers. Members of the Fos family cannot form homo-or heterodimers among themselves but can heterodimerize with Jun proteins. Although there is high sequence homology in the basic leucine zipper domains of AP-1 proteins, there is much less homology in other regions, which contain activation domains. The activation domains of c-Jun and c-Fos, which are essential for transcriptional activation in vitro, have been characterized functionally (33-35) but not structurally and are known to bind other proteins such as the coactivator CBP (36). Despite a wealth of studies, the mechanisms by which c-Jun and c-Fos stimulate transcription are not well understood.
Here we have investigated the mechanism by which human c-Jun activates transcription using a highly purified human RNA polymerase II transcription system. Activation of transcription by recombinant c-Jun was found to be dependent on TAFs. Interactions between c-Jun and TAFs were identified and characterized using protein-protein affinity chromatography, which revealed that c-Jun binds to the N terminus of human TAF II 250. The function of c-Jun in derepressing transcription driven by TFIID and the ability of c-Jun to block TAF II 250 repression of TBP transcription and DNA binding were characterized in vitro. Our results reveal a unique interaction between c-Jun and TAF II 250 that results in derepression of TFIID-driven transcription in vitro. These studies lead to a novel model for activation in which c-Jun interacts with TAF II 250 to relieve repression of TFIID-DNA binding, thereby contributing to transcriptional activation.

EXPERIMENTAL PROCEDURES
Plasmids-pET-Jun, a plasmid for expressing full-length human c-Jun, was a gift from T. Hoey (Tularik, Inc.). In creating this plasmid an NcoI site was generated that changed the fourth base pair in the c-Jun-coding region from AT to GC, resulting in a point mutation in the second amino acid of c-Jun (Thr to Ala). pGEX-Jun was created by digesting pET-Jun with NcoI and EcoRV and subcloning the c-Jun cDNA into the NcoI and StuI sites of pGEX-2TKN (a kind gift of S. Ruppert and R. Tjian).
Plasmid p(AP-1) 5 -E1b-CAT was created by inserting five direct copies of a double-stranded oligonucleotide containing the AP-1 element from the human metallothionein IIA promoter (Ϫ105 to Ϫ96) into the XbaI site of plasmid E1b-CAT (38). Plasmid p(AP-1) 5 -E1b-G-less, used as template DNA in the in vitro transcription assays, was created by inserting a 377-base pair G-less cassette (excised from p(GAL4) 5 -E1b-G-less (a kind gift from M. G. Peterson and R. Tjian) with SacI) into the SacI site of p(AP-1) 5 -E1b-CAT. Plasmid p(GAL4) 5 -AdMLP-G-less (39), which contains five GAL4 sites upstream of the adenovirus major late core promoter (AdMLP; Ϫ53 to ϩ10) and a 380-base pair G-less cassette, was generated by subcloning a HindIII-EcoRI fragment isolated from plasmid G5-MLP-G-less (a kind gift from Mike Carey) into the HindIII and EcoRI site of pBluescript-KS(ϩ).

Expression and Purification of Recombinant c-Jun and GST Fusion Proteins-Recombinant full-length human c-Jun was expressed in
Escherichia coli BL21:DE3. Cultures (500 ml) were grown in LB containing 0.3 mM ampicillin at 37°C until the A 600 nm was 0.4. Expression was induced with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside for 2 h at 37°C. Cells were harvested and resuspended in 10 ml of a solution containing 20 mM Tris (pH 7.9), 1 mM EDTA, 100 mM NaCl, 1 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride, and sonicated 4 times for 15 s. Samples were centrifuged for 30 min in a Sorval SS34 rotor at 18,000 rpm and 4°C. Precipitated material (inclusion bodies) containing c-Jun was resuspended in 10 ml of 5 mM DTT and sonicated 2 times for 30 s. Samples were centrifuged for 10 min in a Sorval SS34 rotor at 15,000 rpm and 4°C. Insoluble material was washed three more times by resuspending in 10 ml of 5 mM DTT, followed by centrifugation. Pellets from the final wash were resuspended in 5 ml of buffer A (20 mM Tris (pH 7.9), 1 mM EDTA, 5 mM DTT, and 7 M urea) containing 0.1 M NaCl. Soluble protein was loaded on an 8-ml column packed with sulfopropyl-Sepharose fast flow (Amersham Pharmacia Biotech) and washed with three column volumes of buffer A containing 0.1 M NaCl followed by three column volumes of buffer A containing 0.2 M NaCl. c-Jun was then eluted with buffer A containing 0.3 M NaCl and subjected to three sequential dialyses in buffer B (20 mM Tris (pH 7.9), 0.1 mM EDTA, 10% glycerol, 5 mM DTT) containing the following additions: 1) 1 M urea and 1 M NaCl; 2) 1 M NaCl; 3) 0.1 M NaCl. After dialysis, the purified c-Jun was separated into aliquots and stored at Ϫ80°C.
GST-Jun was expressed in E. coli strain XA90. Cultures were grown in LB containing 0.3 mM ampicillin at 37°C until the A 600 nm was 0.4. Expression was induced with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h at 37°C. Cells were harvested and resuspended in a solution containing 20 mM Tris (pH 7.9), 20% glycerol, 1 mM EDTA, 5 mM MgCl 2 , 0.1 M NaCl, 5 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride and sonicated 4 times for 15 s. Insoluble material was washed with 5 mM DTT as described for c-Jun above. Pellets from the final wash were resuspended in a solution containing 20 mM Tris (pH 7.9), 1 mM EDTA, 5 mM DTT, and 6 M guanidine-HCl, and protein was extracted by overnight nutation at 4°C. After centrifugation to remove the remaining insoluble material, extracts were diluted to give final concentrations of 200 ng/l of GST-Jun. Samples were subjected to three sequential dialyses in buffer B containing 0.1 M NaCl and guanidine-HCl at 2, 1, and 0.5 M. The guanidine-HCl was then slowly lowered to 50 mM by pumping (over 10 h) buffer B containing 0.1 M NaCl into the third dialysis listed above. After dialysis, samples containing GST-Jun were separated into aliquots and stored at Ϫ80°C. Just before use in proteinprotein interaction assays, GST-Jun was further purified by incubation with glutathione-Sepharose for 1 h at 4°C followed by two washes with 10 column volumes of buffer B containing 1 M NaCl and 0.1% Nonidet P-40 and three washes with the appropriate protein-protein interaction assay buffer (see below and figure legends). The affinity resin contained ϳ1 mg of protein/ml of beads.
GST-TAF II 250-(1-163) and GST were expressed in E. coli strain XA90 as described for GST-Jun above. Harvested cells were resuspended in 1/50 culture volume of lysis buffer containing 20 mM Tris (pH 7.9), 20% glycerol, 1 mM EDTA, 0.1 M NaCl, 1 mM DTT, and 0.2 mM phenylmethylsulfonyl fluoride. Samples were sonicated 4 times for 15 s and centrifuged for 30 min in a Sorval SS34 rotor at 18,000 rpm and 4°C. Supernatants from the centrifugation were incubated with glutathione-Sepharose at 4°C for 1 h with constant nutation. Resin was washed two times with 10 column volumes of buffer B containing 1 M NaCl and 0.1% Nonidet P-40. For use in protein-protein interaction assays, the resin was washed three times with ten column volumes of the appropriate protein-protein interaction assay buffer (see below and figure legends). For elution, resin was washed five times with 10 column volumes of buffer C (100 mM Tris (pH 7.9), 120 mM NaCl). Protein was eluted by incubating resin at 4°C for 10 min in 1 column volume of buffer C containing 20 mM glutathione (reduced) with constant nutation. Eluates were dialyzed overnight at 4°C in a buffer containing 20 mM Tris (pH 7.9), 10% glycerol, 1 mM EDTA, 0.1 M NaCl, and 1 mM DTT. Samples were separated into aliquots and stored at Ϫ80°C.
Protein-Protein Interaction Assays-Each reaction contained 1 g of GST-Jun or GST immobilized on 10 l of glutathione-Sepharose beads (Amersham Pharmacia Biotech). Immobilized proteins and extracts containing target proteins were incubated with micrococcal nuclease at 30°C for 10 min. Target proteins were then incubated with GST or GST-Jun beads at 4°C for 2 h with constant nutation in TGEMD buffer (20 mM Tris (pH 7.9), 10% glycerol, 1 mM EDTA, 5 mM MgCl 2 , and 1 mM DTT) containing 0.1% Nonidet P-40, 0.2 mM phenylmethylsulfonyl fluoride, and 0.1 M NaCl. After incubation, beads were washed at least 3 times with 10 -15 bead volumes of buffer. Protein was eluted from the beads in SDS sample buffer and resolved by SDS-PAGE.
In Vitro Transcription-General transcription factors (TBP, TFII-B, -E, -F, and -H) and RNA polymerase II were expressed and purified as previously described (40). Human TFIID was immunopurified (39). Recombinant human TFIIA was prepared as previously described (41,42). The amounts of general factors and buffer conditions used for transcription were as previously described (39). The amounts of the TFIID (10 ng) and TBP (0.5 ng) used in transcription reactions contained equal moles of TBP, as determined by quantitative Western analysis with polyclonal antibody directed against TBP (39).
To study transcriptional activation (see Fig. 1B) c-Jun was incubated with p(AP-1) 5 -E1b-G-less for 5 min on ice. After the addition of general transcription factors, reactions (20 l) were transferred to 30°C for 25 min before adding nucleoside triphosphates. RNA synthesis was allowed to proceed for 20 min at 30°C. To study derepression (see Fig. 4), c-Jun was incubated with TBP or TFIID in 13-l reactions for 10 min on ice. p(GAL4) 5 -AdMLP-G-less template DNA and the remaining general transcription factors where then added to reactions in a final volume of 20 l.
To study transcriptional repression (see Fig. 5A) GST-TAF II 250-(1-163) and GST were added to transcription reactions containing p(GAL4) 5 -AdMLP-G-less template DNA before the addition of TBP, TFIIB, TFIIF, and RNA polymerase II. In Fig. 5B, c-Jun was incubated with GST-TAF II 250-(1-163) (lanes 3 and 4) in 3 l for 13 min on ice. p(GAL4) 5 -AdMLP-G-less template DNA was then added, followed immediately by the addition of the general transcription factors to a final volume of 20 l.
DNase I Footprinting-DNase I footprinting reactions were performed with a 183-base pair DNA fragment containing the AdMLP (Ϫ53 to ϩ33) 32 P-labeled on the 5Ј-end of the nontemplate strand. The buffer conditions were identical to those used for transcription (before the addition of NTPs). When present, GST-TAF II 250-(1-163) and c-Jun were incubated together in a reaction volume of 2 l for 10 min on ice. These mixtures were then added to reactions containing promoter DNA followed immediately by the addition of TBP at a final concentration of FIG. 2. c-Jun binds TAF II 250 with high specificity. A, protein-protein affinity chromatography was used to identify TAFs that interact with GST-Jun. 35 S-Labeled human TAF II 250 (lanes 1-3), Drosophila TAF II 40 (lanes 4 -6), Drosophila TAF II 30␣ (lanes 7-9), and Drosophila TAF II 30␤ (lanes 10 -12) were expressed by in vitro translation and tested for interaction with immobilized GST-Jun or GST. 10% of the input 35 S-labeled proteins were included in lanes 1, 4, 7, and 10. B, recombinant human TAF II 250 was specifically purified from cell lysates using GST-Jun affinity resin. HA-TAF II 250 was expressed in insect cells using a recombinant baculovirus. Portions of cell lysate containing TAF II 250 were incubated with GST-Jun and control GST resins. After extensive washing, protein bound to the affinity resins was resolved by SDS-PAGE and stained with Coomassie Brilliant Blue. An aliquot of insect cell lysate equal to 100% of the input was included in lane 1.
containing 0.15 units/l of DNase I (Promega) and 10 mM CaCl 2 was added to each reaction. Reactions were stopped, and DNA products were analyzed as described (39).

Activation of Transcription by Human c-Jun in a Highly
Purified in Vitro System Requires TAFs-To study the mechanism of c-Jun transcriptional activation, we developed a human RNA polymerase II transcription system reconstituted from highly purified recombinant and native general transcription factors. This system is useful for studying the roles of individual subunits of the general RNA polymerase II transcription machinery, including the TFIID TAFs, in basal and activated transcription. Recombinant human c-Jun was expressed in E. coli and purified to near homogeneity (Fig. 1A). To test the activity of c-Jun for activating transcription in vitro we created a DNA template, p(AP-1) 5 -E1b-G-less, containing five AP-1 elements upstream of the adenovirus E1b TATA box and a 377-base pair G-less cassette. We detected basal transcription from this template in vitro (Fig. 1B) in reactions reconstituted with equal molar concentrations of either immunopurified native TFIID (lane 1) or recombinant TBP (lane 3). Interestingly, reactions reconstituted with TBP reproducibly showed higher levels of basal transcription than those containing TFIID even though equivalent molar concentrations of TBP and TFIID were used (39). This is likely to be due to activities present in the TFIID complex that inhibit its binding to DNA, such as the N terminus of TAF II 250 (23). c-Jun (30 ng) activated transcription ϳ10-fold in reactions reconstituted with TFIID (Fig. 1B,  compare lanes 1 and 2) but did not stimulate transcription in reactions reconstituted with TBP (Fig. 1B, compare lanes 3 and 4). Thus, in a highly purified in vitro transcription system, TAFs are required for transcriptional activation by human c-Jun.
c-Jun Binds TAF II 250 with High Specificity-The finding that TAFs were required to mediate transcriptional activation by c-Jun in our in vitro transcription system led us to hypothesize that c-Jun interacted with one or more of the TAFs. To identify interactions between c-Jun and TAFs, we performed protein-protein affinity chromatography utilizing immobilized GST-Jun and in vitro translated 35 S-labeled TAFs. TAFs were incubated with GST-Jun or control GST resin, and after extensive washing, proteins bound to the resins were analyzed by SDS-PAGE and autoradiography. These studies revealed that TAF II 250 interacted with immobilized GST-Jun but not control GST ( Fig. 2A, lanes 1-3). Other TAFs, for example dTAF II 40, dTAF II 30␣, and dTAF II 30␤, did not interact with GST-Jun ( Fig. 2A, lanes 4 -12).
To further characterize the affinity and specificity of the c-Jun-TAF II 250 interaction, we utilized extracts of insect cells containing HA-tagged TAF II 250 that was expressed from a recombinant baculovirus. Insect cell extracts were incubated with immobilized GST-Jun and control GST. After extensive washing, bound protein was resolved by SDS-PAGE and visualized by staining with Coomassie Brilliant Blue. As shown in Fig. 2B, the GST-Jun resin specifically purified recombinant HA-TAF II 250 from the crude insect cell extract. The numerous and abundant insect cell proteins in the extract were not retained on the GST-Jun resin.
c-Jun Binds the N-terminal Region of TAF II 250 That Is Involved in Transcriptional Repression-To gain insight into the role of the c-Jun-TAF II 250 interaction in transcriptional activation, we delimited the region of TAF II 250 that is bound by c-Jun. Two C-terminal deletions of TAF II 250 were expressed by in vitro translation and tested for interaction with immobilized GST-Jun: TAF II 250-(1-560) and TAF II 250-(1-163). As shown in Fig. 3A, TAF II 250-(1-560) and TAF II 250-(1-163) each bound GST-Jun at a level comparable with that observed for fulllength TAF II 250 (amino acids 1-1893).
As an additional test of the ability of the N-terminal 163 amino acids of TAF II 250 to bind c-Jun, we expressed and purified a GST-TAF II 250-(1-163) fusion protein and incubated the immobilized GST-TAF II 250-(1-163) with purified c-Jun. The silver-stained SDS gel in Fig. 3B shows GST-TAF II 250-(1-163) bound c-Jun (lane 5). c-Jun was not retained on control GST resin (lane 4). Thus, purified c-Jun can bind directly to GST-TAF II 250- (1-163).
These results revealed that the region of TAF II 250 bound by c-Jun is the N-terminal domain, which is known to inhibit TBP-DNA binding and transcription. Based on this finding, we hypothesized that c-Jun binds the N terminus of TAF II 250 and inhibits its repression of TBP-DNA binding, thereby contributing to transcriptional activation.
c-Jun Derepresses Basal Transcription Directed by TFIID-It is possible that the lower level of basal transcription observed with TFIID compared with recombinant TBP (see Fig.  1) was due to the N-terminal region of TAF II 250 interacting with the DNA binding surface of TBP to repress the binding of TBP to TATA DNA. In this case, c-Jun may bind directly to the N terminus of TAF II 250 in TFIID to derepress transcription. To test this we used a DNA plasmid that lacks known AP-1 sites, p(GAL4) 5 -AdMLP-G-less, which contains the adenovirus major late core promoter (Ϫ53 to ϩ10) upstream of a 380-base pair G-less cassette. TFIID and TBP were incubated separately with increasing concentrations of c-Jun and then added to transcription reactions with p(GAL4) 5 -AdMLP-G-less. As shown in Fig. 4, lanes 7-10, c-Jun stimulated transcription when preincubated with TFIID. The level of transcription with TFIID approached that observed with TBP at the highest concentrations of c-Jun. c-Jun did not increase the level of transcription when preincubated with TBP (Fig. 4, lanes 2-5).
c-Jun Overcomes Repression of TBP-DNA Binding and Transcription Caused by the N Terminus of TAF II 250 -To demonstrate that the N terminus of TAF II 250 can repress in vitro transcription and to directly test whether c-Jun can inhibit TAF II 250-mediated transcriptional repression, we added purified GST-TAF II 250-(1-163) to transcription reactions. The ability of TAF II 250-(1-163) to repress basal transcription was determined by comparing the effects of GST-TAF II 250-(1-163) and GST on basal transcription in a minimal in vitro transcription system consisting of TBP, TFIIB, TFIIF, RNA polymerase II, and negatively supercoiled p(GAL4) 5 -AdMLP-G-less template DNA. As shown in Fig. 5A, the addition of GST-TAF II 250-(1-163) repressed basal transcription by more than 95%, whereas GST did not affect transcription when titrated over the same concentration range. Thus, the N terminus of TAF II 250 can repress basal transcription in a highly purified minimal transcription system. We next tested the ability of c-Jun to block transcriptional repression caused by GST-TAF II 250-(1-163). c-Jun was preincubated with GST-TAF II 250-(1-163) before adding the p(GAL4) 5 -AdMLP-G-less template (lacking known AP-1 sites) and the remaining transcription factors. c-Jun overcame the repression of basal transcription observed with GST-TAF II 250-(1-163) (Fig. 5B, compare lanes 3 and 4 with lane 2). These data demonstrate that c-Jun can bind the N terminus of TAF II 250 to block its ability to repress basal transcription.
It was possible that the transcriptional effects presented above resulted from c-Jun binding the N terminus of TAF II 250 to inhibit its repression of TATA box binding by TBP. To test this, we performed DNase I footprinting with TBP using a DNA fragment containing the adenovirus major late core promoter but lacking AP-1 sites. GST-TAF II 250-(1-163) and c-Jun were added to determine their effects on TBP-DNA binding. Fig. 5C shows plots of relative intensities (y axis) versus position on the gel (x axis) for the region spanning the TATA box (left to right represents top to bottom on the gel). TBP alone protected bands in the TATA box region from DNase I digestion by greater than 90%. GST-TAF II 250-(1-163) almost completely repressed TBP-DNA binding. In the presence of GST-TAF II 250- (1-163), the intensities of bands in the TATA box region were decreased by less than 10% from the control reaction performed in the absence of transcription factors. Incubation of c-Jun with GST-TAF II 250-(1-163) blocked the repression of TBP binding, resulting in ϳ70% of the maximal protection in the TATA box region that was observed with TBP alone. Since the footprinting experiments were performed with recombinant TBP, GST-TAF II 250-(1-163), and c-Jun, we can rule out a requirement for other general transcription factors in repression by GST-TAF II 250-(1-163) and inhibition of this repression by c-Jun. Taken together, the transcription and DNase I footprinting experiments strongly support a model in which c-Jun binding to the N terminus of TAF II 250 inhibits repression of TBP-DNA binding and transcription initiation. DISCUSSION Here we have elucidated one mechanism by which c-Jun activates transcription. c-Jun binds the N terminus of TAF II 250 to derepress RNA polymerase II transcription. Not only did c-Jun derepress TFIID-directed transcription in vitro, but c-Jun was also capable of blocking repression of TBP-DNA binding by the N terminus of TAF II 250. Our results are the first to characterize a mechanism by which c-Jun activates RNA polymerase II transcription in a cell-free system. In addition, we provide the first evidence that activators can directly bind the N terminus of TAF II 250 to derepress transcription in vitro.
Derepression is emerging as a general mechanism for controlling levels of transcription in eukaryotic cells. This is not surprising since derepression is a prevalent mechanism for controlling gene expression in prokaryotes (43). The most widespread example of derepression in eukaryotes is thought to be the alteration of chromatin structure by histone acetyltransferases and ATP-dependent chromatin remodeling factors (3). In small nuclear RNA transcription the snRNA activating protein complex (SNAPc) has a built-in DNA binding damper that is deactivated by the transcriptional activator Oct-1 (44). Another general mechanism of transcriptional repression in eukaryotic cells results from the interaction of the N terminus of the largest TFIID subunit to the concave surface of TBP to prevent binding of TFIID to TATA boxes (23,25). The finding that c-Jun binds the N terminus of TAF II 250 to derepress TFIID-driven transcription in vitro further supports the generality of derepression as a means of regulating eukaryotic transcription. We are left wondering whether other transcriptional activators also bind the N terminus of TAF II 250 to derepress transcription. We note that significant stimulation of transcription from the p(GAL4) 5 -AdMLP-G-less template (which lacks known AP-1 sites) required 400 ng or more of c-Jun. These amounts are far greater than the amounts of transcriptional activators typically used in in vitro transcription experiments. For example, we observe significant activation from p(AP-1) 5 -E1b-G-less with 15 ng of c-Jun. With this low amount of c-Jun we do not observe activation from any template that does not contain known AP-1 sites. Moreover, in early c-Jun activation experiments, we did not titrate c-Jun to high enough levels to observe AP-1 site-independent derepression. Perhaps other transcriptional activators that bind the N terminus of TAF II 250 to derepress transcription will be revealed through a combination of protein-protein interaction assays (with activators and the N terminus of TAF II 250) and in vitro transcription experiments aimed at testing for site-independent transcriptional stimulation in a TFIID-dependent transcription system.
Other models for derepression of TFIID binding have been studied. For example, it has been shown that VP16 and dTAF II 230 competitively bind the concave surface of TBP (25). A point mutation in TBP (L114K) that disrupted the TBP-VP16 interaction prevented VP16 activation in vivo. The same point mutation in TBP also disrupted interaction with dTAF II 230. In addition, when the N terminus of yTAF II 145 was replaced with an acidic activation domain, the fusion protein autoinhibited TFIID in yeast (29). A two-step "hand-off" model has been proposed to explain the function of acidic activation domains in derepressing DNA binding by TFIID (25,29). In the first step, the acidic activation domain replaces the N terminus of dTAF II 230/yTAF II 145 by binding the underside of TBP. In the second step, TBP binds to the TATA box when the acidic activation domain releases its hold on TBP. This proposed mechanism of relieving the internal repression of TFIID binding is quite different from the mechanism of derepression by c-Jun that we have characterized here; however, both modes of derepression serve to increase TFIID-driven transcription.
Since TFIIA has been shown to stimulate TFIID binding to the TATA box and to compete with the largest subunit of TFIID for binding TBP (24,28), it is possible that transcriptional activators could indirectly derepress TAF II 250 inhibition of TATA box binding by helping TFIIA bind TBP. Although TFIIA may play a role in c-Jun derepression, it is not required under all conditions. In our highly purified in vitro transcription system we found that c-Jun blocked the repression of transcription by GST-TAF II 250-(1-163) in the absence of TFIIA. Furthermore, using DNase I footprinting we found that c-Jun blocked the ability of GST-TAF II 250-(1-163) to repress TATA box binding by TBP. Thus, TFIIA and other general transcription factors are not required for the function of c-Jun in derepressing TBP binding in vitro.
TBP has also been shown to dimerize in vitro and in vivo either alone or when part of TFIID (45,46). Dimers form through interactions between the concave surfaces of two molecules of TBP, resulting in a complex that is not capable of binding DNA. The coordination of TBP dimerization and the N terminus of TAF II 250 interacting with TBP is not understood. Perhaps these two distinct mechanisms of repressing TFIID-DNA binding function together to ensure that the repressed state is maintained in the absence of interaction with TFIIA and transcriptional activators.