TATA-binding protein-associated factors enhance the recruitment of RNA polymerase II by transcriptional activators.

Transcription factor (TF) IID, comprised of the TATA-binding protein (TBP) and TBP-associated factors (TAFs), is a general transcription factor required for RNA polymerase II (pol II) transcription on most eukaryotic genes. Recent findings that TAFs may not be globally required for activator-dependent transcription in vivo and in vitro and that both TAF-dependent and TAF-independent promoters are found in yeast suggest that transcriptional activation can occur through at least two different pathways, depending on the presence or absence of TAFs. Using order-of-addition and template challenge assays performed in a human cell-free transcription system reconstituted with recombinant general transcription factors (TFIIB, TBP, TFIIE, TFIIF), a recombinant general cofactor (PC4), and highly purified epitope-tagged multiprotein complexes (TFIID, TFIIH, pol II), we demonstrate that when TBP is used as the TATA-binding factor transcriptional activators such as Gal4-VP16 and human papillomavirus E2 mainly function by facilitating pol II entry to the promoter region. In contrast, when TFIID is used as the TATA-binding factor, promoter recognition by TFIID appears to be the rate-limiting step facilitated by transcriptional activators during preinitiation complex assembly. Using protein-protein pull-down and far-Western analyses, we further show that the presence of TAFs in TFIID facilitates the recruitment of pol II by transcriptional activators, thereby switching the rate-limiting step from pol II entry to promoter recognition. Our findings thus provide distinct molecular mechanisms for TAF-independent and TAF-dependent activation.

In eukaryotes, transcription preinitiation complex (PIC) 1 formation can occur via at least two different pathways. The sequential assembly pathway typically begins with the binding of TFIID to the core promoter region, followed by the entry of TFIIB, RNA polymerase II (pol II)-TFIIF complex, TFIIE, and TFIIH (1,2). In this pathway, any steps of the PIC assembly are likely to be regulated by gene-specific transcription factors. Another pathway for PIC assembly usually takes place via recruitment of a preassembled TFIID-deficient pol II holoenzyme complex, which contains pol II, a subset of general transcription factors (GTFs), such as TFIIB, TFIIE, TFIIF, and TFIIH, as well as proteins involved in other cellular functions (3). In this two-component assembly pathway, either TFIID binding to the promoter region or pol II holoenzyme entry can be facilitated by transcriptional regulators (4 -7). At present, many pol II holoenzyme complexes with distinct protein compositions have been isolated from both yeast and mammalian cells, reflecting a diverse function of pol II in regulating gene activity (3).
To regulate the steps for PIC formation, gene-specific transcription factors often requires additional protein cofactors to efficiently communicate with the general transcription machinery. TFIID, being a multisubunit protein complex comprised of the TATA-binding protein (TBP) and approximately a dozen TBP-associated factors (TAFs), clearly plays a critical role not only as a core promoter-binding factor (8,9), but also as a transducer in conveying the upstream regulatory signals to the downstream general transcription machinery (10 -12). These functional properties of TFIID are in part accounted for by the presence of TAFs, which further modulate the activity of TBP by providing enzymatic activities (13)(14)(15)(16)(17) and additional contact surfaces for protein-protein and protein-DNA interactions (18,19). An important consequence of TAF association with TBP is to allow TFIID, but not TBP, to transcribe chromatin templates (20,21). On the other hand, TAFs may mask the surfaces of TBP available for interaction with the TATA box and other protein regulators, thereby restricting the accessibility of the free form of TBP to the promoter region (22,23). The observation that TAFs may not be globally required for activatordependent transcription in vivo (24 -27) and in vitro (6, 28 -31) suggests that transcriptional activation can occur through at least two different pathways, depending on the presence or absence of TAFs. This hypothesis is further supported by recent findings that both TAF-dependent and TAF-independent promoters are found in yeast (32,33) and that a TBP-sans-TAFs complex containing TBP and unprocessed TFIIA␣␤ precursor is found in undifferentiated mouse embryonic carcinoma cells (34). Clearly, TFIID and TBP have distinct properties in supporting basal and activator-dependent transcription in eukaryotes.
Using human cell-free transcription systems performed in HeLa nuclear extracts (29) or in reconstituted transcription systems (6,30,31), we and others demonstrated previously that both TAF-independent activation (i.e. using TBP as the TATA-binding factor) and TAF-dependent activation (i.e. using TFIID as the TATA-binding factor) could be recapitulated in * This work was supported in part by Grants GM59643 and CA81017 from the National Institutes of Health and in part by Grant RPG-97-135-04-MBC from the American Cancer Society. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  vitro with Gal4 fusion proteins containing different activation domains and also with other activators such as thyroid hormone receptor. To define the molecular mechanisms of TAFindependent and TAF-dependent activation, we further carry out order-of-addition and template challenge experiments in our highly purified in vitro transcription system. Intriguingly, we are able to show that in TAF-independent activation, pol II entry during PIC assembly is the major step regulated by transcriptional activators, whereas in TAF-dependent activation, promoter recognition by TFIID becomes the rate-limiting step enhanced by activators. Using GST pull-down and far-Western analyses, we also demonstrate that the reason why pol II entry is no longer the rate-limiting step when TFIID is used as the TATA-binding factor is because TAFs in TFIID can facilitate the recruitment of pol II by transcriptional activators, thereby switching the rate-limiting step from pol II entry to promoter recognition. These findings not only provide a molecular mechanism by which activation can occur in the absence of TAFs, but also document a novel function of TAFs in recruiting the initiation form of pol II.
In Vitro Transcription-In vitro transcription was typically carried out in a 30-l reaction containing 50 ng of pG 5 MLT, 50 ng of p2E2(IR)⌬53, 10 ng of TFIIB, 1 ng of TBP or an equivalent amount of TFIID (as judged by Western blotting with anti-TBP antibodies), 2.5 ng each of TFIIE␣ and TFIIE␤, 1 ng of TFIIF, 15 ng of TFIIH, and 3ϳ18 ng of pol II following the condition described (6), except using 3.3 mM MgCl 2 for the reactions and including a 20-min preincubation time before initiation of transcription. For activator-dependent transcription, 150 ng of PC4 and 50 ng of Gal4-VP16 or E2 were also included as specified. Reactions were then performed and analyzed as described previously (6). The transcription signals were quantitated by Phospho-rImager (Molecular Dynamics). Unless otherwise specified, fold activation in each set of reactions is defined as the signal intensity from each activator-binding site-containing template relative to that from the same DNA template performed in the absence of Gal4-VP16 and PC4 (i.e. the first lane of each reaction set).
Template Challenge and Order-of-Addition Experiments-A two-step incubation procedure was carried out as described previously (6) with minor modifications. Briefly, 50 ng of pG 5 MLT and 50 ng of p2E2(IR)⌬53 were preincubated with 1 ng of TBP or an equivalent amount of TFIID, alone or together with other proteins, in the absence or presence of 150 ng of PC4, 50 ng of Gal4-VP16, or 50 ng of E2 at 30°C for 20 min. Ribonucleoside triphosphates (NTPs) and the remaining components required for transcription were then added to initiate transcription. Reactions were continued at 30°C for 60 min before analyzed for RNA formation. The challenge template (500 ng of p⌬MLP) was included either at the beginning or at the end of the preincubation period.
Protein-Protein Interaction-To perform GST pull-down assays, we first expressed GST-tagged HPV-11 E2 (GST-E2) and GST protein individually in the Escherichia coli strain BL21(DE3)pLysS and prepared the bacterial lysate as described previously (42). For interaction assays, ϳ0.5 g of GST or GST-E2 protein was first immobilized onto 20 l of glutathione-Sepharose TM 4B beads (Amersham Pharmacia Biotech) at 4°C overnight. The beads were then sequentially washed three times with 1 ml of BLB (20 mM Tris-HCl, pH 7.9 at 4°C, 20% glycerol, 0.2 mM EDTA, 0.5 M NaCl, 10 mM 2-mercaptoethanol, 0.2 mM PMSF, 1 g/ml pepstatin, 1 g/ml leupeptin, 1 g/ml aprotinin) plus 0.1% Nonidet P-40, three times with 1 ml of BC100 (20 mM Tris-HCl, pH 7.9 at 4°C, 20% glycerol, 0.2 mM EDTA, 100 mM KCl, 1 mM DTT, 0.5 mM PMSF), and incubated with individually purified proteins, including 12.5 ng of TBP or an equivalent amount of TFIID, 25 ng of PC4, or 300 ng of pol II purified from nuclear pellets (15), in 100 l of different salt-containing BC buffers (i.e. 20 mM Tris-HCl, pH 7.9 at 4°C, 20% glycerol, 0.2 mM EDTA, 1 mM DTT, 0.5 mM PMSF, plus different mM concentrations of KCl; see Ref. 43) plus 0.1% Nonidet P-40. After incubation at 4°C for 1 h, the beads were washed three times with 1 ml of the specific salt-containing BC buffer plus 0.1% Nonidet P-40, followed by three washes with BC100. The protein-bound beads were finally mixed with 30 l of 2ϫ protein sample buffer in which 10 l of each samples were analyzed by Western blotting. A similar condition was used for the experiments described in the legend to Fig. 5B, except that the incubation was performed at 200 l of BC100 plus 0.1% Nonidet P-40 with 150 ng of pol II or 375 ng of pol II holoenzyme.
Interaction assays with immobilized pol II were conducted by first incubating 10 g of anti-pol II RPB1 antibodies (N20, purchased from Santa Cruz) with 10 l of protein A-Sepharose CL-4B beads (Amersham Pharmacia Biotech) at 4°C overnight. After washing the beads with 100 l of 0.2 M sodium borate, pH 9.0, 100 l of 20 mM dimethylpimelimidate were added to cross-link the antibodies to the beads. The reaction was carried out at room temperature with constant rotation for 30 min and stopped by washing the beads first with 100 l of 0.2 M ethanolamine, pH 8.0, and then incubating with another 100 l of 0.2 M ethanolamine, pH 8.0, at room temperature for 2 h. The antibodyconjugated beads (10 l) were then equilibrated to BC100 and incubated with 3 g of pol II at 4°C overnight. The immobilized pol II beads or beads alone were washed with 200 l of BC100 three times and incubated with either 10 ng of TBP or an equivalent amount of TFIID in 100 l of BC100 plus 0.03% Nonidet P-40 at 4°C for 1 h. The beads were washed with 200 l of BC100 three times and mixed with 30 l of 2ϫ protein sample buffer. Samples were then analyzed by Western blotting with a 2000-fold dilution of anti-TBP SL39a monoclonal antibody (44) or with a 1000-fold dilution of anti-pol II 8WG16 (45), anti-TAF II 250, and anti-TAF II 135 antibodies, following the described protocol (41).
Far-Western Analysis-32 P-Labeled TFIID used for far-Western analysis was prepared by incubating 10 l of TFIID, which contains ϳ80 ng of FLAG-tagged TBP with a heart muscle kinase phosphorylation site linked to the FLAG sequence (42,46), with 40 units of heart muscle kinase (Sigma, catalog number P-2645) and 2 l of [␥-32 P]ATP (ICN, 7000 Ci/mmol) in a 30-l reaction containing 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 12 mM MgCl 2 , and 1 mM DTT at 30°C for 30 -60 min. Labeled TFIID was then separated from the unincorporated free nucleotide by using the Nick TM column (Amersham Pharmacia Biotech). 32 P-Labeled pol II was prepared by incubating 1 g of pol II (15) with 0.3 milliunits of casein kinase II (Calbiochem, catalog number 218698) and 2 l of [␥-32 P]ATP in a 35-l reaction containing 20 mM Tris-HCl, pH 7.5, 25 mM NaCl, 12 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM EGTA, and 1 mM DTT at 30°C for 1 h and similarly purified by using the Nick TM column.

RESULTS
Human Papillomavirus E2 Protein Can Activate Transcription via Both TAF-independent and TAF-dependent Pathways-To study the mechanisms of transcriptional activation mediated by TBP (i.e. TAF-independent activation) and TFIID (i.e. TAF-dependent activation), we established a highly purified in vitro transcription system reconstituted with only recombinant proteins (TFIIB, TBP, TFIIE, TFIIF, and PC4) and epitope-tagged multiprotein complexes (TFIID, TFIIH, and pol II). In this TAF-independent activation system mediated by TBP, PC4 is the only general cofactor required for transcriptional activation by Gal4 fusion proteins with different activation domains (30). To examine whether TAF-independent activation could also be recapitulated in this highly purified in vitro transcription system with other activators such as human papillomavirus (HPV) E2 protein, which modulates the expression of viral E6 and E7 oncoproteins (47,48), we constructed a DNA template p2E2(IR)⌬53 containing HPV type 11 (HPV-11) E2-binding sites linked to the adenovirus major late promoter (MLP) in front of a G-less cassette of ϳ280 base pairs. Two additional DNA templates containing either five Gal4-binding sites (pG 5 MLT) or devoid of any activator-binding sequences (p⌬MLP), previously constructed on the same core promoter elements but linked to different lengths of G-less cassettes, were used for comparison (Fig. 1A). As shown in Fig. 1B, HPV-11 E2 protein activated transcription through its cognate DNA-binding sites either in the presence (lanes 1-4) or in the absence (lanes 5-8) of TAFs, similar to the effects observed previously with Gal4-VP16 (30). This result indicates that TAFs are not absolutely required for E2-mediated activation in our reconstituted transcription system performed with naked DNA templates.

Pol II Entry during PIC Assembly Is the Rate-limiting Step Enhanced by Transcriptional Activators in TAF-independent Activation, whereas Promoter Recognition by TFIID Is the Ratelimiting
Step Facilitated by Transcriptional Activators in TAFdependent Activation-To investigate the mechanism of TAFindependent activation, we performed order-of-addition and template challenge experiments to define the steps of PIC assembly regulated by transcriptional activators. As outlined in Fig. 2A, both pG 5 MLT and p2E2(IR)⌬53 templates were preincubated with TBP, alone or together with PC4, an activator, and other GTFs, for 20 min. The remaining transcriptional components and ribonucleoside triphosphates were then added to initiate transcription. In this assay, a 10-fold excess of challenge template (p⌬MLP) was included, normally after the preincubation step, to test the stability of protein-DNA complex assembled, in the absence or presence of activators, on pG 5 MLT and p2E2(IR)⌬53. As a control experiment, when the challenge template was added during the preincubation step, both basal and activator-dependent transcription from pG 5 MLT and p2E2(IR)⌬53 were reduced, because less protein became available to the activator-binding site-containing templates ( Fig. 2A, compare lanes 1-4 with lanes 5-8). In contrast, the transcription signals from p⌬MLP were dramatically increased, reflecting the fact that more p⌬MLP templates were present in the reactions. As noted previously (49,50), the relatively low concentration of PC4 used in this assay clearly enhanced basal transcription in the absence of an activator ( Fig. 2A, compare lanes 5 and 6). Interestingly, when the challenge template was added after the preincubation step, we found that both E2 and Gal4-VP16 failed to significantly enhance transcription unless TBP, TFIIB, and pol II were included during the preincubation step ( Fig. 2A, lanes 9 -32). This indicates that these transcriptional activators mainly work by facilitating pol II entry during PIC assembly in the absence of TAFs (Fig. 2B). Our data also suggest that TFIIF, although it helps pol II forming a stable preinitiation complex on many promoters examined (51-54), is not essential for pol II recruitment by Gal4-VP16 and E2 in TAF-independent activation ( Fig. 2A, compare lanes 17-20 with lanes 21-24), consistent with the observation that TFIIF appears dispensable for formation of stable preinitiation intermediates on some core promoters (54,55). To explore the possibility whether preincubation of TFIIF with TFIIB, TBP, and DNA templates would suffice pol II recruitment by transcriptional activators, we per-

FIG. 1. TAF-independent and TAF-dependent activation mediated by HPV-11 E2 and Gal4-VP16.
A, DNA templates used for in vitro transcription assays. Constructions of various G-less cassettes driven by five Gal4-binding sites (pG 5 MLT) or 2 HPV-11 E2-binding sites (p2E2(IR)⌬53) were described under "Experimental Procedures." The p⌬MLP template contains the same adenovirus major late (MLP) TATA box and initiator (Inr) element as those present in pG 5 MLT and p2E2(IR)⌬53. B, in vitro transcription assays. Transcription reactions were performed in a highly purified cell-free transcription system reconstituted with recombinant general transcription factors (TFIIB, TBP, TFIIE, and TFIIF) and epitope-tagged protein complexes (TFIID, TFIIH, and pol II), in the absence (Ϫ) or presence (ϩ) of a recombinant general cofactor PC4 or a recombinant transcriptional activator (E2 or Gal4-VP16). Fold activation in each set of reactions is defined as the signal intensity quantitated by a PhosphorImager (Molecular Dynamics) from each activator-binding site-containing template relative to that from the same DNA template performed in the absence of activators and PC4 (i.e. the first lane of each reaction set). formed a similar template challenge experiment by adding TFIIF prior to pol II entry. As shown in Fig. 2C, preincubation of TFIIF, TFIIB, and TBP with DNA templates did not shift the activator-facilitated step (compare lanes 5-8 with lanes 9 -12 and lanes [13][14][15][16]. If pol II recruitment is indeed the step facilitated by transcriptional activators in TAF-independent activation, we should not detect enhancement of transcription when activators were added after pol II entry during PIC assembly. This idea was directly tested by conducting another order-of-addition experiment with DNA templates preincubated with TBP, TFIIB, and pol II, in the absence or presence of PC4, for 20 min. Activators were added either during or after the preincubation step to compare their effect on pol II recruitment. As shown in Fig. 3, TAF-independent activation could only occur when transcriptional activators were added prior to, but not after, pol II entry during PIC assembly (compare lanes 3 and 4 with lanes  5 and 6). This result further demonstrates that activators indeed facilitate pol II entry during PIC assembly in TAF-independent activation. Clearly, the coactivator role of PC4 has to be present at the same time with activators, as inclusion of PC4 after the preincubation step failed to potentiate activator function (Fig. 3, compare lanes 3 and 4 with lanes 7 and 8). For unknown reason, PC4 appeared to increase overall basal tran-scription when added after the preincubation step (Fig. 3, compare lanes 2 and 9). Thus, no activator-facilitated recruitment of pol II entry could be observed if PC4 was added after the preincubation step (Fig. 3, compare lanes 7 and 8 with lane 9).
The template challenge and order-of-addition experiments were also conducted in the same way with TFIID as the TATAbinding factor to define the steps of PIC assembly regulated by E2 and Gal4-VP16. In contrast to the results performed with TBP, a significant enhancement of transcription was already achieved with TFIID included during preincubation (Fig. 4, A  and B), indicating that the activator-facilitated step apparently shifted to promoter recognition by TFIID. Collectively, these data suggest that there are two major steps regulated by transcriptional activators: pol II entry during PIC assembly in TAF-independent activation and promoter recognition by TFIID in TAF-dependent activation.
TAFs Facilitate the Recruitment of the Initiation Form of Pol II by E2-To define the molecular mechanisms by which the presence of TAFs shifts the activator-regulated step from pol II entry to promoter recognition, we carried out protein-protein pull-down assays. As shown in Fig. 5A, pol II, PC4, TBP, and TFIID all interacted with immobilized GST-E2, but not with GST alone, at low salt concentrations. These interactions were disrupted when the salt (KCl) concentrations were increased to  5-32). Lanes 1-4 were standard transcription reactions performed without (Ϫ) template challenge. Fold activation is the same as defined in the legend to Fig. 1B. B, quantitation of the transcription signals in TAF-independent activation. In this graph, fold activation in each set of reactions is presented as the signal intensity from each activator-binding site-containing template relative to that from the same DNA template performed in the presence of PC4 without an activator. C, preincubation of TFIIF with TFIIB and TBP does not suffice pol II recruitment by transcriptional activators in TAF-independent activation. The template challenge and order-of-addition transcription assays were performed as described in A. 300 mM or above during incubation, indicating that E2 interactions with pol II, PC4, TBP, and TFIID, although specific, were relatively weak. This result, similar to the functional properties of Gal4-VP16, which has been shown to interact with all these protein factors, is consistent with previous reports that BPV-1 E2 contacts TBP (56) and HPV-8 E2 interacts with both TBP and the human TAF II 55 component of TFIID (57). In this experiment, the pol II sample was purified from the nuclear pellets of the induced FLAG-tagged pol II cell line and thus contained both the initiation form, which contains the non-or hypophosphorylated (i.e. IIa) form of the largest subunit (RPB1), and the elongation form, which contains the hyperphosphorylated (i.e. IIo) form of RPB1 (15).
To see if TBP or TFIID could enhance the interaction between E2 and pol II, we included an equivalent amount of TBP and TFIID, as normalized by Western blotting with anti-TBP antibodies, individually with immobilized GST-E2 and pol II. As shown in Fig. 5B, immobilized E2 alone retained mainly the elongation form of pol II and only a negligible amount of the initiation form (compare lanes 1 and 2). However, when TFIID, but not TBP or PC4 (used as control), was additionally included in the binding reaction, the initiation form of pol II was more efficiently bound to E2 (Fig. 5B, lanes 2-5, compare the signal ratio between IIa and IIo forms of pol II in each lane). This suggests that the presence of TAFs in TFIID facilitates the recruitment of the initiation form of pol II by transcriptional activators, thereby switching the rate-limiting step from pol II entry to promoter recognition. If this mechanism accounts for the difference between TAF-independent and TAF-dependent activation, we speculated that TAFs might also facilitate the recruitment of pol II holoenzyme by transcriptional activators. Indeed, the presence of TFIID, but not TBP, enhances the interaction between E2 and a TFIID/TBP-deficient human pol II holoenzyme complex (Fig. 5B, lanes 6 -9), which was purified from a HeLa-derived cell line that conditionally expresses the FLAG-tagged RPB9 subunit of human pol II and contains only the initiation form of pol II complex (6).
The GST pull-down assay suggests that TAFs in TFIID may provide additional contact surfaces to enhance the recruitment of the initiation form of pol II, other than the direct interaction between TBP and the nonphosphorylated carboxyl-terminal domain of RPB1 in pol II (58). To explore this, we conducted a coimmunoprecipitation experiment by incubating an equivalent amount of TFIID and TBP (Fig. 6A, lanes 1 and 2, TBP  panel) with immobilized pol II-beads or beads alone. As shown in Fig. 6A, TFIID was more efficiently retained on the pol II-beads than TBP (compare lanes 1 and 2 with lanes 5 and 6,  TBP panel). The presence of TFIID on the immobilized pol II-beads was further demonstrated by using anti-TAF II 250 and anti-TAF II 135 antibodies (Fig. 6A, lanes 2 and 6). In contrast, no TBP or TFIID was found in the bound fractions with the beads alone (Fig. 6A, lanes 3 and 4), indicating a direct interaction between pol II with TBP and likely with additional components of TFIID.
To identify the components of pol II and TFIID bridging these two large protein complexes, we performed far-Western analysis using either 32 P-labeled TFIID or 32 P-labeled pol II as probe. As shown in Fig. 6B, TFIID clearly interacts with the RPB1 and RPB2 subunits of pol II (lane 1), the RAP74 subunit of TFIIF (lanes 2 and 6), HPV-11 E2 (lane 3), and Gal4-VP16 (lane 4). Likewise, pol II interacts with at least a component of TFIID and with RAP74 (Fig. 6C, lanes 1 and 2). Since human TAF II 135 and TAF II 150 comigrate in the gel, the identity of the TFIID subunit that contacts pol II remains to be further investigated. proteins with different activation domains (30). The availability of this highly purified in vitro transcription system provides us with a unique opportunity to further dissect the mechanisms of TAF-independent activation and TAF-dependent activation. We applied order-of-addition to divide the steps of PIC assembly, according to the sequential assembly pathway that was established more than a decade ago (59). The template challenge also allows us to test the stability of promoter-bound protein complexes that do not "commit" transcription on specific templates.
As template commitment is usually the rate-limiting step for PIC assembly (60) and is likely to be regulated by various transcription factors and cofactors, it is not surprising to see that promoter recognition by TFIID is the rate-limiting step enhanced by transcriptional activators. Our result (Fig. 4) is consistent with previous reports that transcriptional activators or non-TAF coactivators may recognize TAFs to increase the recruitment or stability of TFIID binding to the core promoter (10,(61)(62)(63)(64). Unlike TFIID, TBP alone cannot commit transcription to specific templates. Template commitment with TBP only occurs when additional components of the general transcription machinery are also present to stabilize the promoter-TBP complex (Refs. 6 and 54 and this study). Inclusion of a transcriptional activator and PC4 with TBP and DNA templates, in the absence of pol II and other GTFs, is not sufficient to commit TBP for transcription from prebound DNA templates The presence of the proteins on the beads were monitored by Western blotting with different anti-protein antibodies as described previously (15,30,39). The detection of TFIID was conducted with antibodies against the TBP, TAF II 135, TAF II 95, and TAF II 55 components of human TFIID. One-tenth of each purified input (InP) protein used for the interaction assay was loaded on the left of each strip as control for Western blotting. B, E2 recruitment of pol II and pol II holoenzyme is facilitated by TFIID, but not by TBP or PC4. Interaction assays were conducted with 100 mM KCl-containing buffer as described in A, except that PC4, TFIID (D), or TBP (T) was additionally included during the incubation of immobilized GST-E2 beads with either traditionally defined pol II (core-pol II) or pol II holoenzyme (holo-pol II). The anti-pol II carboxyl-terminal domain antibody 8WG16 (45), which recognizes both hyperphosphorylated (IIo) and hypo-or nonphosphorylated (IIa) forms of RPB1 (15), was used in these assays.
FIG. 6. TFIID interacts directly with components of pol II. A, TFIID binds more efficiently than TBP to immobilized pol II. An equivalent amount of TBP (T) and TFIID (D) was incubated, respectively, with immobilized pol II-beads or beads alone as described under "Experimental Procedures." One-third of the bound fractions and one-tenth of the inputs were separated on SDS-polyacrylamide gels and analyzed by Western blotting with different anti-protein antibodies as indicated on the left. B, TFIID interacts directly with RPB1, RPB2, RAP74, HPV-11 E2, and Gal4-VP16. Far-Western analysis was performed with 32 P-labeled TFIID as described under "Experimental Procedures." The tested proteins were first resolved on a 6% (left panel) or 15% (right panel) SDS-polyacrylamide gel and then transferred to a nitrocellulose membrane. After denaturation and renaturation, the membrane was incubated with 32 P-labeled TFIID. The signals were visualized by autoradiography with identified protein signals indicated on the left of each panel. C, pol II interacts with at least a component of TFIID and with RAP74. Far-Western analysis was performed with 32 P-labeled pol II as described under "Experimental Procedures." The tested proteins were separated on a 6% SDS-polyacrylamide gel and processed as outlined in B. (6). The enhanced transcription by E2 and Gal4-VP16 in TAFindependent activation only occurs when pol II was included during the preincubation step (Fig. 2), indicating that these activators mainly work by facilitating pol II entry during PIC assembly in the absence of TAFs. This idea was further substantiated by the experiment demonstrating the loss of TAFindependent transcription when activators were added after pol II entry (Fig. 3). The possibility that the preincubation of pol II may result in an "activation" of pol II is less likely, as promoter recognition by TFIID and pol II entry represent two distinct activator-regulated steps in TAF-dependent and TAFindependent activation, respectively, are also observed in our cell-free transcription system performed with a preassembled TFIID-deficient pol II holoenzyme complex in conjunction with either TFIID or TBP (Ref. 6 and data not shown). Clearly, TAF-independent activation mediated by TBP must take place via a mechanism distinct from TAF-dependent activation mediated by TFIID. Undoubtedly, promoter recognition by TFIID and pol II entry during PIC assembly are two major steps for PIC assembly (Fig. 7), which can occur via either a sequential assembly pathway (TFIID/TBP, TFIIB, pol II/TFIIF, TFIIE, and TFIIH) or a two-component pathway (TFIID/TBP and pol II holoenzyme). This model also supports the view that recruitment of either TFIID or pol II holoenzyme is crucial in activating transcription in both yeast and mammalian cells (4 -7, 65). Obviously, a direct test of the model presented for having two distinct rate-limiting steps that shift depending upon having TBP or TFIID is to perform a kinetic analysis that would test association/dissociation rates of TBP and TFIID to the DNA sequence in the presence and absence of the activators and other components of the general transcription machinery, which will further shed light on the mechanisms of TAF-independent and TAF-dependent activation.
Our findings presented here not only provide a molecular mechanism by which activation can occur in the absence of TAFs, but also document a novel role of TAF function. That is, TAFs can facilitate the recruitment of pol II by transcriptional activators, thereby accounting for the functional difference be-tween TAF-independent and TAF-dependent activation pathways. Our data are also consistent with an earlier observation that TAFs may promote a step of PIC assembly post TFIIB entry by transcriptional activators (66). As TAF-independent activation can occur both in vivo and in vitro, this study thus provides a molecular basis for TAF-independent activation, which represents a new paradigm of eukaryotic gene regulation. FIG. 7. Model for TAF-independent and TAF-dependent activation. The thick arrow indicates the rate-limiting step for preinitiation complex assembly facilitated by transcriptional activators. For simplicity, a preassembled human pol II holoenzyme complex containing pol II and all general transcription factors but TFIID (6) is used for illustration.