Properties of PC4 and an RNA Polymerase II Complex in Directing Activated and Basal Transcription in Vitro *

A human RNA polymerase II (pol II) complex was isolated from a HeLa-derived cell line that conditionally expresses an epitope-tagged RPB9 subunit of human pol II. The isolated FLAG-tagged pol II complex (f:pol II) contains a subset of general transcription factors but is devoid of TFIID and TFIIA. In conjunction with TATA-binding protein (TBP) or TFIID, f:pol II is able to mediate both basal and activated transcription by Gal4-VP16 when a transcriptional coactivator PC4 is also provided. Interestingly, PC4, in the absence of a transcriptional activator, actually functions as a repressor to inhibit basal transcription. Remarkably, TBP is able to mediate activator function in this transcription system. The presence of TBP-associated factors, however, helps overcome PC4 repression and further enhance the level of activation mediated by TBP. Alleviation of PC4 repression can also be achieved by preincubation of the transcriptional components with the DNA template. Sarkosyl disruption of preinitiation complex formation further illustrates that PC4 can only inhibit transcription prior to the assembly of a functional preinitiation complex. These results suggest that PC4 represses basal transcription by preventing the assembly of a functional preinitiation complex, but it has no effect on the later steps of the transcriptional process.

In eukaryotes, transcription of protein-encoding genes requires general transcription factors (GTFs) 1 and RNA polymerase II (pol II), which are assembled on the promoter region to form a preinitiation complex (PIC) capable of producing RNAs when both ribonucleoside triphosphates and energy sources are provided (1,2). In vivo, the PIC assembly is a rate-limiting step for the initiation of transcription and is often facilitated by gene-specific transcriptional activators (3,4). Currently, there are two proposed pathways for the PIC assembly on TATAcontaining promoters. In the sequential pathway, TFIID binding to the TATA box is followed, in a stepwise fashion, by the joining of TFIIB, pol II/TFIIF, TFIIE, and TFIIH, whereas in the two-component pathway, binding of TFIID is accompanied by a preassembled pol II complex that contains pol II, a subset of GTFs, SRBs (suppressors of RNA polymerase B mutations (5,6)), and other proteins involved in chromatin remodeling, DNA repair, or mRNA processing (5)(6)(7)(8)(9)(10)(11)(12)(13). Transcriptional activators, in most cases, are able to increase the level of initiation by enhancing the recruitment of TFIID and/or other components of the basal transcription machinery to the promoter region (3,4). This activation process often requires transcriptional coactivators. Thus far, two major classes of general coactivators required for activator function have been identified in mammalian cell-free transcription systems. One is TBPassociated factors (TAFs) in TFIID (14 -16), and the other is protein cofactors derived from the upstream stimulatory activity (USA) found in the phosphocellulose P11 0.85 M KCl fraction of HeLa nuclear extracts (17,18).
Positive cofactor 4 (PC4) was isolated from a crude USA fraction and was able to substitute for USA to mediate activatordependent transcription in vitro (19 -21). PC4 is a nonspecific DNA-binding protein, which shows a higher affinity toward single-stranded (ss) DNA molecule (20 -22). The ssDNA binding activity of PC4 can replace human ssDNA-binding protein (HSSB, also called replication protein A (RPA)) in supporting the T antigen-catalyzed unwinding of SV40 origin-containing duplex DNA (23). Nevertheless, PC4 cannot substitute for HSSB in other aspects of replication activities mediated by HSSB (23). Likewise, the transcriptional activity of PC4 cannot be replaced by other ssDNA-binding proteins (19). The coactivator function of PC4 seems to correlate with its doublestranded DNA binding activity (21) and its interactions with transcriptional activators and with components of the general transcription machinery such as TFIIA (19). Surprisingly, gene inactivation of a PC4 homologue in yeast does not lead to cell death, indicating that PC4 is nonessential in yeast (24,25). Since yeast PC4 also exhibits distinct properties from that of human PC4 (24,25), it is likely that PC4 may function differentially in various organisms. This remains to be further investigated.
Using an in vitro transcription system reconstituted with either TBP or TFIID and a preassembled pol II complex, we found that PC4 could function as a repressor to suppress basal transcription in the absence of an activator. Interestingly, TBP was able to mediate Gal4-VP16 activation in the absence of TAFs. This finding suggests that human TBP can indeed mediate activator function, as observed in the yeast system (26 -28). To understand the molecular mechanism of PC4 repression, we carried out template challenge and Sarkosyl disruption experiments using our two-component transcription system. The results indicate that PC4 represses transcription by preventing the assembly of a functional preinitiation complex when an activator is not present.
Establishment of a Tetracycline-regulated Clonal Human Cell Line-Ten g of PvuI-linearized pTetCMV-F o :hRPB9 DNA was cotransfected with 50 g of sheared calf thymus DNA and with 0.5 g of SacIlinearized pREP4 (Invitrogen) into HtTA-1 (34), a HeLa-derived cell line that constitutively expresses a tetracycline-controlled transactivator. Detailed procedures for establishing the tetracycline-regulated clonal cell lines had been described (30). A cell line (hRPB9 -3) that showed induced expression of FLAG-tagged hRPB9 in the absence of tetracycline was isolated after screening a dozen hygromycin-resistant cellular clones.
Western Blotting-To detect the presence of individual GTFs as shown in Fig. 1A, 50 l of HeLa nuclear extracts and 200 l of f:pol II were mixed separately with an equal volume of the 2ϫ protein sample buffer. The protein mixture was then separated by 10% SDS-polyacrylamide gel electrophoresis using a preparative mini-gel apparatus (Bio-Rad). After transfer to the nitrocellulose filters, samples were divided into multiple lanes using a Mini-Protean II multiscreen apparatus (Bio-Rad). Each sample lane was then incubated with a primary antibody, usually diluted 1000-fold unless otherwise specified, in a final volume of 600 l. The rest of the procedures for Western blotting were performed as described (30).
Protein Purification-The f:pol II complex was purified from hRPB9 -3 as follows. The hRPB9 -3 cell line was maintained in suspension culture with Joklik medium containing 5% calf serum in the presence of tetracycline (1 g/ml) and selected with G418 (0.6 mg/ml, total weight) for 4 days before expansion for the preparation of nuclear extracts and S100. To induce protein expression, cells were pelleted and washed at least 3 times with 1ϫ phosphate-buffered saline to remove tetracycline. Cells were then resuspended in fresh Joklik medium plus 5% calf serum. Nuclear extracts and S100 were prepared from hRPB9 -3 cells, 4 days after removing tetracycline, as described (35). To purify f:pol II, 10 ml of the hRPB9 -3 nuclear extract or S100 was incubated with 250 -500 l of anti-FLAG M2-agarose beads (Kodak) at 4°C for 6 -12 h. Bound proteins were then washed and eluted with the synthetic FLAG peptide with 100 mM KCl-containing buffer as described previously for TFIID purification (36).
In Vitro Transcription-In vitro transcription was typically carried out in a 25-l reaction mixture containing 50 ng of pG 5 HMC 2 AT, 20 ng of pML⌬53, 42 ng of renatured TFIIA (35 ng of p55 and 7 ng of p12), 10 ng of TFIIB, 10 ng of TBP, or 1 l of FLAG-tagged TFIID (which contains approximately 1 ng of TBP as judged by Western blotting), 20 ng each of TFIIE␣ and TFIIE␤, 28 ng of renatured TFIIF (20 ng of RAP74 and 8 ng of RAP30), 1 l (ϳ7.5 ng) of FLAG-tagged TFIIH, and 0.6 l (ϳ18 ng) of core-pol II or f:pol II using the conditions as described previously (17). For activator-dependent transcription, 100 ng of PC4 and 50 ng of Gal4-VP16 were also included as specified. In the minimal transcription system, 3 l (ϳ90 ng) of f:pol II was used in conjunction with either 2 ng of TBP or 2 l of FLAG-tagged TFIID. The amount of FLAG-tagged Gal4 fusion proteins used in the minimal transcription reaction was: 3 ng of Gal4 (1-94), 6 ng of Gal4-Pro, 6 ng of Gal4-Gln, and 30 ng of Gal4 (1-147). Reactions were then performed and analyzed as described (36).
Template Challenge Experiments-A two-step incubation procedure described by Kaiser et al. (21) was used with a minor modification. Briefly, 50 ng of pG 5 HMC 2 AT and 20 ng of pML⌬53 were preincubated with either 2 l of FLAG-tagged TFIID (or 2 ng of TBP) or 3 l of f:pol II at 30°C for 50 min. Ribonucleoside triphosphates (NTPs) and the other component required for transcription (f:pol II or TFIID or TBP) were added after the preincubation period to initiate transcription. Reactions were continued at 30°C for 60 min and then analyzed for RNA formation (36). Gal4-VP16 and PC4, when added, were used at 50 and 100 ng, respectively. The challenge template (500 ng of pG 5 HMC 2 AT or 500 ng of pML⌬53) was included either at the beginning or at the end of the preincubation period as depicted in Fig. 3A.
Sarkosyl Disruption of Preinitiation Complex Formation-In this experiment, 0.015% Sarkosyl was added either at the beginning or at the end of the preincubation period, which was performed in the presence of TBP, f:pol II, and both pG 5 HMC 2 AT and pML⌬53 DNA templates as described in the template challenge experiments. Ribonucleoside triphosphates were then added to initiate transcription. One hundred nanograms of PC4 were added at various time points as outlined at the bottom of Fig. 4.

Isolation of a TFIID-deficient Human Pol II Complex-A
human pol II complex was purified from a clonal HeLa-derived cell line (hRPB9 -3) that conditionally expresses the FLAGtagged RPB9 subunit of human pol II (see "Experimental Procedures"). The purified FLAG-tagged pol II complex (f:pol II) contains not only pol II subunits as detected by Western blotting with antibodies against RPB1, RPB2, RPB6, RPB8, and RPB9 2 but also a subset of GTFs including TFIIB, TFIIE, TFIIF, and TFIIH (Fig. 1A). The f:pol II complex contains stoichiometric amounts of TFIIB and TFIIF but substoichiometric quantities of TFIIE and TFIIH (Fig. 1A, compare relative signals detected in nuclear extracts and f:pol II), as also evidenced by quantitative Western blotting using purified recombinant proteins as standards. 3 TFIID and TFIIA were not detected in f:pol II at a sensitivity of 1 ng with anti-TBP and anti-TAF II 55 antibodies and at a sensitivity of 0.1 ng with anti-TFIIA p35 antibodies (Fig. 1A). 2 Our f:pol II, enriched approximately 200-fold after immunoaffinity purification, did not seem to contain other transcriptional cofactors such as PC4 and Dr1 and transcriptional activators including Sp1, YY1, USF, p53, pRB, and the p50 subunit of NFkB. 2 f:pol II Is Functional in a Highly Purified in Vitro Transcription System-Recombinant human TFIIA, TFIIB, TFIIE, TFIIF, and FLAG-tagged TFIID and FLAG-tagged TFIIH were used in conjunction with either core-pol II (i.e. traditionally defined pol II) or immunoaffinity-purified f:pol II for transcriptional analysis. Both core-pol II and f:pol II, normalized by the content of RPB2 in the purified complexes, showed comparable levels of transcriptional activities irrespective of whether TBP or TFIID was used as the TATA-binding factor (Fig. 1B, top and  bottom panels, lanes 1 and 9). The pG 5 HMC 2 AT template contains 5 Gal4-binding sites preceding the HIV-1 TATA box and the adenovirus major late (ML) initiator element in front of a G-less cassette of approximately 380 nucleotides, whereas pML⌬53, which lacks the activator-binding sites, has a shorter G-less cassette (ϳ280 nucleotides) driven only by the major late promoter TATA and initiator elements. In our transcription system, TFIIB, TFIIF, and a TATA binding activity (either TBP or TFIID) were essential for transcription by core-pol II (Fig. 1B, lanes 3, 4, and 6), whereas TFIIE and TFIIH, although not necessary for transcription from supercoiled DNA templates (44 -46) (Fig. 1B, lanes 5 and 7), are required for transcription from linearized DNA molecules. 2 Interestingly, transcription from different promoter elements seem to require differential amounts of TFIIE and TFIIH, as leaving out TFIIE and TFIIH affected transcription from pG 5 HMC 2 AT more dramatically than from pML⌬53 (Fig. 1B, lanes 5 and 7). In contrast, transcription by f:pol II required only a TATA-binding factor (Fig. 1B, lanes 9 -16), although leaving out TFIIE significantly reduced basal transcription from both DNA templates. The transcription data not only functionally confirm the identities of GTFs detected in our purified f:pol II (Fig. 1, A and B) but further suggest that a two-component pathway comprised of preassembled f:pol II and a TATA-binding factor is probably sufficient for the assembly of a functional preinitiation complex (see below). Obviously, TFIIA and TAFs were not needed for basal transcription from either DNA template (Fig. 1B, compare top and bottom panels, lanes 1 versus 2 and lanes 9 versus 10).
We also examined the factor requirement for activator-dependent transcription in our highly purified in vitro transcription system. All GTFs, except TFIIA, were necessary, in conjunction with a transcriptional activator (Gal4-VP16) and a coactivator (PC4), for activated transcription by core-pol II (Fig. 1C, lanes 1-10). In contrast, f:pol II only requires TFIID, PC4, and Gal4-VP16 for activated transcription from pG 5 HMC 2 AT (Fig. 1C, lanes 11-20). Again, leaving out TFIIE and TFIIH showed some effect on activated transcription by f:pol II, reflecting their substoichiometric amounts in purified f:pol II (Fig. 1A). Apparently, TFIIA was not required for activated transcription by either core-pol II or f:pol II in this highly purified in vitro transcription system (Fig. 1C, compare lanes 1  and 2 with lanes 11 and 12). Presumably TFIIA is only needed to antagonize the repressive effect from some negative factors that may be present in cruder systems (2). Indeed, we have found that TFIIA becomes essential whenever a partially purified E/F/H or USA fraction is used in the transcription assay. 4 In addition, the requirement for TFIIA can be affected by the amount of TFIID and the promoter structure of the DNA templates used in the reaction (47). Nevertheless, this analysis indicates that, in addition to f:pol II and TFIID, a transcriptional activator (Gal4-VP16) and a coactivator such as PC4 were minimally required for activated transcription in vitro.
TAF-independent and TAF-dependent Transcriptional Activation-To investigate the role of TBP and TAFs in activatordependent transcription, we compared in parallel the transcriptional activities of TBP and TFIID in a transcription system comprised of f:pol II, TBP or TFIID, PC4, and Gal4-VP16. In this assay, TFIID was the only source of TAFs, which were not found in f:pol II as judged by Western blotting (Fig.  1A) and transcriptional assays (Fig. 1B, lanes 12 in upper and  lower panels, and Fig. 1C, lane 14). As expected, no transcription could be detected in the absence of a TATA binding activity provided by either TBP or TFIID ( Fig. 2A, lanes 1 and 6). When TBP or TFIID, containing an equivalent amount of TBP, was also included, basal transcription could be detected from both pG 5 HMC 2 AT and pML⌬53 DNA templates ( Fig. 2A, lanes 2  FIG. 1. f: Transcription reactions were performed as described in B with the addition of PC4 and Gal4-VP16 (Act). and 7). If Gal4-VP16 was added to the system without PC4, only minor if any enhancement of transcription was observed ( Fig. 2A, compare lanes 2 and 3, and 7 and 8), confirming the importance of additional cofactors other than TAFs in mediating activator function (17,36). Surprisingly, the coactivator PC4 in the absence of an activator acts as a repressor to suppress basal transcription mediated by TBP ( Fig. 2A, compare lanes 2 and 4). PC4 repression was not obvious in the case of TFIID (Fig. 2A, compare lanes 7 and 9), indicating that TAFs can overcome PC4 repression. Surprisingly, when Gal4-VP16 was also provided, we observed transcriptional activation mediated by both TBP and TFIID (Fig. 2A, compare lanes 2 and 5,  and 7 and 10). These data suggest that human TBP, in the absence of TFIID TAFs, can also mediate transcriptional activation in a mammalian cell-free transcription system, as previously shown in yeast (5). The presence of TAFs, however, help overcome PC4 repression and further enhance the level of activation mediated by TBP (Fig. 2A, compare lanes 7 and 9,  and 7 and 10).
To see if TBP-and TFIID-mediated activation was unique to the acidic type of activation domains as exemplified by Gal4-VP16, we also tested the ability of other Gal4 fusion proteins in activating transcription in our minimal transcription system. Both Gal4-Pro and Gal4-Gln, which contain proline-rich and glutamine-rich activation domains linked, respectively, to the Gal4 DNA-binding domain, were able to activate transcription mediated by TBP, mainly at the level of antirepression (Fig.  2B, lanes 1-4). In contrast, Gal4-(1-94), which was the portion used in Gal4-Pro and Gal4-Gln, had no effect on transcription (Fig. 2B, compare lanes 2 and 5). Interestingly, Gal4-(1-147), which was used to make Gal4-VP16, was capable of activating transcription to a similar level as seen by Gal4-Pro and Gal4-Gln (Fig. 2B, lanes 1, 3, 4, and 7). This is consistent with the observation that an activation domain was present between amino acids 75 and 147 (19). This might explain why an additional level of transcriptional activation was observed with Gal4-VP16 in our minimal transcription system mediated by TBP (Fig. 2B, lanes 1 and 6); presumably two activation domains were synergistically used to activate transcription. When TAFs were present in our minimal transcription system, all these activators showed activation above the basal level (Fig. 2B, lanes 8 -14). As mentioned earlier ( Fig. 2A, lanes 2 and  4 versus lanes 7 and 9), PC4 did not repress basal transcription in the minimal transcription system when TFIID was substituted for TBP (Fig. 2B, lanes 1 and 2 versus lanes 8 and 9). We should emphasize that TAF-independent antirepression mediated by Gal4-Pro and Gal4-Gln was also an activation process, as it occurred only in the presence of an activator and only on the DNA template (i.e. pG 5 HMC 2 AT) containing the activatorbinding sites.
PC4 Repression Could Also Be Alleviated by Preincubating TBP and f:pol II with the DNA Template-Since PC4 is a nonspecific DNA-binding protein (20 -22), it may inhibit transcription by blocking the access of protein factors to the promoter region. If so, PC4 repression should be alleviated by preincubating components of the general transcription machinery with the DNA template. To test this hypothesis, we first examined the requirement of protein factors for template commitment in our minimal transcription system reconstituted with TBP (or TFIID) and f:pol II. Template commitment is usually the rate-limiting step for the assembly of a functional preinitiation complex (48) and is likely to be regulated by various transcription factors and cofactors. As outlined in Fig.  3A, both pG 5 HMC 2 AT and pML⌬53 DNA templates were preincubated with TBP (or TFIID) or f:pol II, individually or simultaneously, for 50 min. The other transcription components and ribonucleoside triphosphates were then added to initiate transcription. Reactions were continued for an hour before they were analyzed for RNA formation. In this experiment, a 10-fold excess of pG 5 HMC 2 AT template was added either during or after the preincubation period to test the stability of the protein-DNA complex. Demarcating the transcription reaction into two separate steps via order-of-addition did not change the overall yield of transcripts (Fig. 3, B and C, lanes 1 and 2). When excess pG 5 HMC 2 AT was added during the preincubation period, transcription from pML⌬53 was reduced because less protein became available to the pML⌬53 DNA template (Fig. 3,  B and C, compare lanes 2 and 3, lanes 5 and 6, and lanes 8 and  9). If TFIID was preincubated with DNA templates before template challenge, no reduction of pML⌬53 transcription was observed (Fig. 3B, compare lanes 2 and 4). This result was consistent with previous observations that TFIID could stably bind to the promoter region once committed (48 -51). The same results were obtained by using pML⌬53 as the challenge template. 2 In contrast, preincubation of TBP or f:pol II did not resist template challenge (Fig. 3B, compare lanes 5 and 7, and  Fig. 3C, compare lanes 2 and 4, and lanes 5 and 7), indicating that TBP or f:pol II alone could not stably bind to the promoter region. This result is also consistent with the observation that TBP, in the absence of other general transcription components, could not commit transcription to a particular template (51). Interestingly, when TBP and f:pol II were both present during preincubation, this promoter-bound complex became stable and was thus resistant to template challenge (Fig. 3C, compare  lanes 8 and 10). We conclude from these template challenge experiments that in our two-component transcription system only TFIID, but not TBP or f:pol II alone, can stably bind to the promoter region and thereby commit to the transcription process, consistent with previous results performed either with nuclear extracts or with complete transcription systems reconstituted with partially purified components (48 -51).
We also examined if TBP or f:pol II would become stably bound to the promoter in the presence of a transcriptional activator and a coactivator. As shown in Fig. 3D, Gal4-VP16 and PC4 were unable to enhance the binding of TBP or f:pol II (lanes 6 -8 and lanes [11][12][13]. Only when both TBP and f:pol II were present would template commitment occur (Fig. 3D, lanes  16 -18), consistent with the result performed in the absence of Gal4-VP16 and PC4 (Fig. 3C). Interestingly, preincubation of TBP and f:pol II together, but not individually, with PC4 and DNA templates alleviated PC4 repression (Fig. 3D, compare  lanes 1 and 2, 4 and 5, 9 and 10, and lanes 14 and 15). This result suggests that PC4 may inhibit transcription via competition for binding to the promoter region, which is consistent with its role as a nonspecific DNA-binding protein. Once a stable preinitiation complex was formed on the promoter region, PC4 no longer displaced the stable complex.
PC4 Inhibits the Assembly of a Functional Preinitiation Complex-To further define the mechanism of PC4 repression, we divided the transcription reactions into multiple steps and performed an order-of-addition experiment with Sarkosyl challenge. In previous studies (52,53), Sarkosyl was used to prevent the formation of a functional preinitiation complex. At a concentration of 0.015%, Sarkosyl would inhibit the formation of a preinitiation complex but not the preformed complex in transcription systems reconstituted either with partially purified GTFs (52) or with TBP and f:pol II. 2 Therefore, only a single round of transcription should occur, if 0.015% Sarkosyl was added after formation of the preinitiation complex. Comparison of transcription signals obtained in the absence and presence of Sarkosyl (added at t 50 after preincubation, see the time course outlined at the bottom of Fig. 4) indicated that there are approximately two to three rounds of transcription in our two-component system (Fig. 4, compare lanes 3 and 7). If Sarkosyl was included during the preincubation period (t 0 ), transcription was abolished (Fig. 4, compare lanes 3 and 5) as no functional preinitiation complex could be formed in the presence of 0.015% Sarkosyl. We then asked if PC4 would also inhibit later steps of transcription, such as elongation and reinitiation, by adding PC4 at different time points after the assembly of the preinitiation complex. Consistent with the template challenge experiment (Fig. 3D), PC4 repression could be alleviated if transcription components were allowed to assemble on the promoter region before transcription started (Fig. 4, compare lanes 2, 4, and 10). However, the transcription level in the presence of PC4 was not restored to the original level without PC4 even after preincubation of TBP and f:pol II (Fig. 4, compare lanes 3 and 4, 9 and 10), indicating that PC4 might also affect elongation or reinitiation. When Sarkosyl was added to score for a single round of transcription, no inhibition by PC4 was observed (Fig. 4, compare lanes 11 and 12, lanes 15  and 16, lanes 19 and 20). These results suggest that PC4 also inhibits reinitiation but not elongation, in agreement with its role in preventing the assembly of a functional preinitiation complex.

DISCUSSION
The isolation of a TFIID-deficient f:pol II complex provides us with a unique opportunity to investigate the role of TAFs in the transcriptional process via a two-component transcription system reconstituted with f:pol II and either TBP or TFIID. The observations that TAFs indeed contribute to a high level of activation and that TBP can also mediate activation in our transcription system (Fig. 2) indicate that the previous discrepancy between in vivo yeast studies and in vitro mammalian cell-free transcription systems are not due to species variation. It is likely that some components of the general transcription machinery may affect TBP-mediated activation. Indeed we have recently observed that human TFIIH has a significant effect on TBP-mediated activation, 4 which can also be recapitulated in a transcription system reconstituted with individually purified general transcription factors. In addition to GTFs, our f:pol II also contains minor amounts of human SRBs such as cyclin C and CDK8, and some chromatin remodeling factors including GCN5 and BRG1, 2 suggesting that f:pol II may correspond to previously characterized pol II holoenzymes (5-13). The exact relationship, however, remains to be further defined.

FIG. 3. Template challenge experiments.
A, outline of the twostep incubation in a minimal transcription system. Reactions were performed as described under "Experimental Procedures." B, TFIID was used as the TATA binding activity for the template challenge experiment. The protein factors included in the preincubation period are indicated at the top of the panel. The challenge template (pG 5 HMC 2 AT, labeled as "G 5 HM") was included either at the beginning (pre) or at the end (after) of preincubation. (Ϫ) indicates no template challenge was conducted. Lane 1 represents a standard transcription assay performed without the preincubation period. C, TBP was used as the TATA binding activity for the template challenge experiment. D, a template challenge experiment performed in the presence (ϩ) or absence (Ϫ) of Gal4-VP16 and PC4 during the preincubation period. In this experiment, TBP was used as the TATA-binding factor. The challenge template (labeled "high") is 500 ng of pML⌬53, which was included either at the beginning (pre) or at the end (after) of the preincubation period as outlined at the bottom of the gel. The regular amount (20 ng) of pML⌬53 used for the transcription assay is labeled "low." In our two-component transcription system, PC4 plays a dual role in the transcription process. In the absence of an activator, PC4 acts as a negative cofactor to suppress basal transcription, whereas in the presence of an activator, PC4 acts as a coactivator to mediate transcriptional activation. While this paper was under review, the Roeder laboratory (54) also reported that PC4 could inhibit basal transcription in a system reconstituted with only TBP, TFIIB, TFIIF, and pol II. Although the double-stranded DNA binding activity of PC4 has been shown to be essential for its coactivator function (21), it is not clear which PC4 activity is responsible for repression of basal transcription in the absence of an activator. Since deletion of the N-terminal 21 amino acid residues did not affect PC4 repression of basal transcription, it seems that the N-terminal SEAC (serine/acidic-rich) domain is not involved in transcriptional repression by PC4. 2 It is therefore likely that PC4, being a nonspecific DNA-binding protein, competes with TBP or TFIID for promoter binding. In the study of SV40 DNA replication, it was found that PC4 could inhibit the formation of RNA primers required for both leading and lagging strand DNA synthesis (23). This inhibition, however, could be partially reversed by the addition of excess HSSB (23), indicating that PC4 might repress RNA primer synthesis via its nonspecific DNA binding activity. The observations were consistent with our results that PC4 repression of basal transcription could be overcome by preincubation of transcriptional components with the DNA template (Figs. 3D and 4). Alternatively PC4 may interact with components of the basal transcription machinery, thereby titrating out critical protein factors needed for transcription. Indeed, we also observed that PC4 repression of basal transcription could be overcome by increasing amounts of TFIID, TFIIH, or f:pol II in the transcription reactions. 2 It is likely that activators may help stabilize or enhance the formation of a functional preinitiation complex, thereby overcoming repression by negative factors. The activation domain plays an essential role in this process, since the Gal4 DNA-binding domain alone cannot alleviate repression (Fig. 2B). Despite a low level of activator-dependent transcription, which was detected from pG 5 HMC 2 AT in the absence of PC4 in the minimal transcription system containing only Gal4-VP16, f:pol II, and TBP or TFIID (Fig. 2A, compare lanes 2 and 3, and lanes 7 and  8), the presence of PC4 significantly increased the overall level of activation ( Fig. 2A, compare lanes 2, 3, and 5, and lanes 7, 8,  and 10). It remains to be investigated if SRBs found in f:pol II are responsible for this stimulation or if f:pol II may contain specific cofactors that potentiate the activation mediated by Gal4-VP16.