Mediator and TFIIH Govern Carboxyl-terminal Domain-dependent Transcription in Yeast Extracts*

In Saccharomyces cerevisiae, the RNA polymerase II (RNA Pol II) carboxyl-terminal domain (CTD) is required for viability, and truncation of the CTD results in promoter dependent transcriptional defects. A CTD-less RNA Pol II is unable to support transcription in yeast extracts, but basal transcription reactions reconstituted from highly purified general transcription factors are CTD-independent. To reconcile these two findings, we have taken a biochemical approach using yeast extracts and asked whether there is a factor in the cell that confers CTD-dependence upon transcription. By placing a cleavage site for the tobacco etch virus protease prior to the CTD, we have created a highly specific method for removing the CTD from RNA Pol II in yeast whole cell extracts. Using derivatives of this strain, we have analyzed the role of the Srb8-11 complex, Mediator, and TFIIH, in CTD-dependent basal transcription by either mutation or immunodepletion of their function. We have found that Mediator is a direct intermediary of CTD-dependent basal transcription in extracts and that the requirement for Mediator and the CTD in basal transcription originates from their ability to compensate for a limiting amount of TFIIH activity in extracts.

The largest subunit of Saccharomyces cerevisiae RNA Pol 2 II, Rpb1, shares many common features with the catalytic subunits of other eukaryotic and prokaryotic RNA polymerases. A defining feature of the large subunit of RNA Pol II in eukaryotes is a unique C-terminal domain (CTD) that endows the enzyme with many capabilities critical to virtually every aspect of gene expression. The yeast CTD contains 26 repeats of a heptapeptide sequence that is highly conserved in all eukaryotes. The CTD is largely unstructured in the absence of binding partners (1) and is subject to a cycle of phosphorylation and dephosphorylation by multiple kinases and phosphatases (2). The CTD and its phosphorylation state are critical for the coordination and regulation of transcriptional initiation, elongation and termination, DNA repair, mRNA processing, and mRNA export, through interactions with individual factors directly involved in these processes (3). Complete removal of the CTD from RNA Pol II leads to a lethal phenotype in yeast (4), and the truncation of the CTD (to 10 -12 repeats) leads to distinct phenotypes, such as cold sensitivity and the inability to grow on a variety of carbon sources (4,5). The cellular basis for these truncated CTD phenotypes is directly linked to the inability to express certain subsets of genes, which correlates with particular upstream activating sequences in their promoters (6 -8). Despite extensive biochemical and genetic analysis of CTD function, there is still much that is not understood about its fundamental role in transcription initiation.
Studies in yeast extracts have shown that both activated transcription reactions (9,10) and basal (in the absence of a site-specific DNA-bound activator) transcription reactions (9,11) are highly sensitive to either the truncation or complete removal of the CTD. Transcription systems derived from metazoan cells also exhibit a CTD dependence in vitro, but the observed CTD requirement appears to be promoter-specific (for a review, see Ref. 12) and largely dependent on certain sequences in the core promoter (13). Truncation or removal of the CTD does not compromise the nonspecific polymerase activity of RNA Pol II (11,12), indicating that the CTD does not impact the catalytic ability of the polymerase. In both yeast-and metazoa-based transcription systems in vitro, CTD-dependent transcription seems to correlate with the inability of RNA Pol II, with an absent or truncated CTD, to form a stable preinitiation complex (PIC) (10,14). In contrast to reactions using a yeast extract-based system, basal transcription reconstituted from a minimal set of purified yeast general transcription factors is CTD-independent (11). This finding has led to the proposal that the CTD is opposing a negative effector in extracts (11). Unlike basal transcription, activated transcription in vitro is CTD-dependent in assays both in extracts (9) or reconstituted from the minimal set of purified yeast factors required for activated transcription (15). Both the CTD independence of basal transcription and the CTD dependence of activated transcription in the purified yeast system appear to be linked to the action of the global co-activator complex Mediator.
The Mediator complex is a conserved interface between gene-specific regulatory proteins and the general transcription apparatus of eukaryotes at transcription initiation (16 -18). The roles of Mediator and the CTD in transcription have been shown to be intimately associated, through both genetic and biochemical studies. Selection for genomic suppressor mutations that reverse CTD truncation phenotypes identified the SRB genes (5,19). Mutations in several SRB genes, which encode members of a minimal functional Mediator complex (16), resulted in dominant suppressors of CTD truncation (19). Loss of function mutations in the genes SRB8, SRB9, SRB10, and SRB11, which encode a subcomplex sometimes found associated with Mediator (20 -22), resulted in recessive suppressors of CTD truncation. These studies suggested a role for the Srb8-11 complex as a negative effector of transcription (23,24). Srb10 and -11 are a cyclin/kinase pair whose ability to phosphorylate various transcription factor substrates, including TFIIH (25) and the CTD (26), has been suggested to play role in repression. Purified yeast Mediator was shown to associate with RNA Pol II, at least in part, via the CTD (15,23,27). Mediator and the CTD are also mutually required for activated transcription in a purified system (15). Akin to its role in vitro, specific Mediator subunits are required for certain genespecific regulatory factors to activate transcription in vivo (16), and although, like the CTD, Mediator is not required for basal transcription in a purified system, mutations in individual Mediator subunits can lead to defects in the expression of virtually all genes in vivo (28). Consistent with this observation, mutant Mediators in yeast nuclear extracts also lead to defects in basal transcription (10,29) in vitro. Interestingly, the kinase activity of the general transcription factor TFIIH is also required for transcription in extract-based systems but not in systems reconstituted from purified factors (11,30,31). TFIIH, Mediator, and the CTD are functionally associated by virtue of the ability of Mediator to specifically enhance CTD phosphorylation by the TFIIH kinase (15,27,32) and are thought to be physically associated through an interaction of the Gal11 subunit of Mediator with TFIIH (33). Biochemical (11) and genetic studies (34) have proposed that a negative effector causes Mediator and the CTD to function as general transcription factors in the cell or in extracts but not in purified systems. Although genetic studies have suggested candidates for this effector, such as the Srb8,9,10,11 complex and NC2 complex (34), none of these have been biochemically verified as relieving Mediator-or CTD-dependent basal transcription in vitro.
Previous studies of CTD-dependent transcription have lacked the methodology to directly and specifically remove the CTD of RNA Pol II in extracts. We have addressed this problem by placing a cleavage site for the tobacco etch virus (TEV) protease (35) prior to the CTD of Rpb1 in yeast. This strain can be used for the direct removal of the CTD in extracts as well as for the preparation of purified CTD-less RNA Pol II. We have used these reagents to elucidate the molecular requirements for CTD-dependent transcription in extracts. These studies have revealed that deletion of Srb10, a loss of function suppressor of CTDtruncation in vivo, does not relieve CTD-dependent transcription in extracts. We have found, however, that the Mediator complex is a positive acting factor required for CTD-dependent transcription in extracts. Furthermore, we have found that limiting TFIIH activity leads to a requirement for both Mediator and the CTD in basal transcription both in extracts and in a system reconstituted entirely from purified factors.

MATERIALS AND METHODS
Construction of Strains-Two sets of primers were designed in order to introduce the TEV protease cleavage site into the RPB1 gene by PCR. A segment of the S. cerevisiae RPB1 gene from the naturally occurring KpnI site to the beginning of the CTD was amplified with 5Ј-CAATC-GGTACCGGTGCATTT-3Ј and 5Ј-CGGACTGCAGTGCTTCACC-ATAAGCACCAA-3Ј to introduce a PstI site into the coding sequence. A PstI site and the coding sequence for the TEV protease cleavage site were introduced 5Ј to the coding sequence for the CTD using the primers 5Ј-GGAACTGCAGGAGAATCTTTATTTTCAGGGCCCTACA-TCTCCCGGATTTGGA-3Ј and 5Ј-CTTGAAGCTTAGAAGTTGG-ACGGAC-3Ј to amplify the CTD from RPB1. The first PCR product was digested with KpnI and PstI, and the second was digested with PstI and HindIII. A three-part ligation with the cut PCR products and KpnI/ HindIII-cut pSPL1 (36), which contains the full-length RPB1 gene under the control of its own promoter, yielded full-length RPB1 with the TEV protease site inserted. The EcoRI/HindIII fragment from this plasmid was then ligated into EcoRI/HindIII cut Yplac111 (37) to give the pRTV plasmid. A plasmid shuffle was carried out by transforming the strain Z-26 (a ura3-52 leu2-3 leu2-112 his3-⌬300 rpb1-187::HIS3/pRP1121 (URA3 CEN4 RPB1)) with the RPB1-TEV plasmid followed by a 5-fluoroorotic acid selection for the loss of the plasmid bearing wild-type RPB1.
To construct the RPB1-TEV-⌬srb10 strain, the kanMX4 cassette was amplified from pRS400 (38) using the following primers: 5Ј-TTTGCT-TCCCAATTGAATTAAGGCCGCCTAGTTTTGACGGGAGGAG-AGAG-3Ј and 5Ј-CTATCTTCTGTTTTTCTTTCGAGATGGCTCA-TCTGATGCATTGTTTCCTG-3Ј. The PCR product was then used to transform the RPB1-TEV strain and selected for G418 resistance. The srb10 deletion was confirmed by PCR. To construct the RPB1-TEV TFB4-FLAG strain, a triple FLAG epitope tag was placed at the C terminus of TFB4 by amplifying the 3Ј-end of a previously FLAG-tagged copy of TFB4 from the genomic DNA of the SHY365 strain (39) using the following primers: 5Ј-CAGGGAGAGTTGTTGCCGTT-3Ј and 5Ј-GACGAAGGTTACCTGCTTG-3Ј. The PCR product was then used to transform the RPB1-TEV strain and selected for G418 resistance. The correct integration of the FLAG tag was confirmed by PCR and also by immunoblotting for FLAG-tagged Tfb4. To construct the RPB1-TEV SRB5-FLAG strain, a triple FLAG epitope tag was placed at the C terminus of SRB5 by amplifying the 3Ј-end of a previously FLAG-tagged copy of SRB5 from the genomic DNA of the SHY349 strain (40) using the following primers: 5Ј-GGAGGGTTCCTTTTAAAAGCA-3Ј and 5Ј-GAAGCAAATTGCCAAACA-3Ј. The PCR product was then used to transform the RPB1-TEV strain and selected for G418 resistance. The correct integration of the FLAG tag was confirmed by PCR and also by immunoblotting for FLAG-tagged Srb5. Both the SHY349 and the SH-Y365 strains were kindly provided by Steve Hahn.
Preparation of Purified CTD-less RNA Pol II Using the RPB1-TEV Strain-RNA Pol II was purified as described (41) from the RPB1-TEV strain. Tris chloride (pH 7.5) and zinc chloride were replaced by Tris acetate (pH 7.5) and zinc acetate, respectively. An aliquot of the purified polymerase (34.5 g) was then treated with 1000 units of TEV protease (Invitrogen) according to the manufacturer's protocols. The reaction was carried out for 17 h at 4°C. The digest was further incubated with 8WG16 beads (41) to deplete the solution of free and uncleaved CTD. The glycerol concentration of the supernatant was readjusted to 10%, the dithiothreitol concentration was readjusted to 2 mM, and the solution was concentrated in a YM-50 centricon (Millipore Corp.).
Yeast Extract Preparation-Transcription-competent whole cell extracts (WCE) were prepared from the wild type and modified yeast strains using the method of Kong and Svejstrup (42). The procedure was modified such that a special protease inhibitor mixture (40 M pefabloc SC (Roche Applied Science), 2 M pepstatin A) was used in the final dialysis buffer in order not to interfere with TEV protease cleavage. The whole cell extract was diluted to a concentration of 22 mg/ml, and 1.32 mg of whole cell extract was treated with 100 units of TEV protease (Invitrogen). The reaction was carried out at 4°C. Aliquots were removed at 0, 3, and 21 h, respectively, and directly used for immunoblots and transcription assays.
The RY260 (43) strain carrying the rpb1-1 mutation was used to prepare an extract specifically depleted of RNA Pol II activity. Cells were grown at 25°C to an A 600 of 1.5, after which they were rapidly shifted to 37°C and allowed to grow at 37°C for another 1.5 h. DE400 extracts were prepared from a whole cell extract by a single-step gradient fractionation of the WCE on DEAE-Sepharose (Amersham Biosciences) as previously described (44).
For preparation of extracts from the RPB1-TEV SRB5-FLAG and RPB1-TEV TFB4-FLAG strains, cells were grown at 30°C to an A 600 of 1.5, and a whole cell extract and DE400 extract were prepared as above and with the special protease inhibitor mixture (40 M pefabloc SC (Roche Applied Science), 2 M pepstatin A) in the final dialysis buffer.
Immunodepletion from Extracts-Anti-FLAG M2-agarose beads (Sigma) were washed in WCE dialysis buffer (20 mM Hepes-KOH (pH 7.6), 10 mM MgSO 4 , 1 mM EGTA, 20% glycerol, 100 mM (NH 4 ) 2 SO 4 , 0.5 mM dithiothreitol, and special protease inhibitors) and added to DE400 fractions at a ratio of 20 l of beads to 80 l of FLAG-tagged DE400 fraction adjusted to 100 mM (NH 4 ) 2 SO 4 . After incubation for 2 h with gentle inversion at 4°C, beads were removed by microcentrifugation, and an amount of beads equal to that used initially was added to the supernatant. The extracts were incubated for an additional 2 h at 4°C. Beads were removed by centrifugation, and the doubly depleted DE400 fraction was used in immunoblots and transcription assays. Undepleted control DE400 fractions were made by following the same procedure as above but using unmodified agarose beads.
Transcription Assays-Basal transcription reconstituted from purified yeast general transcription factors (41) was measured using the G-less cassette assay as previously described (15) with the following modifications. The final salt concentration in the reaction buffer was 150 mM KOAc, and the reactions were preincubated for 10 min in the absence of nucleotides followed by a 10-min reaction time after the addition of the nucleotides. The general transcription factors used for all reactions in this study came from identical aliquots of a single preparation. The basal transcription reactions based on the transcription-competent whole cell extract or the DE400 extract were performed as previously described (44) with the following modifications. The final salt concentration in the reaction buffer was 100 mM KOAc, and the reactions were preincubated for 10 min in the absence of nucleotides followed by a 10-min reaction time after the addition of the nucleotides. Purified wild type or CTD-less RNA Pol II was added to the DE400based reactions in amounts normalized for their nonspecific activity. RNA Pol II nonspecific initiation/chain elongation assays were performed as described (11). When the nonspecific assay was performed in extracts, the contribution of RNA Pol II to the total polymerase activity was determined by subtracting the signal in the presence of 5.5 mM ␣-amanatin from the signal in the absence of ␣-amanatin.

A TEV Protease Site in Rpb1 Allows for the Specific Removal of the CTD from RNA Pol II in Its Purified Form and in Extracts-
To create a RNA Pol II with a CTD that could be readily removed by a site-specific protease, the seven-amino acid TEV protease consensus cutting sequence was inserted into the coding sequence of the gene RPB1 by replacing 10 amino acids just prior to the CTD (Fig. 1A). The plasmid coding for the modified RPB1 (pRTV) was exchanged into a yeast strain by plasmid shuffle, such that the growth of the new strain (RPB1-TEV) was dependent on the RPB1-TEV gene. The growth characteristics of the modified strain were essentially identical to the parental strain supported by wild type RPB1. To test the ability of the TEV protease to cleave the CTD, we purified RNA Pol II from the RPB1-TEV strain using standard purification protocols (41). The purified RNA Pol II was subjected to cleavage with TEV protease, and the results were monitored by SDS-PAGE using silver staining. Prior to all subsequent assays, we incubated the CTD-less RNA Pol II from the cleavage reaction with 8WG16 (␣-CTD) beads to ensure that free CTD and remaining CTD-bearing RNA Pol II were cleared from the sample. The mobility shift in the Rpb1 subunit caused by the TEV protease treatment corresponded to the change in molecular mass (ϳ21 kDa) expected for loss of the CTD (Fig.  1B). At this point, the TEV protease itself could be removed by size exclusion or by virtue of a His 6 tag on its N terminus in conjunction with Ni 2ϩ -agarose. We found, however, that the TEV protease did not interfere with any subsequent assays of the CTD-less RNA Pol II. Basal transcription assays in a purified transcription system (Fig. 1C) confirmed that the CTD-less RNA Pol II generated from the RPB1-TEV strain had activity equivalent to that of wild type RNA Pol II when normalized for nonspecific polymerase activity.
To confirm that the TEV protease could cleave the CTD from RNA Pol II in a crude yeast extract, we prepared transcription-competent whole cell extracts from the RPB1-TEV strain as described (42). TEV protease was added to one reaction, whereas a second identical control reaction was incubated in parallel in the absence of the protease. Cleavage of the CTD was followed by immunoblotting using the 8WG16 (␣-CTD) antibody and monitoring the band at ϳ200 kDa. Over time, the intensity of the Rpb1 band in the control reaction remained constant, whereas the Rpb1 band in the TEV protease reaction decreased, markedly reflecting CTD cleavage (Fig. 2). Other antibodies directed against a RNA Pol II subunit (Rpb3), a Mediator subunit (Med7), and a nontranscription protein (Kar3) were used as loading controls and did not vary during the course of the digest. The detached CTD peptide appeared to be degraded, since we were unable to detect a new band of the appropriate molecular mass (ϳ21 kDa) with the 8WG16 antibody upon cleavage (data not shown). To show that TEV protease cleavage of the CTD was dependent on insertion of the cleavage site, a control TEV protease digest was performed on an extract prepared from a wild type RNA Pol II strain and showed no removal of the CTD over time (Fig. 2). Bands marked with an asterisk result from impurities in the TEV protease. C, purified basal transcription assays with wild type and CTD-less RNA Pol II purified from the RPB1-TEV strain. The transcription reactions in C are from the same exposure of a single gel, and the lanes are separated and rearranged for the purposes of the figure. All transcription reactions were performed using both the pGCN4:CG-and pJJ470:CG-DNA templates (27).

Direct Removal of the CTD from RNA Pol II in Extracts Results in a
Defect in Basal Transcription-To evaluate the effect of removing the CTD of RNA Pol II in extracts, we performed both nonspecific polymerase assays and specific basal transcription assays using the control and TEV-digested WCE shown in Fig. 2. Previous efforts to use site-specific proteases in transcription-competent extracts have been hampered by the protease having effects on transcription independent of the intended substrate (11). To confirm that the TEV protease did not cause cleavage site-independent transcription defects, we prepared a WCE from the parent strain of RPB1-TEV, in which growth was supported by wild type RNA Pol II. TEV protease treatment of the WCE from the wild type strain showed identical basal transcription in the specific assay to the untreated control sample (Fig. 3A). However, since the CTD is cleaved from RNA Pol II in the RPB1-TEV extracts, basal transcription in the specific assay eventually decreases to ϳ20% of what is observed in the control RPB1-TEV extracts incubated under identical conditions in the absence of the TEV protease (Fig. 3, B and C). During the course of the digest, both the control and the TEV protease-treated sample showed no significant change in nonspecific RNA Pol II activity, indicating that the amount of catalytically active RNA Pol II remained constant despite the absence of the CTD (Fig. 3C). Removal of the CTD by the TEV protease dramatically impairs basal transcription and appears to be highly specific for the TEV protease cognate site, which is not present in the wild type yeast proteome.
Removal of Srb10 Does Not Reverse CTD-dependent Transcription in Extracts-Since a loss of function mutation in the SRB10 gene suppressed the transcriptional defects and growth phenotypes of CTD truncation in vivo (20), we deleted the gene encoding Srb10 from the RPB1-TEV strain to determine the effect of Srb10 on CTD-dependent transcription in extracts. The RPB1-TEV-⌬srb10 strain was viable and had phenotypes, such as "clumps" of cells in liquid culture, associated with SRB10 deletion (46). A WCE was prepared from the RPB1-TEV-⌬srb10 strain, treated with TEV protease, and monitored for basal transcription and nonspecific RNA Pol II activity. Basal transcription in the RPB1-TEV-⌬srb10 extract decreased in correspondence to the loss of the CTD with a profile similar to that of the WCE prepared from the RPB1-TEV strain, whereas nonspecific RNA Pol II activity was again unaffected (Fig. 4). This finding shows that loss of Srb10 function does not "suppress" the transcriptional defects associated with loss of the CTD in yeast extracts.

The Addition of Purified TFIIH to Extracts Relieves CTD-dependent
Basal Transcription-In order to set up a biochemical activity assay to identify the factor in extracts that conferred CTD dependence on basal transcription, we prepared extracts that were depleted of RNA Pol II activity. Using the rpb1-1 temperature-sensitive strain, extracts were prepared from cells that were treated at 37°C for 1.5 h directly before harvesting to inactivate the endogenous mutant RNA Pol II (43). After a crude step fractionation on DEAE-Sepharose (44), the DE400 fraction prepared from the rpb1-1 strain was demonstrated to be defective in both the nonspecific RNA Pol II assay (supplemental Table I) and in the basal transcription assay (Fig. 5A). Similar to what was previously observed in nuclear extracts (11) and in our CTD cleavage experiments (Fig. 3), the addition of purified wild type RNA Pol II, but not CTD-less Immunoblots showing the site-dependent cleavage of the CTD by TEV protease. Yeast extracts made from the RPB1-TEV strain were treated (lanes 2, 4, and 6) or mock-treated (lanes 1, 3, and 5) with TEV protease, and samples were removed at 0, 3, and 21 h. Samples were probed by Western blot using antibodies against the CTD (8WG16), Kar3, Rpb3, and Med7. As a control to show that the CTD cleavage was dependent on the presence of the TEV protease site, yeast extracts made from a wild type (WT) RPB1 strain were treated (lanes 8, 10, and 12) or mock-treated (lanes 7, 9, and 11) with TEV protease and samples removed at 0, 3, and 21 h. Samples were probed by immunoblotting using antibodies against the CTD (8WG16), Rpb3, and Med7.  2, 4, and 6) or mock-treated (lanes 1, 3, and 5) with TEV protease, and samples were removed at 0, 3, and 21 h for direct use in the basal transcription assay. The samples were identical to those analyzed in Fig. 2 (lanes 7-12). B, basal transcription in a WCE prepared from a wild type RPB1-TEV strain was impaired by removal of the CTD using TEV protease. WCE made from the RPB1-TEV strain was treated (lanes 2, 4, and 6) or mock-treated (lanes 1, 3, and 5) with TEV protease, and samples were removed at 0, 3, and 21 h for direct use in the basal transcription assay. The samples were identical to those analyzed in Fig. 2  (lanes 1-6). C, plot of the CTD independence of nonspecific RNA Pol II activity and the CTD dependence of basal transcription in a WCE prepared from the RPB1-TEV strain. Nonspecific RNA Pol II activity was measured for the samples described in B, and the results are plotted as the ratio of the activity of the TEV protease-treated sample to the mock-treated sample, where the ratios were normalized such that the 0 h time point was set to 100% (absolute values for the nonspecific activity are shown in supplemental Table I). The basal transcription shown in B was quantified using a PhosphorImager (Amersham Biosciences), and the results are plotted as the ratio of the activity of the TEV protease-treated sample to the mock-treated sample, where the ratios were normalized such that the 0 h time point was set to 100%. All reactions in Fig. 3 were performed using a single DNA template, pGCN4:CG- (27).
RNA Pol II, fully restored basal transcription to the rpb1-1 extracts (Fig.  5A). The CTD dependence in these reactions could originate from a positive acting factor that enhances basal transcription in a CTD-dependent fashion or a negative acting factor that is antagonized in a CTD-dependent manner. If a negative effector of transcription was present in the extracts, then adding the extracts back to a basal transcription reaction composed from a complete set of purified general transcription factors and CTD-less RNA Pol II should lead to a decrease in transcription. However, when the rpb1-1 extract was added to purified reactions with CTD-less RNA Pol II, little change was observed in the signal (Fig. 5B, lanes 1 and 2). One possible explanation for this result was that the quantity of the putative negative acting factor was insufficient to act on the additional purified general transcription factors. A second, but not mutually exclusive, explanation is that CTD-dependent basal transcription results from a limiting amount or activity of one or more of the general transcription factors. In this scenario, the additional purified factors in the reaction would circumvent the requirement for the CTD as well as any other factors working in conjunction with the CTD. To determine whether one particular purified general transcription factor added to the reaction was reversing CTDdependent basal transcription, we systematically removed each of the additional factors and assayed for CTD-dependent transcription. Since we already had supplemented our DE400 extract reactions in Fig. 5A with recombinant TBP and TFIIB, only purified TFIIF, TFIIE, and TFIIH were candidates. Although the addition of rpb1-1 extracts to purified reactions lacking either purified TFIIE and/or TFIIF restored transcription, it did not lead to CTD-dependent basal transcription. The addition of rpb1-1 extracts to reactions lacking purified TFIIH, however, only restored high levels of basal transcription to the reactions that utilized wild type but not CTD-less RNA Pol II (Fig. 5B, lanes 3 and  4). A comparable result could be achieved by adding back purified TFIIH in the absence of the other general transcription factors (Fig. 5C).
One explanation for the ability of purified TFIIH to restore transcription to extracts containing CTD-less RNA Pol II was that an increase in the total amount of TFIIH in the reactions was circumventing the requirement for the CTD. A second explanation was that the TFIIH in extracts was in a "repressed" conformation that required the CTD for activity, whereas the purified TFIIH was "constitutively active" in the absence of the CTD. To explore these possibilities, we FLAG-tagged the Tfb4 subunit of TFIIH in the RPB1-TEV strain. Using this strain, we prepared DE400 extracts in which TFIIH was depleted in the presence or absence of the CTD on RNA Pol II (Fig. 6A). As judged by both the ␣-FLAG and ␣-Tfb1 antibodies, TFIIH was significantly depleted from these extracts while the amount of Mediator and RNA Pol II stayed constant (Fig. 6A). To obtain an estimate of the relative amounts of TFIIH in entirely purified factor-based and extract-based transcription assays, we developed immunoblots with an ␣-Tfb1 antibody. Although our purified TFIIH has a His 6 tag that causes it to have a slightly lower mobility on SDS-PAGE, the amounts of TFIIH in equal volumes of nondepleted extracts and the purified preparation appeared to be comparable (Fig. 6A, lanes 1 and 6) based on Tfb1. However, since Tfb1 also exists in the repairosome complex (47), the Tfb1 signal in extracts may not accurately represent the exact amount of TFIIH. In an otherwise wild type extract, depletion of TFIIH led to an almost complete cessation of basal transcription that could be restored with the addition of purified TFIIH (Fig. 6B).
Using the TFIIH-depleted extracts shown in Fig. 6A, we added back increasing amounts of purified TFIIH to basal transcription reactions containing either wild type or CTD-less RNA Pol II. The addition of 0.6 l of purified TFIIH restores basal transcription to a TFIIH-depleted extract with wild type RNA Pol II (Fig. 6C, lane 5) but still does not restore basal transcription to reactions with CTD-less RNA Pol II (Fig.  6C, lane 6). Both the level of basal transcription with wild type RNA Pol II and the pattern of CTD dependence were similar to those of the undepleted extracts with no added TFIIH (Fig. 6C, lanes 5 and 6 versus  lanes 1 and 2). In reactions with CTD-less RNA Pol II, the addition of a greater amount of purified TFIIH (2 l) restored basal transcription to either depleted extracts or undepleted extracts (Fig. 6C). The endogenous TFIIH in the control extract combined with purified TFIIH (Fig.  6C, lane 3) did not enhance basal transcription with the wild type RNA Pol II to levels higher than purified TFIIH in the depleted extract (Fig.  6C, lane 7). Furthermore, the addition of as little as 0.2 l of purified TFIIH was able to restore robust basal transcription in depleted extracts with wild type RNA Pol II, indicating that basal transcription under these conditions required only low concentrations of TFIIH. In addition, at no point during a titration of purified TFIIH to CTD-less Pol II reactions was there an amount that restored transcription to the unde-  2, 4, and 6) or mock-treated (lanes 1, 3, and 5) with TEV protease, and samples were removed at 0, 3, and 21 h for direct use in the basal transcription assay. All reactions in Fig. 4 were performed using a single DNA template, pGCN4:CG- (27). B, plot of the CTD independence of nonspecific RNA Pol II activity and the CTD dependence of basal transcription in WCE prepared from the RPB1-TEV-⌬srb10 strain. Nonspecific RNA Pol II activity was measured for the samples described in A, and the results are plotted as the ratio of the activity of the TEV protease-treated sample to the mock-treated sample, where the ratios were normalized such that the 0 h time point was set to 100% (absolute values for the nonspecific activity are shown in supplemental Table I). The basal transcription shown in A was quantified using a PhosphorImager (Amersham Biosciences), and the results are plotted as the ratio of the activity of the TEV protease-treated sample to the mock-treated sample, where the ratios were normalized such that the 0 h time point was set to 100%. pleted extracts but not the depleted extracts. All of the above evidence suggests that there is limiting TFIIH activity in the extracts rather than the TFIIH activity being sequestered in an inactive form released by a CTD-dependent mechanism. The presence of the CTD is essential for basal transcription in extracts at low TFIIH concentrations. We also observed that although the purified TFIIH restored transcription to the depleted CTD-less extract, the signal was much lower than the depleted wild RNA Pol II extract. These findings indicated that a CTD-dependent positive acting factor that compensates for limiting TFIIH could be responsible for CTD-dependent transcription in extracts.
Mediator Is a Positive Acting Factor in Extracts That Is Dependent on CTD and TFIIH-In reactions composed from a minimal set of purified general transcription factors, the addition of purified Mediator leads to an enhancement of basal transcription (15,27), and this enhancement is dependent on the CTD (15). To elucidate the role of Mediator in CTDdependent basal transcription in extracts, we FLAG-tagged the Srb5 subunit of Mediator in the RPB1-TEV strain and prepared DE400 extracts from these strains. Mediator was depleted from these extracts in the presence or absence of the CTD on RNA Pol II (Fig. 7A). As judged by both the ␣-FLAG and ␣-Rgr1 antibodies, Mediator was significantly depleted from these extracts while the amount of TFIIH and RNA Pol II stayed constant (Fig. 7A). Since TFIIH concentration apparently has a strong influence on CTD-dependent basal transcription, we also established that the RPB1-TEV TFB4-FLAG extract and the RPB1-TEV SRB5-FLAG extract had identical amounts of the Tfb1 subunit of TFIIH (Fig. 7A, lanes 5 and 7). Similar to previous observations in nuclear extracts from Mediator mutants (10,29), depletion of Mediator led to a significant decrease in basal transcription in extracts with wild type RNA Pol II (Fig. 7B). The depletion had no discernable effect on basal transcription in extracts with CTD-less RNA Pol II due to the low or nonexistent starting signal (Fig. 7B, lane 2). The addition of purified Mediator restored full basal transcription to the depleted extract with wild type RNA Pol II but had little effect on the undepleted or CTD-less RNA Pol II reactions (Fig. 7B, lanes 5-8).
Mediator appears to function as a CTD-dependent positive acting factor in extracts. Based on this finding, one would predict that if the CTD requires Mediator to have its effect on basal transcription, then Mediator-dependent basal transcription should also be mitigated by the addition of purified TFIIH. When purified TFIIH was added to the Mediator-depleted extracts, we observed that the Mediator dependence of basal transcription was reduced (Fig. 7C, lanes 1 and 3 versus lanes 5  and 7) in a similar manner to the reduction in the CTD dependence of basal transcription (Fig. 7C, lanes 3 and 4 versus lanes 7 and 8). Mediator is not required for high levels of TFIIH to restore basal transcription with CTD-less Pol II (Fig. 7C, lane 6), demonstrating that Mediator is not necessary for the TFIIH bypass of the CTD requirement. These data suggest that Mediator could be directly functioning as an intermediary for CTD-dependent basal transcription in extracts. Alternatively, Mediator could be indirectly serving as an intermediary for CTD-dependent basal transcription in extracts by facilitating the action of a yet unidentified factor. To test whether Mediator functions as a direct intermediary of CTD-dependent basal transcription, we tested the CTD and TFIIH dependence of Mediator's ability to enhance basal transcription in the purified basal transcription system.
Using our minimal purified basal transcription system, we titrated TFIIH into reactions containing either wild type or CTD-less RNA Pol II, in the absence or presence of purified Mediator (Fig. 8A). At the highest TFIIH concentrations (Fig. 8A, lanes 9 -12), neither Mediator nor the CTD were required for basal transcription. At these TFIIH concentrations, however, Mediator does enhance basal transcription by ϳ7-fold in the presence of the CTD but has no effect in its absence as was previously shown (15). At lower concentrations of TFIIH, basal transcription is below the level of detection in all reactions except those containing both wild type RNA Pol II and Mediator (Fig. 8A). At the very least, the enhancement of basal transcription by Mediator has increased by 8 -10-fold at the lower TFIIH concentrations versus the enhancement observed at high TFIIH concentrations. If the level of enhancement had remained the same at the lower TFIIH concentrations, the FIGURE 5. The addition of purified TFIIH reverses CTD-dependent transcription in extracts. A, basal transcription in DE400 extracts prepared from the rpb1-1 strain is CTD-dependent. The rpb1-1 DE400 extract has no basal transcription activity in the absence of added RNA Pol II (lane 1). Adding back wild type RNA Pol II (145 ng) restored basal transcription to the rpb1-1 DE400 extract (lane 2), whereas the addition of a equivalent amount of CTD-less RNA Pol II (145 ng) failed to restore a comparable level of basal transcription to the rpb1-1 DE400 extract (lane 3). B, basal transcription in reactions reconstituted from an entire set of purified general transcription factors and rpb1-1 DE400 extracts was CTD-dependent only when purified TFIIH is excluded. Basal transcription reactions reconstituted from CTD-less RNA Pol II and purified TBP, TFIIB, TFIIE, TFIIF, and TFIIH were not actively repressed by a factor in the rpb1-1 DE400 extract (lanes 1 and 2). Removal of purified TFIIH from these reactions restored CTDdependent basal transcription by disabling the reaction exclusively with CTD-less RNA Pol II but not wild type RNA Pol II (lanes 3 and 4). C, the addition of purified TFIIH restored basal transcription to reactions containing rpb1-1 DE400 extracts and CTD-less RNA Pol II (lanes 4 -6) and had a considerably smaller effect on comparable reactions with wild type RNA Pol II (lanes 1-3). For each part of the figure, the reactions were from the same exposure of a single gel, the lanes were separated and rearranged for the purposes of the figure, and reactions were performed using both the pGCN4:CGand pJJ470:CG-DNA templates (27). transcription signal would have been well within the limits of detection. Akin to what was observed in the extract based system, the purified system shows that under conditions of low TFIIH activity, basal transcription is highly dependent on both Mediator and the CTD. Neither Mediator nor the CTD appear to be able to exert their influence on basal transcription without the other. There is, however, a very small amount of basal transcription that can be reproducibly observed in the reaction with Mediator and CTD-less RNA Pol II under limiting TFIIH conditions (Fig. 8A, lane 8). It is unclear whether this effect is a result of Mediator working inefficiently in the absence of the CTD or a small amount of contaminating wild type RNA Pol II in the CTD-less preparation.
The requirement for Mediator and the CTD in extracts appears to originate from limiting TFIIH activity. To determine whether Mediator exclusively compensates for limiting TFIIH activity, we monitored Mediator enhancement of basal transcription in the purified system with limiting amounts of several other components of the general tran-scription machinery. Limiting TBP (Fig. 8B), TFIIB (Fig. 8C), and TFIIF (Fig. 8C) all lead to an increase in the -fold enhancement of basal transcription by Mediator and an apparent "requirement" for Mediator in transcription under these conditions. In contrast to the other components of the general transcription machinery, titrating down the amount of RNA Pol II did not increase the -fold enhancement of basal transcription by Mediator (Fig. 8D).

DISCUSSION
This work advances both the methodology used to study the CTD in yeast extracts and the knowledge of CTD and Mediator function in transcription initiation. Previous studies have used several different techniques to analyze CTD-dependent transcription in extracts, each with their own limitations (9,10,11,14,43,48). Placing a specific protease site prior to the CTD provides an optimal methodology to prepare extracts from an essentially wild type strain and then look for functional effects directly related to CTD removal. A previous attempt to utilize a FIGURE 6. CTD-dependent basal transcription in TFIIH-depleted extracts is governed by the amount of purified TFIIH added. DE400 extracts prepared from the RPB1-TEV TFB4-FLAG strain were immunodepleted of TFIIH in the presence and absence of the CTD on RNA Pol II and analyzed for basal transcription. A, immunoblotting analysis of purified TFIIH and TFIIH-depleted DE400 extracts prepared from the RPB1-TEV TFB4-FLAG strain containing either wild type (WT) or CTD-less RNA Pol II. Pure TFIIH was prepared from a TFB1-6His strain as described (41). B, basal transcription was reconstituted using RPB1-TEV TFB4-FLAG DE400 extracts, containing wild type RNA Pol II, and an entire set of purified general transcription factors, lacking TFIIH. TFIIH-immunodepleted extract resulted in a deficiency in basal transcription (lane 2) that could be recovered by the addition of purified TFIIH (lane 3). C, two different concentrations of purified TFIIH were added to the TFIIH-depleted extract with CTD-less RNA Pol II and in vitro transcription assays performed as in B (lanes [5][6][7][8]. Basal transcription in the control extracts was also assayed in the absence (lanes 1 and 2) and presence (lanes 3 and 4) of purified TFIIH. For each part of the figure, the reactions or immunoblot were from the same exposure of a single gel, and the lanes were separated and rearranged for the purposes of the figure. All transcription reactions in Fig. 6 were performed using both the pGCN4:CG-and pJJ470:CG-DNA templates (27).
factor Xa site was not amenable to study RNA Pol II transcription in a crude system, since Factor Xa, added to yeast extracts, affected transcription independently of the factor Xa site (11). The use of TEV protease in this study provided an effective manner of removing the CTD from RNA Pol II while avoiding secondary effects on transcription. Given the specificity of TEV protease, it is also reasonable to believe that extracts from the RPB1-TEV strain would be useful for studying other CTD-dependent phenomena, such as mRNA processing, in extracts. The RPB1-TEV strain also provided a viable and highly specific alternative to other methodologies (11,14) for preparing purified CTD-less RNA Pol II. Our observation of basal transcription defects upon CTD cleavage provides the first direct demonstration of CTD-dependent transcription in an otherwise wild type extract. Another advantage of this system is that CTD-dependent functions can be directly monitored in extracts prepared from RPB1-TEV strains that harbor various mutations.
In this study, the deletion of the Srb10 protein, a recessive suppressor of CTD-truncation in vivo, was unable to relieve basal transcription defects associated with CTD cleavage by TEV in extracts. This finding is consistent with gene expression microarray studies indicating that Srb10 most likely exerts its influence on transcription in an upstream activating sequence-dependent manner at specific genes (28) in vivo rather than as a general repressor of basal transcription. This point is further supported by transcription assays in nuclear extracts obtained from a series of CTD truncation mutants, which demonstrated that activated transcription was more sensitive than basal transcription to shortening of the CTD (9). Our finding that Srb10 deletion cannot compensate for the complete loss of the CTD in extracts also indicates that partial truncation of the CTD and complete removal may cause different molecular consequences at a promoter. The existence of CTD-independent promoters in metazoan transcription systems, and CTD-independent basal transcription in reactions composed from a minimal set of purified general transcription factors led to the proposal that the action of the general transcription FIGURE 7. Mediator-dependent basal transcription in extracts is governed by the CTD and the activity of TFIIH. DE400 extracts prepared from the RPB1-TEV SRB5-FLAG strain and were immunodepleted of Mediator in the presence and absence of the CTD on RNA Pol II and analyzed for basal transcription. A, immunoblotting analysis of purified Mediator, Mediator-depleted DE400 extracts prepared from the RPB1-TEV Srb5-FLAG strain containing either wild type (WT) or CTD-less RNA Pol II, and the RPB1-TEV TFB4-FLAG DE400 extract (same sample as shown in Fig. 6A, lane 6). Purified Mediator was prepared as described (41). B, basal transcription was reconstituted from control and ␣-FLAG-immunodepleted RPB1-TEV SRB5-FLAG DE400 extracts with wild type or CTD-less RNA Pol II and no purified TFIIH added (lanes 1-4). Purified Mediator (0.1 l) was added to an identical set of basal transcription reactions (lanes 5-8). C, two sets of basal transcription reactions were reconstituted using control and ␣-FLAG immune depleted RPB1-TEV SRB5-FLAG DE400 extracts, containing either wild type or CTD-less RNA Pol II, and an entire set of purified general transcription factors, lacking TFIIH. The first set of reactions (lanes 1-4) had no added TFIIH, whereas 2 l of purified TFIIH was added to each of the reactions in the second set (lanes [5][6][7][8]. For each part of the figure, the transcription reactions or immunoblot was from the same exposure of a single gel, and the lanes were separated and rearranged for the purposes of the figure. All transcription reactions in Fig. 7 were performed using both the pGCN4: CG-and pJJ470:CG-DNA templates (27).
factors does not directly require the CTD (12). Other experiments suggested that CTD-dependent basal transcription in a yeast extract-based system might be a result of a negative effector of transcription (11). Hence, it was surprising that adding back an extract, which demonstrated CTD-dependent transcription on its own, to an active purified basal transcription system with CTD-less RNA Pol II did not lead to a down-regulation of transcription (Fig. 5B). Omission of purified TFIIH restored CTD-dependent basal transcription in reactions containing both extract and purified general factors. This finding suggested that either the form or amount of TFIIH in extracts requires the CTD to efficiently perform its role in transcription (Fig. 5). If the CTD was "activating" a repressed form of TFIIH in extracts, then one would predict that the amount of "constitutively active" purified TFIIH necessary to restore basal transcription to a TFIIH-depleted extract would be identical for both the CTD-less and wild type RNA Pol II. However, the TFIIH-depleted extract with wild type RNA Pol II required less purified TFIIH to restore basal transcription (equivalent to the level in undepleted extracts) than was required to restore detectable levels of basal transcription to the TFIIH depleted extracts with CTD-less RNA Pol II (Fig. 6C). We conclude that there is very little endogenous TFIIH activity in extracts compared with the amounts typically required for our minimal purified basal transcription reactions. This finding suggests that there is a CTD-dependent positive acting factor in extracts that compensates for limiting amounts of TFIIH. Mediator was a leading candidate for this positive acting factor, since we have previously observed that the complex can enhance basal transcription in a purified transcription system in a CTD-dependent manner (15).
Depletion of Mediator significantly decreased basal transcription in extracts, and adding back the highly purified complex exclusively restored basal transcription to the depleted extract with wild type RNA Pol II but had little effect on either the depleted extract with CTD-less RNA Pol II or the undepleted extract with wild type RNA Pol II. These observations support the concept that Mediator is an intermediary in CTD-dependent basal transcription in extracts (Fig. 7). If Mediator was acting in the same pathway as the CTD in extracts, then Mediator-dependent basal transcription should also be mitigated by the addition of purified TFIIH. Indeed, the same amount of purified TFIIH required to relieve CTD-dependent basal transcription also mitigated Mediatordependent basal transcription in a nearly identical fashion (Fig. 7C). Furthermore, Mediator is not only necessary but is also sufficient to confer this CTD dependence on transcription, since Mediator was required for CTD-dependent basal transcription in purified reactions under TFIIH-limiting conditions (Fig. 8A).
It has been previously demonstrated that Mediator was critical for both the formation of a stable preinitiation complex (10, 29) and a functional reinitiation scaffold (49). Our results show that incorporating and stabilizing TFIIH in preinitiation complexes and reinitiation scaffolds may be a mechanism by which Mediator performs this function. This mechanism most likely involves a direct physical interaction between TFIIH and Mediator, since earlier studies have shown that a direct functional interaction between Mediator, TFIIH, and the CTD leads to enhanced phosphorylation of the CTD by the TFIIH kinase in vitro (15,27,32). Protein binding assays in vitro have also demonstrated an interaction between the Gal11 subunit of Mediator and TFIIH (33). The fact FIGURE 8. Mediator-dependent basal transcription in a purified transcription system is governed by the activity of TFIIH. The effect of Mediator on basal transcription reactions reconstituted from entirely purified general transcription factors and either wild type or CTD-less RNA Pol II was tested under conditions where the amount of the individual components was systematically varied. If not specified in the figure, the standard amount of each factor is as follows: TBP (1 l/reaction), TFIIB (1 l/reaction), wild type RNA Pol II (0.25 l/reaction), TFIIF (2 l/reaction), and TFIIH (1.9 l/reaction). A, increasing amounts of purified TFIIH were added to purified basal transcription reactions containing either wild type or CTD-less RNA Pol II in the presence and absence of Mediator. B, decreasing amounts of TBP were added to purified basal transcription reactions containing wild type RNA Pol II in the presence and absence of Mediator. The reactions in 8B were performed and analyzed in parallel with the samples in 8A. C, decreasing amounts of TFIIF (lanes 3-6) and TFIIB (lanes 3 and 4 and lanes 7-10) were added to purified basal transcription reactions containing wild type RNA Pol II in the presence and absence of Mediator. A set of control reactions with low and high amounts of TFIIH (lanes 1-4) was used to compare the enhancement of basal transcription with those in A and B. D, varying amounts of wild type RNA Pol II were added to purified basal transcription reactions in the presence and absence of Mediator. The reactions in D were run in parallel with the samples in C. For each part of the figure, the transcription reactions were from the same exposure of a single gel, and the lanes were separated and rearranged for the purposes of the figure. Even at longer exposures, no transcription signal could be observed above the noise for A (lanes 1, 3, 4, 5, and 7), B (lanes 3 and 5), and C (lanes 1, 5, 7, and 9). All transcription reactions in Fig. 8 were performed using both the pGCN4:CG-and pJJ470:CG-DNA templates (27). that subunits of TFIIH are among the least abundant components of the general transcription machinery (21) and that the core subunits of TFIIH exist both in a complex dedicated to DNA repair and a complex dedicated to transcription (47) suggests that transcriptionally active TFIIH may be limiting in cells. Hence, the general requirement for Mediator and the CTD for transcription in vivo may also be related to limiting TFIIH activity. The observation that Mediator can also compensate for limiting amounts of other general transcription factors, such as TBP, TFIIB, and TFIIF (Fig. 8), suggests a general mechanism for Mediator to antagonize any repressor that functions by limiting the activity of a general transcription factor. An example of this phenomena may be the suppression of a temperature-sensitive mutation in SRB4, which causes a general defect in RNA Pol II transcription, by a mutation in the BUR6 subunit of the NC2 complex (34). NC2 is thought to function as a repressor of transcription by limiting the activity of TBP by interfering with its ability to interact with TFIIB (50, 51) and/or by stabilizing the binding of a TBP-NC2 complex to non-TATA box DNA (52). At promoters where NC2 can function as a repressor, Mediator may be facilitating transcription by compensating for limiting TBP activity. However, TBP activity may no longer be limiting in NC2 mutants, and the requirement for a fully functional Mediator complex would be relaxed.