A role of the C-terminal part of p44 in the promoter escape activity of transcription factor IIH.

The p44 subunit plays a crucial role in the overall activity of the transcription/DNA repair factor TFIIH: on the one hand its N-terminal domain interacts with and regulates the XPD helicase (, ); on the other hand, as shown in the present study, it participates with the promoter escape reaction. Mutagenesis along with recombinant technology using the baculovirus/insect cells expression system allowed us to define the function of the two structural motifs of the C-terminal moiety of p44: mutations within the C4 zinc finger motif (residues 291-308) prevent incorporation of the p62 subunit within the core TFIIH. Double mutations in the RING finger motif (residues 345-385) allow the synthesis of the first phosphodiester bond by RNA polymerase II, but prevent its escape from the promoter. This highlights the role of transcription factor IIH in the various steps of the transcription initiation process.

The p44 subunit plays a crucial role in the overall activity of the transcription/DNA repair factor TFIIH: on the one hand its N-terminal domain interacts with and regulates the XPD helicase (1, 2); on the other hand, as shown in the present study, it participates with the promoter escape reaction. Mutagenesis along with recombinant technology using the baculovirus/insect cells expression system allowed us to define the function of the two structural motifs of the C-terminal moiety of p44: mutations within the C4 zinc finger motif (residues 291-308) prevent incorporation of the p62 subunit within the core TFIIH. Double mutations in the RING finger motif (residues 345-385) allow the synthesis of the first phosphodiester bond by RNA polymerase II, but prevent its escape from the promoter. This highlights the role of transcription factor IIH in the various steps of the transcription initiation process.
Accurate transcription of class II genes requires formation of a preinitiation complex composed of RNA polymerase II (RNA pol 1 II) and several general transcription factors including TFIIH (3). TFIIH, also involved in DNA repair and cell cycle control (4), plays a central role in the initiation of transcription due to its numerous enzymatic activities: the 3Ј-5Ј XPB helicase opens DNA around the promoter (5-7), whereas the 5Ј-3Ј XPD helicase is more likely devoted to the opening of DNA around a damage (8,9). In transcription the role of XPD seems to be more structural, since it allows the anchoring of CAK (cdk-activating kinase) complex to the core TFIIH, for optimal phosphorylation of RNA pol II (10), and nuclear receptors (11)(12)(13). p62, p52, and p34 have no defined functions, whereas p44 regulates the XPD helicase activity within TFIIH (1,2). Mutations in the two helicases XPB and XPD are associated with three rare genetics disorders: xeroderma pigmentosum, Cockayne syndrome, and trichothiodystrophy (14,15). To understand the role of TFIIH in transcription, we have engaged projects on each subunit within TFIIH.
Mutations in SSL1 (the yeast counterpart of p44) affect the genome stability of the yeast strain (16). Temperature-sensi-tive ssl1 yeast mutants are UV light-sensitive and defective in transcription, DNA repair, and likely in translation (17)(18)(19). These various defects highlight the central role of p44 within TFIIH, with the understanding that this subunit also interacts with almost all the other subunits of the core TFIIH both in human and in yeast (20,21). 2 Having depicted the role of the N-terminal part of p44 which regulates the XPD helicase activity, we here focus our attention on the C terminus end. In the present study, we demonstrate the importance of the C-terminal cysteine-rich motif in preserving TFIIH architecture. We also show that double mutations in the RING finger motif affect the transcriptional activity of TFIIH by preventing its escape from the promoter.

Construction of Recombinant Baculoviruses
Expressing TFIIH Subunits-Baculoviruses expressing the TFIIH subunit were constructed in the pVL1392 or pACAB4 vectors (PharMingen). The cDNAs coding for XPB, XPD, p62, p52, p34 subunits of THIIH, and for cdk7, cyclin H, MAT1 subunits of the CAK complex, were inserted into the pVL1392 expressing vector (22,23). Each p44 single mutant was obtained by PCR mutagenesis using two pairs of oligonucleotides on a wild type plasmid containing the p44 open reading frame: CCTCCTGCTAGCTC-AAGTTCTGAA in combination with either GACACTGTGGAGCGAAA-TAGCCT, CCAAAGTAAGACAGCGATTTTACATTCAA, CTGACATC-CATAAGCAAATCTTTCTCC, GGCACACAGCAGCAACATAAACATG, AGAAAACATTTTGGGCCACAGCACA, GCAACAGTGTAGAGAATCA-GCAACAAAAAC, AGCCAGGGCAACAGGCTAGAGAATC, or GAATA-CAGCCAGGGGCACAGTGTAG (to synthesize the 5Ј part of p44 C291A, C308A, C345A, C360A, C363A, H376A, H380A, and C382A, respectively) and CGGGTCCCAGGAAAGGATCCTCA in combination with either AGGCTATTTCGCTCCACAGTGTC, TTGAATGTAAAATC-GCTGGTCTTACTTTGG, GGAGAAAGATTTGCTTATGGATGTCAG, CATGTTTATGTTGCTGCTGTGTGCC, TGGCTGTGGCCCAAAATGT-TTTCT, GTTTTTGTTGCTGATTCTCTACACTGTTGC, GATTCTCTA-GCCTGTTGCCCTGGCT, or CTACACTGTGCCCCTGGCTGTATTC (to synthesize the 3Ј remaining part of these p44 mutations). The double mutants (C360A/C363A and H376A/H380A) were obtained by the same PCR procedure. The mutated fragment was then substituted to the wild type open reading frame using the BamHI/NheI restriction sites. Both XPB and cyclin H were fused to a 6-histidine tag at their N-terminal extremity. The resulting vectors were transfected with linearized baculovirus DNA (PharMingen) in Sf9 cells (Spodoptera frugiperda 9). The recombinant viruses were purified from isolated plaques, and viral stocks were prepared by a three-step growth amplification. The calculation of the multiplicity of infection of each virus stock solution as well as the determination of the best ratio for the expression of each subunit allowed us to optimize the coexpression of the different subunits stoichiometrically.
Purification of TFIIH Complexes-Cobalt chelate affinity purification: Sf9 cells were infected with combinations of baculoviruses expressing the different subunits of TFIIH as described previously (23). 48 h after infection, cells were collected, washed, and dounced in buffer A (20 mM Tris-HCl, pH 7.8, 20% glycerol, 150 mM NaCl) containing 0.1% Nonidet P40, 5 mM ␤-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, and 1 ϫ protease inhibitor mixture, and then the extracts were centrifuged at 14,000 ϫ g during 30 min. The supernatants were first applied on a heparin-Ultrogel column. After a five-resin volume wash with buffer A containing 0.4 M NaCl, the proteins were eluted with this buffer containing 0.5 M NaCl. After a 3-h dialysis against buffer B (50 mM Tris-HCl, pH 7.8, 20% glycerol, 300 mM KCl) fractions were incubated for 1 h at 4°C with 1 ⁄40 fraction volume of cobalt chelate affinity resin (CLONTECH). After a 20-volume wash with buffer B containing 10 mM imidazole, proteins were eluted with 100 mM EDTA and dialyzed against buffer C (50 mM Tris-HCl, pH 7.8, 20% glycerol, 0.1 mM EDTA, 0.5 mM dithiothreitol) containing 50 mM KCl for 3 h at ϩ4°C. Recombinant CAK was purified as described previously (22).
Immunopurification of rIIH6/p44-C291A and rIIH6/p44-C308A-Infected Sf9 cell extracts were incubated overnight at 4°C with protein A-agarose beads (Amersham Pharmacia Biotech) cross-linked to p44 antibody. The resin was washed twice with buffer C containing 150 mM KCl, and the elution was performed by incubating the resin 6 -7 h at 4°C in 0.15 ml of the same buffer in the presence of 2 mg/ml p44 epitope peptide and 0.2 mg/ml insulin.
Enzymatic Activities of TFIIH-The run-off transcription assay was performed by incubating all general transcription factors (TFII-A, -B, -E, -F, -H, -TBP), RNA pol II, and the adenovirus major late promoter (MLP) for 15 min at 25°C in 50 mM Tris-HCl, pH 7.8, 10% glycerol, 1 mM EDTA, 0.5 mM dithiothreitol, 50 mM KCl, and 5 mM MgCl 2 (24). Transcription was then carried out for 45 min at 25°C in the presence of ribonucleotides including radiolabeled CTP (Amersham Pharmacia Biotech). RNA transcripts were resolved on 5% polyacrylamide/urea gel and analyzed by autoradiography. The abortive initiation as well as the ATPase and the helicase assays were done as described earlier (5,25).
The promoter escape reaction was done as described in Ref. 25; briefly RNA pol II and the basal transcription factors, including mutants or wild type rIIH6, were incubated with a premelted DNA (Ϫ8/ ϩ2) for 30 min at 28°C in the presence of 0.4 mg/ml bovine serum albumin, 5 mM MgCl 2 , and dinucleotide ApG as an initiate substrate. After preinitiation complex formation the RNA synthesis was initiated upon addition of CTP, GTP, [␣-32 P]UTP, cordycepin, and 6.5 mM MgCl 2 for 30 min at 28°C. The samples were applied on a 12% polyacrylamide/ urea gel and analyzed by overnight autoradiography.
To investigate the role of the p44 cysteine-rich motifs in the overall activity of TFIIH, we have generated several points mutations in the conserved cysteines and histidines, among human to yeast, implied in the zinc chelation by changing them to alanine (Fig. 1A). These have been generated by PCR following a classical procedure of site-directed mutagenesis. PCR products were subcloned into the pVL1392 vector and finally inserted into the baculovirus genome by recombination. Sf9 insect cells were thereafter infected with the recombinant viruses able to overproduce each of the wild type subunits of TFIIH, including XPB, p62, p52, p34, and XPD as well as the mutated p44. TFIIH subcomplexes (rIIH6, containing core TFIIH plus XPD, see "Experimental Procedures") from infected insect cell extracts were either immunopurified using monoclonal antibodies directed toward p44 or were loaded on an heparin-Ultrogel column followed by a cobalt chelate affinity column.
To analyze the role of the C4 zinc finger motif, we have mutated cysteines to alanine at position 291 and 308 (Fig. 1A). Immunopurified rIIH6 complexes, carrying wild type or mutated p44, were washed at 150 mM salt and further analyzed by Western blotting (Fig. 2A, lanes 1-3). Although all the rIIH6 subunits are present in the baculoviruses extracts (data not shown), both mutated recombinant IIH6/p44-C291A and IIH6/ p44-C308A complexes lack p62 (lanes 2 and 3). When either p44-C291A or p44-C308A are coexpressed with p62, they are able to interact with this latter one (data not shown). It is worthwhile to mention that even at low salt concentration (150 mM), both C291A and C308A mutations in p44 do not allow integration of p62 within the core TFIIH. Furthermore, as expected in the absence of one TFIIH subunit, none of these two mutated rIIH6 complexes exhibit any transcriptional activity (data not shown).
We then focused our attention on the RING finger motif and generated alanine point mutations in the following residues: Cys-345, Cys-360, Cys-363, His-376, His-380, and Cys-382 (Fig.  1A). The rIIH6 complexes carrying these p44 mutations were purified from infected cell extracts and further analyzed by Western bloting. None of the p44 mutations were found to modulate the composition or the stoichiometry of the purified rIIH6 complexes (Fig. 2B, upper panel, lanes 2-7). The various rIIH6 complexes were then tested in a run-off transcription assay containing, in addition to MLP, RNA pol II, the basal transcription factors, TBP, TFIIA, TFIIB, TFIIE, TFIIF, and CAK. The CAK subcomplex was purified using the ability of the FIG. 1. The C-terminal part of p44. A, single and double mutations (in bold) were introduced in the C4 zinc finger and the RING finger motif of p44. B, ribbon overlay of the ␤-strands in the RING finger motif of p44 (residues 345-385) determined by NMR spectroscopy (26). The cysteines and histidines that are the best candidate for the chelation of the two zinc atoms are indicated. Cysteines 382 and 385 are also indicated.
cyclin H subunit to bind to a cobalt affinity column and was added in the reaction to stimulate the RNA synthesis. The amounts of rIIH6 were adjusted on the basis of p62 and p52 content, according to Western blot analysis. The transcriptional activity of the rIIH6 complexes was not modified by the p44 single mutations using the wild type recombinant IIH6 as a reference (Fig. 2B, lower panel, compare lanes 2-7 with lane 1). It has to be noticed that CAK fully exerts its stimulatory function, thus demonstrating that a single mutation in the RING finger motif of p44 does not prevent CAK anchoring to the core TFIIH (22).
The C-terminal Moiety of p44 Is Involved in Promoter Escape-p44 is a central protein in the core TFIIH complex and has been shown to interact with XPB, XPD, p62, and p34 (20, 21). 2 We have expressed XPB, XPD, p62, or p34 and either wild type or mutated p44 and subsequently immunopurified the binary complexes using p44 antibody, to demonstrate that the p44 double point mutations do not modify the interactions with the subunits of the core TFIIH (data not shown). Having shown that double mutations in the RING finger motif were detrimental for transcriptional activity of rTFIIH, we were then wondering if it could affect its proper enzymatic activities. ATPase activity was then assayed in the presence of DNA by incubating each p44 mutated rIIH6 complex with [␣-32 P]ATP as a substrate. All of the complexes exhibit an ATPase activity similar to that of the wild type (Fig. 3A, compare lanes 1-2 with lanes  3-10).
The N-terminal part of p44 interacts with and regulates the XPD helicase within TFIIH, giving rise to its optimum transcriptional activity. To investigate whether p44 C-terminal mutations do not affect the interaction with the XPD helicase, we carried out immunoprecipitation on Sf9 extracts coinfected with baculovirus overexpressing XPD and either p44/C291A, p44/C308A, p44/C360A/C363A, or p44/H376A/H380A, using monoclonal antibody raised against XPD. After extensive washing at 0.4 M KCl, the immunoadsorbed proteins were analyzed by immunoblotting and tested for their ability to unwind the DNA. First, none of these p44 mutations prevent the interaction with XPD (Fig. 3B, upper panel, compare lane 1  with lanes 2-5). Second, not only p44wt, but each mutated p44 so far tested, are able to regulate similarly the XPD helicase activity (Fig. 3B, lower panel, compare lane 6 with lanes 1-5).
Since the double mutations in the p44 RING finger motif do not affect TFIIH enzymatic activities, but are detrimental for its transcriptional function, we then wondered which step of the transcription is defective. Transcription initiation involves a succession of events, including promoter opening, first phosphodiester bond formation, promoter escape, and RNA elongation. We first investigated the ability of the rIIH6/p44-C360A/ C363A and rIIH6/p44-H376A/H380A double mutant complexes to form the first phosphodiester bond, when added in an assay containing MLP in addition to the other transcription factors as well as the dinucleotide CpA, [␣- 32 FIG. 3. p44 is involved in promoter escape. A, rIIH6/p44-C291A, rIIH6/p44-C308A, rIIH6/p44-C360A/C363A, and rIIH6/p44-H376A/ H380A were tested for their ATPase activity. B, the four mutated p44, in addition to XPDwt, were overexpressed in Sf9 insect cells and immunoprecipitated using antibodies directed toward XPD; the precipitated proteins were analyzed for their protein content (WB) and for their helicase activity. C, increasing amounts of rIIH6/p44-wt, rIIH6/ p44-C630A/C363A, and rIIH6/p44-H376A/H380A complexes were analyzed for the first phosphodiester bond formation (lower panel, synthesis of CpApC) and promoter escape (upper panels, synthesis of 17 and 31 nucleotides). In A and C HeLa TFIIH (lane 11 and 8, respectively) is used as a control. mutated complexes, when compared with the wild type rIIH6, allow the association of CTP to the CpA dinucleotide to form a three-nucleotide-long product, meaning that the p44 RING finger mutations, detrimental for the run-off transcriptional activity of TFIIH, do not prevent the first phosphodiester bond formation (Fig. 3C, lower panel, compare lanes 2-3 with lanes  4 -7). This reaction is highly specific, since in the absence of either TFIIH or TBP, which initiate the formation of the preinitiation complex on the TATA-box, no synthesis occurs (Fig. 3C,  lower panel, compare lane 8 with lanes 9 and 1, respectively).
We then investigated the ability of rIIH6/p44-C360A/C363A and rIIH6/p44-H376A/H380A to allow the escape of RNA pol II from the promoter. This assay monitors the escape by using a premelted DNA template around the start site (from the position Ϫ8 to position ϩ2), which circumvents the promoter opening step (5,7) and consequently the requirement of a functional XPB helicase. To allow the accumulation of products that escape the promoter, we employed the chain terminating ATP analog cordycepin, which stops transcription elongation at thymidines at position ϩ17 and ϩ31 (see Ref. 25 and "Experimental Procedures"). The reaction is TFIIH-dependent and gives rise to 17-and 31-oligonucleotide length products (Fig. 3C, upper panel, compares lane 8 with lane 9). Both rIIH6/p44-C360A/C363A and rIIH6/p44-H376A/H380A do not allow the RNA pol II to escape from the promoter to synthesize longer transcripts (Fig. 3C, upper panel, compare lanes 2 and 3 with lanes 4 and 5 and lanes 6 and 7, respectively). Together these results demonstrate for the first time the implication of p44, and more particularly its RING finger motif in the promoter escape step. DISCUSSION It is now accepted that TFIIH joins the promoter late during the formation of the stable preinitiation complex. Once part of this closed and inactive preinitiation complex, and upon addition of ATP as a source of energy (27,28), the XPB helicase of TFIIH will promote the opening of the promoter around the start site, the first phosphodiester bond formation, and further the promoter escape (5,7,29 and see also Refs. 30 and 31). XPD, the second helicase of TFIIH, does not play a role in the opening step, since present or not, mutated or not in its ATP binding site, RNA synthesis, although at a lower level, might occur. Rather it was suggested that it is its physical presence that is required for optimal RNA synthesis in addition to its ability to maintain the CAK complex within TFIIH. Indeed by binding on the one hand MAT1 and on the other hand p44, XPD likely bridges the CAK subcomplex to the core TFIIH (1,32).
The role of p44, as an essential subunit of the core TFIIH, was further sustained by the discovery that a mutation in its yeast counterpart (SSL1) conferred UV sensitivity due to a nucleotide excision repair defect (18). It was then found that mutations either in the N-terminal moiety of p44 or in the C-terminal end of XPD, by preventing their interaction and thus optimal XPD helicase activity, lead to a destabilization of TFIIH. Besides a role as the regulatory subunit of XPD, p44 presented in its C-terminal moiety a highly conserved cysteinerich domain, which was defined to contain a RING finger motif by NMR studies (26). The present study details the role of the C-terminal part of p44.
First, mutations in the C4 zinc finger of p44 completely destabilize the overall architecture of TFIIH. Indeed mutations of the Cys-291 and Cys-308, which chelate a zinc atom together with the Cys-294 and Cys-305, prevent the formation of the core TFIIH: immunopurification, from infected insect cells extracts, of rIIH6 complexes carrying these mutations shows that these mutants lack the p62 subunit. This is not due to a decrease in the interaction between the mutated p44 and p62, since these proteins coimmunoprecipitate when they are coexpressed in insect cells. It is most likely due to a specific role of this C4 zinc finger motif in the structure of TFIIH, since first a complete depletion of p44, and second mutations in the RING finger motif, do not prevent the incorporation of p62 within TFIIH (this study and Ref. 23).
Second C360A/C363A and H376A/H380A double mutations in p44 inhibit TFIIH transcription activity. It is worthwhile to notice that the single mutation of one of these four residues has no effect on the basal transcription activity of TFIIH, indicating that chelation of a zinc atom by three amino acids (instead of four) is sufficient to maintain the RING finger motif of p44. Further investigations of the transcription reaction show that these double mutations in p44, although they allow the first phosphodiester bond formation, inhibit further RNA synthesis by RNA pol II, emphasizing the crucial role of the RING finger motif in the promoter escape process. Additional experiments have shown that this is not due to a defect in the XPD helicase activity within TFIIH (the present study) nor to the inability of cdk7 to phosphorylate the largest subunit of RNA pol II (data not shown). Indeed, CAK was still able to stimulate all the transcription reaction, when added to the rIIH6 complexes, showing that it is fully integrated within TFIIH upon XPD/p44 interaction. We can then assume that the defect of promoter clearance observed for the double RING finger mutants is not due to a defect of CAK activity, but rather to an inaccurate positioning of an other subunit of TFIIH, such as XPB. Indeed, we cannot exclude that p44 mutations might also prevent some proteinprotein interaction in which the RING motif is required (26).
To conclude, we can hypothesize that p44 could allow the slipping of TFIIH on the DNA during the promoter escape knowing that TFIIH is associated with the RNA pol II during the synthesis of the first 40 -50 nucleotides (see Ref. 33 and for a review, see Ref. 34). Whether or not the RING finger motif of p44 directly binds some transcription factors or the promoter, as suggested by cross-linking experiments 3 or by gel shift assays 4 , remains to be further investigated. Preliminary experiments have illustrated interactions between p44 and other basal transcription factors.