A Novel Zinc Finger Structure in the Large Subunit of Human General Transcription Factor TFIIE*

The zinc finger domain in the large subunit of TFIIE (TFIIEα) is phylogenetically conserved and is essential for transcription. Here, we determined the solution structure of this domain by using NMR. It consisted of one α-helix and five β-strands, showing novel features distinct from previously determined zinc-binding structures. We created point mutants of TFIIEα in this domain and examined their binding abilities to other general transcription factors as well as their transcription activities. Four Zn2+-ligand mutants, in which each of cysteine residues at positions 129, 132, 154, and 157 was replaced by alanine, possessed no transcription activities on a linearized template, whereas, on a supercoiled template, interesting functional asymmetry was observed: although the C-terminal two mutants abolished transcription activity (<5%), the N-terminal two mutants retained about 20% activities. The N-terminal two mutants bound stronger to the small subunit of TFIIF than the wild type and the C-terminal two mutants were impaired in their binding abilities to the XPB subunits of TFIIH. These suggest that the structural integrity of the zinc finger domain is essential for the TFIIE function, particularly in the transition from the transcription initiation to elongation and the conformational tuning of this domain for appropriate positioning of TFIIF, TFIIH, and polymerase II would be needed depending on the situation and timing.

In eukaryotes, transcription initiation of protein-coding genes requires RNA polymerase II (pol II) 1 and its auxiliary five general transcription factors, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH (1,2). Pol II and general transcription factors assemble together on a promoter DNA to form a preinitiation complex (PIC). TATA box-binding protein, a DNA binding subunit of TFIID, binds first to the TATA box. TFIIB then binds and recruits pol II to the complex (3). Pol II comes with TFIIF, which is important for accurate location of pol II in the complex around the transcription initiation site (4). Finally, TFIIE and TFIIH are incorporated to complete PIC formation (3).
During PIC formation, TFIIE interacts with various general transcription factors, pol II, and promoter DNA, recruits TFIIH into the PIC, and regulates the enzymatic activities of TFIIH; serine kinase of the C-terminal domain of the largest subunit of pol II, DNA-dependent ATPase, and DNA helicase activities (5)(6)(7). At transcription initiation, TFIIE binds both to pol II near its active site and to promoter DNA ϳ10 bp upstream (Ϫ10) from the transcription initiation site (ϩ1), where the promoter melting starts (8). There TFIIE conducts pol II to sit at the initiation site, makes pol II processive upon C-terminal domain phosphorylation by stimulating C-terminal domain kinase activity of TFIIH, and simultaneously assists a helicase subunit XPB of TFIIH to start melting at the Ϫ10 position. At the transition from initiation to elongation, TFIIE also plays a direct role in promoter clearance by regulating kinase and DNA helicase activities of TFIIH (6,9).
Human TFIIE (hTFIIE) is a heterotetramer consisting of two ␣ (hTFIIE␣; 57 kDa, 439 aa) and two ␤ (hTFIIE␤; 34 kDa, 291 amino acids) subunits (10 -13). Both subunits possess several structural motifs and characteristic sequences. Although hTFIIE has been functionally characterized, little is known about its structure, especially on hTFIIE␣ (14 -16). Main reason is its low solubility. Although high concentration of protein sample should be prepared for the structural studies, each hTFIIE subunit aggregates and precipitates at the concentration higher than 10 mg/ml. Thus we have been focusing on dissecting functional domains in each subunit. To identify core regions in hTFIIE␣, limited proteolyses were performed and a highly conserved region (residues 113-174) containing a zinc finger motif, which is essential for transcription activity of TFIIE (14,17) was isolated (Fig. 1). In this study, we have * This work was supported by a Collaborative of Regional Entities for the Advancement of Technological Excellence (CREATE) from Japan Science and Technology agency, by a Project of Protein 3000, Transcription and Translation, and by Grants in Aid for Scientific Research from Ministry of Education, Culture, Sports, Science and Technology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains determined the solution structure of this core domain (hTFIIE␣c) of hTFIIE␣ by NMR and provided insight into its role for the TFIIE function.

EXPERIMENTAL PROCEDURES
Construction of hTFIIE␣c Expression Plasmid-For subcloning the nucleotide sequence of the hTFIIE␣c corresponding to the amino acid residues 113-174 into a bacterial expression vector, the oligonucleotide ES1T (5Ј-GTGTACCATATGAGAATTGAGACCGATGAGAGAG-3Ј) was designed to create an NdeI site (underlined) and to add the methionine codon before the R113 residue and the nucleotide ES1B (5Ј-GAACGG-GATCCTCAGCGTGCATCTTTTTTGGGCATTGC-3Ј) was designed to create a BamHI site (underlined) right after the stop codon (bold letters). This nucleotide fragment was amplified by PCR using two oligonucleotides and the p2EA plasmid containing the complete open reading frame of human TFIIE␣ (hTFIIE␣ cDNA) as a template (11), digested with NdeI and BamHI, and subcloned into the NdeI and BamHI restriction sites of the pBluescript SK(Ϫ). The nucleotide sequences of the cloned PCR products were confirmed using an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). The NdeI-BamHI fragment of hTFIIE␣c cDNA was then subcloned into the pET11d vector (Novagen) to construct the N-terminal six histidine (His 6 )-tagged hTFIIE␣c (6His-hTFIIE␣c) expression plasmid.
Purification of hTFIIE␣c-The hTFIIE␣c plasmid was transformed into Escherichia coli BL21(DE3)pLysS (Novagen). The cells were grown at 37°C in LB or in M9 minimal media containing  (22) to detect trans-hydrogen bond 3h J NCЈ couplings. Spectra were processed using NMRPipe (23) and analyzed using the programs PIPP, CAPP, and STAPP (24) and NMRView (25). NMR relaxation rates for amide nitrogens of hTFIIE␣c were measured according to the procedure described previously (26). The 15 N T 1 data were collected using 15 15 N NOEs data were measured by taking the ratio of the peak intensities from experiments performed with and without application of 1 H saturation.
Structure Calculation-Interproton distance restraints derived from NOE intensities were grouped into three distance ranges, 1.8 -2.7 Å (1.8 -2.9 Å for NOEs involving NH protons), 1.8 -3.3 Å (1.8 -3.5 Å for NOEs involving NH protons), and 1.8 -5.0 Å, corresponding to strong, medium, and weak NOEs, respectively. The upper limit was corrected for constraints involving methyl groups, aromatic ring protons, and nonstereospecifically assigned methylene protons. Dihedral angle restraints for were set to Ϫ90°Ͻ Ͻ Ϫ40°for 3 J HN␣ Ͻ 6.0 Hz and Ϫ160°Ͻ Ͻ Ϫ80°for 3 J HN␣ Ͼ 8.0 Hz, based on the short range NOEs and backbone chemical shifts. 1 angles were restrained Ϯ30°for three side-chain rotamers. The zinc was constrained to be tetrahedrally coordinated by four cysteines as follows. Zn-S␥ bond length sets to 2.3 Å, Zn-S␥-C␤ and S␥-Zn-S␥ angles set to 108°and 109°. Hydrogen bond restraints in areas of regular secondary structure were introduced at the final stages of refinement. Structure calculations were performed by the distance geometry and simulated annealing using program X-PLOR (27). A total of 100 structures were calculated, and 69 structures had no NOE violations larger than 0.2 Å and no dihedral angle violations greater than 1°. Structural statistics for the 20 best structures are summarized in Table I (31). A restriction enzyme site was placed in an oligonucleotide to select for properly mutated plasmids as described elsewhere, and the mutants were then checked by sequencing using an ABI Prism 310 Genetic Analyzer (Applied Biosystems). The oligonucleotides used for mutation were listed in Table II. Created enzyme sites were underlined, and the enzyme names were written in parentheses. Mutated sequences are written in bold, and each mutated codon is boxed. For three oligonucleotides creating C157A, D164A, and D164K, restriction enzyme sites were not created and mutated sequences were directly checked by sequencing. The NdeI-BamHI fragment of mutated hTFIIE␣ cDNA was then subcloned into the pET11d vector (Novagen) to construct the N-terminal His 6 -tagged hTFIIE␣ (6His-hTFIIE␣) expression plasmid.
In Vitro Transcription Assays-Recombinant general transcription factors as well as native pol II and TFIIH were purified as described previously (31). In vitro transcription with either a supercoiled or a linearized template was carried out with increasing amounts (12 and 24 ng) of wild type or mutant hTFIIE␣ proteins. The plasmid pML(C 2 AT)⌬-50, which contains the adenovirus 2 major late promoter and gives a 390-nucleotide transcript, was used as a template (31). To prepare the linearized template, pML(C 2 AT)⌬-50 was digested with SmaI. After transcription, radiolabeled transcripts were subjected to urea-PAGE and detected by autoradiography. The transcripts were quantified by a Fuji BAS2500 Bio-Imaging analyzer. Relative transcription activities of the mutant hTFIIE␣ proteins were calculated by defining the transcription activity of 24 ng of wild type hTFIIE␣ as 100%.
GST-Pull-down Assay-GST fusion proteins were performed as previously described (31). The bound proteins were detected by Western blotting with anti-hTFIIE␣ mouse monoclonal antibody (1:8,000 dilution) (Medical Biology Laboratory). Signals were detected using the enhanced chemiluminescence (ECL) detection system (Amersham Biosciences) using RX-U film (Fuji Film).

RESULTS
Identification of the Core Domain of hTFIIE␣-TFIIE␣ is highly conserved among eukaryotes. In addition, TFE from archaea has recently been found to be a homolog of the Nterminal half of hTFIIE␣ (32)(33)(34). To identify the structural domain, the recombinant hTFIIE␣ was digested by limited proteolysis. Three structural domains consisting of the amino acid residues 113-174, 276 -323, and 283-323 and two structural domains consisting of the amino acid residues 106 -188 and 302-342 were identified by trypsin and chymotrypsin digestions, respectively, followed by mass spectrometries. Finally, a central core domain (hTFIIE␣c) was identified to consist of the amino acid residues 113-174 containing a putative zinc finger motif (Fig. 1) (11). 6His-hTFIIE␣c was expressed in E. coli BL21(DE3)pLysS, purified, and subjected to conventional multidimensional NMR measurements.
The coordination of Zn 2ϩ to four cysteine residues and the auxiliary two bifurcated hydrogen bonds fix the t1 and t2 turns rigidly, forming two anti-parallel ␤-sheets approach around one side of the ␣-helix. Fig. 2E shows that a hydrophobic core is formed by hydrophobic amino acids: Ala-141, Leu-144, and Val-161 as well as five phenylalanine residues that are piled up each other and arranged like a capital character, L. This aromatic-aromatic interaction network seems to contribute significantly to the stability of the structure. These residues are well conserved among species (Fig. 1). Electrostatic potential on the molecular surface shows that the hTFIIE␣c has a negative potential cluster, which is formed by conserved Asp-138, Glu- 160, Glu-162, Glu-163, Asp-164, and Glu-165, and is extensively located on a concave surface opposite to the zinc-binding site (Fig. 2F).
N-terminal and C-terminal regions were disordered due to poor distance restraints. Nevertheless only about 40 amino acids can architect the rigid structure. Actually, 15 N-T 1 , T 2 , and 15 N-{ 1 H}heteronuclear NOE values revealed flexibility of backbone in the N-terminal 13 and C-terminal 11 residues (Fig. 3).
Functional Roles of hTFIIE␣c in Transcription-For functional analyses of hTFIIE␣, we substituted each of the four Zn 2ϩ -binding cysteine residues at 129, 132, 154, and 157 with alanine, C129A, C132A, C154A, and C157A, as well as two highly conserved acidic residues, Glu-140 and Asp-164 on the negatively charged surface, with alanine or lysine, E140A, E140K, D164A, and D164K. All these mutants of hTFIIE␣ with His 6 at the N terminus were expressed in E. coli, purified through Ni-NTA-agarose column, and subjected to SDS-PAGE (Fig. 5A). The effects of these point mutants on basal transcription by using a supercoiled adenovirus major late pML⌬-50(C 2 AT) template were analyzed (Fig. 5, B and C). Although C154A and C157A abolished transcription activity (Ͻ5%), C129A and C132A retained 12 and 25% activities, respectively (Fig. 5B). On the other hand, the acidic residue mutants were active in transcription; E140A possessed almost the same activity as the wild type, E140K was also active but with the reduced activity of 50%, and D164A and D164K mutants possessed higher activities at ϳ170 -180% than the wild type (Fig. 5C).
To dissect the effects of these mutants on the transcription step more precisely and to see the transition activity from transcription initiation to elongation, a linearized template was used (44). Most of the transcription activities of the mutants on a linearized template were similar to the activities on a supercoiled template except that the activities of C129A and C132A were completely abolished (Fig. 5D), and the mutant D164A showed further augmented activity of 370% (Fig. 5E). These suggest that hTFIIE␣c can be functionally dissected into two; the C-terminal half involves in transcription initiation alone, but the N-terminal half involves not only initiation but also the transition from initiation to elongation. Although Asp-164 is conserved in eukaryotes as well as even in archaea (Fig.  1), substitutions to alanine and lysine residues, D164A and D164K, rather stimulated transcription both on supercoiled and linear templates (Fig. 5, C and E).
Effects of Mutations on Binding to the General Transcription Factors-We next investigated the binding abilities of these mutants to the general transcription factors by GST-pull-down assays (Fig. 6, A and B). All mutants as well as the wild type gave the similar binding patterns except that C129A and C132A bound to TFIIF␤ about 3-fold stronger than the wild type (Fig. 6A, lane 6) and that C154A completely lost and C157A reduced their binding abilities to the XPB subunit of TFIIH (Fig. 6B, lane 3). The bindings of the mutants to intact pol II were also tested but all, including the wild type failed to bind (data not shown). The stronger binding to TFIIF␤ may contribute to the transcription initiation activities of C129A and C132A on a supercoiled template retained ϳ12-25%. TFIIF is important for transcription initiation and thus will be able to substitute for a part of TFIIE␣ function. The reduced binding of C154A and C157A to XPB of TFIIH may cause abolished transcription on a supercoiled template; C154A and C157A could not recruit XPB at the region on the promoter where promoter melting starts, and thus these mutants are defective in transcription initiation, because XPB functions for promoter opening utilizing its DNA helicase activity.
The Asp-164 mutants of hTFIIE␣c, D164A and D164K, did not show so much difference from the wild type in binding to general transcription factors except that both bound more strongly to the XPB and XPD subunits of TFIIH. In addition, D164A bound to the p44 and Cdk7 subunits of TFIIH (Fig. 6, A  and B). It is intriguing to consider that three TFIIH subunits (XPD, p44, and Cdk7) have been reported to be involved in the transition stage from initiation to elongation. This fits to the present experiment that D164A was more active in transcription, especially, on a linearized template (3.7-fold of the wild type activity) (Fig. 5E) and indicates that TFIIH may bind more rigidly to D164A than the wild type hTFIIE␣ and contributes to the transition stage to elongation. DISCUSSION To elucidate the functions of hTFIIE␣c, we investigated the effects of several point mutants on transcription and on the binding abilities to the general transcription factors (Figs. 5 and 6). For transcription activity assay we used the supercoiled and linearized templates of adenovirus major late promoter. Requirement of TFIIE and TFIIH for the promoter melting depends on promoter DNA (45)(46)(47)(48). On the supercoiled template, TFIIE can melt the promoter independently of TFIIH. However, on the linearized template both TFIIE and TFIIH are necessary for the transition activity from the initiation to elongation. In mutants of Zn 2ϩ ligand cysteine residues at 129, 132, 154, and 157, the interesting finding was the functional asymmetry in hTFIIE␣c. The N-terminal two mutants C129A and C132A retained 12 and 25% activities in transcription with the supercoiled template, whereas the C-terminal two mutants C154A and C157A almost completely abolished the transcription activities (Fig. 5B). On the other hand, all of the cysteine  ). B and C, effect of zinc finger point mutations on the basal transcription activity of hTFIIE␣ with a supercoiled template. In vitro transcription assays with a supercoiled template were carried out with increasing amounts (6, 12, and 18 ng) of wild type hTFIIE␣ (IIE␣ wt) or hTFIIE␣ proteins with point mutations in the zinc finger region. After the transcription reaction, radio-labeled transcripts were subjected to urea-PAGE and detected by autoradiography (lower panels). Relative transcription activities of the mutant hTFIIE␣ proteins (%) were calculated by defining the transcription activity of 18 ng of wild type hTFIIE␣ as 100% and are shown as bars (upper panels). Mutated residues are indicated on the bottom. As a control, transcription was carried out without hTFIIE␣ protein (-␣). Arrows on the right side indicate the position of the 390-nulceotide transcripts. D and E, effects of zinc finger point mutations on the basal transcription activity of hTFIIE␣ with a linearized template. In vitro transcription assays with a linearized template were carried out with increasing amounts (12 and 18 ng) of wild type hTFIIE␣ (IIE␣ wt) or hTFIIE␣ proteins with point mutations in the zinc finger region. Mutated residues are indicated on the bottom. As a control, transcription was carried out without hTFIIE␣ protein (-␣). Arrows on the right side indicate the position of the 390-nulceotide transcripts. mutants showed no transcription activities with the linearized template (Fig. 5D). It is suggested that a defect in transcription initiation should cause no transcription activity with a supercoiled template, whereas a defect in the transition from initiation to elongation could cause some transcription activities with a supercoiled template, but not at all with a linearized template. So C154A and C157A would be deficient in transcription initiation, whereas C129A and C132A would be somewhat deficient in initiation but significantly deficient in the transition stage. The N-terminal zinc finger mutants C129A and C132A showed stronger binding to TFIIF␤ and the C-terminal mutants C154A and C157A, on the other hand, showed diminished binding to XPB (Fig. 6A, lane 6, and Fig. 6B, lane 3). It is noteworthy that the TFIIE deletion mutants of the zinc finger domain still possessed the stimulation activity of TFIIH-mediated C-terminal domain phosphorylation of pol II in the presence of template DNA and other general transcription factors but not in the absence (14). Thus, the structural integrity of hTFIIE␣c might be essential for the TFIIE function, the conformational tuning of hTFIIE␣c for appropriate positioning of TFIIF, TFIIH, and pol II would be needed depending on the situation and timing.
The functional meaning of Asp-164 is still under consideration, because it is still difficult to explain the reasons why the D164A and D164K mutants showed stronger transcription than the wild type. Because Asp-164 is well conserved among species, this acidic character may be essential for association with some ubiquitous transcription factors, for example, TRAP/ Mediator, cofactors, and transcription elongation factors (49). One candidate is p100, a coactivator interacting with EBNA2 and STAT6, because this protein was reported to bind to both subunits of TFIIE (50,51). p100 possesses the Tudor domain, which is supposed to bind to the methylated peptides (52). Recently, transcriptionally active pol II with Ser-5 phosphorylation of the heptapeptide repeat of the largest subunit was found to recruit histone H3 Lys-4-specific methyltransferase Set1 (reviewed in Ref. 53). Thus, it is intriguing to speculate that p100 and TFIIE work for the bridges to connect between methylated histones and pol II. Further studies are necessary to confirm those possibilities for the functional roles of hTFIIE␣c.