Transcription Factor B Contacts Promoter DNA Near the Transcription Start Site of the Archaeal Transcription Initiation Complex*

Transcription initiation in all three domains of life requires the assembly of large multiprotein complexes at DNA promoters before RNA polymerase (RNAP)-catalyzed transcript synthesis. Core RNAP subunits show homology among the three domains of life, and recent structural information supports this homology. General transcription factors are required for productive transcription initiation complex formation. The archaeal general transcription factors TATA-element-binding protein (TBP), which mediates promoter recognition, and transcription factor B (TFB), which mediates recruitment of RNAP, show extensive homology to eukaryal TBP and TFIIB. Crystallographic information is becoming available for fragments of transcription initiation complexes (e.g. RNAP, TBP-TFB-DNA, TBP-TFIIB-DNA), but understanding the molecular topography of complete initiation complexes still requires biochemical and biophysical characterization of protein-protein and protein-DNA interactions. In published work, systematic site-specific protein-DNA photocrosslinking has been used to define positions of RNAP subunits and general transcription factors in bacterial and eukaryal initiation complexes. In this work, we have used systematic site-specific protein-DNA photocrosslinking to define positions of RNAP subunits and general transcription factors in an archaeal initiation complex. Employing a set of 41 derivatized DNA fragments, each having a phenyl azide photoactivable crosslinking agent incorporated at a single, defined site within positions –40 to +1 of the gdh promoter of the hyperthermophilic marine archaea, Pyrococcus furiosus (Pf), we have determined the locations of PfRNAP subunits PfTBP and PfTFB relative to promoter DNA. The resulting topographical information supports the striking homology with the eukaryal initiation complex and permits one major new conclusion, which is that PfTFB interacts with promoter DNA not only in the TATA-element region but also in the transcription-bubble region, near the transcription start site. Comparison with crystallographic information implicates the PfTFB N-terminal domain in the interaction with the transcription-bubble region. The results are discussed in relation to the known effects of substitutions in the TFB and TFIIB N-terminal domains on transcription initiation and transcription start-site selection.

* This work was supported by National Institutes of Health Grant GM41376 and a Howard Hughes Medical Institute Investigatorship (to R. H. E.) and by National Institutes of Health Grant GM42025 and National Science Foundation Grant MCB 96-31093 (to R. A. S.). 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. served sequence block" (CSB); Ref. 33). Substitutions in the conserved sequence block in archaeal TFB indicate a role in transcription-initiation NTP concentration dependence (33), and substitutions in eukaryal TFIIB indicate roles in transcription-initiation efficiency and start-site selection (37)(38)(39)(40)(41)(42). Although the roles of the TFB and TFIIB N-terminal domains in transcription initiation and start-site selection have been well characterized, the structural and mechanistic basis of these roles has remained undefined.
In published work, systematic site-specific protein-DNA photocrosslinking has been used to characterize the structural organization of eukaryal transcription initiation complexes (43)(44)(45), bacterial transcription initiation complexes (46,47), and the part of an archaeal transcription initiation complex at and downstream of the transcription start site (positions Ϫ1 to ϩ20 of the template strand (21)). Here, we report the use of systematic site-specific protein-DNA photocrosslinking to define the structural organization of the part of an archaeal transcription initiation complex upstream of the transcription start site (positions Ϫ40 to ϩ1; nontemplate and template strands), the part that contains binding determinants for RNAP subunits and general transcription factors. Our results confirm the anticipated high degree of homology between the archaeal and eukaryal transcription systems and identify an unanticipated interaction between the TFB N-terminal domain and promoter DNA in the transcription-bubble region, upstream of and at the transcription start site. The results immediately suggest a mechanistic basis for the role of TFB and TFIIB N-terminal domains in transcription initiation and start-site selection.

EXPERIMENTAL PROCEDURES
PfRNAP-Under anaerobic conditions, frozen Pyrococcus furiosus cell paste (15 g, prepared as in Bryant and Adams (48)) was resuspended in 82 ml of ice-cold 50 mM Tris (pH 7.5), 22 mM NH 4 Cl, 10 mM EDTA, and 10% glycerol, and cells were lysed in two passes through a French press operating at 1200 p.s.i. After the addition of Polymin P to 0.4% w/v and stirring for 30 min at room temperature, the sample was centrifuged at 100,000 ϫ g for 4 h at 4°C, the pellet was resuspended in 50 ml of the same buffer, and the sample again was centrifuged at 100,000 ϫ g for 4 h at 4°C. The pellet was extracted 3 times with 100 ml of 50 mM Tris (pH 7.5), 1.2 M NH 4 Cl, 10 mM EDTA, and 10% glycerol, extracts were pooled, ammonium sulfate was added to 70% saturation, the sample was stirred for 1 h on ice, and the sample was again centrifuged at 100,000 ϫ g for 4 h at 4°C. The resulting pellet was aerobically resuspended in 100 ml of buffer A (50 mM Tris (pH 7.5), 50 mM KCl, 10 mM EDTA, and 20% glycerol) and dialyzed against three changes of 1 liter of the same buffer for 12 h at 4°C (12-14-kDa molecular weight cut-off dialysis tubing (SpectraPor)). The sample was centrifuged at 20,000 ϫ g for 1 h at 4°C. The sample was loaded onto a 220-ml DEAE-Sepharose column (Amersham Biosciences) pre-equilibrated in buffer A and eluted with 2 liters of a linear gradient of 100% buffer A to 100% buffer B (50 mM Tris (pH 7.5), 1 M KCl, 10 mM EDTA, and 20% glycerol). Fractions containing DNA-dependent RNA polymerase (RNAP) activity (49,50) were pooled, concentrated, and loaded onto a 30-ml heparin-Sepharose column (Amersham Biosciences) equilibrated with 15% buffer B. The column was eluted with a 300-ml linear gradient of 15 to 100% buffer B. Pooled fractions containing RNAP activity were concentrated and dialyzed against buffer A for 12 h at 4°C. The sample was centrifuged at 20,000 ϫ g for 30 min at 4°C and loaded onto a Mono Q 10/10 column (Amersham Biosciences). The column was eluted with a 100-ml linear gradient of 10 -100% buffer B. Pooled fractions containing polymerase activity were concentrated and stored as aliquots at Ϫ80°C. Yield, ϳ4.7 g.
PfTBP-Plasmid PfTBP/pT7-7, which encodes PfTBP under control of the bacteriophage T7 gene 10 promoter, was constructed by replacing the NdeI-BamHI segment of plasmid pT7-7 (51) with a NdeI-BamHI DNA fragment containing the PfTBP-coding sequence (prepared by add-on PCR using P. furiosus genomic DNA as template (a gift of F. Jenney, University of Georgia, Athens, GA)). Transformants of Escherichia coli strain BL-21 DE3 (Novagen) with plasmid PfTBP/pT7-7 were cultured at 37°C in 4 liters of LB medium (Fisher) to an A 600 of ϳ0.9, induced by the addition of isopropyl ␤-D-thiogalactopyranoside to 0.1 mM, cultured an additional 4 h at 37°C (A 600 ϳ 2.0), and harvested by centrifugation at 300 ϫ g for 20 min at 4°C. Cell pellets (10 -12 g) were frozen and stored at Ϫ80°C until used. Cells were thawed by suspension in 60 ml of lysis buffer (50 mM potassium phosphate (pH 7.8), 10 mM 2-mercaptoethanol, 1 mM phenylmethanesulfonyl fluoride, 1 g/ml leupeptin, and 1 g/ml DNase I) and were lysed by sonication; the lysate was centrifuged at 30,000 ϫ g for 50 min at 4°C. After incubation of the supernatant for 30 min at 70°C, the sample was centrifuged at 30,000 ϫ g for 50 min at 4°C, the supernatant was loaded onto a 2.5 ϫ 10-cm Q-Sepharose benchtop column (Amersham Biosciences), the column was washed with 500 ml of 50 mM potassium phosphate buffer (pH 7.8), and the column was eluted with a 150-ml linear gradient of 0 -600 mM KCl in the same buffer. Pooled fractions containing PfTBP (50 ml; ϳ300 mM KCl) were desalted into 50 mM potassium phosphate (pH 7.8) and concentrated to 2 ml by ultrafiltration at 4°C (200-ml ultrafiltration unit (Amicon) with YM3 membrane (Millipore)). The sample was loaded onto a 1 ϫ 10-cm Mono Q column (Amersham Biosciences), and the column was eluted with a 55-ml linear gradient of 0 -600 mM KCl in 50 mM potassium phosphate (pH 7.0). Pooled fractions containing PfTBP (5 ml; ϳ400 mM KCl) were exchanged into 50 ml of 50 mM potassium phosphate (pH 7.8), 500 mM KCl, concentrated by ultrafiltration through a YM3 membrane (Millipore) at 4°C, and stored in aliquots at Ϫ70°C. SDS-PAGE indicated an ϳ21-kDa protein, corresponding to the expected molecular mass of PfTBP. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry revealed a single molecular ion series of 21,285 atomic mass units (predicted, 21,311 atomic mass units). The yield was ϳ40 mg, and purity was ϳ95%.
PfTFB-Plasmid pMLtfb (52), which encodes PfTFB under control of the bacteriophage T7 gene 10 promoter, was introduced by transformation into E. coli strain BL-21 DE3 (Novagen). Growth, induction, harvesting, lysis, and clarification were performed as described above for PfTBP. The supernatant was loaded onto a 2.5 ϫ 20-cm phosphocellulose column (Whatman), the column was washed extensively with 50 mM potassium phosphate buffer (pH 7.8), and the column was eluted with 150 ml of 50 mM potassium phosphate (pH 7.8) and 500 mM KCl. The sample was concentrated to ϳ2 ml by ultrafiltration through a YM3 membrane (Millipore) at 4°C and loaded onto a 2.6 ϫ 60-cm Superdex 75 column (Amersham Biosciences), and the column was eluted in 5-ml fractions of the same buffer. Pooled fractions containing PfTFB (20 ml) were stored in aliquots of 500 l at Ϫ70°C. SDS-PAGE analysis indicated a ϳ34-kDa protein, corresponding to the expected molecular mass of PfTFB. The yield was ϳ25 mg, and the purity was ϳ95%.
Derivatized Promoter DNA Fragments-M13mp18-gdh and M13mp19-gdh were constructed by replacement of the EcoRI-SphI segments of, respectively, M13mp18 and M13mp19 (New England Biolabs) by EcoRI-SphI DNA fragments containing positions Ϫ75 to ϩ45 of the P. furiosus gdh promoter (prepared by add-on PCR using as template plasmid pLU479 (Ref. 8; a generous gift from M. Thomm, University of Regensburg, Regensburg, Germany). Promoter DNA fragments containing a phenyl azide photoactivable cross-linking agent incorporated at a single, defined DNA phosphate and containing an adjacent 32 P label were prepared using M13mp18-gdh (for analysis of the template strand) and M13mp19-gdh (for analysis of the nontemplate strand) as templates by the method of Naryshkin et al. (47), except that after the primer extension and ligation steps products were digested with FcpI and AvaII, 5Ј overhanging ends were filled in using DNA polymerase Klenow fragment (Promega) and dATP, dGTP, dCTP, and dTTP (Amersham Biosciences), and 5Ј phosphates at ends and nicks were removed using calf intestinal alkaline phosphatase (Promega; 3 units for 45 min at 37°C) (filling of 5Ј overhanging ends and removal of 5Ј phosphates from ends and nicks significantly reduced nonspecific crosslinking by end-and nick-bound PfRNAP 2 ).
Site performed in the presence of PfRNAP by the addition of 3 l of 583 g/ml heparin and incubation for 2 min at 70°C. Crosslinking, nuclease digestion, and product analysis were performed by a modification of the procedures of Naryshkin et al. (47). Reaction mixtures were UV-irradiated for 60 s (11 mJ/mm 2 at 365 nm) in a Spectrolinker XL-1000 UV crosslinker (Fisher). Reaction vessels were siliconized polypropylene microcentrifuge tubes contained within borosilicate glass culture tubes (16 ϫ 100 mm) filled with water pre-equilibrated to 70°C. Crosslinked polypeptides were identified by performing nuclease digestion (the addition of 2 l of 100 mM CaCl 2 and 10 units/l DNase I (Sigma) and incubation for 10 min at 37°C followed by the addition of 1 l of 10% SDS and incubation for 5 min at 70°C followed by the addition of 1 l of 30 units/l S1 nuclease (Roche Applied Science) and incubation for 10 min at 37°C), SDS-PAGE, and autoradiography.

RESULTS
Site-specific Protein-DNA Photocrosslinking-The procedure used consists of four steps ( Fig. 1A; Refs. 43-47) as follows: (i) Chemical and enzymatic reactions were used to prepare a DNA fragment containing a phenyl azide photoactivable crosslinking agent ("probe") and an adjacent 32 P radiolabel incorporated at a single, defined DNA phosphate (with a 9.7-Å linker between the reactive atom of the probe and the phosphorus atom of the phosphate and with an ϳ11-Å maximum "reach" between potential crosslinking targets and the phosphorus atom of the phosphate). (ii) The multiprotein-DNA complex of interest was formed using the site-specifically derivatized DNA fragment, and the multiprotein-DNA complex was UV-irradiated, initiating covalent crosslinking with polypeptides in direct physical proximity to the probe. (iii) Extensive nuclease digestion was performed, eliminating uncrosslinked DNA and converting crosslinked DNA to a crosslinked, radiolabeled 3-5 nucleotide "tag." (iv) The tagged polypeptides were identified using denaturing polyacrylamide gel electrophoresis and autoradiography.
For analysis of the P. furiosus transcription initiation complex, we constructed 41 derivatized DNA fragments, each having a probe incorporated at a single, defined phosphate of the P. furiosus glutamate dehydrogenase promoter (gdh; each second phosphate on each strand from positions Ϫ40 to ϩ1;  [43][44][45][46][47]. Lines 1 and 2, chemical and enzymatic reactions are used to prepare a promoter DNA fragment with a photoactivable crosslinking agent (R) and an adjacent radiolabel (*) incorporated at a single, defined site. Line 3, UV irradiation of the derivatized protein-DNA complex initiates crosslinking. Nuclease digestion eliminates uncrosslinked DNA and converts crosslinked DNA to a crosslinked, radiolabeled 3-5-nucleotide tag. B, sequence of the P. furiosus gdh promoter, with probe sites in the 41 derivatized DNA fragments analyzed in this work, indicated by arrows. The transcription start site, the TATA element, and the TFB recognition element (BRE) are indicated by shading; the transcription-bubble region is indicated by a black rectangle.
formed the complex, UV-irradiated the complex to initiate crosslinking, and identified crosslinked polypeptides.
Experiments were performed at 70°C. In vitro transcription assays indicate that the PfRNAP-PfTFB-PfTBP-promoter complex exhibits optimal activity at 70°C (with high activity at 60 -90°C (8-10)). Potassium permanganate footprinting experiments indicate that the PfRNAP-PfTFB-PfTBP-promoter complex is an open complex at 70°C (with a single-stranded "transcription bubble" extending from at least position Ϫ7 to at least position ϩ3 at 70°C (53)). Heparin-challenge experiments further indicate that the PfRNAP-PfTFB-PfTBP-promoter complex is a promoter-specific and transcriptionally competent complex at 70°C. 3 Complications due to crosslinking within nonspecific, nonproductive complexes were avoided by inclusion of heparin (which disrupts nonspecific and non-productive complexes on double-stranded DNA (54)) and by inclusion of filling-in and dephosphorylation steps in DNA-fragment preparation (which prevents formation of heparin-resistant nonspecific and nonproductive complexes at DNA ends and DNA nicks 3 ). Identical results were obtained using an alternative procedure in which complications due to crosslinking within nonspecific and nonproductive complexes were avoided by use of immobilized templates and stringent washing (see the procedures described in Kim et al. (45,46)). 3 As a direct control for specificity of crosslinking, three derivatized promoter DNA fragments with probes well outside the range of reported protein-DNA interactions in the eukaryal initiation complex (44,45), i.e. at positions Ϫ70 and ϩ45 on the nontemplate strand and Ϫ66 on the template strand, were constructed and analyzed. No crosslinking was detected using the three negative control derivatized promoter DNA fragments either in experiments with the PfTFB-PfTBP-promoter complex or in experiments with the PfRNAP-PfTFB-PfTBP-promoter complex (Fig. 2). Representative data for the PfTFB-PfTBP-promoter complex and the PfRNAP-PfTFB-PfTBP-promoter complex are shown in Fig. 2, and the results are summarized in Fig. 3.
PfTFB Photocrosslinking-PfTFB crosslinked in the TATAelement region (positions Ϫ40 to Ϫ14; strong crosslinks at positions Ϫ40 to Ϫ30 and Ϫ22 to Ϫ14 of the nontemplate strand), both in the PfTFB-PfTBP-promoter complex and in the PfRNAP-PfTFB-PfTBP-promoter complex (Figs. 2 and 3). PfTFB also crosslinked in the transcription-bubble region (crosslinks between positions Ϫ10 and ϩ1; strong crosslinks between positions Ϫ10 and Ϫ2), both in the PfTFB-PfTBPpromoter complex and in the PfRNAP-PfTFB-PfTBP-promoter complex (Figs. 2 and 3). Two crosslinks in the transcriptionbubble region in immediate proximity to the transcription start site (crosslinks at positions Ϫ4 and Ϫ2 of the template strand) exhibited a marked dependence on PfRNAP presence, being weak or absent in the PfTFB-PfTBP-promoter complex but strong in the PfRNAP-PfTFB-PfTBP-promoter complex (Fig. 2,  A and B, last two lanes below the promoter sequence; Fig. 3). DISCUSSION Our photocrosslinking results define the positions of PfRNAP subunits, PfTBP and PfTFB relative to positions Ϫ40 to ϩ1 of promoter DNA in the PfTFB-PfTBP-promoter and PfRNAP-PfTFB-PfTBP-promoter complexes in solution. In conjunction with the results of Bartlett et al. (21), which define positions of PfRNAP subunits relative to positions Ϫ1 to ϩ20 of promoter DNA, our results provide an essentially complete description of protein-DNA interactions in the archaeal transcription initiation complex (Supplemental Fig. 1A).
RNAP-Only 2 of the 12 PfRNAP subunits were observed to crosslink to positions Ϫ40 to ϩ1 of the PfRNAP-PfTFB-PfTBPpromoter complex: subunit AЈ and subunit B (Figs. 2B and 3B). Subunit AЈ crosslinked primarily in the transcription-bubble region close to the transcription start site. Subunit B crosslinked both in the transcription-bubble region and upstream of the transcription-bubble region. Our results for the archaeal initiation complex are similar to published results for the bacterial and eukaryal initiation complexes, for each of which only two RNAP subunits were observed to crosslink to positions Ϫ40 to ϩ1: the largest subunit (the homolog of subunit AЈ) and the second largest subunit (the homolog of subunit B) (Supplemental Fig. 1; Refs. 44 -46). We infer that, in the archaeal initiation complex, RNAP interacts with promoter DNA in a manner equivalent to that in the bacterial and eukaryal initiation complexes. The RNAP active-center cleft, comprising determinants of subunit AЈ and subunit B, interacts with the transcriptionbubble region and transcription start site (cf. Refs. 3, 46, 57, and 58), and the RNAP flap or wall, comprising determinants of subunit B, interacts with DNA upstream of the transcription-bubble region (cf. Refs. 3, 46, 57, and 58).
TBP-PfTBP crosslinked exclusively within the TATA ele- ment, both in the PfTFB-PfTBP-promoter complex and in the PfRNAP-PfTFB-PfTBP-promoter complex (Figs. 2 and 3). The results are as expected based on the crystallographic structures of TBP-DNA and TBP-TFBc-DNA complexes (27,28) and on DNase I footprinting experiments with archaeal TBP-DNA and TBP-TFBc-DNA complexes (7,11,13). The results are similar to published results for eukaryal initiation complexes (43,44,59), supporting the homology between the two transcription systems.
TFB-PfTFB crosslinked in the TATA-element region yielding strong crosslinks immediately upstream and downstream of the TATA element, both in the PfTFB-PfTBP-promoter complex and in the PfRNAP-PfTFB-PfTBP-promoter complex (Figs. 2 and 3). TFB-DNA crosslinking in the TATA-element region is as expected based on crystallographic structures of TBP-TFBc-DNA complexes (27,28) and on footprinting experiments with TBP-TFBc-DNA complexes (7,11,13,60,61). TFB-DNA crosslinking immediately upstream of the TATA element is as expected for sequence-specific interaction between the BH4Ј-BH5Ј helix-turn-helix motif of the TFB Cterminal domain and the TFB recognition element (Refs. 13, 28, 59, 61; see also Ref. 29), and TFB-DNA crosslinking immediately downstream of the TATA element is as expected for interaction between the BH2-BH3 loop of the TFB C-terminal domain and the DNA minor groove downstream of the TATA element (13; see also Ref. 29). The pattern of TFB-DNA crosslinking in the TATA-element region of the archaeal transcription initiation complex is similar to the pattern of TFIIB-DNA crosslinking in the TATA-element region of the eukaryal RNAPII transcription initiation complex (Supplemental Fig. 1; Refs. 43-45 and 59), supporting the homology between the two transcription systems.
PfTFB also crosslinked in the transcription-bubble region, yielding crosslinks close to and at the transcription start site, both in the PfTFB-PfTBP-promoter complex and in the PfRNAP-PfTFB-PfTBP-promoter complex (Figs. 2 and 3). TFB-DNA crosslinking in the transcription-bubble region spanned positions Ϫ10 to ϩ1 and involved both DNA strands. TFB-DNA crosslinking close to the transcription start site on the template strand exhibited a pronounced dependence on RNAP, being weak or absent in the PfTFB-PfTBP-promoter complex but strong in the PfRNAP-PfTFB-PfTBP-promoter complex (Fig. 2). We infer that residues of TFB contact or closely approach essentially the entire transcription-bubble region, including the transcription start site. Furthermore, we infer that residues of TFB contact or closely approach both strands of the melted transcription bubble within the PfRNAP-PfTFB-PfTBP-promoter complex and, thus, most likely are located between the two strands of the melted transcription bubble, within the RNAP active-center cleft in the PfRNAP-PfTFB-PfTBP-promoter complex. Finally, we infer that RNAP modulates proximity between TFB and DNA close to the transcription start site through interactions with TFB and/or through interactions with DNA (e.g. melting of the transcription bubble).
In analysis of eukaryal transcription initiation complexes, no crosslinks were observed between TFIIB and the transcription start site region (i.e. no TFIIB-DNA crosslinks were observed downstream of position Ϫ6 (44,45)). However, we emphasize that this negative result permits no strong conclusions for two reasons: (i) the azidophenacyl crosslinking chemistry used is chemoselective (with an essentially absolute requirement that a probe-proximal segment of a target polypeptide contains a lysine, histidine, or tyrosine residue for efficient reaction 4 ) and (ii) the eukaryal complexes analyzed were closed complexes (in the absence of ATP) or mixtures of closed complexes and open complexes (in the presence of ATP) (with a predominance of closed complexes 5 ). We favor the view that, in accord with the high sequence and structural homology between the eukaryal and archaeal initiation complexes, both archaeal TFB and eukaryal TFIIB closely approach the transcription start site region in their respective open initiation complexes but that for technical reasons the interaction is readily detectable for TFB but not readily detectable for TFIIB.
Structural Implications of PfTFB Downstream Crosslinking-Comparison of the observed TFB-DNA crosslinks (Figs. 2 and 3) with the crystallographic structures of complexes of TBP, TFB C-terminal domain, and DNA (27,28) allows us to ascribe all TFB-DNA crosslinking in the TATA-element region (positions Ϫ40 to Ϫ14) to interactions between the TFB Cterminal domain and DNA (Supplemental Fig. 2).
However, we also observe TFB-DNA crosslinking in the transcription-bubble region, close to and at the transcription start site (positions Ϫ10 to ϩ1; Figs. 2 and 3). Based on crystallographic structures of complexes of TBP, TFB C-terminal domain, and DNA (27,28), this crosslinking cannot be ascribed to interactions between the TFB C-terminal domain and DNA. (Crystallographic structures of complexes of TBP, TFB C-terminal domain, and DNA show no interactions between TFB C-terminal domain and DNA downstream of position Ϫ16 (Supplemental Fig. 2).) There is only one structurally plausible explanation for the observed TFB-DNA crosslinking in the transcription-bubble region, near the transcription start site; i.e. a part of TFB other than the TFB C-terminal domain contacts or closely approaches the transcription-bubble region, near the transcription start site. We infer that residues of the TFB N-terminal domain and/or the TFB interdomain linker contact or closely approach the transcription-bubble region, near the transcription start site.
The proposed interaction of residues of the TFB N-terminal domain and/or interdomain linker with the transcription-bubble region, near the transcription start site, immediately suggests a structural and mechanistic explanation for roles of the TFB and TFIIB N-terminal domains in post-recruitment reactions involving the transcription start site: transcription-initiation NTP concentration dependence (33), transcription-initiation efficiency (33, 38 -40), and start-site selection (33,(37)(38)(39)(40)(41)(42). We propose that residues of the TFB and TFIIB N-terminal domains contact or closely approach the transcription-bubble region near the transcription start site and mediate, through direct interactions near the transcription start site, roles in these post-recruitment reactions.
Consistent with the above proposals, footprinting experiments indicate that TFB induces DNase I hypersensitivity within the transcription-bubble region (at positions Ϫ6 and Ϫ5 of the nontemplate strand) and that induction of DNase I hypersensitivity does not require the TFB zinc ribbon but does require the TFB CSB. 6 In the bacterial transcription initiation complex, a region of the initiation factor contacts or closely approaches the transcription-bubble region near the transcription start site: region 3.2 (R3.2, also known as the 3 -4 linker (18,19)). R3.2 consists of two sub-regions: an extended segment that binds within the RNAP RNA exit channel (18,19,57) and a hairpin loop that binds within the RNAP active-center cleft between the two strands of the melted transcription bubble, near the transcription start site (18,19). Deletion of the R3.2 hairpin loop affects transcription-initiation NTP concentration dependence and transcription-initiation efficiency (18,62). The R3.2 hairpin loop can be crosslinked to the initiating NTP (63). The R3.2 hairpin loop has been proposed to interact directly with the initiating NTP and thereby to facilitate de novo, unprimed initiation of RNA synthesis (18,62). Thus, in transcription initiation, region 3.2 performs functions analogous to those of TFB and TFIIB N-terminal domains.
Based on our crosslinking results (Figs. 2 and 3), and on documented roles of TFB and TFIIB N-terminal domain CSBs in transcription-initiation NTP concentration dependence, transcription-initiation efficiency, and start-site selection (33,(37)(38)(39)(40)(41)(42), we suggest that the TFB and TFIIB N-terminal domains in the archaeal and eukaryal transcription initiation complexes occupy positions analogous to those of R3.2 in the bacterial transcription initiation complex. Specifically, we suggest that the TFB or TFIIB N-terminal-domain zinc ribbon binds in or near the RNAP RNA exit channel (as documented by protein-protein photocrosslinking and protein affinity cleaving (64)) and that the TFB and TFIIB N-terminal domain CSBs bind in the RNAP active-center cleft, between the two strands of the melted transcription bubble, near the transcription start site. We further suggest that the TFB and TFIIB N-terminaldomain CSBs make direct interactions with the initiating NTP that facilitate de novo, unprimed initiation of RNA synthesis. Consistent with the above proposals, in the recently determined crystallographic structure of a complex between yeast TFIIB and yeast RNAP II, the TFIIB N-terminal domain zinc ribbon is located in the RNAP RNA exit channel, and the TFIIB N-terminal domain CSB is located in the RNAP active-center cleft. 7