transcription factor e is a part of transcription elongation complexes

Transcription Factor E Is a Part of TranscriptionElongation Complexes*□S

Received for publication, September 4, 2007, and in revised form, October 4, 2007 Published, JBC Papers in Press, October 5, 2007, DOI 10.1074/jbc.M707371200

Sebastian Grunberg‡, Michael S. Bartlett§1, Souad Naji‡2, and Michael Thomm‡3

From the ‡Lehrstuhl fur Mikrobiologie, Universitaet Regensburg, 93053 Regensburg, Germany and the §Department of Biology,Portland State University, Portland, Oregon 97207

Ahomologueof theN-terminal domainof the� subunit of thegeneral eukaryotic transcription factor TFE is encoded in thegenomes of all sequenced archaea, but the position of archaealTFE in transcription complexes has not yet been defined. Weshow here that TFE binds nonspecifically to single-strandedDNA, and photochemical cross-linking revealed TFE binding toa preformed open transcription bubble. In preinitiation com-plexes, the N-terminal part of TFE containing a winged helix-turn-helix motif is cross-linked specifically to DNA of the non-template DNA strand at positions �9 and �11. In complexesstalled at �20, TFE cross-linked specifically to positions �9,�11, and �16 of the non-template strand. Analyses of tran-scription complexes stalled at position �20 revealed a TFE-de-pendent increase of the resumption efficiency of stalled RNApolymerase and a TFE-induced enhanced permanganate sensi-tivity of thymine residues in the transcription bubble. Theseresults demonstrate the presence of TFE in early elongationcomplexes and suggest a role of TFE in stabilization of the tran-scription bubble during elongation.

The archaeal transcriptional machinery is a simplified ver-sion of the polymerase II (polII)4 machinery. For basal in vitrotranscription ofmost promoters, two transcription factors TBPand TFB are sufficient, orthologues to the eukaryotic transcrip-tion factors TBP and TFIIB (reviewed in Refs. 1–3). Archaealgenomes encode a protein with a sequence similar to theN-ter-minal part of the � subunit of TFIIE. The N-terminal part of

TFIIE-� is sufficient for basal transcription in the polII system(4). This finding suggests that the archaeal version of TFIIE ismuch like the evolutionary precursor of TFIIE possessing thecore functions of this transcription factor.TFIIE is a general RNA polII transcription factor that stabi-

lizes the preinitiation complex (PIC) in concert with TFIIH(reviewed by Ref. 5). TFIIE binding to polII recruits TFIIH tothe preinitiation complex (4, 6, 7). It binds to single-strandedDNA (8) and is required for ATP-dependent promoter open-ing by the helicase activity of TFIIH. Negative supercoiling ofthe template and short mismatched heteroduplex DNAaround the initiation sites in topologically relaxed templatesbypass the requirement for ATP, TFIIH, and TFIIE, indicatingan important role of TFIIE and TFIIH in open complex forma-tion (7). TFIIE stimulates both the helicase and the kinase activ-ity of TFIIH (8) responsible for open complex formation andphosphorylation of the C-terminal domain of polII. TFIIE,TFIIH, and ATP play an essential role in promoter clearanceeven on negatively supercoiled templates (7, 10). Thus, TFIIE,which is released before position �10 from early elongationcomplexes (11), also plays an essential role in the transitionfrom initiation to elongation.TFIIE in human cells consist of two subunits � (57 kDa) and

� (34 kDa) that form an �2 �2 heterotetramer (9, 12). Photo-chemical cross-linking results revealed interactions of bothsubunits of TFIIE with promoter DNA in and immediatelydownstream of the region of the open complex (13) and inanother study upstream of the transcription start site (14).Recently, the TFIIE binding site has been located at the Rpb1clamp domain in the PIC (15).Open complex formation in the archaeal PIC occurs inde-

pendent of TFIIH that is not encoded in archaeal genomes.Therefore, analysis of the core functions of TFE is not compli-cated by interactions with the complex multisubunit factorTFIIH in the archaeal system. Archaeal TFE is not essential forin vitro transcription but stimulates the initiation rate at someweakly expressed promoters (16), and mutagenesis of theTATA box or low concentrations of TBP also make the tran-scription of strong promoters dependent upon the presence ofTFE (17). Little is known on the mechanism by which TFEstimulates transcription. TFE can compensate for TFBmutantsin promoter-directed transcription assays and stabilizes PICscontainingTFBmutants in electrophoreticmobility shift assays(18). TFE stimulated RNAP recruitment to heteroduplex tem-plates not bound efficiently by TBP-TFB-RNAP complexesalone (18). This finding led to the speculation that TFE stabi-lizes PICs by enhancing DNA melting and DNA loading. A

* This work was supported by grants from the priority program “Genomefunction and regulation” of the Deutsche Forschungsgemeinschaft (toM. T.) and the American Heart Association Northwest Affiliate to (M. S. B.).The costs of publication of this article were defrayed in part by the pay-ment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734 solely to indi-cate this fact.

□S The on-line version of this article (available at http://www.jbc.org) containsa supplemental figure.

1 To whom correspondence may be addressed: Dept. of Biology, PortlandState University, P.O. Box 751, Portland, OR, 97207. Tel.: 503-725-3858; Fax:503-725-3888; E-mail: [email protected].

2 Present address: The Scripps Research Institute, Dept. of Cell Biology, IMM-10, 10550 North Torrey Pines Rd., La Jolla, CA 92037.

3 To whom correspondence may be addressed: Lehrstuhl fur Mikrobiologie,Universitaet Regensburg, 93053 Regensburg, Universitaetsstrasse 31,D-93053 Regensburg, Germany. Tel.: 49-941-943-3160; Fax: 49-941-943-2403; E-mail: [email protected].

4 The abbreviations used are: polII, polymerase II; RNAP, RNA polymerase;recRNAP, reconstituted RNAP; endRNAP, endogenous RNAP; TF, transcrip-tion factor; PIC, preinitiation complex; NTCB, 2-nitro-5-thiocyanobenzoicacid; nt, nucleotides; TBP, TATA-binding protein; T, template strand; NT,non-template strand.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 282, NO. 49, pp. 35482–35490, December 7, 2007© 2007 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

35482 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 282 • NUMBER 49 • DECEMBER 7, 2007

by guest on August 30, 2016

http://ww

w.jbc.org/

Dow

nloaded from


http://ww

w.jbc.org/

Dow

nloaded from


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

http://www.jbc.org/

http://www.jbc.org/

direct stimulatory effect of TFE on open complex formation, inparticular in the upstream part of the bubble, has been shownrecently by permanganate footprinting (19). The stimulatoryeffect of TFE on transcription depends upon the presence of theE�F heterodimer corresponding to polII subunits Rpb7/4 (20),and the interaction of TFE with RNAP is mediated mostly bysubunit E� (19). Here, we demonstrate that TFE induces ahigher resumption efficiency of stalled elongation complexesand show the presence and position of TFE in initiation andelongation complexes by photochemical cross-linking. Thefindings reported here suggest an unexpected role of TFE in thestabilization of elongation complexes.

EXPERIMENTAL PROCEDURES

Immobilized in Vitro Transcription Assays of Paused TernaryComplexes—C-minus cassettes of the gdh promoter, biotiny-lated at the 5�-end of the template DNA strand, attached tostreptavidin-coated magnetic beads, were used as DNA tem-plates for pausing of early elongation complexes at position�20 relative to the transcription start site �1. Standard reac-tion conditions were used as described in Ref. 19 in a reactionvolume of 35 �l. 168 fmol of immobilized template DNA wereincubated for 30 min at 70 °C in 40 mM NaHEPES, pH 7.3, 250mM NaCl, 5 mM �-mercaptoethanol, 0.1 mM EDTA, 2.5 mMMgCl2, 0.1mg/ml bovine serum albumin, 40�MATP andGTP,2 �M UTP, 0.15 MBq of [�-32P]UTP (110 TBq/mmol), 204 nMTBP, 42 nM TFB, and 45 nM reconstituted RNAP (recRNAP) or23 nM endogenous RNAP (endRNAP). When indicated,endRNAP or recRNAP were replaced by 50 nM RNAP�E�F(19). TFE or RNAP subunits E�Fwere added at various amountsas specified in the legend for Fig. 5. The immobilized templateswere purified as described (21) and resuspended in 35 �l oftranscription buffer preheated to 70 °C. The reactions weresplit in 2 � 15.5 �l; after volume compensation to 20 �l withpreheated (70 °C) transcription buffer, one half was directlydenatured for 3 min at 95 °C with loading dye (21). The volumeof the second half was compensated to 20 �l with preheatedtranscription buffer containing 40 �M all NTPs. The reactionswere incubated for 3min at 70 °C and stoppedby the addition ofloading dye followed by denaturation. Reactions were analyzedby electrophoresis in 28% urea-polyacrylamide gels. Abortivetranscripts from immobilized templates were isolated from thesupernatant of purified ternary transcription complexes asdescribed (21).KMnO4 Footprinting of Initiation and Elongation Com-

plexes—The DNA templates used in footprinting experimentsresembled the templates used for immobilized in vitro tran-scription assays but additionally were labeled with [�-32P]ATPon the free 5�-end of the non-template DNA strand asdescribed previously (21). Footprinting reaction conditionsessentially equaled the conditions described in Ref. 19. 70 nMRNAP, 286 nMTBP, 59 nMTFB, 10�MATP, GTP, 1.5�MUTP,and 500 nM TFE (when indicated) were incubated for 5 min at70 °C in KMnO4 transcription buffer omitting [�-32P]UTP,bovine serum albumin, and �-mercaptoethanol. Complexeswere isolated (21) and washed with 50 �l of KMnO4 transcrip-tion buffer (70 °C). Pellets were resuspended in either 25 �l ofKMnO4 transcription buffer (70 °C) or 25 �l of KMnO4 tran-

scription buffer (70 °C) containing 500 nM TFE as specified. Allreactions were incubated at 70 °C for 2 min. Then, KMnO4 wasadded to a final concentration of 23 nM. The samples were incu-bated for 5 min at 70 °C, stopped, and exposed to piperidinetreatment as described previously (21).Photochemical Protein-DNA Cross-linking of Paused Tran-

scription Complexes—Paused elongation complexes wereassembled as described (19). DNA templates used in photo-chemical cross-linking experiments were internally �-32P-la-beled C-minus�20 cassettes of the gdh promoter DNA con-taining azidophenacylated phosphorothioate at specificlocations, prepared and radiolabeled by DNA polymerase-di-rected incorporation of one radioactive dNTP 3� to the phos-phorothioate, essentially as described (22). 2 nM template DNAwere incubated with 20 nM TBP, 60 nM TFB, 46 nM RNAP, and140 or 500 nMTFE as specified in 12.5�l of transcription buffercontaining 40 mM NaHEPES, pH 7.3, 250 mM NaCl, 2.5 mMMgCl2, 0.1 mM EDTA, 600 ng of nonspecific competitor DNA(herring spermDNA), and 40 �MATP, GTP, and UTP. NTP-Cmix was omitted when indicated. Complexes were formed at70 °C; specific competitor DNA (gdh promoter DNA bp �164to �113 at 400 nM) was added after 3 min. Reactions wereUV-irradiated for 7 min at 70 °C 2 min after the addition ofspecific competitor DNA. FollowingUV irradiation, complexeswere treated with nucleases and finally analyzed on 4–19% or8–19% gradient polyacrylamide-SDS gels, essentially asdescribed (23). Radiolabeled proteins were visualized using animage plate and image analyzer (FLA-500, Fuji, Japan). Photo-chemical protein-DNA cross-linking of initiation complexesexperiments were performed as described in Ref. 24, and TFEwas added at 140 nM as specified.Chemical Cleavage of Cross-linked TFE—Ten 12.5-�l tran-

scription initiation complex reactions containing TBP, TFB,TFE, and RNAP were assembled using gdhP derivatized withAPB at �9 and radiolabeled at �8 and �7 of the non-tran-scribed strand, as described above. Following cross-linking, thenuclease-treated reactions were pooled, and proteins were sep-arated by SDS-PAGE8–24%gradient. The band correspondingto TFE was cut out, and TFE was eluted overnight at 37 °C in 1ml of 10 mM Tris, pH 8.0, with bovine serum albumin (25�g/ml). Eluted TFE was concentrated by microfiltration(10,000 molecular weight cut-off), and aliquots were treatedwith NTCB (50 mM), which cleaves at cysteines, essentially asdescribed (24). Mock treatments were performed under thesame conditions but in the absence of NTCB. The 25-�l cleav-age reactions were diluted to 500 �l with equilibration buffer(20 mM Tris, pH 8.0, 300 mM NaCl, and 0.01% Tween 20) andconcentrated by microfiltration (3,000 molecular weight cut-off) followed by a second addition of 500 �l of equilibrationbuffer and concentration to �20 �l. Samples were split, withhalf representing “input,” and half mixedwithNi2�-conjugatedparamagnetic beads for 20 min at room temperature. Superna-tant was saved (“super”), and the beads were washed two timeswith 200 �l of equilibration buffer. Protein was eluted from thebeads by incubation for 10 min in equilibration buffer plus 250mM imidazole (“elute”). Input, super, and elute were analyzedby SDS-PAGE 10–24% gradient, and dried gels were visualizedby phosphorimaging.

TFE in Elongation Complexes

DECEMBER 7, 2007 • VOLUME 282 • NUMBER 49 JOURNAL OF BIOLOGICAL CHEMISTRY 35483


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Photochemical Protein-DNA Cross-linking of TFE to Single-stranded DNA—Nonspecific TFE/single-stranded DNA cross-links were performed using six single-stranded DNA oligonu-cleotides containing the sequence of the non-template strandofthe Pyrococcus gdh promoter DNA from �30 to �13. Eacholigonucleotide contained an azidophenacylated phosphoro-thioate at one specific location (at positions �26, �21, �16,�10,�6,�1, and�1 relative to the transcription start site) and

was radiolabeled at the 5�-endwith [�-32P]ATP (21). Equal con-centrations of radiolabeled and photoreactive oligonucleotideswere incubated in binding reactions at a final concentration of 1�M as described under “Photochemical Protein-DNA Cross-linking of Paused Transcription Complexes” in the absence ofNTPs. Templates were incubated with 1.9 �M TFE, rpoF orrpoE�, or rpoL at 70 °C for 5 min. Then, the reactions wereUV-radiated at 70 °C. After 5min of further incubation, loading

dye was added; the reactions weredenatured for 3 min at 95 °C andloaded onto 4–19% gradient poly-acrylamide SDS-gels.PhotochemicalProtein-DNACross-

linking of TFE and RNAP to DNAwithin a Preopened TranscriptionBubble—For cross-linking of TFEand RNAP to DNA containing apreopened transcription bubble,templates were prepared essentiallyas described (22) with the followingchanges. 5� immobilized oligonu-cleotides containing an phosphoro-thioate at one specific location wereradiolabeled and extended to fulltemplate length basically asdescribed in Ref. 22. Then, the twostrands were separated via NaOHtreatment, and the strand carryingthe phosphorothioate with adjacentradiolabel was hybridized to a longoligonucleotide with mismatchregion comprising positions �6 to�6 (19) relative to the transcriptionstart site �1 and azidophenacylatedas described (22). For native RNAPshift experiments, 2 nM photoreac-tive DNA template were incubated

FIGURE 1. TFE binds to a mismatched bubble. A, sequence and features of the preopened transcriptionbubble DNA construct after EcoRI treatment, comprising a mismatch region from position �6 to �6 relative tothe transcription start site �1 (boxed) with azidophenacylated phosphorothioate at position �3 (circled, flashsign) and adjacent radiolabel at position �1 (circled, radioactive sign) on the non-transcribed strand. B, RNAPcan bind to photoactive preopened bubble DNA in the absence of TBP/TFB. DNA (2 nM) was incubated withvarious amounts of RNAP (21, 42, 105, and 210 nM) at 70 °C and UV-irradiated before loading for electrophore-sis. C, TFE and RNAP bind to the photoactive, preopened DNA template independently. TFE was added at 1.9�M (lane 2) to photocross-linking reactions. RNAP was added at 105 nM (lane 3), and RNAP and TFE were addedat 105 nM and 1.9 �M (lane 4), respectively.

FIGURE 2. TFE cross-links to the non-template strand in PICs. A, TFE cross-linking at position �9 occurs only on the non-transcribed strand of the gdhpromoter. After transcription complexes were formed with RNAP, transcription factors as specified, and promoter probe photoactivated at position �9 on thenon-transcribed strand (left panel) or the transcribed strand (right panel), cross-links were induced by UV irradiation, processed with nuclease, and resolved bySDS-PAGE as described under “Experimental Procedures.” Cross-linked proteins were identified by their relative electrophoretic mobilities and indicated at theright side of the gels. B, cross-linking of TFE, TFB, and RNAP in the region from �4 to �21 of the non-template strand of the gdh promoter. Cross-linking toprobes specifically photoactivated at the positions indicated in the figure was performed as in A.




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

in the dark with various amounts of RNAP as indicated in thelegend for Fig. 1 in transcription buffer at 70 °C for 5 min andUV-irradiated for 7min at 70 °C. Following electrophoresis in a5% native gel (22), samples were visualized using an image plate

and image analyzer. For cross-link-ing experiments, templates (and 105nM RNAP when specified) wereincubated at 70 °C. After 3min, TFEwas added to a final concentrationof �1.9 �M. Samples were incu-bated for an additional 2 min at70 °C followed by UV radiation,nuclease treatment, and SDS-gelelectrophoresis as describedunder “Photochemical Protein-DNA Cross-linking of PausedTranscription Complexes.”

RESULTS

TFE Interacts with DNAUpstream of the Open Complex inPreinitiation Complexes—Eucary-otic TFIIE is involved in promoterbinding andmelting (5), andTFIIE�is cross-linked with promoter DNAat position �10, immediatelyupstream of the open complex (25).To investigate the interaction ofTFE from Pyrococcus furiosus withthe promoter region in the PIC, weconducted photochemical cross-linking assays with single-strandedDNA, a premelted heteroduplexmimicking an open complex, andwe also analyzed cross-linking tothe gdh promoter at positionsupstream of the transcription startsite in the PIC.We used six single-stranded

DNA templates extending fromposition �30 to �4 of the gdh pro-moter with azidophenacylatedphosphorothioate at positions �26,�21, �16, �10, �6, and �1. TFEcross-links to these single-strandedprobes were observed after SDS-PAGE (data not shown). In controlreactions, no cross-linking wasobserved with purified RNAP sub-units F and L, but subunit E �wasalso cross-linked to this probe (datanot shown). This finding indicatesthat archaeal TFE can bind to sin-gle-stranded DNA. A heteroduplexwith misspaired DNA in the regionfrom�6 to�6 (Fig. 1A) and a cross-linkable derivatization at position�3 was bound by RNAP in electro-

phoreticmobility shift assays (Fig. 1B) at enzyme toDNA ratiosranging from20- to 100-fold (lanes 3–5).When binding of TFE,RNAP, and a combination of TFE and RNAP to this templatewas studied by photochemical cross-linking, TFE, RNAP, and

FIGURE 3. A, NTCB cleavage and purification of cross-linked TFE. �9NT cross-linking reactions containing TBP,TFB, TFE, and RNAP were separated by SDS-PAGE, the TFE band was excised, and the labeled protein was elutedfollowed by NTCB treatment. Half of the reactions was directly loaded onto an SDS-PAGE gel (lane 2). The otherhalf of the reactions were bound to TALON (Ni2�) beads. Unbound TFE in the supernatant of a binding reactionwith Ni2� beads was analyzed in lane 4, and Ni2�-bound TFE was eluted with imidazole and analyzed in lane 6.In control reactions, the NTCB treatment was omitted (lanes 1, 3, and 5), respectively. B, predicted cleavageprofile for N-terminal His-tagged TFE. Since NTCB cuts at cysteine residues, three single hit products are pos-sible, each inducing one long cleavage product (15–17 kDa) containing a winged helix domain and one smallercleavage product (6 –9 kDa) containing the C-terminal region.

FIGURE 4. A, TFE stimulates synthesis of short RNA products. The addition of various amounts of TFE to animmobilized transcription assay containing endogenous RNAP (lanes 1–3) leads to a significant increase ofsynthesis of short RNA products. TFE was added at 36 (lane 2) or 357 nM (lane 3). B, subunits E�F are importantfor significant stimulation of abortive transcription by TFE. Transcription reactions were conducted asdescribed under “Experimental Procedures” and contained 50 nM RNAP�E�F. 260 nM E� and 405 nM F wereadded in lanes 2 and 6, whereas 660 nM E� and 1 �M F were added to the reactions analyzed in lane 3. Increasingamounts of TFE were added as indicated in the figure (71 nM, lane 4; 357 nM, lanes 5 and 6). C, purification ofimmobilized ternary complexes (21) and of released abortive transcripts analyzed in panels A and B.




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

both in combination were found to be cross-linked to position�3 of the non-template strand after SDS-PAGE (Fig. 1C). Aprobe containing duplex DNA of the gdh promoter derivatizedat position �9 was incubated in binding reactions with RNAP,TBP, TFB, and TFE, and cross-linking to TFB and RNAP sub-units B, A�, and A� was observed both on the template and onthe non-template strand (Fig. 2A, lanes 2–4). By contrast,cross-linking of TFE was only foundat the non-template strand (Fig. 2A,lanes 5 in the left and right panel).Cross-linking of TFE to probes pho-tolabeled at the non-template standbetween position �4 and �21 wasalso studied. Strong cross-linking ofTFE at position�11 (Fig. 2B, lane 4)and a weak cross-linking signal atposition �4 were observed (Fig. 2B,lanes 4 and 8). No cross-linking wasobserved at positions �21 and �5(Fig. 2B, lanes 2 and 6). This findingindicates that archaeal TFE interactsspecifically with the non-templatestrand of promoter DNA mainly inthe region upstream of or at theupstream end of the transcriptionbubble. Cross-linking of TFE wasonly observed in the presence ofRNAP and of TBP/TFB (Fig. 2A).This finding indicates that a com-plete PIC and the recruitment ofRNAP to the promoter are a prereq-uisite for loading of TFE to the PIC.In the N-terminal domain of TFE

from Sulfolobus solfataricus, thecrystal structure of a characteristicwinged helix-turn-helix motif hasbeen resolved (26; PDB number191h) that is also conserved in Py-rococcus TFE (see supplementalmaterial). To identify the domain ofPyrococcus TFE responsible forinteraction with the non-templatestrand of the promoter DNA, N-ter-minal His6-tagged TFE was isolatedfrom gel-purified PICs cross-linkedto the non-template strand at posi-tion �9 and cleaved by a reagentspecific for cysteine residues(NTCB). The cleavage products ofTFE consisting of 15.1, 17.3, and17.6 kDa are expected to contain theN-terminal part of TFE harboringthe winged helix-turn-helix motif(Fig. 3B). The purified NTCB cleav-age products were incubated withTALON-Ni2�-coated beads, andproteins attached to the beads werepurified and eluted from the beads

with imidazole. Both the non-bound fraction in the superna-tant of binding reactions with Ni2�-coated beads and the frac-tion eluted with imidazole from the beads were analyzed bySDS-PAGE (Fig. 3A). NTCB-treated cross-linked TFE wascleaved to a major product of �15 kDa (Fig. 3B, lane 2). Thisfragment was enriched in the fraction eluted fromNi2�-coatedbeads (Fig. 3A, lane 6) and not in the supernatant from binding




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

reactions containing the Ni2�-coated beads (Fig. 4A, lane 4).This finding indicates that the N-terminal part of TFE contain-ing the winged helix-turn-helix motif is involved in binding ofTFE to DNA in the PIC.TFE Stimulates Abortive Transcription—When transcrip-

tion complexes were stalled at position �20 on immobilizedtemplates containing the gdh promoter (21) and the complexeswere purified by magnet particle separation, small amounts of5-nt abortive transcripts were found in the supernatant of thereaction (see Fig. 4C for the experimental setup) and weretherefore released from the complexes (Fig. 4A, lane 1). Abor-tive transcripts are generally thought to be formed in repeatedcycles by promoter-associated RNAP before the enzyme disen-gages from the promoter (reviewed by 27). The synthesis ofabortive transcripts was �3-fold increased in the presence ofTFE (Fig. 4A, lanes 2 and 3). This finding indicates that TFEstabilizes these early promoter-bound transcription com-plexes. TFE interaction with RNAP has been shown to bedependent upon RNAP subunits E�F (19, 20). Therefore, westudied the effect of TFE on transcription by the reconstitutedarchaeal core enzyme lacking these subunits (19). As expected,TFE had only a weak stimulatory effect on abortive transcriptionby the core enzyme, and synthesis of the released 5-nt transcriptwas significantly increasedwhen the reactionswere supplementedwith E� and F or subunit E� alone (Fig. 4B, lanes 6 and 7).TFE Increases the Resumption Efficiency of Stalled Complexes

of the Post-escaped State—The archaeal RNAP starts transloca-tionwhenRNA reaches a length of 10 nt (21) and structural andbiochemical studies of ternary polII complexes suggest that ter-nary complexes reach their full stability when the growing RNAis 14–15nt in length (28, 29). To investigate the effect of TFEonlater steps of transcription, we analyzed the synthesis of a tran-script stalled at position �20 and the resumption of stalledtranscription complexes in the presence and absence of TFE bythe reconstituted Pyrococcus RNAP. TFE had amoderate effecton the formation of stalled ternary complexes (Fig. 5A, comparelanes 1, 3, and 5, and 5C, lanes 1 and 3; a quantification of theresults is shown in Fig. 5A in the right panel) consistent with a�1.5-fold stimulatory effect on the synthesis of run-off tran-scripts from the gdh promoter (19). We next analyzed theresumption of ternary complexes by the addition of a completeset of NTPs to stalled complexes and studied the decrease ofthe 20-nt RNA in ternary complexes after the NTP chase. Theresumption of stalled complexes formed by the reconstitutedRNAP in the absence of TFE was �50% (Fig. 5A, lane 2). WhenTFE was added to initiation reactions, it increased resumption

from 50 to �75% (Fig. 5A, compare lanes 2, 4, and 6). Consid-ering that the maximum resumption efficiency is 100%, thisfinding indicates a significant TFE-induced increase in theresumption of stalled complexes. A quantification of theresumption is shown in Fig. 5A, in the left panel (below the gel).When TFE was not present during initiation reactions butadded later to stalled complexes (Fig. 5C, lanes 5 and 6), it didnot exert an effect on elongation of stalled complexes. Thisfinding suggests that TFE is recruited only to the PIC and not tostable ternary complexes. To further investigate the role of TFEon the elongation of stalled transcripts, we next purified com-plexes of the core enzyme stalled at position �20. Consistentwith the previously published result (19, 20) that TFE contactsRNAP via E�F, TFE had only a moderate effect (�10%) onresumption of ternary complexes formed by the core enzyme(Fig. 5B, lanes 11 and 12) but dramatically stimulated elonga-tion of stalled RNAP from 10 to �30 or 35% when E� (Fig. 5B,

FIGURE 5. A, TFE increases the resumption of stalled elongation complexes. The reaction mixture contained DNA template, TBP, TFB, and 43 nM recRNAP asindicated. TFE was added to 36 nM (reaction analyzed in lanes 3 and 4) or 357 nM (reaction analyzed in lanes 5 and 6). Ternary transcription complexes werestalled at position �20 by the use of a C-minus cassette of the gdh promoter (21) in reactions not containing CTP. Ternary complexes were purified by the useof streptavidin-coated beads, and resumption was analyzed by the addition of a complete set of NTPs (Chase). Transcripts in lanes 2, 4, and 6 represent thefraction of stalled ternary elongation complexes that were unable to resume transcription after the addition of NTPs. The total activity in stalled ternarycomplexes is shown in the diagram on the right of the gel, and the resumption of stalled complexes is shown in the diagram below the gel. B, subunit E� isessential for the TFE-induced increase of the resumption efficiency of stalled complexes. TFE was added at a molar concentration of 357 nM TFE (lanes 5 and 6),F at 405 nM, and E� at 260 nM as indicated. C, TFE can be recruited only to PICs. Stalled transcription complexes were formed in the absence (lanes 1 and 2 and5 and 6) and the presence of 500 nM TFE (lanes 3 and 4). Complexes were washed with preheated transcription buffer and resuspended, and one half of thereaction was complemented with a complete set of NTPs (lanes 2, 4, and 6), whereas the other half was directly denatured (lanes 1, 3, and 5). To chase thereactions analyzed in lane 6, TFE was added to purified ternary complexes at a final concentration of 500 nM. Following incubation for 3 min at 70 °C, reactionswere denatured and analyzed on 28% urea-PAGE. The total activity in stalled complexes is shown on the right, and the resumption efficiency of stalledcomplexes is shown below the figure.

FIGURE 6. TFE stabilizes the transcription bubble in an elongation com-plex. Lanes 2– 4 show a mixed population of open DNA sites in initiation andelongation complexes (paused at position �20 relative to the transcriptionstart site �1 on the immobilized gdh-C20 template). Thymine residues wereanalyzed by the single strand-specific reagent KMnO4 on a 10% sequencinggel. Reactions containing TBP/TFB and endogenous RNAP (when indicated)were incubated for 5 min at 70 °C, washed, resuspended, and again incubatedfor 2 min at 70 °C followed by KMnO4 treatment as described under “Experi-mental Procedures.” Reactions shown in lane 3 contained 500 nM TFE addedto PICs, whereas in the reactions analyzed in lane 4, TFE was added to stalledternary complexes. Significantly modified bases in the region of the elonga-tion complex are highlighted with black spots (lane 3). Boxed bands in lanes 2,3, and 4 were used for quantification of permanganate sensitive signals. Thequantification of permanganate footprints in elongation complexes is shownto the right.




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

lanes 5 and 6) and E�F (Fig. 5B, lanes3 and 4) was added to initial tran-scription reactions. These findingssuggest that TFE added prior to ini-tiation affects the conformation ofthe elongation complex in the post-escaped state.To investigate the effect of TFE

on the resumption of stalled com-plexes in more detail, we analyzedthe permanganate sensitivity of thetranscription bubble in both pro-moter-bound binary complexes andelongated ternary complexes. Tothis end, theNTP concentrationwasoptimized to obtain mixed popula-tions of promoter-bound and stalledcomplexes (see “Experimental Pro-cedures”). When TFE was addedbefore the transcription reactionswere started with NTPs, both thesignals corresponding to the opencomplex at the promoter and the�11, �14, and �16 signal corre-sponding to the bubble in elonga-tion complexes were intensified(Fig. 6, compare lanes 2 and 3). Thisfinding indicates that the addition ofTFE to initiation reactions alsoaffects the stalled elongation com-plex. This effect is not due to anincreased formation of stalled com-plexes in the presence of TFEbecause TFE increased the forma-tion of stalled complexes only by20% (Fig. 5A), whereas the stimula-tory effect on bubble opening in theternary complexwas�4-fold (Fig. 6,right panel). This finding suggeststhat TFE increases resumption ofstalled complexes by stabilizing thetranscription bubble. When TFEwas added to reactions containingstalled complexes and these reac-tions were incubated for a further 2min, only the modification of thy-mine residues located in the pro-moter region was stimulated (Fig. 6,lane 4). This finding confirms ourconclusion inferred from analyses ofthe resumption efficiency of stalledcomplexes that TFE is unable tobind to stable elongation complexes.TFE Cross-links in Stalled Elon-

gation Complexes to the Non-tem-plate Strand of the TranscriptionBubble—Permanganate footprint-ing revealed that TFE-induced sta-




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

bilization of the transcription bubble in stalled elongation com-plexes occurs at positions �11, �14, �15, and �16 (Fig. 6). Toinvestigate the molecular basis for the TFE interaction withelongation complexes inmore detail, the immobilized gdh tem-plate was photoactivated at positions �9, �11, �16, and �21,both at the template and at the non-template DNA strand andcross-linking of TFE and RNAP to these probes, was studied(Fig. 7). In the panels showing cross-linking of transcriptionalcomponents to position �9, �11, and on to the non-templatestrand of position �21, a control reaction lacking NTPs wasdone to reveal the cross-linking pattern in preinitiation reac-tions (right lane in these panels). For each control reaction, nocross-linking of TFE was observed. This finding indicates thatthe cross-linking signals obtained after the addition of NTP totranscription reactions are specific for stalled elongation com-plexes. As predicted from published work (22), cross-linkingof RNAP subunits B, A�, and A� was observed in the PICs atpositions �9, �11, and �16 but not at position �21 (Fig. 7,last lane each in panel, �9, �11, and in the right part ofpanel, �21).When TFE was added to transcription reactionscontaining NTPs, TBP and TFB cross-linking of TFE wasobserved in panels �9, �11, and �16, specifically on the non-template strand (see lanes 6 and 7 in the right part of eachpanel). No cross-linking of TFE to the non-template strandcould be detected at position �21 and on all studied positionsof the template strand. Cross-linking of RNAP subunits B andA� was increased in the presence of TFE in some instances (Fig.7B, panel �9, template strand, and panel �21, non-templatestrand) and cross-linking of A�decreased by TFE (see panel�11, template strand). Taken together, these findings suggestthat TFE affects the conformation of the elongation complexand indicate binding of TFE to the non-template strand in thetranscription bubble in elongation complexes.

DISCUSSION

In the first part of this study, we have identified the positionof archaeal TFE in the PIC. Archaeal TFE contains a wingedhelix domain, suggesting a potential for TFE to bind DNA,although the putative DNA recognition helix is negativelycharged and binding of this domain toDNAwas not detected inelectrophoretic mobility shift assays (26). We show here bycross-linking that TFE such as eukaryotic TFIIE (8) can bind tosingle-stranded DNA (data not shown) and to a preopenedbubble both in the presence and in the absence of DNA-bound RNAP (Fig. 1). In addition, TFE is cross-linked spe-cifically to the non-template DNA strand upstream of the

transcription start site in PICs (Fig. 2), and cleavage of thecross-linked TFE molecule revealed that the N-terminaldomain harboring the winged helix fold is involved in DNAbinding (Fig. 3). These data indicate that this domain isinvolved in interactions of TFE with DNA of the non-templateDNA strand in the PIC mainly in the region upstream of theopen complex. Intriguingly, TFE has been shown to promoteopen complex formation in particular in the upstream part ofthe transcription bubble (19), and the findings reported heresuggest that this stabilization of the open complex is mediatedby direct interaction of TFEwith the non-templateDNA strand(Fig. 2). TFIIE� andTFIIE�were cleaved by probes in promoterDNA downstream of the TATA box and overlapping the tran-scription start site in the yeast PIC (30). Cleavage of TFIIE�mapped to the winged helix domain. The cross-linking datapresented here suggest that archaeal and eukaryotic TFIIE�interact with a similar region of DNA in the PIC as predicted onthe basis of the remarkable structural similarity of the wingeddomain of eukaryotic TFIIE� and archaeal TFE (26).

Archaeal TFE interacts directly with TBP and RNAP (17).This finding and the absence of TFIIF and of TFIIH in thearchaeal transcriptional machinery suggest a major role of TFEas bridging factor between RNAP and TBP (26). The stabilizingeffect of TFE on TBP binding and RNAP recruitment may bealso responsible for the stimulatory effect of TFE on the syn-thesis of a 5-nt abortive transcript (Fig. 4). Intriguingly, thesynthesis of a 4-nt abortive transcript is a major feature ofunstable early transcription complexes in the pre-escape stateof polII (reviewed by Ref. 27). The findings reported here andpublished data (19) support the conclusion that TFE is involvedin the synthesis of these abortive transcripts by stabilizing thetranscription bubble in promoter-bound complexes in bothsystems.Eukaryotic TFIIE has a clearly established role in the TFIIH-

driven transition from initiation to elongation (reviewed by Ref.5), but the release of TFE from early elongation complexesbefore the formation of a 10-nt transcript was reported (11).Several lines of evidence presented in this report indicate acompletely unexpected additional role of archaeal TFE at laterstages of elongation. First, TFE enhances the resumption effi-ciency of transcription complexes stalled at position �20 (Fig.5). The finding that the addition of TFE to purified stalled com-plexes has no effect on resumption (Fig. 5c, lanes 5 and 6) indi-cates that TFE recruitment can occur only during formation ofthe PIC and not at later stages of the transcription cycle. Sec-

FIGURE 7. A, sequence and photolabeling of the gdh promoter. The probe used for cross-linking contained bp �39 to �35, relative to the transcription start site,indicated by the bent arrow. The TATA box and the pausing position (Pos.) �20 are boxed, and the locations of photoactive labels (phosphorothioate derivativecontaining azidophenacyl bromide) on the template and the non-template strand, respectively, are indicated by asterisks. B, stalled transcription complexeswere formed with photoactive DNA templates and challenged with nonspecific (nonsp.) competitor DNA and specific promoter DNA (see “ExperimentalProcedures”). Gdh promoter DNA was added 2 min prior to UV radiation and nuclease treatment (Dnase I and S1 nuclease in combination or Dnase I and S7nuclease in combination). The addition of specific competitor DNA minimized the amount of free transcriptional components, and therefore, nonspecificcross-linking. The cross-linked proteins were analyzed in 4 –19% gradient SDS-PAGE. The position of the azidophenacylated phosphorothioate in eachtemplate is indicated at the left side. The left panel shows the resulting cross-links on the template strand, whereas cross-links on the non-template strand areshown in the right panel. Proteins added to each reaction are indicated on the top of the gels. Triangles indicate an increase of TFE from 140 to 500 nM TFE in thereaction. Cross-linked proteins were identified by their relative electrophoretic mobilities and are indicated at the right-hand side of each gel. For controlpurposes, some reactions were incubated without NTPs (�9T and �9NT, lanes 8; �11T and �11NT, lanes 9; �21NT, lane 7). Cross-linked reactions for position�9T and �21NT were proceeded with S7 nuclease, which can be radiolabeled autocatalytically (the position of the labeled S7 nuclease is indicated). Theposition of undigested DNA identified in control lane 1, not containing proteins, is indicated. C, schematic diagram of the experimental design for photochem-ical DNA-protein cross-linking of paused ternary elongation complexes.




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

ond, permanganate footprinting revealed a strong TFE-in-duced stabilization of the transcription bubble in elongationcomplexes stalled at position�20 (Fig. 6). This finding suggeststhat TFE translocates with RNAP following initiation and actsby stabilizing the transcription bubble in elongating complexes.The physical association of TFE with the elongation complexhas been clearly shown directly by photochemical cross-linking(Fig. 7). Analysis of immobilized complexes containing RNAPin both the pre-escaped and the post-escaped state revealedthat TFE was specifically cross-linked to the non-templatestrand of the transcription bubble from position �9 to �16(Fig. 7). The cross-linking pattern of the RNAP subunits A�, A�,and B was also modified in TFE containing ternary complexes.These findings provide evidence that TFE is a part of earlyelongation complexes and also affects the conformation ofthe ternary elongation complexes. It is tempting to speculatethat the presence of TFE in elongation complexes in vivoincreases processivity and decreases pause site occupancy.For example, in particular on long transcripts or operonswith several genes, the presence of TFE could have a verylarge cumulative effect. A direct role of eukaryotic TFIIE inelongation has not been shown, but a potential effect on theprocessivity of polII by modification of subunits has beendiscussed (reviewed by Ref. 5). Detailed structural studies ofelongation complexes stalled in various registers and cross-linking analyses of transcription complexes at later stages oftranscription are required for a deeper understanding of thefunction of TFE in elongating complexes. Our results reveala novel core function of TFE in elongation that might also berelevant in eukaryotic transcription complexes lackingTFIIH (31).

Acknowledgments—We thank Jaimie Powell and Michael Micorescufor assistance in cloning and expression of recombinant TFE.

REFERENCES1. Bartlett, M. S. (2005) Curr. Opin. Microbiol. 8, 677–6842. Geiduschek, E. P., and Ouhammouch, M. (2005) Mol. Microbiol. 56,

1397–14073. Thomm, M. (2007) in Archaea Molecular and Cellular Biology (Cavic-

chioli, R. ed) pp. 139–157, American Society for Microbiology, Washing-ton, DC

4. Ohkuma, Y., Hashimoto, S.,Wang, C. K., Horikoshi,M., and Roeder, R. G.(1995)Mol. Cell. Biol. 15, 4856–4866

5. Ohkuma, Y. (1997) J. Biochem. 122, 481–4896. Maxon, M. E., Goodrich, J. A., and Tjian, R. (1994)Genes Dev. 8, 515–5247. Holstege, F. C. P., van der Vliet, P. C., and Timmers, H. Th. M. (1996)

EMBO J. 15, 1666–16778. Kuldell, N. H., and Buratowski, S. (1997)Mol. Cell. Biol. 17, 5288–52989. Ohkuma, Y., and Roeder, R. G. (1994) Nature 368, 160–16310. Goodrich, J. A., and Tjian, R. (1994) Cell 77, 145–15611. Zawel, L., Kumar, K. P., and Reinberg, D. (1995)Genes Dev. 9, 1479–149012. Ohkuma, Y., Sumimoto, H., Horikoshi, M., and Roeder, R. G. (1990) Proc.

Natl. Acad. Sci. U. S. A. 87, 9163–916713. Kim, T.-K., Ebright, R. H., and Reinberg, D. (2000) Science 288,

1418–142114. Robert, F., Forget, D., Li, J., Greenblatt, J., and Coulombe, B. (1996) J. Biol.

Chem. 271, 8517–852015. Chen, H.-T., Warfield, L., and Hahn, S. (2007) Nat. Struct. Mol. Biol. 14,

696–70316. Hanzelka, B. L., Darcy, T. J., and Reeve, J. N. (2001) J. Bacteriol. 183,

1813–181817. Bell, S. D., Brinkman, A. B., van der Oost, J., and Jackson, S. P. (2001)

EMBO Rep. 21, 133–13818. Werner, F., and Weinzierl, R. O. J. (2005)Mol. Cell. Biol. 25, 8344–835519. Naji, S., Grunberg, S., and Thomm, M. (2007) J. Biol. Chem. 282,

11047–1105720. Ouhammouch, M., Werner, F., Weinzierl, R. O. J., and Geiduschek, E. P.

(2004) J. Biol. Chem. 279, 51719–5172121. Spitalny, P., and Thomm, M. (2003) J. Biol. Chem. 278, 30497–3050522. Bartlett, M. S., Thomm,M., and Geiduschek, E. P. (2000)Nat. Struct. Biol.

7, 782–78523. Bartholomew, B., Tinker, R. L., Kassavetis, G. A., and Geiduschek, E. P.

(1995)Methods Enzymol. 262, 476–49424. Bartlett, M. S., Thomm, M., and Geiduschek, E. P. (2004) J. Biol. Chem.

279, 5894–590325. Forget, D., Langelier, M.-F., Therien, C., Trinh, V., and Coulombe, B.

(2004)Mol. Cell. Biol. 24, 1122–113126. Meinhart, A., Blobel, J., and Cramer, P. (2003) J. Biol. Chem. 278,

48267–4827427. Dvir, A. (2002) Biochim. Biophys. Acta 1577, 208–22328. Dvir, A., Tan, S., Conaway, J. W., and Conaway, R. C. (1997) J. Biol. Chem.

272, 28175–2817829. Dvir, A., Conaway, R. C., and Conaway, J. W. (1997) Proc. Natl. Acad. Sci.

U. S. A. 94, 9006–901030. Miller, G., and Hahn, S. (2006) Nat. Struct. Mol. Biol. 13, 603–61031. Parvin, J. D., and Sharp, P. A. (1993) Cell 73, 53–540




http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Supplemental material:

PF0491 --MGRD----KKNTALLDIARD-----------IGGDEAVEVVKALEKKG-EATDEELAE 42

PH0619 --MGGEQRMSKRNKELLEIGRD-----------IGGDEAVEIIKALEKKG-EATDEELAE 46

MJ0777 MRSVQSMRKEKKIERIYEMLNDPLVQEVLFNIFEGDEKGFEVIDVLLEKG-ETTEEEIAK 59

SSO0266 --------MVNAEDLFINLAKS-----------LLGDDVIDVLRILLDKGTEMTDEEIAN 41

: : :: .. .:. .::: * .** * *:**:*:

PF0491 LTGVRVNTVRKILYALYDAKLATFRRVRDDETGWYYYYWRIDTKRLPEVIRTRKLQELEK 102

PH0619 ITGIRVNTVRKILYALYDAKLADFRRVKDDETGWYYYYWHIETKRLPEIIRARKMQELEK 106

MJ0777 ELGVKLNVVRKLLYKLYDARLVDYKRWKDEDTNWYSYTWLPTLEKLPYVVKKKINELIKD 119

SSO0266 QLNIKVNDVRKKLNLLEEQGFVSYRKTRDKDSGWFIYYWKPNIDQINEILLNRKRLILDK 101

.:::* *** * * : :. ::: :*.::.*: * * .:: :: : :..

PF0491 LKQMLQEETSETYYHCGTPGHPKLTFDEAFEYGFQCPICGEILYEYDNSKIIEELKKRIE 162

PH0619 LKKMLQEETSEVYYHCGNPDHPKLTFDEAFEYGFVCPICGEILHQYDNSAVIEELKKRIE 166

MJ0777 LEKKLEFEKNNMFFFCPN-CNVRFTFEEAMDYGFSCPGCGNMLQEFDNSELIKDLEEQIK 178

SSO0266 LKTRLEYEKNNTFFICPQ-DNSRYSFEEAFENEFKCLKCGSQLTYYDTDKIKSFLEQKIR 160

*: *: *..: :: * : : :*:**:: * * **. * :*.. : . *:::*.

PF0491 ELEIELGLRSPPKEEKPKKATRRKKSRSGKKKK 195

PH0619 ELEIELGLRAPPKKEKKGKKSK-KRSKKSKKK- 197

MJ0777 FLKEELKNNPFLK-------------------- 191

SSO0266 QIEEEIDKETKLGANKNH--------------- 178

:: *: ..

��

��

��

�� wing

The winged helix domain of Pyrococcus furiosus. Alignment of the Pfu TFE winged helix

domain amino acid sequence (PF0491) with corresponding regions in other archaeal TFE

sequences (Pyrococcus horikoshii, PH0619; Methanocaldococcus jannaschii, MJ0777;

Sulfolobus solfataricus, SSO0266). Predicted secondary structure elements for Pfu TFE are

indicated above the sequence (�-helices, cylinders; �-strands, arrows). Naming of the

secondary structures follows the naming suggested in reference 26 for Sulfolobus solfataricus.

Amino acids boxed in red are identical in all sequences in the alignment, green boxes present

amino acids with conserved substitutions. Semi-conserved substitutions were observed in

amino acids boxed in yellow. Sequences from the UCSC Archaeal Genome Browser were

aligned with ClustalW (32). The PSIPRED Protein Structure Prediction Server was used for

secondary structure predictions.

References

32. Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994) Nucleic Acids Res. 22, 4673-4680

Sebastian Grünberg, Michael S. Bartlett, Souad Naji and Michael ThommTranscription Factor E Is a Part of Transcription Elongation Complexes

doi: 10.1074/jbc.M707371200 originally published online October 5, 20072007, 282:35482-35490.J. Biol. Chem.

10.1074/jbc.M707371200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

http://www.jbc.org/content/suppl/2007/11/06/M707371200.DC1.html

http://www.jbc.org/content/282/49/35482.full.html#ref-list-1

This article cites 30 references, 18 of which can be accessed free at


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M707371200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;282/49/35482&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/282/49/35482

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=282/49/35482&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/282/49/35482

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/suppl/2007/11/06/M707371200.DC1.html

http://www.jbc.org/content/282/49/35482.full.html#ref-list-1

http://www.jbc.org/

transcription factor e is a part of transcription elongation complexes

Documents