gku146.pdf - Oxford Academic

Domain movements of the enhancer-dependentsigma factor drive DNA delivery into the RNApolymerase active site insights from singlemolecule studiesAmit Sharma1 Robert N Leach1 Christopher Gell1 Nan Zhang2 Patricia C Burrows2

Dale A Shepherd1 Sivaramesh Wigneshweraraj2 David Alastair Smith13

Xiaodong Zhang2 Martin Buck2 Peter G Stockley1 and Roman Tuma1

1Astbury Centre for Structural Molecular Biology University of Leeds Leeds LS2 9JT UK 2Department of LifeSciences Sir Alexander Fleming Building Imperial College London SW72AZ UK and 3School of Physics andAstronomy University of Leeds Leeds LS2 9JT UK

Received November 26 2013 Revised January 28 2014 Accepted January 30 2014

ABSTRACT

Recognition of bacterial promoters is regulated bytwo distinct classes of sequence-specific sigmafactors p70 or p54 that differ both in their primarysequence and in the requirement of the latter foractivation via enhancer-bound upstream activatorsThe p54 version controls gene expression inresponse to stress often mediating pathogenicityIts activator proteins are members of theAAA+ superfamily and use adenosine triphosphate(ATP) hydrolysis to remodel initially auto-inhibitedholoenzyme promoter complexes We havemapped this remodeling using single-moleculefluorescence spectroscopy Initial remodeling isnucleotide-independent and driven by binding bothssDNA during promoter melting and activatorHowever DNA loading into the RNA polymeraseactive site depends on co-operative ATP hydrolysisby the activator Although the coupled promoterrecognition and melting steps may be conservedbetween p70 and p54 the domain movements ofthe latter have evolved to require an activatorATPase

INTRODUCTION

Gene transcription is a tightly regulated cellular processThe basic features of the transcription machinery areconserved through all kingdoms of life with the multi-subunit RNA polymerase (RNAP) playing a central role(12) Bacteria use a sequence-specific DNA-bindingprotein the sigma (s) factor which together with fivecore subunits (a2bb0o) forms the RNAP holoenzymethat performs promoter recognition and DNA unwinding(34) during transcription initiation Two principal classesof s factors the s70 and s54 class are known The formermediates transcription of house-keeping genes (s70-RNAP) whereas the latter activates gene expression inresponse to defined environmental cues (s54-RNAP) (5)On binding to the core RNAP subunits s70 reorganizes

its auto-inhibitory domains Regions 11 and 31ndash42 withrespect to a central Region 12ndash24 leading to the recog-nition of the 10 promoter element (6) resulting in spon-taneous isomerization of the closed promoter complex(RPc) to the open promoter complex (RPo) (34) Incontrast the s54-dependent transcription system encoun-ters a kinetic barrier to RPo formation (7) Binding of theRNAP to the promoter at the 1224 conserved se-quences leads to the formation of a stable RPc thatrarely isomerizes spontaneously to the open promoter

To whom correspondence should be addressed Tel +44 1133 433092 Fax +44 1133 437897 Email pgstockleyleedsacukCorrespondence may also be addressed to Roman Tuma Tel +44 1133 433080 Fax +44 1133 437897 Email rtumaleedsacukPresent addressesRobert N Leach Covance Laboratories Ltd Otley Road Harrogate HG3 1PY UKDale A Shepherd Physical and Theoretical Chemistry Laboratory Department of Chemistry University of Oxford Oxford OX1 3QZ UKSivaramesh Wigneshweraraj Centre for Molecular Bacteriology and Infection Flowers Building of Imperial College London SW7 2AZ UKD Alastair Smith Avacta Group Plc Unit 706 Avenue E West Thorp Arch Estate Wetherby LS23 7GAUK

Published online 19 February 2014 Nucleic Acids Research 2014 Vol 42 No 8 5177ndash5190doi101093nargku146

The Author(s) 2014 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby30) whichpermits unrestricted reuse distribution and reproduction in any medium provided the original work is properly cited

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

complex (Figure 1A) Bacterial Enhancer Binding Proteins(bEBPs) belonging to the ATPases associated with variouscellular Activities (AAA+) family such as NtrC and PspF(Phage Shock Protein F) bind to upstream enhancer-binding sequences and couple the energy associated withATP hydrolysis to remodel the RPc leading to the forma-tion of the transcriptionally competent RPo (Figure 1A)(8) bEBPs have conserved domain architecture with acentral AAA+ domain followed by a HTH DNAbinding domain AAA+ ATPase domains bind tosigma54 via its GAFTGA loop motif (910)Biochemical and mutagenesis experiments have estab-

lished a number of functional regions of s54 (I II and IIIFigure 1B) and their roles in the various stages of tran-scription initiation (5) Region III (residues 108ndash477) isprimarily involved in binding to the promoter at severalsites with the strongest interaction being between theRpoN box (residues 454ndash463) and the consensus 24GG promoter element (1112) Region II is variable inamino acid composition across different species and com-pletely absent in Rhodobacter capsulatus suggesting it is

nonessential Region I (residues 1ndash56) plays a crucial regu-latory role in maintaining the RPc by forming a networkof interactions both with the 12 promoter element andwith Region III thus preventing spontaneous conversionfrom RPc to RPo (13ndash16)

Cryo-EM reconstructions of the s54-holoenzymerevealed the presence of a lsquobridging densityrsquo (Db)believed to be a part of Region I that occupies theDNA-binding cleft between the b and b0 pincers of thecore polymerase thus potentially preventing loading ofthe promoter template DNA into the RNAP active site(17) A large-scale domain reorganization of s54 occurswhen the holoenzyme interacts with PspF in thepresence of ADPAlFx suggesting that ATP hydrolysisis key to removal of the lsquoroadblockrsquo posed by theRegion I en route to RPo formation (1718) Similarlythe DNaseI footprint of RPc on the glnAp2 promoter issmaller (34 to+2) than that of RPo ( 34 to+23) (19)suggesting that significant activator-mediated structuralreorganization of s54 domains relative to the promoteris also required for productive RPo formation (1320)

Figure 1 Experimental design and description of components for the smFRET assays (A) Cartoon illustration of the pathway leading to opencomplex formation in s54-dependent transcription machinery (B) Schematic description of the domain architecture of s54 showing the differentdomains of the molecule and the positions of dye-labeling (left) and cryo-EM reconstruction of the Es54PspFADPAlFx complex depicting theapproximate locations of the regions of s54(right) (C) Schematic description of nifH promoter DNAs used for smFRET assays Positions ofdye-labels used in TIRFM measurements are shown

5178 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

through a series of well-orchestrated but largelyunmapped intermediate steps

To map these steps we have used complementary singlemolecule fluorescence (SMF) techniques alternating laserexcitation (ALEX) total internal reflection fluorescencemicroscopy (TIRFM) and fluorescence correlationspectroscopy (FCS) to probe the domain movements ins54 on holoenzyme formation in the closed promotercomplex on binding to a pre-melted nifH promotermimicking the open complex on PspF-mediated RPi (atthe transition state of ATP hydrolysis) and RPo formationand finally on RNA primed initiation and subsequenttranscript elongation

In contrast to s70-RNAP the s54 enzyme progressivelyrepositions its auto-inhibitory domain in response tobinding both single-stranded template DNA and activa-tor ATP hydrolysis by the latter is then required to fullyremove the roadblock to template entering the active site

MATERIALS AND METHODS

Proteins

Klebsiella pneumoniae s54 wild-type and variants (C20C463 C20C463 and R336AC20C463) and the wild-type and T86A versions of the catalytic domain ofPspF(1ndash275) were expressed and purified as described pre-viously (20) Labeling of the C20C463 and R336AC20C463 dual-Cys variants and the single-Cys variants C20 orC463 of s54 was carried out by addition of a 10-fold molarexcess of Alexa Fluor 488 andor 594-C5 maleimide dyeover protein in the presence of 25mM HEPES 74100mM NaCl and 1mM Tris(2-Carboxy Ethyl)Phosphine (TCEP) at room temperature (20ndash22C) for25 h with further incubation at 4C overnight Thesample was then quenched with 10mM b-mercaptoethanol (BME) and kept at 22C for 15minPurification of labeled protein was carried out on aSuperdex-75 gel filtration column equilibrated in 25mMHEPES 74 100mM NaCl and 10mM BME Elutedsamples were then analyzed on a 15 (wv) SDS-PAGEfor purity of the labeled species and the degree of labelingwas calculated for each dye using e=94000 cm1M1forAlexa Fluor 594 and e=71000 cm1M1 for Alexa Fluor488 dye Escherichia coli RNA polymerase apo-enzymewas purchased from Epicentre Inc

Promoter DNA template

Synthetic oligonucleotides corresponding to the 38 to+6 (44 bp) and 43 to + 17 (60 bp) regions of the nifHpromoter carrying base pair mismatches at 1211sites to mimic the closed promoter configuration (early-melted) and mismatched between 10 to 1 sites toresemble the open promoter (late-melted) were purchasedfrom IDT DNA(Coralville USA) (Figure 1C) (21) ForTIRF measurements the 38 to+6 oligo was used withAlexa Fluor 488 (Life Technologies Ltd Paisley UK) orbiotin labels at the 50-end ALEX and FCS measurementswere carried out on the 60-bp oligonucleotide DuplexnifH promoter DNA was formed by mixing and heating

the appropriate templates at 95C for 5min followed bycooling to 22C

Short-primed RNA assay

Functional analysis of labeled s54 was carried out asdescribed in (18)

ATPase assay

The ATPase activity of PspF1ndash275 was measured in thepresence of Es54 and late-melted nifH promoter (10 to1wt) using the NADH-coupled ATPase assay (22) Inall 100 nM holoenzyme (Es54) was mixed with 200 nMlate-melted nifH promoter and 10 mM PspF1ndash275 andvarying concentrations of ATP (01ndash5mM) at 30C inthe presence of 25mM Tris (pH 80) 100mM KCl10mM MgCl2 1mM DTT 10Uml pyruvate kinase10mM phosphoenolpyruvate 20Uml lactate dehydro-genase and 1mM NADH Data points were collectedfor 120min and the rate of steady-state ATP hydrolysiscalculated by observing the change in A340nm after nor-malization with PspF concentration

Fluorescence anisotropy

Single-Cys variants of s54(C463 or C20) labeled witheither AF488 or 594 dye were used for anisotropy meas-urements on a Horiba JobinndashYvon Fluorolog at concen-trations identical to that used for ALEX measurementsbefore dilution and at a temperature of 30CAex=590 nm Aem=617 nm for AF594 Dex=488 nmDem=517 nm for AF488) Anisotropy was calculatedusing the equation r= (IvvGIvh)(Ivv+2GIvh) Thepolarization correction factor G was determined usinghorizontal polarization excitation

Binding affinity determination

In all experiments 100 nM of apoenzyme (E) was mixedwith 10 nM of dual-labeled s54 (DL) in 1 STA buffer(25mM Tris acetate pH 80 8mM Mg Acetate 10mMKCl 1mM DTT and 35 (wv) poly-ethylene glycol6000) (17) and allowed to incubate for 5min at 30CThe sample was then excited at 480 nm and an emissionspectrum (510ndash650 nm) collected with the ExEm slits setto 5 and 10 nm respectively The solution was titratedwith unlabeled promoter DNA and the spectra wererecorded at 30C allowing an 5-min equilibrationperiod after each addition Fluorescence intensities werecorrected for dilution and normalized with respect to themaximum donor fluorescence Normalized intensities wereplotted as a function of promoter concentration and fittedusing a single site-model and the non-linear least-squaresmethod in GraphPad Prism (GraphPad Software wwwgraphpadcom)

FCS data collection and processing

FCS measurements were performed as described in (23) ontranscription complexes formed on the 60 bp length ofAlexa488 labeled early- or late-melted nifH promoter at30C Transcription complexes were formed on 1 nMearly- or late-melted nifH with concentrations of the

Nucleic Acids Research 2014 Vol 42 No 8 5179

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

other components kept identical to that used in theALEX measurements AF488-labeled C463 s54 wasused to measure the hydrodynamic radius of theholoenzyme

Sample preparation and ALEX measurements

For observation of dual-labeled s54 alone 1 mM proteinwas diluted into 1 STA buffer (17) supplemented with01mgml BSA to a final concentration of 100 pMHoloenzyme (Es54) was prepared by mixing 100 nMdual-labeled s54 with 200 nM core polymerase(Epicentre Inc) and incubating for 5min at 37C Theeffects of holoenzyme binding to early- or late-meltednifH promoter DNA (60 bp) were assessed by mixing100 nM of either oligo to the holoenzyme and incubationof the reaction at 37C for 5min The intermediatetrapped complex was prepared by the addition of04mM ADPAlFx to the reaction mixture carryingEs54 PspF1ndash275 and early- or late-melted nifH promoter(24) The effects of ATP hydrolysis on RPc were assessedby the addition of 3mM ATP to the transcription mixcarrying the holoenzyme and early-melted promoter atconcentrations mentioned above RPo formation wasinvestigated at various ATP concentrations in thepresence of holoenzyme and late-melted promoter afterincubation at 37C for 30min subsequently the samplewas diluted into buffer carrying heparin (01mgml) for1min All the above-mentioned reactions were diluted to afinal concentration of 100 pM labeled s54 in 1 STAbuffer carrying an excess of the other reaction componentsand 01mgml BSAAbortive initiation and elongation reactions were

carried out by addition of 05mM UpG dinucleotideprimer and GTP(025mM) to the activated RPo to formthe initial transcribing complex(RPitc4) or the full com-plement of NTPs (GTPCTPUTPATP) to form theelongation complex Samples were incubated at 37C for5ndash7min to allow formation of these complexes Thesewere diluted to a final concentration of 100 pM labeledspecies in 1 STA buffer carrying an excess of the otherreaction components and 01mgml BSA

Alternating laser excitation system and data processing

A custom-made inverted confocal microscope setupcoupled with two lasers a diode-pumped 488 nm laserand a He-Ne 594 nm laser were used to provide excitationat both donor and acceptor fluorophore wavelengths tostudy FRET Alternation of laser excitation was achievedby electro-optical modulators that are controlled bysoftware developed in the LabView graphicalprogramming environment (LabViewTM 71 ProfessionalDevelopment System for Windows National InstrumentsAustin TX USA) (25) After passing through the polar-izers a spatial filter comprising two 50-mm lenses and a15-mm pinhole was used to obtain Gaussian beam profilesA series of mirrors guide the beams to a dual banddichroic mirror that reflects it into the objective Samplecontained in the confocal volume is illuminated byalternating lasers and the fluorescence signal from thesample is guided through pinhole and filters onto two

avalanche photodiodes(APD) where a TTL-type signal isproduced for photon counting

The laser alternation period was set to 100ms (duty cycleof 40) with intensity for the 488-nm laser (Dex) at90 mW and the 594-nm laser intensity (Aex) set to60 mW (continuous wave mode) As each labeledmolecule passes through the confocal volume a singlefluorescence burst is generated which is used to generatefour photon streams fDem

Dex fAemDex f

DemAex f

AemAex where fYX rep-

resents the photon count for a single fluorescence burst onexcitation at wavelength X (acceptor excitation (Aex)or donor excitation (Dex)) resulting in emission at awavelength Y (acceptor emission(Aem) or donoremission (Dem)) (26) In our experiments the identifica-tion of bursts was performed by setting the threshold forburst selection to 3 kHz and the number of photonsconstituting a burst to 35ndash50 photons to removeany false positives due to background After the identifi-cation of bursts the uncorrected ratiometricobservables E and S are calculated using the followingequations (25)

ErawPr frac14 f Aem

Dex =ethfAemDex+f Dem

Dex THORN eth1THORN

Sraw frac14 ethf AemDex+f Dem

Dex THORN=ethfAemDex+f Dem

Dex+f AemAex THORN eth2THORN

A 2D ErawPr -S

raw histogram is then plotted and the donorleakage factor (Lk)= f Aem

Dex =fDemDex is computed using

Sraw gt 08 and the direct excitation factor(Dir)= f Aem

Dex =fAemAex is calculated using Sraw lt 033 The

values for ErawPr and Sraw are then subtracted for cross-talk

contributions from donor-leakage (Lk) and directexcitation of the acceptor (Dir) to provide the cross-talk-corrected EPr and S using the equations

EPR frac14 FFRET=ethFFRET+F DemDex THORN eth3THORN

S frac14 ethFFRET+F DemDex THORN=ethF

FRET+F DemDex+F Aem

Aex THORN eth4THORN

where FFRET frac14 F AemDex LkDir

By plotting the EPR versus 1S for two standard poly-proline chains (6 or 12 residues long) with AF488 andAF594 dyes conjugated to the ends one obtains theslope() and the intercept () from which the detection-correction factor (g) is derived (eq 5) (25)

frac14 eth 1THORN=eth+ 1THORN eth5THORN

The detectionndashcorrection factor g was found to varybetween 075ndash12 in this work (27) Finally the accurateFRET efficiency E is calculated using the followingequation (25)

E frac14 EPR=frac12 eth 1THORNEPR eth6THORN

All data collection and processing scripts were coded inLabView graphical programming environment(LabViewTM 71 Professional Development System forWindows National Instruments USA) and Origin 70and were provided to us by Dr Ted Laurence LawrenceBerkeley National Laboratories USA (25)

5180 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

Immobilization of transcription complexes and TIRFmeasurement

Glass coverslip surfaces (BDH Laboratory SuppliesPoole UK) were prepared following the method describedin (28) Coverslips were incubated with Vectabond reagent(Vector Laboratories Burlingame USA) according to themanufacturerrsquos instructions after which they were coatedwith 25 (wv) polyethyleneglycol succinimidyl ester(PEG-NHS Rapp Polymere Tubingen Germany) and025 (wv) biotinylated-PEG-NHS (Rapp Polymere) in01M sodium bicarbonate (pH 83) for 3 h to eliminatenon-specific adsorption of proteins The functionalizedsurface was then rinsed with 10mM Tris pH 74 andincubated with 02mgml Immunopure-streptavidin(Pierce Biotechnology Rockford USA) in the samebuffer for 1 h and then rinsed with 3 400 ml of bufferA (25mM Tris acetate pH 80 with 8mM magne-sium acetate 10mM KCl 1mM dithiothreitol and 35(wv) poly-ethylene glycol 6000) Excess buffer was blottedoff and Alexa 488-labeled biotinylated DNA (10 pM inbuffer A) was applied to the surface and allowed to bindfor 30min Excess unbound DNA was then removed byblotting and rinsing using 3 400 ml of buffer A

Transcription complexes were assembled as described in(20) before being diluted to a final concentration of100 pM in buffer A supplemented with an anti-photo-bleaching cocktail onto the DNA-derivatised surfaceUse of the nucleotide hydrolysis transition stateanalogue ADPAlFx creates activator bound speciesdescribed as lsquotrappedrsquo (16) allowing study of a putativeintermediate complex (IC) equivalent to those studied bycryoEM (1729)

Images were acquired using a custom-built invertedmicroscope with objective illumination (30) and processedusing an in-house developed software

TIRF data acquisition

Alexa488-DNA fluorescence was excited by 488-nm lightin a through-the-objective total internal reflection (TIR)scheme Our custom-built microscope described fully in(30) operated in the following manner laser light(3mW) was focused on the back focal plane of a 100145 numerical aperture oil immersion microscope object-ive (Alpha-Plan Fluar Zeiss) The emerging light wasincident on the coverslipsample interface above thecritical angle for TIR generating an evanescent fieldthat efficiently excites labeled molecules near (lt250 nm)the surface The combined fluorescence emission fromthe fluorescent dyes Alexa488 (direct excitation thedonor dye) and Alexa594 (due to FRET the acceptordye) was collected by the same objective The light wasthen separated by a short-pass dichroic mirror (555DCSXall filters Chroma Technology Corp USA) The donorand acceptor emission signals were then filtered (HQ52550M and HQ62030M respectively) and brought backinto close but not overlapping spatial registration usinga long-pass dichroic mirror (565DCLP) The two signalswere then focused onto opposite quadrants of the samecamera (iXon EMCCD Andor) The camera wasoperated at maximum gain at a temperature of 70C

with an exposure time of 100ms and frame-transferactivated Movies were recorded full-frame from thecamera and streamed for up to 60 s

TIRF data analysis

Movies were then exported and analyzed using algorithmswritten in Igor Pro (Wavemetrics USA) complexes ofAlexa488-late-melted promoter with Alexa594-s54-RNAP were identified by manually selecting redspots of fluorescence from the acceptor molecules Thetrajectories of fluorescence intensity versus time for thesespots (averaged over a 3 3 matrix of pixels) and for thecorresponding region in the green image (for which aknown offset exists) were then examined for single-stepphoto-bleaching Up to 50 out of all fluorescentfeatures in the images demonstrated this behavior Thetwo fluorescence intensity trajectories from each pair ofdonor and acceptor molecules were then extracted andsaved for further analysisThe FRET efficiency corresponding to a given mol-

ecule was obtained from each pair of recorded trajectoriesin the following manner first the background levels(Bd and Ba for the donor and acceptor respectively)were calculated by averaging over several seconds of thedata present at the end of the traces ie after both dyeshad photobleached and these values subtracted fromeach trajectory Next the fluorescence signal ofdonor dye Id for complex without photoactive acceptordye was obtained by averaging over the donor trajectoryafter the acceptor has bleached Finally the donorsignal in the presence of the acceptor Ida wasdetermined by averaging over a region in the donortrajectory after the laser was turned on but before theacceptor bleaches In our experiments 50 oftrajectories exhibiting the single-step donor photobleach-ing showed this useful bleaching sequence Note thatbecause of the long lifetime of the complexes studiedhere the disappearance of acceptor signal in trajectoriesrepresents acceptor bleaching rather than complex disas-sociation with very high probability Typically at leastseveral seconds of the trajectories could be averaged todetermine Id and Ida for each molecule essen-tially eliminating shot noise effects The FRET efficiencywas then calculated using the donor-bleaching method(30ndash32)

EFRET frac14 1 Ida=Ideth THORN eth7THORN

FRET efficiency values were calculated for gt100 mol-ecules obtained from at least three separate surface regionsfor assays with the same components Histograms (bin size0025) were then generated from the FRET efficiency (allanalyses carried out in Igor Pro) FRET histograms werethen fitted to a function comprising the sum of up to twoGaussian functions of the form

fethEFRETTHORN frac14 y0+Xn

ifrac141

areai

widthiffiffiffiffiffiffiffiffi=2p exp

2ethEmeaniTHORN2

width2i

eth8THORN

Nucleic Acids Research 2014 Vol 42 No 8 5181

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

where the number of species fit n was either 1 or 2 Themean width and area are the mean FRET efficiency thewidth and the area from the Gaussian fit to species i re-spectively y0 is the baseline All errors quoted are the fiterrors (plusmn 1 SD) Akaikersquos Information Criterion Test(Origin 86TM) was used to determine the statisticallyrelevant number of Gaussians for fitting

RESULTS

Experimental design and description of probes

To investigate domain-specific spatial reorganizationswithin s54 during the molecular events leading toactivator-driven RPo formation and subsequenttranscription initiation we introduced cysteine residuesat positions Q20 (in the activator-binding domainRegion I) and E463 (in the DNA-binding domainRegion III) (20) and labeled these sites with donor(Alexa Fluor 488 AF488) and acceptor (Alexa Fluor594 AF594) fluorophores (Figure 1B see lsquoMaterials andMethodsrsquo section and Supplementary Figure S1A)Unlabeled 60-bp DNA fragments of the test nifHpromoter (Figure 1C and Supplementary Table S1)mimicking the conformations in the closed (RPc twobases unpaired downstream of the promoter GC) oropen (RPo unpaired for 10 to 1) promoter complexeswere used as DNA templates (20) The DL s54 wasfunctionally similar to the wild-type Es54 holoenzyme(1833) confirming that the dyes do not interfere signifi-cantly with its activity (Supplementary Figure S1B)Anisotropy measurements established that rotationalfreedom of the dyes was not impeded and that restraininginteractions are unlikely (Supplementary Table S2)allowing conversion of FRET efficiencies into relativedye separationsThe binding affinities of the DL Es54 to the early-

and late-melted promoters (60-bp early- or late-meltednifH) was estimated by donor fluorophore quenchingunder conditions similar to that used in single moleculemeasurements (Supplementary Figure S1C) The Kd

values obtained for DL Es54 binding to the early-melted promoter (022plusmn004 nM) and to the late-melted promoter (025plusmn003 nM) are similar to thatobtained previously for the binding of the unlabeledEs54 to the homoduplex nifH promoter [085 nM (34)]In all the solution single molecule measurementsDL was present at 100 pM whereas all the othercomponents (including the nifH promoter) were in atleast 1000-fold molar excess Thus according to themeasured equilibrium constants gt99 of all DL Es54

was in the desired complex Additional confirmation ofcomplex formation was obtained by monitoring thehydrodynamic radius of species in solution using sin-gle-molecule fluorescence correlation spectroscopy(smFCS) (23) of singly labeled promoter or sigma54equivalents (Supplementary Table S3)In addition to looking at inter-domain reorganiza-

tion between Regions I and III of s54 we alsoinvestigated the spatial movement of these domains

relative to the leading or trailing edges of thenifH promoter fragments For these investigations weused single cysteine variants (Q20C or E463C) labeledwith Alexa Fluor 594 C5-maleimide dye (20) Early- andlate-melted promoters (44-bp long) were end-labeledeither at the + 6 (leading edge end) or the 38(trailing edge end) positions with Alexa Fluor 488(Figure 1C) (20)

Activator- and DNA-binding domains of sigma54 are inproximity and conformationally lsquolockedrsquo in the RPc

To investigate relative distance changes occurring betweenthe two fluorophores in DL we used ALEX (2635) in aconfocal microscope setup that allows accumulation ofsingle molecule FRET (smFRET) data from freelydiffusing labeled molecules in the solution(Supplementary Figure S2A and B) SupplementaryFigure S2C illustrates the reproducibility of the equilib-rium data while Supplementary Figure S2D points outvariations in FRET populations due to batch to batchdifference in efficiency of the ATP-driven activation ieas a result of the irreversible step

Small angle X-ray scattering (SAXS) measurementsalong with cryo-EM of s54 suggest that the molecule isanisometric in its native unbound form (1036) We firstsought to estimate the distance between Regions I and IIIof DL using smFRET (Figure 2) FRET efficiencies ex-hibited a bimodal distribution indicating co-existence oftwo main conformational populations Fitting to twoGaussian curves is consistent with a major sub-population(73) with a mean FRET efficiency (E) of 074 corres-ponding to a relative domain separation R of 51 A anda minor sub-population (27) with E=049 and R of61 A (Figure 2A and Supplementary Table S3) Nativeion-mobility mass spectrometry confirmed the presence oftwo conformations (Supplementary Figure S3A and B)These FRET signals are from native s54 because indenaturing conditions (6M urea) E decreased from075 to 008 (Supplementary Figure S3C)

Upon binding to core the DL FRET showed a singledistribution (E=071 and R 52 A) close to that of thedominant conformer in free s54 (Figure 2B) Becauseholoenzyme formation apparently has a relatively smalleffect we confirmed that complex formation hadoccurred using FCS The hydrodynamic radius (Rh) of27 nm for Alexa Fluor 488 labeled s54 (SupplementaryTable S3) is consistent with the Rg (radius of gyration)value of 34 nm derived from SAXS (36) Addition of thecore polymerase increased the FCS Rh to 53 nm confirm-ing complex formation by gt80 of the sigma factor

During initiation the holoenzyme (Es54) binds to the1224 sites on s54-dependent promoters leading to theformation of a transient nucleation site where DNAmelting occurs (37) To investigate domain movementduring this process we used the early-melted promotera nifH duplex having mismatches at 1211 mimickingthe promoter DNA conformation adopted during forma-tion of the RPc and which is separated from the RPo by ahigh-energy barrier in vitro (15) The RPc showed a singleFRET population (E=075 and R 500 A) similar to

5182 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

the dominant solution conformation although with a sig-nificantly reduced peak width implying a complex inwhich the domain positions are more rigidly defined(Figure 2C) consistent with earlier data suggesting thatRegions I and III interact with the 24 and 12 sitesforming a stable inhibitory complex (14)

Following addition of PspF to the RPc mimic there islittle change in the mean FRET efficiency implying thatthe inter-domain separation in DL is unchanged(Figure 2D) When 3mM ATP was added to thismixture there was no change in FRET (Figure 2E) inagreement with previous observations suggesting that the

tight binding of s54 at the 1211 fork-junction preventsRPo formation in vitro (15) We conclude that there is nosignificant change in the relative positions of Regions Iand III of s54 in the RPc mimic which appear lockedinto their positions by core This complex cannot be effi-ciently remodeled in the context of the RPc FCS con-firmed that complex formation between holoenzyme andtemplate occurred in these samples and was stable to thepresence of PspF with and without ATP (SupplementaryTable S3)

Large-scale remodeling of p54 occurs for RPo formationaided by activator binding

To gain insights into the s54 reorganization needed forssDNA delivery to the RNAP active site we analyzedholoenzyme binding events on the late-melted promotermimicking the melted DNA within the RPo Binding ofthe holoenzyme to this DNA fragment resulted in abimodal FRET distribution corresponding to two con-formations of the DL These species represent 43 and57 of the total population with E=029 and R 69 Aand E=073 and R 51 A respectively (Figure 3A) Thehigher FRET species appears similar to the dominant con-former seen for free sigma but the lower FRET state rep-resents a more extended conformation The Rh value forthe Es54 LM complex is similar to that for Es54 alone(Supplementary Table S3) but this may reflect the flexi-bility of a duplex with such an extended single-strandedopening rather than a lack of complex formation given thesub-nanomolar Kd values Clearly s54 is able to exist inmultiple conformations in the holoenzyme and binding tothe late-melted promoter stabilizes a form of s54 in whichthe relative separation of Regions I and III has increasedby 18 A In the absence of PspF-mediated remodelingthe holoenzyme-late melted promoter should be heparin-sensitive (38) and the low FRET state disappeared when01mgml heparin was added (Supplementary FigureS4A) demonstrating reversibility These results suggestthat partial remodeling of Regions I and III occurs onbinding the promoter DNA RPo mimic presumablybecause of the recognition of the single-stranded 10region but not to the point that the complex is heparin-resistantWe then analyzed the effect of adding apo-PspF to the

holoenzyme late-melted promoter complex Although thisdoes not change the FRET efficiencies of the two popula-tions the lower FRET population (E=027) becomesdominant (78 of the total) This together with theincreased Rh of the complex formed (SupplementaryTable S3) suggests that the apo-PspF binds and stabilizesthe more extended s54 conformation (Figure 3B) This is anovel finding because no stable interactions between apo-PspF and the holoenzyme have been reported previouslyTo confirm this effect we used a Loop 1 variant (T86A) ofPspF that hydrolyzes ATP but does not interact withRegion I and thus fails to provide mechano-chemicalcoupling (9) The population histogram of the holoenzymelate-melted promoter in the presence of the T86A PspFvariant closely resembles that in the absence of wild-typePspF (Supplementary Figure S4B) implying no

Figure 2 Activator and DNA-binding domains of s54 are in proximityand conformationally lsquolockedrsquo in the RPc Population histogramsprobing domain movement of s54 in the closed promoter complex(RPc) (A) Dual-labeled s54 alone in solution (B) s54-holoenzymecomplex (C) binding of the s54-holoenzyme to the early-meltedpromoter (D) addition of PspF to the RPc and (E) addition of ATPto the RPc

Nucleic Acids Research 2014 Vol 42 No 8 5183

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

Figure 3 Large-scale remodeling of s54 occurs in events leading to RPo formation Population histograms obtained from ALEX measurementsshowing change in inter-domain FRET efficiencies in s54 (A) On binding to the late-melted nifH promoter (B) after addition of PspF (C) Afteraddition of ADPAlFx and (D) After addition of ATP and heparin Green trace in the panels depicts the newly formed FRET species whereas thered trace shows the initial conformation adopted by s54

5184 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

productive contact in the absence of T86 Clearly thePspF Loop1 engages with s54 before any nucleotide-driven transactions Interestingly TIRF assays thatmeasure the relative separation of the dyes between eachs54 region and the adjacent promoter (trailing or leading)edge do not show bimodal populations or changes inresponse to these events implying either that they do notalter local DNA contacts before ATP hydrolysis (Figure4A and B first and second panels from the top) or thatone of the two domains in one conformation is too farfrom the respective DNA edge to contribute to FRET

Nucleotide hydrolysis is required for productivedownstream movement of Regions I and III and loads thetemplate strand into the RNAP active site

Because nucleotide hydrolysis by PspF is essential for pro-ductive engagement of the template DNA with the DNA-binding cleft of RNAP we assessed the spatial reorgan-ization of Regions I and III during this step ADPAlFx anucleotide analog believed to trap the activator at the

point of hydrolysis transition state (24) was added toform an intermediate complex RPi containing PspFRNAP s54 and LM DNA This led to a furtherincrease in relative domain separation (by 10 A)compared with the nucleotide-free PspF complex(Figure 3C) In contrast to earlier events in whichRegions I and III maintained their positions along thepromoter DNA nucleotide analog binding leads torelative movement of Region I toward the leading edgeby 9 A (a higher FRET state appearing at E=072R=51 A Figure 4A third panel) The ratio of high tolow FRET states in the TIRF matches that of the twopopulations seen in solution consistent with both assaysreporting similar molecular events Region III in thiscomplex remains similarly placed with respect to thetrailing edge dye (Figure 4B third panel) Howeverthese changes do not survive heparin challenge implyingthat the DNA has not fully moved into the active site andor RNAP closure around the DNA is incomplete This isconsistent both with the Rh value which barely changes

Figure 4 TIRFM FRET measurements showing the relative domain movement of s54 with respect to the leading and the trailing edge of the late-melted promoter DNA (A) The events recorded for the movement of Region I with respect to the leading edge of the late-melted promoter duringstages leading to RPo formation From top to bottom the histograms show the population distributions obtained on the formation of Es54LMcomplex (top) Es54LMPspF complex (second) Es54LMPspFADPAlFx (third) and Es54LMPspFATP with heparin (bottom) (B) The popu-lation histograms obtained from TIRF measurements for the movement of Region III of s54 with respect to the trailing edge of promoter DNA Tophistogram shows the population distribution obtained on binding of Es54LM complex (top) Es54LMPspF complex (second)Es54LMPspFADPAlFx complex (third) and Es54LMPspFATP complex with heparin (bottom) The number of events recorded for eachexperiment is indicated by the value (N)

Nucleic Acids Research 2014 Vol 42 No 8 5185

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

(Supplementary Table S3) and a more limited use of tran-scription start sites in RPi compared with RPo(18)In contrast to the RPi ATP hydrolysis results in a low

FRET species at an E=022 that is stable to heparinchallenge The E value suggests that post hydrolysis thedomains are slightly closer to each other than in the RPi(Figure 3D) Concomitantly Region I reorganizes withrespect to the leading edge the change in FRET efficiencysuggesting that it moves further downstream (Figure 4Abottom panel) Region III also relocates with respect tothe promoter DNA trailing edge also moving down-stream to be closer to Region I (E=042 Figure 4Bbottom panel) FCS data show that following ATP hy-drolysis the Rh of the complex decreases(Supplementary Table S3) consistent with bending ofthe DNA as the initiation site becomes lodged in theactive site of RNAP similar to observations in the caseof the s70 polymerase (3940)Activator-mediated nucleotide hydrolysis for RPo

formation is efficiently circumnavigated by the R336Avariant of s54 which results in transcript formationfrom the late-melted promoter in the absence of the activa-tor (1841) The smFRET efficiency with the dual-labeled

R336A s54 variant holoenzyme when bound to thelate-melted promoter (Supplementary Figure S6C) wasessentially identical (E=023) to that obtained for thewild-type s54 after activator binding and ATP hydrolysis(Figure 3D) This confirms the interpretation that R336Ais a bypass mutant for activation-dependent initiation andfurther establishes that the s54 domain separation isrequired for this process The domain disposition ins54R336A alone (Supplementary Figure S6A) or incomplex with the core (Supplementary Figure S6B) issimilar to that of the wild-type s54 (Figure 2A andB)and only changes following binding to late-meltedpromoter Therefore domain movement during activationmust be driven andor maintained by DNA recognition ofthe single-stranded promoter DNA element and becoupled to DNA melting The importance of promoterDNA conformation is further confirmed by the lack ofs54 domain movements in the presence of PspF bindingand after ATP hydrolysis in either s54 alone or in RNAPholoenzyme in the absence of promoter DNA(Supplementary Figure S7)

We also investigated the dependence of RPo formationon ATP concentration (Supplementary Figure S5) using

Figure 5 Relative domain positions of s54 remain largely unaltered on abortive transcription and elongation (A) Population histogram showing thechange in distance between Regions I and III of s54 on addition of UpG dinucleotide primer to the heparin-resistant RPo (B) On addition of GTPto form the abortive product UpGpGpG and (C) on addition of the full complement of rNTPs to the RPo

5186 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

the appearance of the heparin-resistant domain separationas a measure of remodeling Although the Kd for ATPbinding to PspF is 03mM (42) similar to theapparent KM 05mM (Michaelis constant) seen here(Supplementary Figure S5G) the ATP hydrolysis-drivenremodeling required higher ATP concentrations(Supplementary Figure S5AndashF) yielding an apparent KM

of 15mM This suggests that multiple PspF subunitsmust engage and hydrolyse ATP for productiveisomerization to occur a situation similar to the kineticcooperativity seen for some hexameric helicases (43)

The relative positions of Regions I and III remain largelyunaltered during early transcription and elongation

Following loading of the template strand and delivery ofthe+1 start site into the active site of RNA polymerasethe transcriptional machinery is set to form 2ndash12 nt longabortive transcripts before entering the elongation phase(424445) We used the smFRET assay to investigate therelative domain movement during these processes Weadded UpG dinucleotide the 1+1 primer to the

heparin-resistant RPo The s54 FRET signals hadsimilar maxima to the two states seen previously in theisomerised RPo (cf Figure 5A with Figure 3A andSupplementary Figure S5D) Next a short transcript(UpGpGpG) was formed by addition of GTP(Figure 5B) followed by elongation in the presence of allNTPs (Figure 5C) In both cases the peak FRETefficiencies remain largely unchanged However the Rh

value declines slightly with each addition and transcriptextension This could be due to either partial loss of s54 orto the transcribing template scrunching into the active siteThe former explanation is more consistent with therelative increase in the high FRET state reflecting thedominant conformation of the free sigma (Figure 2A)

DISCUSSION

Based on previous structural and biochemical work it wasanticipated that Regions I and III of s54 undergo signifi-cant repositioning within the holoenzyme uponATP-driven remodeling by PspF (172046) Recently a

Figure 6 Tracking conformation dynamics of s54 during the transcription cycle Model shows the domain disposition adopted by the Regions I andIII of s54 with respect to each other and also in the context of the nifH promoter DNA beginning from free s54 in solution (Step I) holoenzymeformation (Step II) closed promoter complex (RPc Step III) activator engagement and RNAP isomerization (Steps IV and V) intermediatecomplex (RPi) formation (Step VI) full template loading (Step VII) initial transcript formation (Steps VIII and IX) and elongation complex(Step X)

Nucleic Acids Research 2014 Vol 42 No 8 5187

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

multistep kinetic model has been proposed for this activa-tion based on single molecule observations (47) Here wecan complement these studies by assigning the kineticsteps to defined structural rearrangements of Regions Iand III with respect to each other and to the promoterIn accordance with the kinetic scheme (47) the irrevers-ible ATP hydrolysis-driven step is preceded by tworeversible steps No domain separation occurs during thefirst step (step II to III Figure 6) modeled here by Es54

binding to the early-melted promoter However thesecond step which is modeled by Es54-late-meltedpromoter complex formation leads to significant separ-ation between Regions I and III (steps III and IVFigure 6) but without DNA repositioning Because thisdomain movement only happens on the partially meltedpromoter it is most likely driven by ssDNA recognitionas in the case of s70 where sequence-specific ssDNAbinding is coupled to promoter unwinding (3) Thedomain separation is further stabilized by interactionwith the activator PspF even in the absence of ATP(Figure 6 IV and V) In the presence of the transitionstate analog ADPAlFx s54 domains become furtherseparated and Region I moves toward the leading edgeof the promoter (Figure 6 V and VI) However onlyATP hydrolysis makes the protein rearrangement irrevers-ible (Figure 6 VI to VII) the domain separationdecreasing to the one seen in the nucleotide-free state (Vin Figure 6) but with both regions closer to the leadingedge This most likely reflects loading of the templateDNA strand into the active site resulting in bending ofthe promoter and polymerase bb0 clamp closure (48)DNA footprints of the Es54 RPc and RPo show littledifference at the upstream trailing edge whereas in RPothe downstream footprint is extended the interaction withthe downstream fork junction is changed and the start siteDNA is well within RNAP Thus changes in trailing edgeFRET with the Region III label may reflect increasedupstream wrapping of DNA in RPo and changes in s54

Regions I to III separations between RPo and RPc aredominated by a movement of Region I relative to amore static Region IIIIt is interesting to note that Regions I and III are sig-

nificantly more separated in the transition state complex(ADPAlFx) than in the final state after ATP hydrolysisThe NtrC1 activator has been shown to convert from aheptamer to a hexamer which subsequently undergoes alarge conformational change in response to cooperativeATP analog binding yielding an asymmetric gappedhexamer (49) In particular its GAFTGA loop regionresponds to binding of different nucleotides The largestconformational change as with PspF occurs withADPAlFx (50) Our data show that the pre-hydrolysistransition state of ATP results in a lsquopower-strokersquo thatmoves the Region I further toward the downstreampromoter element most likely due to movement of theGAFTGA loop out of the plane of the hexamericATPase In the next step after release of ADP andorPi Region I moves closer to Region III presumably dueto withdrawal of the GAFTGA loop and the complex isrendered heparin-insensitive The hydrodynamic radius ofthe complex also decreases suggesting that at this point

the template strand is fully loaded into the DNA bindingcleft and as a consequence is bent At this step polymeraseclamp closure also occurs loading the template strandinto the active site and enabling transcription (4850)

The high KM for the formation of the heparin-resistantcomplex as judged from the relative separation of RegionsI and III is indicative of cooperativity in which severalsubunits within the PspF hexamer need to bind andorhydrolyse ATP before performing the mechanicalactivity This is consistent with the proposed cooperativeformation of gapped hexamer of the related AAA+acti-vator protein NtrC1 (49) Because ATP hydrolysis repos-itions both Regions I and III with respect to the promoterDNA and PspF has extensive contact with DNA it isplausible that multiple ATP bindinghydrolysis eventsare needed to sequentially remodel s54 domains and re-position DNA

Once the domains of s54 have reorganized to allow thetemplate strand to access the active site of RNAP thedisposition of Regions I and III remains largely un-changed both in the RPo and within the subsequenttranscribing complexes that retain s54

Finally we suggest that whereas for s70 proteins wherethe Region 11 is repositioned from the RNAP DNA-binding channel to allow RPo formation (or the RNAPclamp must open further) s54 Region I is activelyrelocated by the ATPase during hydrolysis In contrastfor s70 RPo formation signals from DNA may induceopening and closing of the RNAP clamp causingRegion 11 to move away from the RNAP mainchannel Hence for s54 RPo formation is coupled to itscognate enzymatic ATPase but for s70 to promoter DNAmotions

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online

ACKNOWLEDGEMENTS

The authors thank Dr Thomas Wilkop and Dr OpasTojira for their assistance with aspects of ALEX data col-lection and processing and Dr Alexander Borodavka forinitial assistance with the FCS measurements

FUNDING

BBSRC [BBH0112341 to PGS and RT] and [BBH0122491 to XZ MB SW] University of Leeds forsupporting the Single Molecule Facility and WellcomeTrust Joint Infrastructure Facility [062164] and [090932Z09Z] PCB was supported by a Wellcome Trustproject grant [WT093044MA to MB] and NZ wassupported by BBSRC project grants [BBJ0028281 andBBG0012781 to MB] Funding for open accesscharge BBSRC and RC UK

Conflict of interest statement None declared

5188 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

REFERENCES

1 EbrightRH (2000) RNA polymerase structural similaritiesbetween bacterial RNA polymerase and eukaryotic RNApolymerase II J Mol Biol 304 687ndash698

2 WernerF (2008) Structural evolution of multisubunit RNApolymerases Trends Microbiol 16 247ndash250

3 FeklistovA and DarstSA (2011) Structural basis for promoter-10 element recognition by the bacterial RNA polymerase sigmasubunit Cell 147 1257ndash1269

4 ZhangY FengY ChatterjeeS TuskeS HoMX ArnoldEand EbrightRH (2012) Structural basis of transcriptioninitiation Science 338 1076ndash1080

5 BuckM GallegosMT StudholmeDJ GuoY and GrallaJD(2000) The bacterial enhancer-dependent sigma(54) (sigma(N))transcription factor J Bacteriol 182 4129ndash4136

6 CallaciS HeydukE and HeydukT (1998) Conformationalchanges of Escherichia coli RNA polymerase sigma70 factorinduced by binding to the core enzyme J Biol Chem 27332995ndash33001

7 PophamDL SzetoD KeenerJ and KustuS (1989) Functionof a bacterial activator protein that binds to transcriptionalenhancers Science 243 629ndash635

8 ZhangX ChaneyM WigneshwerarajSR SchumacherJBordesP CannonW and BuckM (2002) MechanochemicalATPases and transcriptional activation Mol Microbiol 45895ndash903

9 BordesP WigneshwerarajSR ChaneyM DagoAEMorettE and BuckM (2004) Communication betweenEsigma(54) promoter DNA and the conserved threonine residuein the GAFTGA motif of the PspF sigma-dependent activatorduring transcription activation Mol Microbiol 54 489ndash506

10 RappasM SchumacherJ BeuronF NiwaH BordesPWigneshwerarajS KeetchCA RobinsonCV BuckM andZhangX (2005) Structural insights into the activity of enhancer-binding proteins Science 307 1972ndash1975

11 BurrowsPC SeverinovK IshihamaA BuckM andWigneshwerarajSR (2003) Mapping sigma 54-RNA polymeraseinteractions at the -24 consensus promoter element J BiolChem 278 29728ndash29743

12 DoucleffM MalakLT PeltonJG and WemmerDE (2005)The C-terminal RpoN domain of sigma54 forms an unpredictedhelix-turn-helix motif similar to domains of sigma70 J BiolChem 280 41530ndash41536

13 BurrowsPC SeverinovK BuckM and WigneshwerarajSR(2004) Reorganisation of an RNA polymerase-promoter DNAcomplex for DNA melting EMBO J 23 4253ndash4263

14 GallegosMT and BuckM (1999) Sequences in sigmaNdetermining holoenzyme formation and properties J Mol Biol288 539ndash553

15 GuoY LewCM and GrallaJD (2000) Promoter opening bysigma(54) and sigma(70) RNA polymerases sigma factor-directedalterations in the mechanism and tightness of control Genes Dev14 2242ndash2255

16 WigneshwerarajSR ChaneyMK IshihamaA and BuckM(2001) Regulatory sequences in sigma 54 localise near the start ofDNA melting J Mol Biol 306 681ndash701

17 BoseD PapeT BurrowsPC RappasM WigneshwerarajSRBuckM and ZhangX (2008) Organization of an activator-boundRNA polymerase holoenzyme Mol Cell 32 337ndash346

18 BurrowsPC JolyN and BuckM (2010) A prehydrolysis stateof an AAA+ATPase supports transcription activation of anenhancer-dependent RNA polymerase Proc Natl Acad SciUSA 107 9376ndash9381

19 WangJT and GrallaJD (1996) The transcription initiationpathway of sigma 54 mutants that bypass the enhancer proteinrequirement Implications for the mechanism of activation JBiol Chem 271 32707ndash32713

20 LeachRN GellC WigneshwerarajS BuckM SmithA andStockleyPG (2006) Mapping ATP-dependent activation at asigma54 promoter J Biol Chem 281 33717ndash33726

21 CannonW WigneshwerarajSR and BuckM (2002)Interactions of regulated and deregulated forms of the sigma54

holoenzyme with heteroduplex promoter DNA Nucleic AcidsRes 30 886ndash893

22 NorbyJG (1988) Coupled assay of Na+K+-ATPase activityMethods Enzymol 156 116ndash119

23 BorodavkaA TumaR and StockleyPG (2012) Evidence thatviral RNAs have evolved for efficient two-stage packaging ProcNatl Acad Sci USA 109 15769ndash15774

24 ChaneyM GrandeR WigneshwerarajSR CannonWCasazP GallegosMT SchumacherJ JonesS ElderkinSDagoAE et al (2001) Binding of transcriptional activators tosigma 54 in the presence of the transition state analog ADP-aluminum fluoride insights into activator mechanochemicalaction Genes Dev 15 2282ndash2294

25 LeeNK KapanidisAN WangY MichaletXMukhopadhyayJ EbrightRH and WeissS (2005) AccurateFRET measurements within single diffusing biomolecules usingalternating-laser excitation Biophys J 88 2939ndash2953

26 KapanidisAN LaurenceTA LeeNK MargeatE KongXand WeissS (2005) Alternating-laser excitation of singlemolecules Acc Chem Res 38 523ndash533

27 SchulerB LipmanEA SteinbachPJ KumkeM andEatonWA (2005) Polyproline and the lsquolsquospectroscopic rulerrsquorsquorevisited with single-molecule fluorescence Proc Natl Acad SciUSA 102 2754ndash2759

28 HaT RasnikI ChengW BabcockHP GaussGHLohmanTM and ChuS (2002) Initiation and re-initiation ofDNA unwinding by the Escherichia coli Rep helicase Nature419 638ndash641

29 BurrowsPC WigneshwerarajSR and BuckM (2008) Protein-DNA interactions that govern AAA+ activator-dependentbacterial transcription initiation J Mol Biol 375 43ndash58

30 GellC BrockwellD and SmithA (2006) Handbook of SingleMolecule Fluorescence Spectroscopy Oxford University PressOxford

31 SabanayagamCR EidJS and MellerA (2005) Usingfluorescence resonance energy transfer to measure distances alongindividual DNA molecules corrections due to nonideal transferJ Chem Phys 122 061103

32 CleggRM (1992) Fluorescence resonance energy transfer andnucleic acids Methods Enzymol 211 353ndash388

33 CannonWV GallegosMT and BuckM (2000) Isomerizationof a binary sigma-promoter DNA complex by transcriptionactivators Nat Struct Biol 7 594ndash601

34 VogelSK SchulzA and RippeK (2002) Binding affinity ofEscherichia coli RNA polymerasesigma54 holoenzyme for theglnAp2 nifH and nifL promoters Nucleic Acids Res 304094ndash4101

35 KapanidisAN LeeNK LaurenceTA DooseS MargeatEand WeissS (2004) Fluorescence-aided molecule sorting analysisof structure and interactions by alternating-laser excitation ofsingle molecules Proc Natl Acad Sci USA 101 8936ndash8941

36 SvergunDI MalfoisM KochMH WigneshwerarajSR andBuckM (2000) Low resolution structure of the sigma54transcription factor revealed by X-ray solution scattering J BiolChem 275 4210ndash4214

37 MorrisL CannonW Claverie-MartinF AustinS and BuckM(1994) DNA distortion and nucleation of local DNA unwindingwithin sigma-54 (sigma N) holoenzyme closed promotercomplexes J Biol Chem 269 11563ndash11571

38 TintutY WangJT and GrallaJD (1995) A novel bacterialtranscription cycle involving sigma 54 Genes Dev 9 2305ndash2313

39 CraigML SuhWC and RecordMT Jr (1995) HO andDNase I probing of E sigma 70 RNA polymerasendashlambda PRpromoter open complexes Mg2+ binding and its structuralconsequences at the transcription start site Biochemistry 3415624ndash15632

40 DarstSA KubalekEW and KornbergRD (1989) Three-dimensional structure of Escherichia coli RNA polymeraseholoenzyme determined by electron crystallography Nature 340730ndash732

41 ChaneyM and BuckM (1999) The sigma 54 DNA-bindingdomain includes a determinant of enhancer responsiveness MolMicrobiol 33 1200ndash1209

Nucleic Acids Research 2014 Vol 42 No 8 5189

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022

42 JolyN and BuckM (2010) Engineered interfaces of an AAA+ATPase reveal a new nucleotide-dependent coordinationmechanism J Biol Chem 285 15178ndash15186

43 LisalJ and TumaR (2005) Cooperative mechanism of RNApackaging motor J Biol Chem 280 23157ndash23164

44 CarpousisAJ and GrallaJD (1980) Cycling of ribonucleic acidpolymerase to produce oligonucleotides during initiation in vitroat the lac UV5 promoter Biochemistry 19 3245ndash3253

45 GoldmanSR EbrightRH and NickelsBE (2009) Directdetection of abortive RNA transcripts in vivo Science 324927ndash928

46 CannonW GallegosMT CasazP and BuckM (1999) Amino-terminal sequences of sigmaN (sigma54) inhibit RNA polymeraseisomerization Genes Dev 13 357ndash370

47 FriedmanLJ and GellesJ (2012) Mechanism of transcriptioninitiation at an activator-dependent promoter defined by single-molecule observation Cell 148 679ndash689

48 ChakrabortyA WangD EbrightYW KorlannYKortkhonjiaE KimT ChowdhuryS WigneshwerarajSIrschikH JansenR et al (2012) Opening andclosing of the bacterial RNA polymerase clamp Science 337591ndash595

49 SysoevaTA ChowdhuryS GuoL and NixonBT (2013)Nucleotide-induced asymmetry within ATPase activator ringdrives sigma54-RNAP interaction and ATP hydrolysis GenesDev 27 2500ndash2511

50 ChenB DoucleffM WemmerDE De CarloS HuangHHNogalesE HooverTR KondrashkinaE GuoL andNixonBT (2007) ATP ground- and transition states of bacterialenhancer binding AAA+ATPases support complex formationwith their target protein sigma54 Structure 15 429ndash440

5190 Nucleic Acids Research 2014 Vol 42 No 8

Dow

nloaded from httpsacadem

icoupcomnararticle42851771073214 by guest on 05 February 2022