GalR represses galP1 by inhibiting the rate-determining open complex formation through RNA polymerase contact: a GalR negative control mutant

doi:10.1016/j.jmb.2004.09.070 J. Mol. Biol. (2004) 344, 609–618

GalRRepressesgalP1by Inhibiting theRate-determiningOpen Complex Formation Through RNA PolymeraseContact: A GalR Negative Control Mutant

Siddhartha Roy1,2, Szabolcs Semsey1, Mofang Liu1, Gary N. Gussin3 andSankar Adhya1*

1Laboratory of MolecularBiology, National CancerInstitute, National Institutes ofHealth, Bethesda, MD 20892USA

2Department of BiophysicsBose Institute, Calcutta 700054India

3Department of BiologicalSciences, University of IowaIowa City, IA 52246, USA

0022-2836/$ - see front matter Published

Abbreviation used: 2-AP, 2-aminoE-mail address of the correspond

[email protected]

GalR represses the galP1 promoter by a DNA looping-independentmechanism. Equilibrium binding of GalR and RNA polymerase to DNA,and real-time kinetics of base-pair distortion (isomerization) showed thatthe equilibrium dissociation constant of RNA polymerase-P1 closedcomplexes is largely unaffected in the presence of saturating GalR,indicating that mutual antagonism (steric hindrance) of the regulator andthe RNA polymerase does not occur at this promoter. In fluorescencekinetics with 2-AP labeled P1 DNA, GalR inhibited the slower of thetwo-step base-pair distortion process. We isolated a negative control GalRmutant, S29R, which while bound to the operator DNA was incapable ofrepression of P1. Based on these results and previous demonstration thatrepression requires the C-terminal domain of the a subunit (a-CTD) ofRNA polymerase, we propose that GalR establishes contact with a-CTD atthe last resolved isomerization intermediate, forming a kinetic trap.

Published by Elsevier Ltd.

Keywords: repression; fluorescence; transcription; contact; mutation

Introduction

In the traditional model of transcription repres-sion in prokaryotic systems, the promoter sequenceoverlaps with the binding site of the repressorprotein.1 Repressor binding to DNA inhibitsinitiation by sterically hindering RNA polymerasebinding to the promoter. This model is supportedby in vitro experiments in a few promoters inEscherichia coli.2,3 However, discovery of dual-function (repression and activation) generegulatory proteins and promoters with complexregulation suggests that alternate mechanisms forrepression must exist.4 Indeed there are repressorproteins that act by trapping RNA polymerase at apost-RNA polymerase binding step resulting inrepression. Repression of the P1 promoter of theE. coli gal operon by GalR does not exclude RNApolymerase binding and is dependent upon thepresence of the C-terminal domain of the a-subunitof RNA polymerase.5,6 p4 protein of Bacillus subtilisphage F29 represses the early A2c promoter at thepromoter clearance step in which the interaction

by Elsevier Ltd.

purine.ing author:

with the C-terminal domain of the a-subunit ofRNA polymerase with p4 plays an important role.7

Repression of transcription at a post-RNA poly-merase binding step also includes the action ofPhoP repressor at the E. coli mgtA P1 promoter,8 IclRrepressor at the E. coli AceB promoter,9 phage 434repressor at the phage PR promoter in E. coli,10 andphage P22 Arc repressor at the phage immIpromoter in Salmonella typhimurium.11 The detailedbiochemical mechanisms by which repressionoccurs in these cases remain to be elucidated.Here we address the issue in the GalR-mediatedrepression of the P1 promoter by investigating akinetic trap.

The E. coli gal operon is normally transcribed fromtwo promoters, P1 and P2, which are repressed byGalR-mediated formation of a DNA loop encom-passing the promoters.12 Loop formation by GalRrequires the histone-like protein HU and super-coiled DNA.13 In the absence of HU or DNAsupercoiling the P1 promoter is repressed by alooping-independent mechanism by binding ofGalR to an operator (OE) located upstream of P1.12,14

Formation of an RNA polymerase-GalR-OE ternarycomplex at P1 was demonstrated by EMSA,6

suggesting that GalR does not repress P1 byexcluding RNA polymerase binding. Moreover,

610 Contact Inhibition of Transcription Initiation

DNase protection results showed that RNA poly-merase forms a heparin-resistant intermediatebetween the closed and open forms at P1 whenGalR is bound to OE.5 To further elucidate themechanism of repression at P1, it is imperative thatthe steps of transcription initiation at the promoterbe resolved and the rates of interconversion of theintermediates be measured by direct assays.Previously we demonstrated that the adenine baseanalog 2-aminopurine (2-AP) when placed atspecific positions of the 14 bp segment presumedto be distorted during isomerization at P1 can beused to resolve the isomerization into several stepsand to assay their rates in real-time.15–19 In thisresearch, we used fluorescence polarization todetermine the dissociation constant of the RNApolymerase-P1 closed complex. This methodtogether with the 2-AP fluorescence real-timeassays allowed us to study the effect of GalR ondifferent steps of transcription initiation. We reportthat GalR kinetically traps a rate-limiting inter-mediate of RNA polymerase complex at P1, thusincreasing the energetic barrier to open complexformation.

Results

Previous studies with GalR suggested that therepressor inhibits P1 by a mechanism other than byinterfering with RNA polymerase binding to thepromoter.5 One way to test the model of exclusionof RNA polymerase binding to promoters byrepressors is to determine the dissociation constantof RNA polymerase-promoter closed-complex inthe absence and in the presence of the repressor.However, this relatively simple task is complicatedby possible binding of RNA polymerase to nearbypromoters, promoter-like tight binding sites, or the

Figure 1. Oligonucleotide sequences used in this study. Fopositions K19 through K16 with respect to C1 (underlinesequences correspond to the extended K10 region of P1. ForOR1, OR2 and OR3, respectively. The underlined nucleotides aone deleted in the D34 template. The oligonucleotide containmutation K7C50 as well (colored magenta), to further reduce

ends of DNA fragments. Thus, we chose conditionsthat eliminated or minimized the alternate bindingmodes. It has been shown that end-binding by RNApolymerase is minimized in buffers at lower pH.20

Since pH 6.4 buffers reduced end-binding withoutaltering promoter binding significantly (data notshown), we carried out all our equilibrium bindingexperiments at pH 6.4.

Figure 1 shows the sequences of the non-templatestrand for each of the gal DNA templates that wereused in this study. The P1 and P2 promoters belongto the extended K10 class in which instead of theK35 consensus sequence a TG motif at K14/K15plays a critical role in transcription initiation.21,22 P1and P2 are five base-pairs apart and thus are out-of-phase with each other on the DNA cylinder.Mutations were identified that debilitate P2 withouthaving a major effect on P1.22 However, it was notclear whether such mutations knocked out RNApolymerase binding to P2 or affected a later step inopen complex formation. In order to cause an RNApolymerase binding defect at P2 we chose tointroduce mutations at several positions, K14,K13, K12 and K11, of P2 (K19, K18 K17 andK16 with respect to P1) in the template denotedP1C. In addition to the P2 mutations, a similar set ofmutations were introduced into the same templateat K14, K11, K10, K9, K8 and K7 of P1 to createthe template that we called P1K. Figure 2 shows thebinding of RNA polymerase to the P1C and P1K

templates as determined by fluorescence polariz-ation. Whereas binding to the P1C templateoccurred at low RNA polymerase concentrationsand saturated quickly, binding to P1K was weak.When fitted to a single–site binding equation (aftercorrecting for the active fractions of RNA poly-merase), the derived dissociation constants were1.01!10K8 M and 4.46!10K10 M for P1K and P1C.

r the gal oligonucleotides, the red colored sequences ared) of P1 (K14 through K11 of P2). The green coloredlPRM green, blue, and red colored nucleotides representre the K35 and K10 elements of PRM. The yellow t is thes only a part of the promoter PR and contains a PR-downthe possibility of binding of polymerase at PR.

Figure 2. RNA polymerase binding at 12 8C to P1CP2K

(C) and P1KP2K (B) templates. The double-strandedtemplates were end-labeled with fluorescein using a C-6linker (see Materials and Methods). The buffer was50 mM Mes (pH 6.4), containing 0.2 M NaCl, 10 mMMgCl2, 100 mg/ml BSA, 1 mM DTT, 1 mM EDTA and 10%glycerol.

Figure 3. RNA polymerase binding at 4 8C to theP1CP2K template in the presence of 200 nM GalR (B)and in its absence (6). The templates were end-labeledwith fluorescein using a C-6 linker. The buffer was 50 mMMes (pH 6.4), containing 0.2 M NaCl, 10 mM MgCl2,100 mg/ml BSA, 1 mM DTT, 1 mM EDTA and 10%glycerol. The DNA concentration was 0.8 nM.

Contact Inhibition of Transcription Initiation 611

Since these experiments were done at 12 8C and0.2 M NaCl, it is likely that these binding constantsreflect formation of closed complexes.23

Table 1. RNA polymerase binding to promoters influorescence polarization assays

Promoter RegulatorDissociation con-

stant at 4 8Ca

P1CP2K 1.14!10K9

P1CP2K GalR (200 nM) 1.17!10K9

PRM 1.9!10K9

PRM CI (56 nM) 7.5!10K10

PRMD34 1.9!10K9

PRMD34 CI (56 nM) 7.5!10K9

PRE 5.2!10K10

PRE CII (45 nM) 1.45!10K10

a Dissociation constants were obtained from fluorescencepolarization assays.

RNA polymerase binding and effect of regulators

RNA polymerase was shown to bind to the galpromoters at 4 8C without significant DNA distor-tion.23,24 A complex at this temperature representsthe closed-complex.25–27 If repression at P1 occursby exclusion of RNA polymerase binding to thepromoter, then the presence of saturating concen-trations of GalR should significantly weaken RNApolymerase binding. Figure 3 shows the isothermsof RNA polymerase binding to P1C in the absenceand in the presence of saturating concentrations ofGalR. The equilibrium dissociation constants inthe absence and in the presence of GalR were(1.14G0.21) !10K9 M and (1.17G0.27)!10K9 M(average of three readings), respectively, indicatingthat the binding of GalR and RNA polymerase tothe P1C promoter are not mutually exclusive.However, it would be instructive to measure thebinding of RNA polymerase in the absence and inthe presence of regulatory proteins to promotersthat either are repressed by antagonistic binding oractivated by agonistic binding of a regulatoryprotein and compare them to the fluorescenceanisotropy of P1. We investigated three cases.

(i) Phage l regulatory protein, CI, bound to thephage OR2 operator is an activator of the cognate

PRM promoter.28,29 In this role, CI should not havean antagonistic effect on RNA polymerase bindingto PRM. In a fluorescence polarization assay, PRM

bound RNA polymerase with high affinity and wasnot weakened by the presence of CI. The equili-brium dissociation constants for RNA polymerasebound to free or CI saturated PRM DNA were1.9!10K9 M and 7.5!10K10 M, respectively(Table 1). Although it is believed that PRM activationby CI occurs at the isomerization step,30,31 theresults indicate a modest enhancement of RNApolymerase binding by CI.

(ii) It was shown that the deletion of the K34Tbase-pair (PRMD34) causes CI to become a repressorof PRM.32 It is believed, although not proven, thatincreased overlap of RNA polymerase and CI

612 Contact Inhibition of Transcription Initiation

binding sites, PRM and OR2, respectively, leads toantagonistic binding.32 We determined the equili-brium constant for binding of RNA polymerase tothe PRMD34 promoter in the absence and in thepresence of CI. The dissociation constant of theRNA polymerase-PRMD34 complex in the absenceof CI was 1.9!10K9 M, which was exactly the sameas in the wild-type. However, in the presence of CI,the affinity decreased to 7.5!10K9 M, indicatingsignificant weakening of RNA polymerase binding(Table 1). This weakening is likely the result of anantagonism of binding of CI and RNA polymeraseto the PRMD34 template. The residual binding ofRNA polymerase in the presence of CI may be dueto binding to DNA ends.

(iii) The CII protein from phage l is a transcrip-tion activator at the phage l promoter PRE.33,34 ThePRE promoter is intrinsically very weak and isactivated many fold in the presence of CII at 37 8C.A substantial part of this activation is attributed toincreased binding of RNA polymerase in thepresence of CII.33 We thus studied binding ofRNA polymerase to the wild-type PRE template. Ina fluorescence polarization assay the binding ofRNA polymerase in the presence of CII wasdistinctly tighter. The dissociation constants ofthe PRE–RNA polymerase complex in thepresence and absence of CII were 1.45!10K10 Mand 5.2!10K10 M, respectively, showing that CIIenhances the binding of RNA polymerase to PRE

(Table 1). However, the threefold enhancement ofRNA polymerase binding by CII measured in thisexperiment at 4 8C was significantly less than it wasreported earlier at 37 8C.

The results with CI and CII above showed that itis possible to detect antagonistic and synergisticbinding of regulatory proteins and RNA polymer-ase to a promoter using the fluorescence anisotropyassay. Hence, the result of non-antagonistic bindingof GalR and RNA polymerase to the P1 promotersuggests a repression mechanism, which is differentfrom that of inhibition of RNA polymerase bindingto the promoter by GalR.

Figure 4. (A) Fluorescence intensity changes on theP1CP2K template labeled at C3 with 2-AP upon mixingwith RNA polymerase in the presence (B) and in theabsence (6) of 3 mM GalR at 25 8C. (B) Fluorescenceintensity changes of P1CP2K template labeled at K7 with2-AP upon mixing with RNA polymerase and 3 mM GalRin the presence (B) and in the absence (C) of 6 mMD-galactose at 37 8C. The buffer was 0.05 M Tris–HCl (pH7.5), containing 0.1 M NaCl and 10 mM MgCl2. The finalconcentration of the template and the RNA polymerasewas approximately 0.83 mM.

Mechanism of galP1 repression

An alternative mechanism of P1 repressionmediated by a contact of GalR with RNA poly-merase at an intermediate state was suggested byDNaseI protection assays.5 We investigated apotential kinetic trap by examining the rates ofisomerization upon RNA polymerase binding in theabsence and in the presence of GalR by the use of2-AP.15–19 We previously used 2-AP to resolve theisomerization pathway of the P1 promoter intomore than one step.19 Two major steps wereobserved in the base-pair distortion phase at P1.The first involved the distortion of the base-pairs atK12, K10, K4, K2,C1 and C3 and the second rate-limiting step involved distortions at positions K7,K5 and C2. This step was enhanced by thecAMP receptor protein (CRP), which is an activatorof P1.

We investigated base-pair distortion at represen-tative positions from each of the two groups, C3and K7, by 2-AP fluorescence to explore the effectof GalR on the kinetics of base-pair distortion.Figure 4(A) shows the fluorescence increase at 25 8Cupon RNA polymerase binding to the P1C templatelabeled with 2-AP at C3. The fluorescence increasefollowed similar initial kinetics both in the absenceand in the presence of GalR with rate constants1.59!10K3 sK1 and 2.14!10K3 sK1, respectively.However, the amplitude of the fluorescenceappeared to be greater in the presence of GalRthan in its absence.

Figure 4(B) shows the effect of GalR on the

Contact Inhibition of Transcription Initiation 613

fluorescence kinetics when the template waslabeled with 2-AP at position K7. Since this rate-limiting step is very slow at 25 8C on this template,accurate measurement of the isomerization rate isdifficult. Hence, these experiments were carriedout at 37 8C. The rate difference in the presence(4.4!10K3/s) and absence (1.7!10K3/s) of theinducer D-galactose at 6 mM was about 2.5-fold,which is comparable to the degree of repressionobserved on linear templates in in vitro transcrip-tion assays.35 A similar picture was obtained also at25 8C (data not shown). The rate constants werecorrected as described19 after obtaining the kapp

from the 37 8C data in Figure 4(B).

Negative control mutants

It appears from the data presented above, thatGalR specifically inhibits the last and the sloweststep of isomerization at P1. As mentioned before,the repression of P1 by GalR is a-CTD-dependent.Deletion of the a-CTD abolished repression byGalR.5 Amino acid residues in a-CTD defective inrepression were also identified.6 Thus, P1 repres-sion results from the establishment of a GalR-a-CTD contact at the last step of isomerizationresulting in a kinetic trap. A strong prediction ofsuch a model of repression is that it would bepossible to isolate a negative control (nc) GalRmutant that could bind well to the cognate operatorOE, but would be incapable of repressing P1. We

searched for such a mutant by mutagenizing theN-terminal part of the galR gene by error-pronePCR amplification and screening for an ncphenotype as described in Materials and Methods.Of four mutants isolated, the most promising onecontained a substitution of arginine for serine atposition 29.

Figure 5(A) (left) shows the repression of tran-scription of P1 by wild-type and S29R GalR in vitroin the absence of DNA looping.36 Activity of the P1and P2 promoters as a function of repressorconcentration is shown in Figure 5(B). It is clearthat 50 nM wild-type GalR repressed P1 efficiently,while S29R was incapable of causing any repressioneven at 100 nM. S29R GalR appeared to be effectivein repression when DNA looping was allowed inthe presence of HU, suggesting that the mutantGalR was functional in DNA binding. However,while 50 nM of the wild-type GalR protein wassufficient for complete repression, 200 nM wasrequired from the S29R mutant (Figure 5(A), right,lane 11 versus lane 14).

In order to confirm that the S29R mutant iscapable of binding to OE, DNase I protection assaysusing primer extension were carried out underexactly the same solution conditions that were usedfor in vitro transcription. Figure 6 shows the resultsof protection assays with wild-type and S29R GalR.Since looping repression by the S29R GalR mutantwas slightly less at 50 nM, 100 nM and 200 nM ofthe mutant repressor was used in the protection

Figure 5. Effect of wild-type andS29R GalR on transcription fromthe P1 and P2 promoters in vitro.(A) Results of in vitro transcriptionassays performed on supercoiledpSA886 ½OC

EOKI � (left panel) and

pSA850 ½OCEO

CI � (right panel) DNA

in the presence of wild-type (lanes2–4 and 9–11) or S29R mutant (lanes5–7 and 12–14) GalR. GalR wasused at 50 nM (lanes 2, 5, 9 and12), 100 nM (lanes 3, 6, 10 and 13)and 200 nM (lanes 4, 7, 11 and 14)concentration. Reactions in theright panel contained 80 nM HUprotein to allow looping.13 TheRNA1 transcript was used as aninternal control between lanes. (B)Activity of the P1 and P2 promotersin non-looping conditions as afunction of repressor concentration.RNA bands were quantified usingthe ImageQuante Phosphorimager(Molecular Dynamics).

Figure 6. DNase I protection of supercoiled gal DNA. Left panel: Complexes of wild-type and S29R GalR andsupercoiled pSA886 DNA were formed under conditions of the in vitro transcription reaction. After probing theDNA–protein complex with DNase I, the nicked DNA was purified and used as a template for primer extension with a32P-labeled oligonucleotide primer. Results shown are from the top strand of the DNA. An A-specific sequencingreaction was used as a standard to identify bands (lane 1). Wild-type GalR protein (lane 3) was used at 50 nM. S29R GalRwas used at 100 nM (lane 4) and 200 nM (lane 5). Nucleotide positions on the left side indicate distances from thetranscription start site of P1. Right panel: Protection of the individual bases in the operator region in the presence of50 nM wild-type GalR (black line, filled circles) and 200 nM GalR S29R (grey line, open circles). The filled bar representsthe 16 bp OE operator. Individual band intensities were corrected with lane intensity differences outside the protectedregion and were divided by the intensity of the corresponding band in the GalRK reaction (lane 2 in the left panel).

614 Contact Inhibition of Transcription Initiation

assays. Wild-type GalR (50 nM) very effectivelyprotected OE. GalR S29R showed similar occupancyat 200 nM. Binding of S29R to the operator was alsoindicated by the appearance of a hypersensitive siteat position K55. The S29R mutant may have analtered contact with the right half-site of theoperator since the hypersensitive site was notobserved in the presence of wild-type GalR. Thisresult suggests that the GalR interface for DNA andRNA polymerase contacts may overlap. This issimilar to the observation in the phage F29 p4regulatory protein, in which mutations affectingp4-RNA polymerase interaction also affected DNAbending property of the regulator.37 When the GalRmutant did occupy the operator (Figure 6), itshowed no ability to repress P1 (Figure 5(A) and(B)). This suggests that the S29R substitutionimparts an nc defect in GalR.

The effect of S29R GalR in the absence and in thepresence of the inducer D-galactose on the slowestof the open complex formation steps, as detected by2-AP fluorescence, was also tested with a templatelabeled at position K5, at 37 8C. The rate offluorescence increase by the wild-type GalRincreased about 2.7-fold in the presence ofD-galactose similar to the results shown inFigure 4(B). When S29R GalR was substituted forwild-type GalR, somewhat lower values of theserate constants were obtained. However, no signifi-cant difference in rate of fluorescence increasewas found in the absence or in the presence ofD-galactose (data not shown). These experiments

suggest that S29R very likely lost the ability tocontact RNA polymerase at the slowest rate-determining step. Why S29R GalR lowers the rateof fluorescence increase (in the presence and in theabsence of D-galactose) is unclear. It is possible thatan increase in positive charge density (S29R) nearthe DNA binding interface increases the non-specific binding, thus, causing a non-specificrepression.

Discussion

In transcription initiation, if a repressor acts byantagonizing binding of RNA polymerase to thepromoter, then the dissociation constant for RNApolymerase-promoter closed complex will besignificantly decreased by the binding of therepressor. We investigated the binding of twoligands to the same molecule by determining theequilibrium binding parameters of the ligands.Since RNA polymerase-promoter closed complexformation is followed by an isomerization step, thelatter was prevented by incubating the reactions at4 8C. We used fluorescence anisotropy for measure-ment of the equilibrium binding constants of RNApolymerase and GalR to the promoter under thiscondition. The dissociation constant for RNApolymerase and P1 obtained by this method wasaround 10K9 M. This reflected significantly tighterbinding than KB

K1 derived from abortive initiationexperiments,38 which may have resulted from the

Contact Inhibition of Transcription Initiation 615

difference in temperatures. The abortive initiationexperiments were conducted at 37 8C, while theequilibrium binding experiments were at 4 8C.

Based on transient DNase protection experimentsat 19 8C, it was shown that an initial contact of RNApolymerase with the K10 segment is followed by anenhanced interaction with the K35 element withoutbasepair distortion.26 If this pattern holds, it ispossible that at 4 8C the initial binding with the K10region is followed by steps that make additionalprotein-DNA contacts (binding to the extended K10 region) at P1, without opening the base-pairs.This additional interaction may be responsible forthe tighter binding of RNA polymerase to P1observed in the absence of GalR when comparedto KB

K1. Thus, the 4 8C complex may not be the firstbinary complex in the pathway, but may be theresult of further isomerization without stranddistortion. GalR does not weaken the observedcomplex at 4 8C (Figure 3). It is unlikely that GalRcould antagonize the initial RNA polymerase–P1binary complex without weakening the subsequentcomplex with additional RNA polymerase-DNAcontacts, as the probability of steric clash is higherin the complexes with more extensive protein–DNAinteractions.

Since GalR apparently did not antagonize closedcomplexes formation at P1, we studied twodifferent base-pair opening rates using 2-APsubstitutions at the corresponding positions in oraround the K10 region.19 2-AP substitution at C3position detected a faster and an earlier step, while2-AP substitution at K7 revealed a slower and alatter step. GalR inhibited the step that wasidentified by 2-AP substituted at K7 but not atC3. Based on the isolation of the GalR S29R ncmutant and previous demonstration that therepression of P1 is dependent upon residues inthe a-CTD of RNA polymerase,5 we propose thatthe inhibition of the latter step and consequentlytranscription repression is achieved by

Figure 7. Steps and hypothetical free energy profile oftranscription initiation at the P1 promoter. Steps of opencomplex formation are as described:19 (C) closed com-plex; (i) interaction of K11A and very fast destabilizationat C2/C3; (ii) fast opening of base-pairs at K4, K2 andC3; (iii) slow opening of the remaining bases; and (O)open complex. The red line represents the trappedintermediate by GalR.

establishment of a direct contact between theS29R-containing surface of the DNA-bound GalRand the a-CTD of RNA polymerase. This is unlikethe protein–protein contact-mediated repression byCytR,39 which involves antiactivation by contactbetween CytR repressor and CRP activator atpromoters. Since GalR has very little influence onthe initial binding step, or on the faster step ofisomerization reported by 2-AP labeled at C3, it islikely that the contact is established at the later stepof base-pair distortion. The model is depicted inFigure 7.

Conclusions

The two proposed mechanisms of repression oftranscription initiation are steric hindrance andcontact inhibition. An advantage of the contactinhibition model in repression is that the kinetics ofinduction of transcription by inactivating therepressor by an inducer can be much faster.Induction in the steric hindrance model wouldrequire dissociation of the repressor from thepromoter followed by target location by RNApolymerase, both of which could be slow processes.Contact inhibition would overcome the entropicbarrier of RNA polymerase binding, offering amechanism for making rapid induction possible.

Due to a greater understanding of geneticregulatory circuits, there is now a resurgence ofinterest in quantitative modeling of genetic net-works, which has led to the design and constructionof synthetic networks involving gene regulators.40

Elucidation of mechanisms of activation andrepression is critical for proper design andconstruction of such networks. If kinetic outcomesof the two modes of gene regulation, repression oractivation, are different, regulatory systemsemploying these different mechanisms must beincorporated into the synthetic networks for properanalysis.

Materials and Methods

Materials

RNA polymerase holoenzyme was purchased fromEpicentre (Madison, WI) and used without furtherpurification. Synthetic oligodeoxynucleotides containing2-AP substitutions or fluorescein label were purchasedfrom Sigma-Genosys (The Woodlands, TX), TrilinkBioTechnologies (San Diego, CA) or FCRF-NCI(Frederick, MD). The 106 nt long polymers were purifiedby polyacrylamide gel electrophoresis. Complementaryoligonucleotides were annealed before use.

Proteins

GalR was purified as described by Majumdar et al.41 CIIand CI were as described by Shih and Gussin33 andJohnson et al.,42 respectively.

616 Contact Inhibition of Transcription Initiation

Fluorescence anisotropy

All fluorescence anisotropy measurements were donein a Panvera Beacon 2000 fluorescence polarizationspectrometer with temperature control (PanVera, Madi-son, WI) using a halogen lamp and 490 nm excitationfilter and 530 nm emission filter. Measurements weredone in 6 mm!50 mm disposable borosilicate glasstubes. The sample volume was typically 100 ml. Measure-ments were carried out at 4 8C unless mentionedotherwise in 50 mM Mes (pH 6.4), 0.2 M NaCl, 10 mMMgCl2, 100 mg/ml BSA, 1 mM DTT, 1 mM EDTA, 10%(v/v) glycerol. The buffer, proteins and oligonucleotideswere pre-equilibrated at 4 8C, then mixed and incubatedfor three minutes before fluorescence polarizationmeasurements. Each point was measured ten times andan average polarization value was obtained.

2-Aminopurine fluorescence

Fluorescence measurements were performed in aPerkin Elmer Luminescence Spectrometer LS 50B andan RTC-2000/SE spectrofluorometer from PhotonTechnology International (Lawrenceville, NJ). Sampleswere excited at 320 nm (10 nm slit width) and emission at370 nm (15 mm slit width) was measured. Prior to themeasurements, the temperatures of all the reactioncomponents and instruments were equilibrated at 25 8C.Fluorescence was measured at 25 8C, except for thetemplate with 2-AP at position K7, which was assayedat 37 8C. Binary complex formation between the promoterand RNA polymerase was initiated by adding 10 ml RNApolymerase (5 mM) to 50 ml DNA template (1 mM)previously mixed with 3 mM of wild-type or S29R GalR.The buffer was 0.05 M Tris–HCl (pH 7.5), 0.1 M NaCl,10 mM MgCl2. As a control, 50 ml of the same DNA (1 mM)was mixed with 10 ml of the RNA polymerase storagebuffer. Immediately after mixing (time 0), both reactionswere transferred to cuvettes (Hellma Cells, 0.300). Theemission from both samples were recorded alternatelyevery 15 seconds for 2000 seconds by employing amotorized four position turret cuvette holder. Tocompensate for machine error, such as machine drift,the fluorescent values of the experiment (DNACRNApolymerase) were divided by those of the control (DNACRNA polymerase storage buffer) to calculate thefluorescence ratio.

We assumed that binary complex formation is veryrapid at the concentrations used. This is likely to be thecase as the concentration of RNA polymerase is manyfold greater than the Kd for RNA polymerase binding tothe P1 promoter, and the on-rate measured for mostpromoters are around 108 MK1sK1.43 Under theseconditions, there is a rapid formation of the initialcomplex that then isomerizes by the first-order kinetics.Sigma plot 4.0 was used to fit the fluorescence data. Therate constants for open-complex formation were calcu-lated by fitting the fluorescent ratio to the equationfZy0Ca (1KeKbt), where f is the ratio at time t, y0 is theinitial ratio, a is the observed amplitude, and b is the rateconstant (kf) of the first-order reaction.

Mutagenesis of the galR gene

To create a PvuII site in the galR gene a silentsubstitution (Q110 CAA KOQ110 CAG) was introducedinto plasmid pSEM1029.44 The region between the newlycreated PvuII site and the upstream XhoI site wasamplified by error-prone PCR in the presence of 600 mM

MnSO4 and 40 mM dGTP.45 The amplified DNA frag-ments were inserted between the XhoI and PvuII sites ofthe original plasmid.

Screening of mutant GalR

Mutagenized GalR plasmids were transformed intothe tester strain DM0026. DM0026 cells contain achromosomal OC

EOKI P1CP2K wlacZ translational fusion,

an OCEO

CI P1KP2C wgusA transcriptional fusion, and a

deletion of the galR gene.46 In these cells, non-contactmutants with normal DNA binding affinity showed P1derepression while P2 repression by DNA loopingremained unchanged. Transformed cells were plated onindicator plates (LB agar containing 30 mg/ml X-gal,100 mM phenylethyl b-D-thiogalactoside, 1 mM cAMP,50 mg/ml zeocin, 50 mg/ml ampicillin, 30 mg/mlchloramphenicol). Blue colonies showing higher levelsof b-galactosidase activity were selected for furtheranalysis. To measure the effect of the mutations onlooping-mediated repression by GalR, b-glucuronidaseactivities were measured from cell extracts. Cells weregrown overnight in LB medium and diluted 50-fold forfurther growth in M63 supplemented with 0.4% (w/v)D-fructose, 0.1% (w/v) Casamino acids and 0.004% (w/v)vitamin B1. At various times, aliquots of cells wereremoved, pelleted, and resuspended in M63 mediumcontaining 100 mg/ml chloramphenicol and stored onice.47 The optical densities of the cells were measured at600 nm. The activity of b-glucuronidase from theP2wgusA fusion was determined by the SOFTMAXmicroplate spectrophotometer system (Molecular DevicesCorp., Sunnyvale, CA). Aliquots of 50 ml of permeabiliza-tion buffer (100 mM Tris (pH 7.8), 32 mM sodiumphosphate, 8 mM dithiothreitol, 8 mM CDTA, 4% tritonX-100) containing 200 mg/ml polymyxin B48 were placedin the wells of a microtiter plate, followed by the additionof 100 ml of cells. Cells were allowed to permeabilize atroom temperature for 15 minutes before 50 ml aliquots ofGUS assay buffer (0.5 mM dithiothreitol, 1 mM EDTA,50 mM sodium phosphate, pH 7) containing 1.25 mMa-p-nitrophenyl b-D-glucuronide (PNPG) were added.47

The rate of b-glucuronide hydrolysis was determined at405 nm at 37 8C.

In vitro transcription

Transcription reactions were performed as described.49

The reaction mixtures (50 ml) contained 20 mM Trisacetate (pH 7.8), 10 mM magnesium acetate, 200 mMpotassium glutamate, 2 nM DNA template, and 20 nMRNA polymerase (USB). After incubation of the reactionsat 37 8C for five minutes, transcription was started by theaddition of 1.0 mM ATP, 0.1 mM GTP, 0.1 mM CTP,0.01 mM UTP, and 5 mCi of [a-32P]UTP (3000 Ci/mmol).Reactions were terminated after ten minutes by additionof an equal volume of transcription loading buffer(0.025% bromophenol blue, 0.025% xylene cyanol,0.01 M EDTA, 90% deionized formamide). After heatingat 90 8C for three minutes, the samples were loaded onto8% (w/v) polyacrylamide-urea DNA sequencing gels forresolving the RNA products.

DNase footprinting

Transcription reaction mixture (100 ml) containing 4 nMsupercoiled DNA template was incubated with orwithout GalR protein at 37 8C for five minutes. After theaddition of 5 mM CaCl2 and one unit of RQ1 DNase

Contact Inhibition of Transcription Initiation 617

(Promega, Madison, WI), reactions were incubated forone minute and stopped by phenol-chloroform extractionfollowed by ethanol-precipitation. The purified DNA wasdissolved in 5 ml of 1! sequencing buffer containing250 mM of each of the four deoxyribonucleoside triphos-phates, one unit of sequencing grade Taq polymerase(Promega), and the 32P end-labeled primer GGTAACCAGAACTCTCATAATTCGCTCCATTAG, which wascomplementary to the top strand of the downstreamregion of the OE operator. DNAs were denatured byincubation at 94 8C for five minutes and annealed byslowly cooling to 50 8C. Primer extension was carried outat 70 8C for five minutes. Reactions were terminated bythe addition of 5 ml of transcription loading buffer andanalyzed on 8% polyacrylamide-urea DNA sequencinggels.

References

1. Jacob, F. & Monod, J. (1961). Genetic regulatorymechanisms in the synthesis of proteins. J. Mol. Biol.3, 318–356.

2. Hawley, D. K., Johnson, A. D. & McClure, W. R. (1985).Functional and physical characterization of transcrip-tion initiation complexes in the bacteriophage lambdaOR region. J. Biol. Chem. 260, 8618–8626.

3. Schlax, P. J., Capp, M. W. & Record, M. T., Jr (1995).Inhibition of transcription initiation by lac repressor.J. Mol. Biol. 245, 331–350.

4. Roy, S., Garges, S. & Adhya, S. (1998). Activation andrepression of transcription by differential contact: twosides of a coin. J. Biol. Chem. 273, 14059–14062.

5. Choy, H. E., Park, S. W., Aki, T., Parrack, P., Fujita, N.,Ishihama, A. & Adhya, S. (1995). Repression andactivation of transcription by Gal and Lac repressors:involvement of alpha subunit of RNA polymerase.EMBO J. 14, 4523–4529.

6. Choy, H. E., Hanger, R. R., Aki, T., Mahoney, M.,Murakami, K., Ishihama, A. & Adhya, S. (1997).Repression and activation of promoter-bound RNApolymerase activity by Gal repressor. J. Mol. Biol. 272,293–300.

7. Monsalve, M., Calles, B., Mencia, M., Salas, M. & Rojo,F. (1997). Transcription activation or repression byphage psi 29 protein p4 depends on the strength of theRNA polymerase–promoter interactions. Mol. Cell, 1,99–107.

8. Yamamoto, K., Ogasawara, H., Fujita, N., Utsumi, R.& Ishihama, A. (2002). Novel mode of transcriptionregulation of divergently overlapping promoters byPhoP, the regulator of two-component system sensingexternal magnesium availability. Mol. Microbiol. 45,423–438.

9. Yamamoto, K. & Ishihama, A. (2003). Two differentmodes of transcription repression of the Escherichiacoli acetate operon by IclR. Mol. Microbiol. 47, 183–184.

10. Xu, J. & Koudelka, G. B. (2001). Repression oftranscription initiation at 434 P(R) by 434 repressor:effects on transition of a closed to an open promotercomplex. J. Mol. Biol. 309, 573–587.

11. Smith, T. L. & Sauer, R. T. (1996). Role of operatorsubsites in Arc repression. J. Mol. Biol. 264, 233–242.

12. Adhya, S., Geanacopoulos, M., Lewis, D. E., Roy, S. &Aki, T. (1998). Transcription regulation by represso-some and by RNA polymerase contact. Cold SpringHarbor Symp. Quant. Biol. 63, 1–9.

13. Aki, T. & Adhya, S. (1997). Repressor induced site-specific binding of HU for transcriptional regulation.EMBO J. 16, 3666–3674.

14. Aki, T., Choy, H. E. & Adhya, S. (1996). Histone-likeprotein HU as a specific transcriptional regulator:co-factor role in repression of gal transcription byGAL repressor. Genes Cells, 1, 179–188.

15. Ujvari, A. & Martin, C. T. (1996). Thermodynamic andkinetic measurements of promoter binding by T7RNA polymerase. Biochemistry, 35, 14574–14582.

16. Sullivan, J. J., Bjornson, K. P., Sowers, L. C. &deHaseth, P. L. (1997). Spectroscopic determinationof open complex formation at promoters forEscherichia coli RNA polymerase. Biochemistry, 36,8005–8012.

17. Bandwar, R. P. & Patel, S. S. (2001). Peculiar2-aminopurine fluorescence monitors the dynamicsof open complex formation by bacteriophage T7 RNApolymerase. J. Biol. Chem. 276, 14075–14082.

18. Liu, M., Gupte, G., Roy, S., Bandwar, R. P., Patel, S. S.& Garges, S. (2003). Kinetics of transcription initiationat lacP1. Multiple roles of cyclic AMP receptorprotein. J. Biol. Chem. 278, 39755–39761.

19. Roy, S., Lim, H. M., Liu, M. & Adhya, S. (2004).Asynchronous basepair openings in transcriptioninitiation: CRP enhances the rate limiting step.EMBO J. 23, 869–875.

20. Shaner, S. L., Melancon, P., Lee, K. S., Burgess, R. R. &Record, M. T., Jr (1983). Ion effects on the aggregationand DNA-binding reactions of Escherichia coli RNApolymerase. Cold Spring Harbor Symp. Quant. Biol. 47,463–472.

21. Keilty, S. & Rosenberg, M. (1987). Constitutivefunction of a positively regulated promoter revealsnew sequences essential for activity. J. Biol. Chem. 262,6389–6395.

22. Ponnambalam, S., Chan, B. & Busby, S. (1988).Functional analysis of different sequence elements inthe Escherichia coli galactose operon P2 promoter. Mol.Microbiol. 2, 165–172.

23. Burns, H. D., Belyaeva, T. A., Busby, S. J. & Minchin,S. D. (1996). Temperature-dependence of open-complex formation at two Escherichia coli promoterswith extended -10 sequences. Biochem. J. 317, 305–311.

24. Kamali-Moghaddam, M. & Geiduschek, E. P. (2003).Thermoirreversible and thermoreversible promoteropening by two Escherichia coli RNA polymeraseholoenzymes. J. Biol. Chem. 278, 29701–29709.

25. Chamberlin, M. J. (1974). The selectivity of transcrip-tion. Annu. Rev. Biochem. 43, 721–725.

26. Li, X. Y. & McClure, W. R. (1998). Characterization ofthe closed complex intermediate formed duringtranscription initiation by Escherichia coli RNA poly-merase. J. Biol. Chem. 273, 23549–23557.

27. Tsodikov, O. V., Craig, M. L., Saecker, R. M. & Record,M. T., Jr (1998). Quantitative analysis of multiple-hitfootprinting studies to characterize DNA confor-mational changes in protein–DNA complexes: appli-cation to DNA opening by Esigma70 RNApolymerase. J. Mol. Biol. 283, 757–769.

28. Hawley, D. K. & McClure, W. R. (1983). The effect of alambda repressor mutation on the activation oftranscription initiation from the lambda PRMpromoter. Cell, 32, 327–333.

29. Bushman, F. D., Shang, C. & Ptashne, M. (1989).A single glutamic acid residue plays a key role inthe transcriptional activation function of lambdarepressor. Cell, 58, 1163–1171.

618 Contact Inhibition of Transcription Initiation

30. Hawley, D. K. & McClure, W. R. (1982). Mechanism ofactivation of transcription initiation from the lambdaPRM promoter. J. Mol. Biol. 157, 493–525.

31. Hwang, J. J., Brown, S. & Gussin, G. N. (1988).Characterization of a doubly mutant derivative of thelambda PRM promoter. Effects of mutations onactivation of PRM. J. Mol. Biol. 200, 695–708.

32. Woody, S. T., Fong, R. S. & Gussin, G. N. (1993). Effectsof a single base-pair deletion in the bacteriophagelambda PRM promoter. Repression of PRM byrepressor bound at OR2 and by RNA polymerasebound at PR. J. Mol. Biol. 229, 37–51.

33. Shih, M. C. & Gussin, G. N. (1984). Role of cII proteinin stimulating transcription initiation at the lambdaPRE promoter. Enhanced formation and stabilizationof open complexes. J. Mol. Biol. 172, 489–506.

34. Hoopes, B. C. & McClure, W. R. (1985). A cII-dependent promoter is located within the Q gene ofbacteriophage lambda. Proc. Natl Acad. Sci. USA, 82,3134–3138.

35. Chatterjee, S., Zhou, Y. N., Roy, S. & Adhya, S. (1997).Interaction of Gal repressor with inducer andoperator: induction of gal transcription fromrepressor-bound DNA. Proc. Natl Acad. Sci. USA, 94,2957–2962.

36. Lewis, D. E. A. & Adhya, S. (2002). In vitro repressionof the gal promoters by GalR and HU depends on theproper helical phasing of the two operators. J. Biol.Chem. 277, 2498–2504.

37. Mencia, M., Monsalve, M., Salas, M. & Rojo, F. (1996).Transcriptional activator of phage phi 29 latepromoter: mapping of residues involved in inter-action with RNA polymerase and in DNA bending.Mol. Microbiol. 20, 273–282.

38. Lavigne, M., Herbert, M., Kolb, A. & Buc, H. (1992).Upstream curved sequences influence the initiation oftranscription at the Escherichia coli galactose operon.J. Mol. Biol. 224, 293–306.

39. Kallipolitis, B. H., Norregaard-Madsen, M. & Valentin-Hansen, P. (1997). Protein–protein communication:structural model of the repression complex formed byCytR and the global regulator CRP. Cell, 89, 1101–1109.

40. Elowitz, M. B. & Leibler, S. (2000). A syntheticoscillatory network of transcriptional regulators.Nature, 403, 335–338.

41. Majumdar, A., Rudikoff, S. & Adhya, S. (1987).Purification and properties of Gal repressor: pL-galRfusion in pKC31 plasmid vector. J. Biol. Chem. 262,2326–2331.

42. Johnson, A. D., Pabo, C. O. & Sauer, R. T. (1980).Bacteriophage lambda repressor and cro protein:interactions with operator DNA. Methods Enzymol.65, 839–856.

43. Knaus, R. & Bujard, H. (1990). In Principles governingthe activity of E. coli promoters (Eckenstein, F. & Lilley,D. M. J., eds) Nucleic Acids andMolecular Biology, Vol. 4,pp. 110–112, Springer-Verlag, Berlin.

44. Semsey, S., Geanacopoulos, M., Lewis, D. E. & Adhya,S. (2002). Operator-bound GalR dimers close DNAloops by direct interaction: tetramerization andinducer binding. EMBO J. 21, 4349–4356.

45. Cadwell, R. C. & Joyce, G. F. (1992). Randomizationof genes by PCR mutagenesis. PCR Methods Appl. 2,28–33.

46. Lewis, D. E., Geanacopoulos, M. & Adhya, S. (1999).Role of HU and DNA supercoiling in transcriptionrepression: specialized nucleoprotein repressioncomplex at gal promoters in Escherichia coli. Mol.Microbiol. 31, 451–461.

47. Wilson, K. J., Hughes, S. G. & Jefferson, R. A. (1992). InThe Escherichia coli gus operon: induction andexpression of the gus operon in Escherichia coli andthe occurrence and use of the gus in other bacteria(Gallagher, S. R., ed) Gus Protocols: Using the Gus Geneas a Reporter of Gene Expression, pp. 7–22, AcademicPress, San Diego.

48. Schupp, J. M., Travis, S. E., Price, L. B., Shand, R. F. &Keim, P. (1995). Rapid bacterial permeabilizationreagent useful for enzyme assays. Biotechniques, 19,18–20.

49. Geanacopoulos, M., Vasmatzis, G., Lewis, D. E., Roy,S., Lee, B. & Adhya, S. (1999). GalR mutants defectivein repressosome formation. Genes Dev. 13, 1251–1262.

50. Dubendorff, J. W., deHaseth, P. L., Rosendahl, M. S. &Caruthers, M. H. (1987). DNA functional groupsrequired for formation of open complexes betweenEscherichia coli RNA polymerase and the lambda PR

promoter. Identification via base analog substitutions.J. Biol. Chem. 262, 892–898.

Edited by R. Ebright

(Received 3 August 2004; received in revised form 23 September 2004; accepted 23 September 2004)