Proc. Nati. Acad. Sci. USAVol. 86, pp. 8457-8461, November 1989Genetics

AbrB, a regulator of gene expression in Bacillus, interacts with thetranscription initiation regions of a sporulation gene and anantibiotic biosynthesis geneJEFFREY B. ROBERTSON*t, MARTIN GOCHTt, MOHAMED A. MARAHIELt, AND PETER ZUBER*t§*Department of Botany and Microbiology, Oklahoma State University, Stillwater, OK 74078; tDepartment of Biochemistry and Molecular Biology, LouisianaState University Medical Center, Shreveport, LA 71130; and tInstitut for Biochemie und Molekulare Biologie, Technische Universitdt Berlin,Franklinstrasse 29, 1000 Berlin 10, Federal Republic of Germany

Communicated by Jesse C. Rabinowitz, June 23, 1989 (received for review March 9, 1989)

ABSTRACT The abrB gene of Bacillus subtilis is believedto encode a repressor that controls the expression of genesinvolved in starvation-induced processes such as sporulationand the production of antibiotics and degradative enzymes.Two such genes, spoVG, a sporulation gene of B. subtilis, andtycA, which encodes tyrocidine synthetase I of the tyrocidinebiosynthetic pathway in Bacilus brevis, are negatively regu-lated by abrB in B. subtilis. To examine the role of abrB in therepression of gene transcription, the AbrB protein was purifiedand then tested for its ability to bind to spoVG and tycApromoter DNA. In a gel mobility shift experiment, AbrB wasfound to bind to a DNA fragment containing the sequence from-95 to +61 of spoVG. AbrB protein exhibited reduced affinityfor DNA of two mutant forms of the spoVG promoter that hadbeen shown to be insensitive to abrB-dependent repression invivo. These studies showed that an upstream A+T-rich se-quence from -37 to -95 was required for optimal AbrBbinding. AbrB protein was also observed to bind to the tycAgene within a region between the transcription start site and thetycA coding sequence as well as to a region containing theputative tycA promoter. These fridings reinforce the hypothesisthat AbrB represses gene expression through its direct inter-action with the transcription initiation regions of genes underits control.

When cells of the spore-forming bacterium Bacillus subtilisencounter a nutritionally poor environment, characteristic ofa culture that has entered stationary phase of growth, genesthat function in such diverse processes as the production ofdegradative enzymes and antibiotics, competence develop-ment, and sporulation are induced through mechanisms thatoperate at the transcriptional level (1-5). The transcription ofmany stationary phase-induced genes requires the product ofthe spoOA gene (1-5). The amino acid sequence of the spoOAgene product shows homology with the activator class of thetwo-component regulatory proteins, which supports the hy-pothesis that spoOA is part of a pathway that senses thenutritional environment (6, 7). One of the primary functionsof SpoOA is to repress the transcription of the abrB gene,whose product is believed to be a negative regulator ofstationary phase-induced genes (8, 9). The amino acid se-quence of the abrB product shares homology with the DNA-binding regions of known transcriptional regulatory proteins,suggesting that AbrB negatively affects transcription througha direct interaction with promoter DNA (9). The currentmodel of spoOA and abrB function is in keeping with earlierstudies that showed that abrB is the site of mutations thatpartially suppress the pleiotropy ofspoOA mutations (10-13).To repress abrB transcription is not the sole function of

spoOA since a spoOA abrB double mutant is unable tosporulate.Two genes that are subject to abrB-dependent repression

are spoVG (14, 15), a sporulation gene, and tycA, a Bacillusbrevis gene that encodes tyrocidine synthetase I, an enzymeof the biosynthetic pathway for the peptide antibiotic tyro-cidine (3, 16). Both are transcriptionally induced in B. subtilisunder conditions of nutritional depletion but differ withrespect to their requirements for transcription initiation andregulation. The initiation of spoVG transcription is carriedout by the minor (0.H) form ofRNA polymerase holoenzyme(17, 18), whereas the tycA promoter is recognized by themajor (0.A) holoenzyme (3). Unlike tycA, the transcription ofspoVG is induced, through an unknown mechanism, by thetreatment of cells with the sporulation-inducing agent decoy-inine (8), a drug that causes a rapid decrease in the intracel-lular concentration of GTP (19). Despite these differences,transcription from each promoter requires the product of thespoOA gene, a requirement that is overcome by mutations inabrB (3, 20). The finding that a mutation upstream of thespoVG promoter within an A+T-rich region (upstream acti-vating sequence, UAS) also overcomes the requirement forspoOA suggests that abrB-dependent control acts near thespoVG transcription initiation region (8). In this report, wedescribe the purification of the abrB product and presentevidence that the AbrB protein interacts directly with DNAthat contains the promoter of either tycA or spoVG.

MATERIALS AND METHODSBacterial Strains and Plasmids. Escherichia coli strain

AG1574 [araD139 A(ara leu)7697 AlacX74 galU,K hsr- m+strA recA56 srL] was described previously (4). E. coli strainK-38 (HfrCA, ref. 20) was obtained from A. Harker (Okla-homa State University) and S. Tabor (Harvard UniversityMedical School). E. coli strain XL1-blue {recAl endAlgyrA96 thi hsdRJ7 (rk- mk-) supE44 relAl lac- [F' proABlacP ZAM15 TnWO (tet9)]} was obtained from Stratagene andwas used as a host for bacteriophage M13. Plasmid pTV20was obtained from P. Youngman (21), and its constructionand use in the molecular cloning ofB. subtilis DNA have beendescribed previously (21). Plasmid pZSP6-189 (22) contains a217-base-pair (bp) fragment that bears the spoVG promoterregion. Plasmids pZL207 (23) and p42/327 (8) are pZA327 (24)derivatives that contain a 157-bp HindIII-Sal I fragment ofthe spoVG and mutant spoVG42 promoters, respectively.Plasmid pZA217 (24) is a pZL205 (23) derivative that containsthe mutant spoVG2J7 promoter on a 98-bp HindIII-Sal Ifragment. Plasmid pGEM3-38 (3) contains the tycA promoter

Abbreviation: UAS, upstream activating sequence.§To whom reprint requests should be addressed at: Department ofBiochemistry and Molecular Biology, Louisiana State UniversityMedical Center, 1501 Kings Highway, Shreveport, LA 71130-3932.

8457

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

1, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989)

region on a 888-bp Pvu II-Xmn I fragment. Plasmids pGEM3and pGEM4 were purchased from Promega Biotec. PlasmidpGP1-2 (25), which carries a kanamycin-resistance determi-nant and the gene encoding 17 RNA polymerase, was ob-tained from A. Harker and S. Tabor. The T7 RNA polymer-ase gene is controlled by the A PL promoter and a plasmid-encoded, heat-sensitive cI repressor (25). Phage M13 cloningvectors mp9 (26) and um3O and um31 (International Biotech-nologies) were used in the molecular cloning of DNA fornucleotide sequence analysis.DNA Manipulation. Plasmid DNA from E. coli cells was

purified by using the method of Birnboim and Doly (27) withmodifications (28). Restriction enzymes and T4 DNA ligasewere purchased from Bethesda Research Laboratories andPromega Biotec and were used according to protocols pro-vided by the vendors. M13 recombinant phage clones weresubject to nucleotide sequencing by using the method ofSanger et al. (29).AbrB Expression System. The plasmid pG3-516RH, a de-

rivative of pGEM3, contains an EcoRI-HindIII fragmentbearing the abrB coding sequence and ribosome binding site(see Fig. 1). The coding sequence lies downstream from a T7phage promoter. This plasmid was introduced by transfor-mation into a derivative of strain K-38 that carries pGP1-2.Transformants were grown at 30°C on LB medium containingampicillin (25 ,ug/ml) and kanamycin (25 ,ug/ml). Individualtransformants were grown in 2 ml of 2 x YT medium con-taining ampicillin and kanamycin at 30°C to late exponentialphase of growth (OD595 = 1.5). The temperature was raisedto 45°C to induce the synthesis of T7 RNA polymerase, andincubation was continued for 4 hr. The cells of each culturewere collected by centrifugation, suspended in electropho-resis sample buffer (30), boiled, and subjected to SDS/PAGEin a 15% polyacrylamide gel (30:1, acrylamide to N,N'-methylene-bisacrylamide) (30). Proteins were visualized bystaining the gel with Coomassie blue. Plasmid-encoded pro-teins were further examined by radiolabeling. This wasaccomplished by taking advantage of the resistance of T7RNA polymerase to the antibiotic rifampicin (25). Cultures ofcells that carry both pG3-516RH and pGP1-2 were grown tolate logarithmic phase and then simultaneously treated withrifampicin (200 ,ug/ml) and 35S-labeled methionine and cys-teine (10 ,Ci; 1 Ci = 37 GBq; ICN). The cells were incubated30 min, collected by centrifugation, and analyzed by PAGEas described above.

Purification of AbrB Protein. E. coli strain K-38/pGP1-2cells bearing plasmid pG3-516RH were grown to OD6w = 1.5at 30°C in 2 liters of2x YT with kanamycin and ampicillin (25,g/ml). The temperature was then raised to 45°C for 2-3 hr.The cells were harvested by centrifugation, resuspended in 20ml of G-50 buffer (20 mM Tris-HCI, pH 7.5/1 mM EDTA/1mM dithiothreitol), and sonicated for a total of 4 min (four1-min pulses between which the cells were incubated on icefor 1 min). The lysate was subjected to centrifugation in aSorvall SA-600 rotor at 10,000 rpm for 10 min. Protein in thesupernatant was subjected to fractionation with (NH4)2SO4 at35% and 50o of saturation. The final supernatant was com-bined with (NH4)2SO4 to 60% of saturation and placed on iceovernight. The precipitate was collected by centrifugationand dissolved in G-50 buffer. The solution was applied to agel-filtration Sephadex G-50 (Pharmacia) column (2 x 80 cm),and protein within the fractions was monitored by measuringOD280 and by SDS/PAGE. Fractions enriched for AbrBprotein were applied to a Heparin-agarose affinity column(0.5 x 10 cm) that was equilibrated with G-50 buffer. Exten-sive washing with G-50 buffer resulted in the release ofAbrBprotein that was substantially pure as judged by SDS/PAGEanalysis (see Fig. 3). The estimated recovery of the proteinwas 30-40%. Samples of AbrB protein were Iyophilized andstored at -80'C.

DNA-Binding Studies. Restriction endonuclease-cleavedspoVG DNA and lyophilized AbrB protein were dissolved inbuffer A (31) (20 mM Tris HCl, pH 7.5/10 mM MgCI2/l mMEDTA/0.3 mM dithiothreitol/3% glycerol). Protein concen-tration was determined by the Bradford assay (32). Thebinding reaction mixtures contained AbrB protein (13-130pmol) and DNA (1 prnol) in 30 gl of buffer A and wereincubated for 30 min at 250C. Two microliters of bromophe-nol blue in 15% glycerol was added to the reactions, whichwere then subjected to electrophoresis on a 6% polyacryl-amide (30:1) gel. DNA was visualized by ethidium bromidestaining and UV irradiation. Conditions for gel mobility shiftassay involving tycA DNA are described in Fig. 6.

RESULTSA transposon insertion mutation in the abrB locus wasisolated from a library of Tn917 insertions (33, 34). Theprocedure developed by Youngman et al. (21) was used toisolate the DNA flanking the putative abrB: :Tn9I7 insertion.Plasmid p2OKR1 (Fig. la) contains 3-kilobases (kb) of chro-mosomal DNA containing part of the abrB gene, one-half ofTn917, and pBR322 DNA. A 1.9-kb Cla I fragment ofp2OKR1chromosomal sequences adjacent to the transposon wasinserted into pGEM4, thus giving rise to plasmid pGC516R(Fig. lb). A 0.7-kb EcoRI-HindIII fragment of pGC516Rcontaining the end ofTn917 and flankingDNA was subjectedto sequence analysis. It was discovered that the Tn917insertion mutation was identical to that characterized byPerego et al. (9), who found that the transposon had insertedbetween the promoter and the coding sequence of abrB. Anopen reading frame encoding a protein of 10,760 Da wasidentified whose sequence was identical to the abrB codingsequence (9).An expression plasmid (pGEM3) bearing a phage T7 pro-

moter was utilized to express abrB in E. coli. Two pGEM3derivatives were constructed. Plasmid pG3-516R (Fig. lb)contains the 1.9-kb Cla I fragment of p2OKR1 that bears theabrB coding sequence. Plasmid pG3-516RH (Fig. 1) containsthe 0.7-kb EcoRI-HindIII fragment of pGC516R. Each plas-mid was used to transform E. coli strain K-38/pGP1-2 (25).A control strain contained pGC516R, which bears the abrBcoding sequence in the opposite orientation with respect tothe T7 promoter of pGEM3 (Fig. 1). Cells of the plasmid-bearing strains were grown in liquid culture and subjected toheat treatment to induce T7 promoter/abrB expression.

a p2OKR I

EcoRI Sall Cal CI EcoRII | 'Tn917 &arB

b To abrB

Clal Dml BhXI HinDIll Deal Clal

I-

p

T7

pGC5I6R

T7

W~X~X~X~N pG3-516R

PT7

WzmmpG3-516RH

KB

FIG. 1. Molecular cloning of the abrB coding sequence. (a) Mapof plasmid p2OKR1, which contains 3 kb of B. subtilis DNA imme-diately flanking abrB::Tn917. (b) Insert of pGC516R. Indicated is thelocation ofthe coding sequence ofabrB and a putative transcriptionalterminator. Also shown are the restriction maps for the inserts ofpG3-516R and pG3-516RH.

8458 Genetics: Robertson et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

1, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989) 8459

12 3 4 5 6 apSP6S

PvuII HiadIll EcoR I

I ixxx

-4 B 4- A-

.- D --

18,400

-0 .tj 1....

14.300

F F ~~~~6.200

FIG. 2. Autoradiograph of the radiolabeled products of abrBplasmids in the K-38/pGP1-2 cells. The labeled products of K-38/pGP1-2 cells containing pGC516RH (lanes 1 and 4), pGEM3(lanes 2 and 5), and pG3-516RH (lanes 3 and 6). Lanes: 1-3,35S-labeled proteins from noninduced cells; 4-6, labeled proteins ofthe induced cells. The arrow shows the radiolabeled product iden-tified as the abrB gene product.

Analysis of the T7 promoter/abrB plasmid-encoded proteinsby PAGE revealed a prominent band corresponding to aprotein of approximately 10 kDa (data not shown). This bandwas absent from the protein profiles of cells bearing either theparent pGEM3 vector or the plasmid-borne abrB sequence inthe opposite orientation to that of the T7 promoter. Theproducts of the plasmids were specifically radiolabeled inorder to more precisely examine the proteins encoded by theabrB DNA. Only the products of the DNA transcribed by therifampicin-resistant T7 RNA polymerase were radiolabeledin this procedure (25). A product was observed in the extractof the pG3-516RH-bearing cells with the same molecularmass as the 10-kDa protein observed in the Coomassieblue-stained profile (Fig. 2). Another protein was also de-tected, but the same species was observed in cells of thecontrol strain containing the parent pGEM3 vector. Severalproteins of molecular mass higher than the putative AbrBprotein were produced in cells with plasmids containing theabrB insert DNA in the opposite orientation with respect tothe T7 promoter. Since the 10-kDa protein was the onlyradiolabeled protein attributed to the T7 promoter/abrBsequences, this protein was identified as the product of theabrB coding sequence.The AbrB protein was produced in the E. coli K-38 cells

carrying pG3-516RH and purified by a procedure involvingammonium sulfate fractionation, gel filtration, and heparinaffinity gel column chromatography (see Materials and Meth-ods). Fig. 3 shows a SDS/PAGE profile of the final productthat was used in the DNA-binding experiments. Gel mobilityshift experiments were performed with the AbrB protein andspoVG promoter DNA. Plasmid pZSP6-189 was cleaved withEcoRI, HindIII, and Pvu II to generate a 182-bp Pvu II-HindIII fragment (fragment B in Fig. 4) of the vector (pSP65)DNA, a 215-bp EcoRI-HindIII fragment (fragment A in Fig.

n.,

- 1 ,0

I14,300

-0.- 6,200

3,400; 2,300

FIG. 3. Gel profile of the purified AbrBprotein (left lane) along with molecular sizemarkers (right lane). Details of the purifica-tion and SDS/PAGE analysis are given inthe Materials and Methods, the Results, andFig. 2.

b

A-B-

A -B -

C 1 2 3 4 5 6 7 8

I& i c

C

D

FIG. 4. The binding of AbrB to spoVG promoter DNA. (a)Restriction map of plasmid pZSP6-189 in the vicinity of the spoVGpromoter-bearing EcoRI-Hindlll fragment (fragment A). Alsoshown are locations of the spoVG DNA of plasmids pZL207 andp42/327 (fragment C) and pZA217 (fragment D). The X in the linerepresenting fragment C indicates the location of the spoVG42mutation. Fragment B contains pSP65 vector DNA. The photographsshow the ethidium bromide-stained polyacrylamide gels of bindingreactions between restriction endonuclease-cleaved plasmid DNAand AbrB protein. (b) Lane 1, molecular size markers (Dpn l-cleavedpGEM4, Promega Biotec); lane 2, EcoRI-HindIII-Pvu 11 cleavedpZSP6-189; lane 3, same as lane 2 plus 13 pmol of AbrB; lane 4, sameas lane 2 plus 65 pmol of AbrB; lane 5, same as lane 2 plus 130 pmolof AbrB. (c) AbrB-DNA binding reactions using mixtures ofHindIII-Sal I-cleaved pZL207 and pZA217 DNA (lanes 1-4) andHindfII-Sal I-cleaved p42/327 and pZA217 (lanes 5-8). Lanes 1 and5, no AbrB protein; lanes 2 and 6, 33 pmol ofAbrB protein per pmolof DNA; lanes 3 and 7, 66 pmol of AbrB per pmol of DNA; lanes 4and 8, 132 pmol of AbrB per pmol of DNA.

4) containing the promoter region of spoVG from -97 to+ 118, and the remaining vector DNA. A fourth fragment(fragment A-B in Fig. 4) is a partial digestion product com-posed of fragments A and B. Samples of the restrictionendonuclease-digested plasmids were combined with AbrBprotein (13, 65, and 130 pmol per pmol of DNA). The threelow molecular weight DNA species (fragments A, B, andA-B) were observed in the polyacrylamide gel profile of thesample to which no protein was added (Fig. 4b). Shifts in thepositions of bands A and A-B were observed in the gelprofiles of the samples that contained the AbrB protein,indicating that AbrB binds preferentially to the fragmentscontaining spoVG DNA.Experiments were undertaken to examine the binding

affinity ofAbrB for two mutant forms of the spoVG promoterthat have been shown to be insensitive to abrB-dependentrepression in vivo (ref. 8; D. Frisby and P.Z., unpublishedresults; ¶). One of these is the spoVG42 promoter, whichbears a C to T transition at position -58 within the spoVGUAS (Fig. 5). The other is a spoVG deletion derivative,spoVG217, which is lacking sequences from -37 to -95. Inthe experiment summarized in Fig. 4c, both spoVG andspoVG217 promoter-bearing fragments were mixed and com-bined with AbrB protein. The protein-DNA complexes werethen subject to polyacrylamide gel electrophoresis. At lowprotein concentration, AbrB was found to bind preferentiallyto the fragment bearing the wild-type promoter DNA. At a2-fold higher concentration, binding to the deletion fragment

Zuber, P., 10th International Spores Conference, March 1988,Woods Hole, MA.

Genetics: Robertson et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

1, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989)

AGCTTTATGACCT

UASAATTGTGTAACTATATCCTATTTTTTCAAAAAATATTTTA

T SpoVG42

spoVG217

AAAACGAGCAGGATTTCAGAAAAAATCGTGGCCAATTGATA

CACTAATGCTTTTATATAGGGAAAAGGTGGTGAACTACT

spoVG217, spoVG42met glu val thr asp val ar leu g val ass thr aspGTG GAA GTT ACT GAC GTA AGA TTA CGCC CC CTGC AAT ACC GAT

gly arg met arg ala ile ala ser iHe ithr lea asp his giuGGT CGC ATG AGA GCG ATT GCA TCC ATC ACC CTG GAT CAC GAA IT

FIG. 5. Nucleotide sequence of the spoVG promoter region (8,23, 24). The nucleotide sequence of the EcoRI-HindIII fragment ofpZSP6-189 is shown. The transcriptional start site is indicated by thehorizontal arrow. The location of the nucleotide positions that havebeen shown to be important for the aH RNA polymerase interactionis indicated by asterisks below the sequence. The ends of fragmentscontaining the spoVG217 deletion mutation promoter as well as thedownstream end point of the DNA fragments containing the mutantspoVG42 promoter are indicated by the brackets above the se-quence. The location of the UAS is indicated by the single horizontalline above the sequence. The ribosome binding site is shown by thedouble horizontal lines above the sequence.

was observed. When the spoVG42 promoter fragment wascombined with spoVG2J7 DNA and AbrB protein, the twomutant promoter fragments were observed to bind AbrB withnear equal affinity. Binding of AbrB to spoVG42 occurred ata concentration of protein that was 2-fold higher than thatrequired for AbrB to interact with the wild-type spoVGpromoter DNA. These results indicate that AbrB shows ahigher affinity for the wild-type spoVG promoter than foreither of its mutant derivatives.The protein was also tested for its ability to bind to the

transcription initiation region of another abrB-controlledgene, tycA (Fig. 6). The plasmid pGEM3-38 (3) contains thepromoter region of tycA and the leader region that liesbetween the transcription start site and the beginning of thetycA coding sequence (3, 35). pGEM3-38 DNA (600 ng) wascleaved with Pvu II and Hae II and combined with AbrB atconcentrations of 25-125 ng per 30-1.d reaction mixture.Binding ofAbrB to the DNA was observed to the 598-bp PvuI-Hae II fragment containing sequences at +39 to a point 309bp downstream within the leader region of tycA. At higherAbrB concentrations, the protein was observed to bind to the231-bp fragment containing the promoter region of tycAbetween +38 and -193. The migration of the Hae II-Pvu Ifragments of vector plasmid DNA in the polyacrylamide gelwas not altered by the addition of the AbrB protein.

DISCUSSIONThe abrB gene encodes a protein that is believed to functionas a negative regulator of genes that are transcriptionallyinduced in cells that have entered stationary phase ofgrowth.A logical explanation for the observed effect ofAbrB on geneexpression is that the protein binds directly to the promoterand/or regulatory sequences of the genes that are underabrB-dependent negative control. The data presented hereshow that the abrB product binds to DNA near the transcrip-tion initiation region of both spoVG and tycA, although atdifferent regions with respect to their transcription start sites.A region upstream of the spoVG sequence where RNApolymerase interacts is required for optimal AbrB binding. Amutation in this region, spoVG42 (located in the UAS) and adeletion mutation, spoVG2J7 (which removes the UAS) both

IcA

p 58 % IH H 598 P 231 H 49755 60 21

370 H 1012 P

1 2 3 4 5 6

I

1012.958-598-497-370231-

6055,

FIG. 6. The binding of AbrB protein to the regulatory region oftycA. (Upper) Pvu I and Hae II map of plasmid pGEM3-38 and thepromoter fragment. The map also indicates the location of the tycApromoter (PtycA) on the 231-bp fragment and the direction of tran-scription. (Lower) DNA binding was detected by electrophoresismobility shift assay as follows: the assay was carried out in a 25-glreaction mixture containing a total of 600 ng of digested DNA and25-125 ng of purified AbrB protein in 20 mM Tris HCl, pH 8.6/5 mMMgCl2/50 mM KCI. The reaction mixture was incubated at roomtemperature for 30 min and then applied to a polyacrylamide gel. Thegel was run at 4TC for 5 min at 400 V and then at 200 V for 1 hr. Thegel was stained with ethidium bromide. Lanes: 1, no AbrB; 2, 25 ngof AbrB; 3, 50 ng of AbrB; 4, 75 ng of AbrB; 5, 100 ng of AbrB; 6,125 ng of AbrB. Open arrowheads indicate the positions of the 598-and 231-bp fragments prior to AbrB addition. Solid arrowheadsindicate the position of the gel-retarded 598-bp and 231-bp fragments.Numbers indicate the sizes of corresponding fragments (in bp). P,Pvu I; H, Hae II.

render spoVG transcription insensitive to AbrB-dependentcontrol. In vitro, AbrB protein exhibits less affinity for theDNA of the two mutant promoters than to the wild-typepromoter. Though binding to the two mutant promoter frag-ments is observed, it is not known if this is a result ofnonspecific binding of AbrB. The data suggest that there maybe more than one binding site in the spoVG promoter regionor that AbrB binds as a multimer, as evidenced by theappearance of partial complexes (Fig. 4b, lane 3) that migrateto positions between those of the unbound DNA (Fig. 4b,lane 2) and the slowest migrating AbrB-DNA band (Fig. 4b,lane 5).

Binding of AbrB to tycA was also demonstrated in gelmobility shift assays. In this case, the protein bound to aregion at least 37 bp downstream from the transcriptionalstart site and upstream from the tycA coding sequence. Athigher protein concentrations, AbrB was observed to bind tothe DNA containing the putative promoter of tycA. The exactlocation of the AbrB binding regions in both the spoVG andtycA promoters awaits more precise methods for localizingprotein-DNA interactions such as DNase protection analysisor footprinting. Comparison of the spoVG and tycA nucleo-tide sequences in the regions where AbrB binds showshomology at several locations that contain an abundance ofadenine and thymine residues. A region at -64 and -54 ofspoVG (TTTTTCAAAAA, which lies within the UAS andcontains the mutation site of spoVG42), is identified as apotential target for AbrB binding. The leader region of thetycA gene contains sequences of repeated GAAAA, a se-quence that has been suggested as a site important for AbrBinteraction (47).These studies provide evidence that is consistent with the

idea that AbrB can serve as a repressor of gene transcriptionby directly binding to the transcription initiation region,thereby impairing the interaction between RNA polymeraseand promoter DNA. This interaction would seem to beindependent of the form ofRNA polymerase that utilizes thepromoter, as the spoVG promoter is recognized by the 0.H

-j --<t4

8460 Genetics: Robertson et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

1, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989) 8461

form of holoenzyme, whereas the tycA promoter is probablyutilized by the aA holoenzyme. In the case ofspoVG, the siteof AbrB interaction includes the UAS (36). This region issimilar to UASs that are associated with powerful promoterssuch as those of ribosomal RNA operons (37). If this regionstimulates transcription through a direct interaction withRNA polymerase, then one could imagine that this interac-tion would be disrupted when AbrB is bound. Anotherpossible mechanism relates to the finding that A+T-richsequences impart curvature in the axis of the DNA helix (38,39) and that this, somehow, exerts a stimulatory effect ontranscription initiation (40). AbrB could alter the conforma-tion of the UAS, perhaps by reducing DNA curvature, insuch a way as to render the UAS incapable of enhancingtranscription initiation. In the case of tycA, AbrB binds to asequence that is spatially separated from the region of RNApolymerase-promoter interaction. It is now well establishedthat several transcriptional repressors in prokaryotes canexert their effect on initiation from a distance (41-43). AbrBprotein also interacts with DNA near the tycA promoterregion, which raises the possibility that two AbrB-DNAcomplexes repress transcription by interacting with oneanother. Such "communication" between DNA-repressorcomplexes in prokaryotes has been demonstrated (44-46).The abrB gene product is but one of several factors that

regulate B. subtilis genes that are induced in the transitionstage when exponential growth ends and stationary phasebegins. Cells of an abrB mutant sporulate normally, and thestationary phase induction of spoVG and aprE appear to beunaltered by an abrB mutation (5, 8). It has been proposedthat AbrB represses vegetative phase expression of thesegenes, but, since there is little change in the pattern ofexpression in an abrB mutant, there probably exist othermechanisms for ensuring that stationary phase-induced genesremain repressed during vegetative growth. In contrast, tycAtranscription appears to be regulated principally by spoOAand abrB, as evidenced by the constitutive expression of atycA-IacZ fusion in an abrB mutant (3). This suggests thatthere are a collection ofgenes in B. subtilis that are repressedin vegetative phase solely through an abrB-dependent mech-anism. The promoters of several of these genes have recentlybeen identified. 11 It is through the study ofthese genes and thecontinued examination of AbrB-promoter interactions thatthe involvement of abrB in the regulation of gene transcrip-tion and in the developmental cycle ofB. subtilis can be betterunderstood.

IZuber, P., New England Region Bacillus Spores Conference,November 1988, Boston.

We thank S. Tabor, A. Harker, A. Grossman, P. Youngman, andJ. Hoch for strains that were used in these studies and T. Henkin forcritical reading of the manuscript. We also appreciate the advice andhelpful discussion provided by M. Nakano and R. Losick. Supportwas provided through Public Health Service Grant GM39479-02 fromthe National Institute of General Medical Sciences to P.Z., fromOklahoma Center for the Advancement of Science and TechnologyGrant HR8-3342 to P.Z., and from the Deutsche Forschungsgemein-schaft (Sfb9-D6) to M.A.M.

1. Losick, R., Youngman, P. & Piggot, P. J. (1986) Annu. Rev.Genet. 20, 625-669.

2. Albano, M., Hahn, J. & Dubnau, D. (1987) J. Bacteriol. 169,3110-3117.

3. Marahiel, M. A., Zuber, P., Czekay, C. & Losick, R. (1987) J.Bacteriol. 147, 432-442.

4. Nakano, M., Marahiel, M. & Zuber, P. (1988) J. Bacteriol. 170,5667-5668.

5. Ferrari, E., Henner, D. S., Perego, M. & Hoch, J. A. (1988) J.Bacteriol. 170, 289-295.

6. Ferrari, F. A., Trach, K., LeCoq, P., Spencer, J., Ferrari, E.

& Hoch, J. A. (1985) Proc. Natl. Acad. Sci. USA 82, 2647-2651.

7. Nixon, B. T., Ronson, C. W. & Ausubel, F. M. (1986) Proc.Natl. Acad. Sci. USA 83, 7850-7854.

8. Zuber, P. & Losick, R. (1987) J. Bacteriol. 169, 2223-2230.9. Perego, M., Speigelman, G. B. & Hoch, J. A. (1988) Mol.

Microbiol. 2, 689-699.10. Guespin-Michel, J. F. (1971) J. Bacteriol. 108, 241-247.11. Ito, J., Mildner, C. & Spizizen, J. (1971) Mol. Gen. Genet. 112,

104-109.12. Ito, J. (1973) Mol. Gen. Genet. 124, 124, 97-106.13. Trowsdale, J., Chen, S. M. H. & Hoch, J. A. (1979) Mol. Gen.

Genet. 8, 61-70.14. Segall, J. & Losick, R. (1977) Cell 11, 751-761.15. Ollington, J. F., Haldenwang, W. G., Huynh, T. V. & Losick,

R. (1981) J. Bacteriol. 147, 432-442.16. Marahiel, M. A., Krause, M. & Skarpeid, H.-J. (1985) Mol.

Gen. Genet. 201, 231-236.17. Carter, H. L., III, & Moran, C. P., Jr. (1986) Proc. Natl. Acad.

Sci. USA 83, 9438-9442.18. Dubnau, E., Weir, J., Nair, G., Carter, L., Moran, C. P. &

Smith, I. (1988) J. Bacteriol. 170, 1054-1062.19. Freese, E., Heinze, J. & Galliers, E. M. (1979) J. Gen. Micro-

biol. 115, 193-205.20. Russell, M. & Model, P. (1984) J. Bacteriol. 159, 1034-1039.21. Youngman, P. J., Perkins, J. B. & Losick, R. (1984) Mol. Gen.

Genet. 195, 424-433.22. Zuber, P., Healy, J. B. & Losick, R. (1987) J. Bacteriol. 169,

461-469.23. Zuber, P. & Losick, R. (1983) Cell 35, 275-283.24. Zuber, P. (1985) in Molecular Biology of Microbial Differenti-

ation, eds. Hoch, J. A. & Setlow, P. (Am. Soc. Microbiol.,Washington, DC), pp. 149-156.

25. Tabor, S. J. & Richardson, C. C. (1985) Proc. Natl. Acad. Sci.USA 82, 1074-1078.

26. Messing, J. & Vieira, J. (1982) Gene 19, 269-276.27. Birnboim, H. C. & Doly, J. (1979) Nucleic Acids Res. 7,

1513-1522.28. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

29. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.- Acad. Sci. USA 74, 5463-5467.

30. Adams, L. D. (1987) in Current Protocols in Molecular Biol-ogy, eds. Ausubel, F. M., Brent, R., Kingston, R. E., Moore,P. B., Seidman, J. G., Smith, J. A. & Struhl, K. (Wiley Inter-science, New York), pp. 10.2.1-10.2.7.

31. Ikeuchi, T., Tsunasawa, S. & Sakiyama, F. (1987) Eur. J.Biochem. 167, 233-238.

32. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.33. Zuber, P., Marahiel, M. A. & Robertson, J. (1988) in Genetics

and Biotechnology of Bacilli, eds. Ganesan, A. T. & Hoch,J. A. (Academic, New York), Vol. 2, pp. 123-127.

34. Sandman, K., Losick, R. & Youngman, P. (1987) Genetics 117,603-617.

35. Weckermann, R., Furbass, R. & Marahiel, M. A. (1988) Nu-cleic Acids Res. 16, 11841.

36. Banner, C. D. B., Moran, C. P. & Losick, R. (1983) J. Mol.Biol. 168, 351-365.

37. Gourse, R. L., deBoer, H. A. & Nomura, M. (1986) Cell 44,197-205.

38. Koo, H.-S., Wu, H.-M. & Crothers, D. M. (1986) Nature(London) 320, 501-506.

39. Diekmann, S. (1987) Nucleic Acids Res. 15, 247-265.40. Bossi, L. & Smith, D. M. (1984) Cell 39, 643-652.41. Dandenell, G. & Hammer, K. (1985) EMBO J. 4, 3333-3338.42. Dunn, T., Hahn, S., Ogden, J. & Schleif, R. (1984) Proc. Natl.

Acad. Sci. USA 81, 5017-5020.43. Irani, M. H., Orocz, L. & Adhya, S. (1983) Cell 32, 783-788.44. Galla, J. D. (1989) Cell 57, 193-195.45. Haber, R. & Adhya, S. (1988) Proc. Natl. Acad. Sci. USA 85,

%83-9687.46. Hothschild, A. & Ptashne, M. (1988) Nature (London) 336,

353-357.47. Smith, I. (1989) in Regulation ofProkaryotic Development, eds.

Smith, I., Slepecky, R. A. & Setlow, P. (Am. Soc. Microbiol.,Washington, DC), in press.

Genetics: Robertson et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

1, 2

021