Vol. 175, No. 10JOURNAL OF BACTERIOLOGY, May 1993, p. 3182-31870021-9193/93/103182-06$02.00/0Copyright © 1993, American Society for Microbiology

ComA, a Phosphorylated Response Regulator Protein ofBacillus subtilis, Binds to the Promoter Region of srfA

MANUELA ROGGIANIt AND DAVID DUBNAU*Public Health Research Institute, 455 First Avenue, New YorkA New York 10016

Received 15 January 1993/Accepted 8 March 1993

ComA is a response regulator protein of Bacillus subtilis which is required for the transcription of severalgenes which are involved in late-growth expression and in responses to environmental stress. Among thesegenes are degQ, gsiA, and srfA. The last is an operon needed for the development of genetic competence,surfactin production, and normal sporulation. We show here that partially purified ComA protein, isolatedfrom an overproducing Escherichia coli strain, is phosphorylated in vitro by incubation with acetyl phosphateand that ComA could bind specifically to a DNA fragment containing the promoter of srfA and associatedsequences. The binding affinity is enhanced when ComA is phosphorylated. DNase I protection analysisidentified two protected sites located upstream from the srA promoter. The presence of DNase I-hypersensitive

bonds induced by ComA binding which are located between the protected sequences is consistent with a modelfor ComA action involving the bending of DNA.

The development of genetic competence in Bacillus sub-tilis is controlled by several regulatory gene products whichact as components of a signal transduction cascade (2).comP, comA, comQ, and spoOK are thought to encodeproteins that are able to sense carbon and nitrogen levels inthe medium (9, 25) as well as population density (4, 16). Thepredicted amino acid sequence of ComP shows similarity tothe histidine kinase class of signal transduction proteins ofprokaryotes in its carboxy-terminal domain and possesses apredicted amino-terminal membrane-spanning domain (25).The ComA amino-terminal sequence shows similarity tothose of the response regulator proteins and contains severalamino acid residues which are invariant in these proteins (20,24). In particular, it has a conserved aspartate residue (D55)which is a likely site for phosphorylation (24). Geneticevidence suggests that ComP and ComA are partners in asignaling pathway (23, 25) which may involve the transfer ofa phosphoryl moiety from ComP to ComA. In addition, thecarboxy-terminal third of the predicted ComA sequencecontains a possible helix-turn-helix DNA binding motifwhich resembles the DNA binding sequences of a family oftranscriptionally active proteins, suggesting for ComA therole of transcriptional activator (2, 24). This family includesseveral response regulator proteins (e.g., DegU, FixJ, andNarL) and other activators that are not in the responseregulator category (e.g., MalT, GerE, LuxR, and SigB).comP and comA are both necessary for the expression ofsrfA, an operon required for production of the peptideantibiotic surfactin as well as for efficient sporulation and thedevelopment of competence (6, 12, 22). Both ComA andComP are also needed for the regulated expression of degQ(9) and gsiA (10). These genes encode products which areexpressed in response to nutrient starvation but are notrequired for competence.Bypass experiments with Pspac-driven srfA expression

indicate that the main roles in competence for comP, comA,spoOK, and comQ are to activate the transcription of srfA (5,

* Corresponding author.t Present address: Department of Microbiology, Mayo Memorial

Building, University of Minnesota, Minneapolis, MN 55455-0312.

14). We have proposed that the sensing of environmentalconditions results in autophosphorylation of ComP and thesubsequent transfer of the phosphate group to ComA, whichthen directly activates the expression of srfA. Deletionanalysis of the srfA promoter has identified upstream se-quences which are necessary for its expression and whichcontain two palindromic sequences (13). Similar sequencesare also present upstream from the promoter regions ofdegQ(9) and gsiA (10), both of which require comP and comA fortheir expression. It was suggested by Msadek et al. andMueller et al. that the palindromic sequences (ComA boxes)may be binding sites for ComA. Furthermore, since in allcases ComP is also required, the phosphorylated form ofComA was postulated to be the activator of srfA transcrip-tion. Further mutational evidence strongly supports thehypothesis that the ComA boxes provide sites for the actionof ComA, whether direct or indirect (15).

In this study, we demonstrate that phospho-ComA bindsdirectly to the promoter region of srfA in vitro and that thisprotein binds to sites on the DNA corresponding to thepreviously identified ComA boxes.

MATERIALS AND METHODS

Plasmids and strains. A fragment of DNA containing thecomA coding sequence and ribosomal binding site wasobtained by polymerase chain reaction (PCR) with PftIDNA polymerase (Stratagene). The PCR product was di-gested at the XbaI and BamHI sites, which had beenincluded in the primers, and cloned between the XbaI andBamHI sites of the T7 expression vector pET3b (21). Theresulting plasmid (pET-comA) was moved into the Esche-richia coli host BL21 DE3 pLysS (21) to create the strainED111.

Overexpression of ComA. ED111 cells were grown inLuria-Bertani broth containing 5 ,ug of chloramphenicol perml and 100 ,ug of ampicillin per ml. The uninduced level oftranscription ofcomA from this plasmid, which was found inindependent experiments to be harmful to E. coli (data notshown), was repressed in the BL21 DE3 pLysS host (21).Expression of ComA was induced in mid-log phase by theaddition of 1 mM IPTG (isopropyl-p-D-thiogalactopyrano-

3182

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

ComA BINDING TO THE srfA PROMOTER REGION 3183

side). Cells were harvested 3 h after induction, washed oncein buffer A (50mM Tris-HCl [pH 8.0], 2mM EDTA, 1 mMphenylmethylsulfonyl fluoride) and resuspended in buffer Acontaining 7.5% glycerol. The cells were lysed by freezingand thawing and then passed twice through a French press at1,200 lb/in'. After centrifugation at 45,000 rpm in a 50 Tirotor at 20C, the bulk of ComA was recovered in the solublefraction, as judged by sodium dodecyl sulfate-polyacryl-amide gel electrophoresis (SDS-PAGE), on 15% (wtlvol)gels.

Partial purification of ComA. To remove DNA, treatmentof the crude extract containing ComA with 0.9% (wt/vol)streptomycin sulfate was followed by centrifugation andrecovery of the supernatant. Subsequently, ComA wasprecipitated by the addition of (NH4)2SO4 to 40% (wt/vol).The resulting pellet was dissolved in a small volume of bufferB (50 mM Tris-HCl [pH 8.0], 2 mM EDTA, 10 mM MgCl2,100 mM KCI, 7.5% glycerol, 1 mM phenylmethylsulfonylfluoride). As judged by SDS-PAGE (15%) and scanningdensitometry of the Coomassie-stained gel, about 50 to 70%of this preparation consisted of ComA. The concentratedprotein solution was divided into small portions and stored at-700C.Phosphorylation of ComA by acetyl-PO4. Dilithium acetyl-

32pO4 was synthesized essentially as described by Stadtman(18) to a specific activity of approximately 8.5 mCi/mmol.The Stadtman procedure was modified by the use of 0.45rather than 4.5 ml of 4 N LiOH. ComA (0.5 mg/ml) wasincubated at 370C in buffer C (50 mM Tris-HCl [pH 7.0], 5mM MSCl2, 1 mM dithiothreitol) containing 50, 10, or 2 mMacetyl- 2p04. Aliquots were withdrawn after 5, 10, 30, and60 min of incubation, and the reaction was stopped by theaddition of SDS sample buffer. Samples were analyzed bySDS-PAGE on 15% gels and autoradiographed. For DNA-protein-binding assays (see below), ComA-PO4 was obtainedby incubating the protein for 60 min at 37°C in buffer Ccontaining 10 mM unlabeled acetyl-PO4 (Sigma) with nofurther purification.

Efficiency of phosphorylation of ComA by acetyl-32P04. Analiquot of a ComA preparation containing 1.25 mg of totalprotein was phosphorylated in the presence of 50 mMacetyl-32PO4 at 37°C for 60 min. The reaction was stopped bythe addition of EDTA to 10 mM and SDS to 0.5% (wtlvol).The unreacted acetyl-32PO4 was removed with a gel filtrationcolumn (G-25 superfine; Sigma) previously equilibrated withbuffer (50 mM Tris-HCl [pH 7.0], 0.5% SDS). The elutedfractions were analyzed by SDS-PAGE (15%) followed byCoomassie staining, and those containing ComA werepooled. Radioactivity was measured in a scintillationcounter, and the protein concentration was determined withthe Bio-Rad DC protein assay reagent.

DNA-protein-binding assay. The DNA binding activity ofComA was measured by gel retardation assays. The probewas a 290-bp DNA fragment containing 286 bp from the srfApromoter region (from positions -154 to +132), includingthe promoter itself and the two putative ComA boxes (13).The DNA fragment was obtained by PCR, and the upstreamprimer carried a unique EcoRI restriction site. The DNAfragment was digested with EcoRI and end labeled by fillingin with Klenow enzyme (Pharmacia) in the presence of[a-32P]dATP. The binding reaction volume of 20 ,ul con-tained 0.2 ng of radiolabeled DNA, 2 ,g of DNA poly(dI-dC) poly(dI-dC) (Pharmacia), and 1 to 30 ,g (total protein)of either the ComA or ComA-PO4 preparation in bindingbuffer (10 mM HEPES [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid] [pH 7.6], 2 mM MgCl2, 0.1 mM EDTA,

1 8.4_1*

FIG. 1. Coomassie-stained gel of partially purified ComA. Theposition of the ComA protein is indicated by an arrow. The positionsof molecular mass markers (in kilodaltons) are also shown. About 5fg of total protein was loaded.

200 mM KCl, 10% glycerol, 5 mM dithiothreitol). Thebinding reaction was started by the addition of probe. Afterbeing incubated at room temperature for 20 min, sampleswere loaded on a 5% nondenaturing polyacrylamide gelwhile being run at 300 V. The gel was then electrophoresedat 120 V for 2 h, dried, and subjected to autoradiography.DNase I footprinting. The probe used was a PCR fragment

containing the srfA promoter region from positions -154 to-1. The upstream and downstream primers contained EcoRIand BamHI restriction sites, respectively. After digestionwith BamHI, the top strand was end labeled by being filled inwith Klenow enzyme in the presence of [a- 2P]dGTP. Thebottom strand was labeled by the polynucleotide kinase(Boehringer) reaction in the presence of [_y-32P]ATP andsubsequently digested with EcoRI. Both probes were puri-fied by electroelution from 8% polyacrylamide gels. Bindingreactions were performed as described above, but in a 50-iulvolume. Bovine serum albumin was added instead of ComAor ComA-PO4 to controls. Probe containing approximately30,000 cpm was added to each reaction mixture. After thebinding reaction, 1 mM MgCl2 and 0.5 mM CaCl2 wereadded to the mixture. After 1 min at room temperature, RQ1DNase I (Promega) was added; 1 min later, the reaction wasstopped with 140 RI of stop solution (192 mM sodiumacetate, 32 mM EDTA, 0.14% SDS, 64 Rg of yeast tRNA perml). The samples were extracted with phenol-chloroform,and the DNA was ethanol precipitated and then resuspendedin 4 RI of loading buffer. After being heated at 90'C for 1 min,samples were electrophoresed on 8% polyacrylamide gelscontaining urea. The gels were dried and subjected toautoradiography. G+A sequencing reactions (17) were runwith each experiment to locate sequence positions andprotected regions.

RESULTS

ComA overproduction and partial purification. ComA wasoverproduced in E. coli by using a T7 RNA polymeraseexpression system (21). The induction of ComA expressionreduced cell growth considerably (data not shown). Afterlysis and centrifugation, the bulk of ComA was found in thesupernatant fraction in soluble form. Upon precipitation ofDNA and some cellular proteins with streptomycin sulfate,most of the ComA was again recovered in the supernatantand was then completely precipitated by ammonium sulfateat 40% saturation. From Coomassie blue staining (Fig. 1) andscanning densitometry of an SDS-polyacrylamide gel, ComAwas estimated to represent 50 to 70% of the total protein inthis preparation.

Phosphorylation of ComA by acetyl-PO4. Some response

VOL. 175, 1993

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3184 ROGGIANI AND DUBNAU

50 mM 10 mM 2mM5 10 30 60 5 10 30 60 5 10 30 60

-43

-18.4

FIG. 2. Phosphorylation of ComA by acetyl-32P04. The threesections exhibit the kinetics of ComA phosphorylation with 50, 10,and 2 mM acetyl_32p04, carried out as described in Materials andMethods. Samples were taken at 5, 10, 30, and 60 mi. The sampleswere resolved by 15% SDS-PAGE followed by autoradiography.The positions of molecular mass markers (in kilodaltons) are alsoshown on the right.

regulator proteins have been shown to be selectively andspecifically phosphorylated in the presence of various low-molecular-weight 'hosphodonor molecules (3, 7). Wewished to explore this possibility for ComA and to obtainComA-P04 for use in DNA binding assays. As shown in Fig.2, after ComA was incubated with acetyl1_32P04, a radioac-tive band with the expected mobility for ComA was de-tected. The radioactive band appeared after 5 min of incu-bation in the presence of 50 mM acetyl- 04 and increasedin intensity with time.TO facilitate the interpretation of binding experiments', the

percentage of ComA-Po4 present in our preparation wasestimatedx. tP04-ComA was separated from unreactedacetyl-32P04 On a sizing column. The CoinM-containingfractions were pooled, and total protein and radioactivitywere measured. SDS-PAGE analysis demonstrated thatsome further purification had occurred on the, sizing column(data not shown) and that ComA constituted about 80 to 90%of the recovered protein. On the basis of the approximatespecific activity of the acetyl-P04rthe radioactivit o-ered, and the concentration of recovered ComA, we esti-mated that about 30% of the ComA in our preparation wasphosphorylated.

Binding of ComA to the srfA promoter region. The tran-scription of srfA in vivo is dependent on the presence offunctional ComA and ComP (5, 14). Deletion analysis of thesrfA promoter suggested that a region of approximately 160bp upstream of the transcriptional start site was necessaryfor srfA expression (13). To test whether the srfA promoteris a direct target for ComA binding, the in vitro interactionsof srfA DNA with ComA and ComA-P04 Were studied. Theelectrophoretic mobility of the srfA DNA fragment wasretarded following incubation with ComA (Fig. 3). Thisretardation was observed in spite of the presence of a largeexcess. by weight (1,000-fold) of competing poly(dI-dC).poly(da-dC), suggesting that ComA binds directly and spe-cifically to the srfA promoter. About 50% of the input DNAwas in the complexed state when 5 rdgof the ComA-Po4preparation was present, whereas 30 edgof the nonphospho-rylatedComtApreparation bound less than half of the DNAprobe. These data suggest that the phosphorylation of ComAenhances the efficient recognition and binding of this proteinto its target DNA.DNase I protection by ComA. To confirm that ComA

recognizes specific sites near the srfA promoter and toidentify these binding sites, DNase I footprinting studieswere carried out. A DNA fragment of srfA containing thetwo putative ComA boxes (13) as well as the promoter itselfwas used as a probe. As shown in Fig. 4 and as summarized

ComA (gg) 30 2010 5 1 - - --

ComA-PO4(@g) - - - - - 30 25 20 15 10 5 2.51

f f"'~-

_ q~ 4 -

FIG. 3. Gel retardation of a 32P-labeled srfA promoter fragmentby ComA. For each lane, the total amount of phosphorylated(ComA-P04) or unphosphorylated (ComA) protein used is indi-cated. The binding and gel-running conditions are described in thetext. Also, as noted in the text, the ComA-P04 is not all phosphor-ylated.

in Fig. 5, several distinct regions were protected in both thetop and bottom strands. The protected regions (from posi-tions -122 to -105 and from positions -78 to -61 in the topstrand and from positions -116 to -105 and from positions-68 to -54 in the bottom strand) include the previouslyidentified ComA boxes (13). These results confirm thatComA recognizes specific sites upstream from the srfApromoter and that these sites correspond to those predictedfrom the comparison of sequences of ComA-regulatedgenes, as well as from genetic studies (10, 13, 15). Moreover,the protection from DNase I digestion afforded by ComAbinding is much weaker when the protein is not phosphory-lated, confirming that ComA-DNA complex formation isenhanced by phosphorylation of ComA. In addition, andunexpectedly, protection was noted on both strands over-lapping the -35 sequence of srfA. This sequence includes aCGG triplet which resembles a half-ComA box. Finally,protection is also apparent on the bottom strand at positions-76 to -79.

It is also noticeable that the presence of ComA-PO4 causespartial protection from DNase I attack and hypersensitivityto DNase I at certain positions; these are indicated in Fig. 5and in most cases can also be recognized in the photographin Fig. 4. A likely explanation for this hypersensitivity isdistortion of the DNA helix due to binding of ComA-P04,resulting in the increased accessibility to DNase I attack ofbonds exposed in the minor groove. The protection notedabove at positions -76 to -79 on the bottom strand may alsobe related to this putative distortion.

DISCUSSION

Sequence analysis has identified two palindromic seg-ments located upstream of the srfA promoter which resem-ble each other and which also show similarity to sequenceslocated upstream of thegsiA and degQ promoters (9, 10, 13).Mutational analysis has confirmed the importance of thesetwo so-called ComA boxes in the transcriptional regulationof srfA (13, 15), and ComA box mutations also preventinduction ofgsiA (11). The present study was undertaken totest the hypothesis that phosphorylated ComA binds directlyto the ComA boxes upstream of the srfA promoter, presum-ably to turn on transcription.Both gel retardation and DNase I footprinting experiments

confirm that this binding occurs in a sequence-specific man-ner. The binding affinity of ComA-PO4 to the srfA promoterfragment is higher than that of unphosphorylated ComA. It

J. BACTERIOL.

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

ComA BINDING TO THE srfA PROMOTER REGION 3185

TOP STRAND

5u DNAse 10 u DNAse

- 50 - < - - 100 50 <

+ +

50 - - O 100 50 - - - c

BOTTOM STRAND

a) 5u DNAse0

CQ - 50 - <aQ +

50 - (D

10 u DNAse

- - 10050 - <+

10050 - (D

!ww t:v -I v t 's s. 9 w ww 1 w

_ __ of.*::: * 4xi X L:w +

ro:ji - wxr.a::: z w, Or*

* * An*- * F3E =

ar _H***o - * F st ** * * 2 ';; <6

*:* .7.s

_ hiw _As,

.. + 'I IN8gX 75 A, * X+ *|*a ,,

'pii

FIG. 4. Footprinting analysis of ComA binding to the srfA promoter region. DNase I protection footprinting was carried out as described

in the text. The left and right halves of the autoradiograph show results obtained with the labeled top (coding) strand and the bottom strand,

respectively. The sequences of the strands are shown in Fig. 5. For each strand, results obtained with two amounts of DNase I (5 and 10 U)and with various input amounts (in micrograms) of ConA_32pO4 and ComA are shown. As noted in the text, the ConmA-P04 is not all

phosphorylated. Also included are lanes containing untreated probe (no DNase I), G+A sequencing ladders, and probe plus DNase I (no

ComA). The positions of the protected regions are marked with brackets. 1, 2, and -35, box 1, box 2, and promoter -35 region protected

sequences, respectively. DNase I-hypersensitive and partially protected sites are indicated by arrowheads (large and small for strongly and

weakly hypersensitive positions, respectively).

appears that our ComA preparation was about 50% pure andthat about 30% of the ComA was phosphorylated in our

experiments. This would suggest a very approximate valueof 1 to 2 1±M for the apparent ComA-PO4 concentrationgiving half-maximum protection from DNase I. The weakerbut detectable binding of unphosphorylated ComA is inaccord with in vivo data showing that overexpression ofComA can activate srfA transcription in the absence ofComP (25).As predicted, the regions protected by ComA from DNase

I cleavage include the ComA boxes. We did not use subsat-urating concentrations of ComA for our footprinting exper-iments to measure the relative affinities for boxes 1 and 2.Nevertheless, it appears that the unphosphorylated ComAexhibited limited protection of the box 1 region but little or

no evident protection of box 2 (Fig. 4). PhosphorylatedComA may also bind with higher affinity to box 1 than to box2. This agrees with the genetic analysis (15), which suggests

BOX-1

GGAAATGATTGCGGCgATCCCGCAAAAAATATTGCTGTAAATAAACTGCCTTTACTAACGCCGTAGGGCGTTTTTTATAACGACATTTATTTGAC* I % 1 *+ *A4o

-120 -110 -100 -90 -80

BOX-2.*._.~ ~ ~ ~ ~ ~r V v 35

GAATCTTZQQGCATCC TGAAACTTTTCACCCATTTTTCGGTGATAAAAACACTTAGAAAGCCGTAGGGCGTACTTTGAAAAGTGGGTAAAAAGCCACTATTTTTGT

-70 -60 -50 -40 -30

FIG. 5. Summary of footprinting results. The sequence foundupstream of the srfA promoter (including the -35 portion of thepromoter) is shown (13). The positions of the two ComA boxesinferred from sequence comparisons and from genetic data (13) are

indicated by heavy horizontal bars. The protected regions on eachstrand are indicated by brackets, and the DNase I-hypersensitiveand partially protected bonds are indicated by arrows (large for

strongly hypersensitive and small for weakly hypersensitive bonds).

a)-00~Q.E ComA-PO

4MO ComA

VOL. 175, 1993

I:

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3186 ROGGIANI AND DUBNAU

that ComA binds to box 1 of the srfA promoter with highaffinity and then binds to the lower-affinity box 2 site,possibly with the formation of a DNA loop (see below). Infact, modification of box 2 to make it identical to box 1renders the upstream copy of box 1 dispensable for srfAtranscription (15). Binding to box 2 is required for theactivation of transcription, presumably by establishing anappropriate contact between ComA and RNA polymerase.Box 2 is located within the upper limit for the range ofdistances from the + 1 nucleotide residue typically occupiedby proximal binding sites for positively acting transcriptionfactors, many of which have been previously shown to actby contacting RNA polymerase (1). The DNase I-hypersen-sitive and partially protected sites evident in Fig. 4 andsummarized in Fig. 5, as well as the protection on the bottomstrand of residues -76 to -79, are consistent with thisnotion ofDNA looping, since they may indicate the presenceof localized distortions of the helix due to bending.The centers of dyad symmetry of the two ComA boxes are

separated by 45 bp. A looping model would require thatComA bind to the same face of the helix at the 2 positions,in this case separated by about four turns, suggesting that inthis case the periodicity of the DNA deviates slightly fromthe canonical 10.5 bp per turn. Nakano and Zuber (15) haveshown that the insertion of 5 bp between boxes 1 and 2results in a decrease in srfA promoter activity, whereas theinsertion of 10 bp results in a twofold enhancement ofactivity. This experiment is consistent with a looping modeland illustrates the importance of the proper relative presen-tation of the ComA boxes.Unexpected protected sequences on both strands overlap-

ping the -35 promoter region of srfA were noted (Fig. 4and 5). This protected region bears some resemblance toa half-ComA box. We do not know whether this resem-blance, and hence the protection, is fortuitous or reflects anin vivo regulatory interaction. The position of this bindingsite is within the range of overlap observed for proximalsites involved in repression as well as for sites involvedin activation (1). Alternatively, high levels of ComA-PO4may exert a negative effect on srfA transcription. In thisconnection, note that ComA closely resembles the responseregulator DegU throughout its length, including within thepresumed carboxy-terminal DNA binding domain. Hy-perphosphorylated DegU exerts a negative effect on thedevelopment of competence (8, 19), acting to reduce thetranscription of srfA (5). It is possible that a high concentra-tion of DegU-P04 permits binding to srfA, resulting inrepression.

It appears that ComA can indeed bind to the expectedsequences upstream from the srfA promoter, that this bind-ing is enhanced by the phosphorylation of ComA, and that itprobably proceeds by cooperative interactions with boxes 1and 2, concomitant with DNA bending. These conclusionsestablish an important link in the circuitry of competenceinduction. However, it has not yet been determined whetherComA binding is sufficient to activate the transcription ofsrfA, and further work will be required in order to addressthis point.

ACKNOWLEDGMENTS

We thank members of our laboratory and I. Smith for valuablediscussions. We also thank M. Nakano and P. Zuber for usefuldiscussions and for communicating results prior to publication.

This work was supported by NIH grant A110311.

REFERENCES1. Collado-Vides, J., B. Magasanik, and J. D. Gralla. 1991. Control

site location and transcriptional regulation in Escherichia coli.Microbiol. Rev. 55:371-394.

2. Dubnau, D. 1991. Genetic competence in Bacillus subtilis.Microbiol. Rev. 55:395-424.

3. Feng, J., M. R. Atkinson, W. McCleary, J. B. Stock, B. L.Wanner, and A. J. Ninfa. 1992. Role of phosphorylated meta-bolic intermediates in the regulation of glutamine synthetasesynthesis in Escherichia coli. J. Bacteriol. 174:6061-6070.

4. Grossman, A. D., K. Ireton, E. F. Hoff, J. R. LeDeaux, D. Z.Rudner, R. Magnuson, and K. A. Hicks. 1991. Signal transduc-tion and the initiation of sporulation in Bacillus subtilis. Semin.Dev. Biol. 2:31-36.

5. Hahn, J., and D. Dubnau. 1991. Growth stage signal transduc-tion and the requirements for srfA induction in development ofcompetence. J. Bacteriol. 173:7275-7282.

6. Jaacks, K. J., J. Healy, R. Losick, and A. D. Grossman. 1989.Identification and characterization of genes controlled by thesporulation-regulatory gene spoOH in Bacillus subtilis. J. Bac-teriol. 171:4121-4129.

7. Lukat, G. S., W. R. McCleary, A. M. Stock, and J. B. Stock.1992. Phosphorylation of bacterial response regulator proteinsby low molecular weight phospho-donors. Proc. Natl. Acad.Sci. USA 89:718-722.

8. Msadek, T., F. Kunst, D. Henner, A. Klier, G. Rapoport, and R.Dedonder. 1990. Signal transduction pathway controlling syn-thesis of a class of degradative enzymes in Bacillus subtilis:expression of the regulatory genes and analysis of mutations indegS and degU. J. Bacteriol. 172:824-834.

9. Msadek, T., F. Kunst, A. Klier, and G. Rapoport. 1991. DegS-DegU and ComP-ComA modulator-effector pairs control ex-pression of the Bacillus subtilis pleiotropic regulatory genedegQ. J. Bacteriol. 173:2366-2377.

10. Mueller, J. P., G. Bukusoglu, and A. L. Sonenshein. 1992.Transcriptional regulation of Bacillus subtilis glucose starva-tion-inducible genes: control ofgsi4 by the ComP-ComA signaltransduction system. J. Bacteriol. 174:4361-4373.

11. Mueller, J. P., and A. L. Sonenshein (Tufts University). 1993.Personal communication.

12. Nakano, M. M., R. Magnuson, A. Myers, J. Curry, A. D.Grossman, and P. Zuber. 1991. srfA is an operon requiredfor surfactin production, competence development, and effi-cient sporulation in Bacillus subtilis. J. Bacteriol. 173:1770-1778.

13. Nakano, M. M., L. Xia, and P. Zuber. 1991. Transcriptioninitiation region of the srfA operon, which is controlled by thecomP-comA signal transduction system in Bacillus subtilis. J.Bacteriol. 173:5487-5493.

14. Nakano, M. M., and P. Zuber. 1991. The primary role of ComAin establishment of the competent state in Bacillus subtilis is toactivate the expression of srfA. J. Bacteriol. 173:7269-7274.

15. Nakano, M. M., and P. Zuber. 1993. Mutational analysis of theregulatory region of the srfA operon in Bacillus subtilis. J.Bacteriol. 175:3188-3191.

16. Rudner, D. Z., J. R. LeDeaux, K. Ireton, and A. D. Grossman.1991. The spoOK locus of Bacillus subtilis is homologous to theoligopeptide permease locus and is required for sporulation andcompetence. J. Bacteriol. 173:1388-1398.

17. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual. Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, N.Y.

18. Stadtman, E. R. 1957. Preparation and assay of acetyl phos-phate. Methods Enzymol. 3:228-231.

19. Steinmetz, M., F. Kunst, and R. Dedonder. 1976. Mapping ofmutations affecting synthesis of exocellular enzymes in Bacillussubtilis. Mol. Gen. Genet. 148:281-285.

20. Stock, J. B., A. J. Ninfa, and A. M. Stock. 1989. Proteinphosphorylation and regulation of adaptive responses in bacte-ria. Microbiol. Rev. 53:450-490.

21. Studier, F. W., A. H. Rosenberg, J. J. Dunn, and J. W.Dubendorf. 1990. Use of 17 RNA polymerase to direct expres-sion of cloned genes. Methods Enzymol. 185:60-89.

J. BACTERIOL.

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

ComA BINDING TO THE srfA PROMOTER REGION 3187

22. van Sinderen, D., S. Withoff, H. Boels, and G. Venema. 1990.Isolation and characterization of comL, a transcription unitinvolved in competence development of Bacillus subtilis. Mol.Gen. Genet. 224:396-404.

23. Weinrauch, Y., and D. Dubnau. Unpublished data.24. Weinrauch, Y., N. Guillen, and D. Dubnau. 1989. Sequence and

transcription mapping of Bacillus subtilis competence genes

comB and comA, one of which is related to a family of bacterialregulatory determinants. J. Bacteriol. 171:5362-5375.

25. Weinrauch, Y., R. Penchev, E. Dubnau, I. Smith, and D.Dubnau. 1990. A Bacillus subtilis regulatory gene product forgenetic competence and sporulation resembles sensor proteinmembers of the bacterial two-component signal-transductionsystems. Genes Dev. 4:860-872.

VOL. 175, 1993

on February 19, 2018 by guest

http://jb.asm.org/

Dow

nloaded from