resd signal transduction regulator of aerobic respiration in bacillus subtilis: ctaa promoter...

Molecular Microbiology (2000) 37(5), 1208±1219

ResD signal transduction regulator of aerobicrespiration in Bacillus subtilis: ctaA promoter regulation

Xiaohui Zhang and F. Marion Hulett*

Laboratory for Molecular Biology, Department of

Biological Sciences University of Illinois at Chicago, 900

S. Ashland Avenue (M/C 567), Chicago, IL 60607, USA.

Summary

A two-component signal transduction system com-

posed of a sensor kinase, ResE, and a response

regulator, ResD, encoded by resD and resE genes of

the res operon (resABCDE), has a regulatory role in

both aerobic and anaerobic respiration. In terms of

aerobic respiration, resD functions upstream of ctaA,

a gene required for haem A biogenesis and hence

for the synthesis of haem A-containing cytochrome

terminal oxidases. Although ResD is probably a

transcription factor, there was no direct evidence

that ResD protein, either phosphorylated or unphos-

phorylated, interacts directly with regulatory regions

of ResD-controlled genes. Here, we report the over-

expression and purification of ResD and ResE and

their role in gene activation. ResD can be phosphory-

lated by ResE in vitro and is a monomer in solution in

either the phosphorylated or unphosphorylated state.

The binding activity of ResD to the ctaA promoter was

examined by gel shift assays and DNase I footprinting

assays. DNase I footprinting showed both unphos-

phorylated and phosphorylated ResD binding to

the ctaA promoter and showed that there are three

binding sites (A1, A2 and A3), two (A1 and A2)

upstream of the 235 promoter region and one (A3)

downstream of the 210 of the promoter. The role of

each site in ctaA promoter activity and ResD binding

was characterized using deletion analysis, followed

by the DNase I footprinting and in vivo transcription

assays of promoter±lacZ fusions. Our results

showed that the concentration of ResD required to

bind at each site is different and that ResD binding at

the A1 site is independent of the other two ResD

binding sites, but that the concentration of ResD,P

required to protect site A2 is reduced when site A3 is

present. In vivo transcription assays from promoter±

lacZ fusion constructs showed that DNA containing

ResD-binding site A2 was essential for promoter

activity and that promoter constructs containing both

binding sites A2 and A3 were sufficient for full

promoter activity.

Introduction

ResD and ResE were identified in the Bacillus genome

sequencing project (Sorokin et al., 1993) as genes

encoding a response regulator and sensor histidine

kinase of prokaryotic two-component signal transduction

systems. They share similarity to the OmpR family of

regulators. Within Bacillus subtilis regulators, they belong

to a small subclass (based on properties of the sensor

kinase) that also includes PhoP and PhoR, which regulate

the phosphate deficiency response (Pho response), and

YycF and YycG, which are essential two-component

regulators of unknown function (Fabret and Hoch, 1998;

Fabret et al., 1999).

ResD and ResE are global regulators of genes involved

in aerobic and anaerobic respiration. Originally, they were

shown to be required for transcription of the resABCDE

operon (resABC encode proteins similar to those involved

in cytochrome c biogenesis), of ctaA (responsible for

conversion of haem O to haem A) and of petCBD

(encoding subunits of the cytochrome bf complex) and

to be essential for anaerobic growth of B. subtilis in the

presence of nitrate (Sun et al., 1996). ResD and ResE are

also required for ctaB transcription (responsible for haem

B conversion to haem O) and for full induction of the Pho

response via a mechanism which involves a positive

feedback loop, resulting in the amplification of both Res

and Pho regulon induction during phosphate deprivation

Birkey et al., 1998). Genes involved in nitrate respiration

which require ResD and ResE include fnr, hmp and

nasDEF (nitrite reductase) (Nakano et al., 1996, 1998;

Nakano and Hulett, 1997).

Although ResD has been shown to be required for

transcription of the above genes in vivo, it has yet to be

shown that ResD acts directly. We chose to examine the

role of ResD in the activation of the ctaA promoter. The

ctaA gene encodes a membrane-bound protein that

catalyses the conversion of haem O to haem A (Svensson

et al., 1993, 1996; Svensson and Hederstedt, 1994) and,

as such, is essential for production of aa3 and caa3

terminal oxidases.

Q 2000 Blackwell Science Ltd

Accepted 26 June, 2000. *For correspondence. E-mail [email protected];Tel. (11) 312 996 5460; Fax (11) 312 413 2691.

Results

Overexpression and purification of ResD and ResE

ResD was overexpressed and purified for investigation of

its functions in vitro. We used IMPACT (Intein-Mediated

Purification with an Affinity Chitin-binding Tag, New

England Biolabs) as an affinity protein purification system

and obtained purified recombinant ResD protein from

Escherichia coli within 24 h by a single chromatographic

step. The ResD protein was successfully expressed in

E. coli grown at 308C, and the soluble protein obtained

from the supernatant fraction of a cell lysate after

centrifugation at 120 000 g for 1 h was applied to a chitin

affinity column. After the column was washed to remove

non-binding protein, the fusion protein was induced to

undergo an intein-mediated self-cleavage on the column

by overnight incubation at 48C in the presence of

dithiothreitol (DTT). The ResD protein was released

while the intein chitin-binding domain remained bound to

the column. The fractions of the eluted ResD exhibited

greater than 98% homogeneity, as determined by SDS±

PAGE with Coomassie brilliant blue staining (Fig. 1A,

lanes 6 and 7).

Purified ResE was necessary for phosphorylation of

ResD. As the insoluble nature of the complete ResE

protein complicated the in vitro studies, we constructed an

N-terminal truncated ResE, *ResE, in a glutathione-S-

transferase (GST) fusion vector. This construct provides a

simplified purification of the soluble form of the protein.

Having its N-terminus 226 amino acids deleted, *ResE

retains the majority of its extended cytoplasmic domain

and the complete conserved catalytic domain. The GST±

*ResE protein was successfully overexpressed in E. coli,

and it accounted for 35% of the total cellular protein

(Fig. 1B). Approximately 70% of the GST±*ResE protein

expressed at 308C was in the soluble fraction after

centrifugation at 120 000 g for 1 h. The soluble fusion

protein was applied to a glutathione agarose column and

the column was washed to remove unbound protein. The

glutathione agarose beads with a bound fusion protein

were mixed with 10 U of thrombin. After thrombin

cleavage, 95% of the truncated *ResE protein was

collected in the eluted fraction from the beads (Fig. 1B,

lanes 6 and 7).

Phosphorylation of ResD by *ResE

To determine whether the N-terminal truncated form of

ResE protein functions as an autokinase, the purified

*ResE was mixed with [g-32P]-ATP. *ResE showed

autophosphorylation (Fig. 1C, lane 1). To determine

whether ResD could be phosphorylated by *ResE,

*ResE was incubated with ResD in the presence of

[g-32P]-ATP. Figure 1C, lane 3, shows that ResD could be

phosphorylated by *ResE. Unlike *ResE, ResD was not

phosphorylated by ATP alone (Fig. 1C, lane 2). This

result indicates that ResE is a kinase capable of ResD

phosphorylation.

Fig. 1. A±C. Purification of ResD and ResE for in vitrophosphorylation studies. The fractions from each step of eachpurification were analysed by SDS±PAGE, and the gel was stainedwith Coomassie brilliant blue. The amount of protein loaded in eachlane is given below in parentheses.A. Purification of intein±ResD from E. coli. Arrows mark thecalculated molecular weight for the ResD±intein±CBD fusionprotein and the position of the purified ResD protein. Lanes: 1,prestained protein standards (Bio-Rad); 2, sonication lysate fromuninduced cells (30 mg protein); 3, sonication lysate from inducedcells (35 mg protein); 4, clarified crude extract from induced cells(35 mg protein); 5, column flow-through (35 mg protein); 6 and 7,fractions of eluted by cleavage buffer with DTT (2 and 5 mg proteinrespectively).B. Purification of overexpressed GST±*ResE from E. coli. Arrowsmark the calculated molecular weight for the GST±ResE fusionprotein and the position of the purified ResE. Lanes: 1, prestainedprotein standards (Bio-Rad); 2, sonication lysate from uninducedcells (5 mg protein); 3, sonication lysate from induced cells (10 mgprotein); 4, pellet fraction (10 mg protein); 5, clarified crude extractfrom induced cells (35 mg protein); 6 and 7, elution collected afterthrombin cleavage on glutathione agarose beads (2.8 and 7 mgprotein respectively).C. Phosphorylation of ResD by *ResE in the presence of ATP.ResD (9.6 mM) and/or *ResE (5.9 mM) were incubated with 1 mCiof [g-32P]-ATP at room temperature for 15 min. The samples wereanalysed by SDS±PAGE. The gel was dried and exposed to an X-ray film. Lanes: 1, ResE incubated with [g-32P]-ATP; 2, ResDincubated with [g-32P]-ATP; 3, ResD incubated with ResE and[g-32P]-ATP.

ResD phosphorylation, oligomerization and DNA binding 1209

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1208±1219

The ResD is a monomer, unphosphorylated or

phosphorylated

Certain response regulators are known to oligomerize

upon phosphorylation. We used gel filtration to determine

whether ResD oligomerizes upon phosphorylation. Puri-

fied ResD was considered unphosphorylated as the

conditions for overexpression of the protein do not favour

the formation of ResD,P. ResD and ResD phos-

phorylated by GST±*ResE were individually applied to a

Superdex 75 gel filtration column. Unphosphorylated

ResD was eluted at a relative molecular mass of

27 000, a value in good agreement with that calculated

from the deduced amino acid sequence of resD. The

elution pattern of phosphorylated ResD from the column

showed that the protein peak for the unphosphorylated

ResD and the major radioactivity peak for ResD,P were

coincident at the same elution volume, indicating that

ResD,P, like unphosphorylated ResD, is a monomer in

solution.

Binding of the ctaA promoter by ResD and ResD,P

We have shown that the ctaA promoter is subject in vivo

to regulation by ResD, the response regulator of the

ResD±ResE two-component system (Sun et al., 1996).

To determine whether ResD binds to the ctaA promoter in

vitro, ResD and ResD,P were incubated with the ctaA

promoter for gel shift assays. With increasing ResD

concentrations, we observed that ctaA promoter mobility

decreased in the presence or absence of ATP (all

reactions contained *ResE), suggesting that both unphos-

phorylated ResD and phosphorylated ResD bind to the

promoter (Fig. 2A and B). A slight shift was apparent with

either ResD or ResD,P at 74 nM protein and became

more pronounced at increased ResD concentrations. The

gel shift pattern was complex, suggesting that additional

ResD molecules bound the ctaA promoter at higher

protein concentrations and that several DNA protein

complexes could be formed.

ResD or ResD,P binds to three sites within the ctaA

promoter

The binding specificities of ResD and ResD,P to the ctaA

Fig. 2. A and B. Gel shift assay of the ctaA promoter by (A) ResDor (B) ResD,P. ResD (74 nM, 148 nM, 296 nM, 593 nM, 1.18 mM,2.37 mM and 4.74 mM) was incubated with (A) or without (B) *ResEin the presence of ATP. The 406 bp ctaAB end-labelled promoterprobe was added. After the binding reaction, the samples wereloaded onto a 5% native polyacrylamide gel to separate the freeDNA from the DNA±protein complex. The amount of ResD proteinadded to the reaction is indicated above each lane.

Fig. 3. DNase I footprinting assays of the ctaA promoter bound byResD and ResD,P. Various amounts of ResD incubated with*ResE in the presence or absence of ATP were mixed with the406 bp ctaAB promoter probe, and DNase I footprintingexperiments were performed on both the coding strand and non-coding strands (see Experimental procedures). The concentrationsof ResD or ResD,P used in each reaction are, from left to right,0 nM, 9 nM, 18 nM, 37 nM, 74 nM, 148 nM, 296 nM, 594 nM and0 nM. Increasing concentrations of ResD or ResD,P are indicatedby a wedge. F, ResD-free lane; G, G sequencing reaction. Thebold lines represent the ResD and ResD,P binding regions. Thehypersensitive site is indicated by an arrowhead.

1210 X. Zhang and F. M. Hulett


promoter were examined by DNase I footprinting using

both strands of the ctaA promoter region from pES68. The

data in Fig. 3 show that regions of the ctaA promoter from

2209 to 2179, from 2108 to 255 and from 22 to 143

(named A1, A2 and A3 respectively) on the coding strand

and essentially the same three sites on the non-coding

strand were protected by ResD or ResD,P. The three

binding sites differ from each other in their discrimination

between ResD and ResD,P for binding and in the actual

concentration of ResD or ResD,P required for protection.

Binding site A1, from bp 2209 to 2179, had the highest

ResD affinity, with DNA protection starting at an 18 nM

concentration of ResD or ResD,P. At binding site A2,

from bp 255 to 2108, ResD,P showed strong protection

at 296 nM and the presence of a hypersensitive site on

the non-coding strand. Unphosphorylated ResD showed

only limited protection at 296 nM, with better protection at

594 nM ResD concentrations at site A2. The A3 binding at

site, from bp 22 to 143, showed nearly complete

protection by 296 nM ResD,P, but less protection by

unphosphorylated ResD at concentrations twofold higher.

The DNase I footprint titration data for each of the three

ResD binding sites were fitted to the Langmuir isotherm

equation (Brenowitz et al., 1996) that was transformed to

the equation Y � 100%/(1 1 KD/[P]) to obtain directly the

equilibrium dissociation constant KD (Table 1). The ResD

binding data are plotted (Fig. 4) as per cent base

protection vs. concentration of ResD or ResD,P ([P]).

Exact sequences of the DNA protected by ResD on each

DNA strand are shown in Fig. 5A.

ResD binding at any one of the three ctaA promoter

binding sites is independent of the other sites; binding site

two is essential for promoter activity whereas binding site

three is required for full promoter activity

Five deletion constructions of the ctaA promoter were

made (Fig. 5B) to assess the role of each ResD binding

site on (i) ResD binding at another site and (ii) promoter

activation. Using JH642 genomic DNA as a template,

primers (Fig. 5A and Experimental procedures) were

used to amplify DNA fragments which contained only a

single ResD binding site or two of the three binding sites

in the ctaA promoter (Fig. 5B). Each of the five amplified

fragments were cloned into pCR2.1(Invitrogen). Each

deleted promoter was used in footprinting studies and

each was cloned into pDH32 constructing a promoter±

lacZ fusion (Table 2) which was integrated in single copy

into the chromosome of JH642 (or the isogeneic resD

mutant strain MH5260; Sun et al., 1996) at the amyE

locus via homologous recombination (Table 2). Figure 6

shows a footprint for ResD,P binding at any one of the

three sites alone similar to that observed at that site in the

complete promoter (Fig. 3), albeit the protection by

Res,P at a concentration of 296 nM was not as

extensive for A2 or A3 as in the complete promoter

(Fig. 3). The promoter activity from the lacZ fusion

constructs containing either only A2 (bp 2152 to 111)

in B. subtilis MH5915 or only A3 (bp 250 to 172) in B.

subtilis MH5916 was compared with the promoter activity

Table 1. ResD or ResD,P dissociation constants for sites A1, A2and A3.

KD ^ s.e.

Site Unphosphorylated Phosphorylated

A1 12.6 ^ 3.3 8.9 ^ 1.4A2 230 ^ 40.1 200 ^ 32.3A3 180 ^ 29.9 47 ^ 7.6

Fig. 4. ResD and ResD,P protection at binding sites A1, A2 and A3 differ in the concentration of protein required for binding site protectionand in the discrimination between ResD and ResD,P. Per cent protection vs. nM protein for ResD (open symbols) and ResD,P (closedsymbols).



Fig. 5. ResD binding sites on the ctaAB intercistronic promoter sequence and diagrams of various promoter subclones.A. The intercistronic ctaAB promoter sequence and 5 0 coding sequence of for CtaA showing the ResD and ResD,P binding sites. The codingand non-coding sequences of the fragment are shown. The binding sites for both ResD and ResD,P are represented by bold solid linesbelow (coding) the sequence. The transcriptional start site is shown in bold and is labelled. Primers used for amplification of the completectaAB promoter and the various ctaAB promoter deletions are shown by arrows (pXH43, FMH385 and FMH372; pXH28, FMH254 andFMH370; pXH37, FMH385 and FMH384; pXH38, FMH383 and FMH372; pXH41, FMH254 and FMH384; pES68, FMH254 and FMH255).B. Diagrams of the various ctaAB promoter-amplified fragments, the pCR2.1 plasmids containing each fragment and the name of the straincontaining a lacZ±promoter fusion construct of each fragment. The ResD binding sites are marked as black boxes. The 5 0 and 3 0 bp of eachpromoter fragment are labelled relative to the transcription start site 11.



from the complete ctaA promoter (bp 2335 to 172) in

MH5909. Strains containing all promoter fusions were

grown in 2� SSG containing 0.5% glucose and b-

galactosidase-specific activity was assayed (Fig. 7).

Strain MH5915 containing the promoter fusion with only

ResD binding site A2 showed less than 50% of the activity

expressed by the full-length promoter whereas the strain

containing only binding site A3 in the promoter±lacZ

fusion was not functional, although that promoter fusion

retained both the 210 and 235 regions (Fig. 5A and B).

To determine whether adjacent DNA containing either the

A1 (5 0) or A3 (3 0) binding site might enhance promoter

expression from the promoter fusion containing only ResD

binding site A2, b-galactosidase expression from MH5919

Table 2. Bacterial strains and plasmids.

B. subtilis strains Genotype or antibiotic resistance Source or reference

JH642 pheA1 trpC2 J. A. HochMH5906 pheA1 trpC2 amyE::pXH31 Cmr This studyMH5909 pheA1 trpC2 amyE::pXH34 Cmr This studyMH5915 pheA1 trpC2 amyE::pXH39 Cmr This studyMH5916 pheA1 trpC2 amyE::pXH40 Cmr This studyMH5919 pheA1 trpC2 amyE::pXH42 Cmr This studyMH5921 pheA1 trpC2 amyE::pXH44 Cmr This study

Plasmids Description Source

pCR2.1 Vector for cloning PCR products (Ampr Kanr) InvitrogenpDH32 Vector for construction of promoter± lacZ fusion (Ampr Cmr) Shimotsu (1986)pES68 Ampr Kanr (4.3 kb) ctaA (A1 1 A2 1 A3) promoter (406 bp) in pCR2.1 This studypXH6 Ampr Kanr (4.6 kb) resD gene (720 bp) in pCR2.1 This studypXH22 Ampr (8.0 kb) resD gene (720 bp) in pTYB2 This studypXH1 Ampr Kanr (5.0 kb) resE gene (1.14 kb) in pCR2.1 This StudypXH3 Ampr (6.1 kb) resE gene (1.14 kb) in pGEX-2T This studypXH28 Ampr Kanr (4.2 kb) ctaA promoter (A1) (227 bp) in pCR2.1 This studypXH37 Ampr Kanr (4.0 kb) ctaA promoter (A2) (163 bp) in pCR2.1 This studypXH38 Ampr Kanr (3.9 kb) ctaA promoter (A3) (122 bp) in pCR2.1 This studypXH41 Ampr Kanr (4.2 kb) ctaA promoter (A1 1 A2) (346 bp) in pCR2.1 This studypXH43 Ampr Kanr (4.1 kb) ctaA promoter (A2 1 A3) (224 bp) in pCR2.1 This studypXH34 Ampr Cmr (10.3 kb) ctaA±lacZ fusion (A1 1 A2 1 A3)(406) in pDH32 This studypXH31 Ampr Cmr (10.1 kb) ctaA±lacZ fusion (A1) in pDH32 This studypXH39 Ampr Cmr (10.0 kb) ctaA±lacZ fusion (A2) in pDH32 This studypXH40 Ampr Cmr (10.0 kb) ctaA±lacZ fusion (A3) in pDH32 This studypXH42 Ampr Cmr (10.2 kb) ctaA±lacZ fusion (A1 1 A2) in pDH32 This studypXH44 Ampr Cmr (10.1 kb) ctaA±lacZ fusion (A2 1 A3) in pDH32 This study

Fig. 6. ResD or ResD,P protection ofpromoter fragments containing a single ResDbinding site. The probe for binding site A1was the insert in pXH28, for A2 the insert inpXH37 and for A3 the insert in pXH38. Onlythe coding strand is labelled. Theconcentrations of ResD used in each reactionare, from left to right, 0 nM, 37 nM, 74 nM,296 nM, 1.18 mM, 2.37 mM and 0 nM. Theamounts of *ResE and ATP in each reactionwere the same as those used for footprintingboth the coding and non-coding strands of thewild-type ctaA promoter shown in Fig. 3. Thelabels are also the same as those in Fig. 3.



and MH5921 was assayed (Fig. 7). Expression from the

promoter fusion with sites A1 and A2 (MH5919) was not

appreciably different from the fusion containing only A2

(MH5915) whereas expression from the promoter fusion

with sites A2 and A3 (MH5921) was similar to the full-

length promoter (MH5909), suggesting that site A3 was

required for full promoter expression. The in vivo

expression data from the five promoter fusion strains

were consistent during three independent experiments.

No expression was obtained from any of these five

promoter fusions in isogeneic strains that carried a null

mutation in resD (data not shown).

The ResD protected region on DNA fragments contain-

ing binding site A2 with either site A1 or site A3 showed no

change in the size of the protected DNA areas at any site

(Fig. 8). The only change in the pattern noted was the

appearance of a small unprotected region apparent within

ResD binding site A1, perhaps indicating some conforma-

tional change in this area when site 3 was deleted from

the whole promoter (Fig. 8). However, the enhanced

affinity of ResD for site A2 by the presence of site A3 on

the same DNA fragment (Fig. 8) correlates with the

enhanced promoter activity from the promoter-containing

sites A2 and A3 compared with the lacZ promoter fusion

containing only site A2 (Fig. 7).

Discussion

ctaA is one of several genes involved in aerobic or

anaerobic respiration which require ResD for expression.

To determine whether the activation of these genes by

ResD involves direct promoter binding and to explore the

mechanism by which ResD regulates gene transcription,

we used biochemical analysis to characterize the ResD

protein with respect to phosphorylation, oligomerization

and DNA-binding activity. The ResD used in this study is

identical to the wild-type protein except for an additional

glycine residue at the C-terminal. The ResE protein used

in these studies is a soluble protein as the sequences

encoding the N-terminal hydrophobic domains and a

portion of the C2 domain were deleted (236 amino acids

of the 589 total amino acids) from the resE±GST fusion

construction.

ResD binding site A2 (bp 2108 to 255) and A3 (bp 22 to

143) are required for full ctaA promoter activity

The ResD target DNA used in these studies included the

ctaA promoter and the divergent promoter for ctaB. We

Fig. 7. Expression of the ctaA promoter is dependent on ResDbinding sites A2 and A3. Parental, resD mutant strains containingthe various ctaA±lacZ promoter fusions were grown in 2� SSG, aspreviously described (Sun et al., 1996). Cell growth (OD540) and b-galactosidase activity were analysed during the 11 h growth period.Growth (filled symbols) and b-galactosidase (open symbols) fromstrains (X) MH5909 (bp 2335 to 172; A1 1 A2 1 A3); (B),MH5915 (bp 2152 to 111; A2); (O) MH5916 (bp 250 to 172; A3);(V) MH5921 (bp 2152 to 172; A2 1 A3) and (P) MH5919 (bp2335 to 111; A1 1 A2).

Fig. 8. ResD and ResD,P DNase I footprinting protection of ctaApromoter fragments containing two of the three ResD binding sites,A1 1 A2 or A2 1 A3. The ctaA promoter inserts in pXH41 andpXH43 were used as probes. Only the coding strand was labelled.Labels and the amounts of ResD, *ResE and ATP in each reactionwere the same as those used for footprinting assays of ctaApromoter fragments containing a single ResD binding site (Fig. 6).



had previously shown that in vivo transcription of ctaA

required ResD (Sun et al., 1996). Recently, it has been

shown that the ctaB promoter requires ResD for post-

exponential transcription (Liu and Taber, 1998). Three

binding sites for ResD were located upstream of ctaA.

Deletion of the most 5 0 site did not influence the

expression of ctaA in vivo, suggesting that site A1 may

be involved in regulation of ctaB. When DNA containing

site A1 and A2 was deleted, there was no ctaA in vivo

expression, although the promoter fusion contained the

235 and 210 regions as well as binding site A3. This

phenotype, which matched that of a resD null mutant

strain exactly, indicates that DNA containing the ResD

binding site A2 is essential for ctaA expression in vivo.

Further, we have shown that DNA between A1 and A2

has little importance for ctaA promoter activity as

promoter fusion constructs starting at the 5 0 end of the

A2 site (bp 2108 to 172) retained in vivo promoter activity

similar to that shown in Fig. 7 for the complete promoter

fusion in MH5909 (X. Zhang and F. M. Hulett, unpub-

lished). 3 0 deletions of the complete promoter, removing

DNA 3 0 of 111 and thus part of site A3 (bp 22 to 143),

caused a 2.5- to threefold reduction in ctaA in vivo

expression under the conditions tested. Thus, the A3

region is necessary for full promoter activity.

All promoter fusions showed induction at approximately

the same post-exponential stage of growth followed by a

decline in promoter activity after 4±5 h, except in the

fusion that contained only the A2 binding site (strain

MH5915; Fig. 7) in which expression failed to decline.

This observation was reproducible but is not understood.

The decrease in ctaA promoter activity correlates with a

decrease in resA transcription (Sun et al., 1996),

suggesting that decreasing ResD±ResE protein concen-

trations may account for a ctaA transcription turn off.

However, this does not explain the prolonged transcription

of the A2 promoter fusion.

A consensus recognition sequence for ResD within the

ResD binding sites is proposed

Certain members of the OmpR family of response

regulators, including Pho regulon regulators PhoP of B.

subtilis (Liu and Hulett, 1998; Eder et al., 1999) or PhoB of

E. coli (Kasahara et al., 1991; Makino et al., 1991; Kim

et al., 1993), bind highly conserved DNA sequences

TT(A/T/C)ACA N4±5 TT(A/T/C)ACA and TGTCATA(A/

T)A(T/A)TGTCA(C/T), respectively, such that analysis of

binding sites from multiple promoters predict a reliable

binding consensus sequence (Makino et al., 1996; Eder

et al., 1999). The binding of PhoB to the bases of the

predicted sequence (TGTCA) has been demonstrated by

nuclear magnetic resonance (NMR) structural analysis of

the PhoB/DNA complex (Okamura et al., 2000). In

contrast to PhoP or PhoB, other OmpR class regulators,

such as OmpR of E. coli, show only weak similarity

between the DNA sequences bound by the protein, a fact

which led to proposals of different consensus sequence

alignments (Harlocker et al., 1995; Huang and Igo, 1996).

In the case of OmpR, DNA affinity cleavage analysis was

required to construct a DNA/OmpR interaction model.

These data provided evidence that repeated 6 bp

sequences, spaced 4 bp apart, each bind one OmpR

molecule (Harrison-McMonagle et al., 1999). Both PhoB

and OmpR bind to the target DNA in tandem. Considering

the results from PhoP, PhoB and OmpR DNA-binding

studies collectively, we might expect to find a consensus

of 5 or 6 bp repeats separated by 4 or 5 bases within the

ResD binding sequences. Reviewing the three ResD

binding sites in the ctaA promoter and those determined

elsewhere (Nakano et al., 2000) for anaerobically induced

genes requiring ResD, a 9 bp sequence, TTTGTGAAT, is

observed in all promoters. This sequence is repeated in

the ctaA3 binding site (6/9 and 7/9 matches respectively),

the ctaA1 binding site (8/9 matches each), the nasD

binding site (6/9 matches each) and the hmp site (6/9 and

7/9 respectively). Half-sites (a single 9 bp consensus) are

found in the ctaA2 ResD binding site (8/9), the fnr binding

site (8/9 matches on the complementary strand) and in

the resA binding site (6/9 matches) (X. Zhang and F. M.

Hulett, unpublished). Preliminary evidence has suggested

that certain base pairs within this sequence are important

for fnr promoter activity. The relevance of this putative

ResD binding consensus sequence to the ResD activation

of these various promoters remains to be tested.

ResD is a monomer and binds target DNA independent of

phosphorylation

Activation of response regulators is generally believed

to be indirectly correlated with phosphorylation of the

receiver domain relieving a mutual functional inhibition of

the N- and C-terminal domains, allowing oligomerization

and/or binding of target DNA by the C-terminal domain

(Fiedler and Weiss, 1995). Indeed the crystal structure of

NarL supports this idea as the N-terminal regulatory

domain and the C-terminal DNA binding domain are

positioned such that DNA binding to the recognition helix

would be precluded (Baikalov et al., 1996, 1998).

Although the structure for only the DNA-binding domain

of the OmpR has been solved, it was determined that the

N-terminus of OmpR could not be positioned with respect

to the C-terminus in a manner similar to those domains

in NarL (Martinez-Hackert and Stock, 1997). Still, it

was concluded based on biochemical characterization

of OmpR that the unphosphorylated N-terminal domain

interferes with DNA binding, but that the surfaces

responsible for inhibition are unknown. For PhoB of



E. coli, another OmpR family response regulator, it has

been shown that deletion of the C-terminal domain mimics

phosphorylation activation of the N-terminus such that

the N-terminal domain dimerizes, a dimer for which the

structure and the interaction surfaces are known (Sola

et al., 1999). Studies reported here show that ResD is a

monomer in solution and is able to bind target DNA either

in the phosphorylated or unphosphorylated state, albeit

the phosphorylated form usually binds at somewhat lower

protein concentrations. Our data suggest that ResD is a

member of a third class of response regulators in which

the N-terminus and the C-terminus function indepen-

dently. Similar results were shown for a second B. subtilis

OmpR family response regulator which has significant

identity in protein sequence with ResD,PhoP; except in

the case of PhoP, the dimeric state was independent of

phosphorylation and phosphorylation did not result in

further oligomerization nor was it essential for PhoP

binding to target DNA (Liu and Hulett, 1997). Other

response regulators which both bind DNA without

phosphorylation and do not change oligomeric state

upon phosphorylation include E. coli UphA (Dahl et al.,

1997) and Bordetella pertussis BvgA (Boucher et al., 1997).

For this class of proteins, phosphorylation may function

only to change the conformation of the protein, thereby

effecting interaction with target DNA and/or the RNA

polymerase and resulting in activation of transcription.

Experimental procedures

Bacterial strains and plasmids

E. coli DH5a was used as the host for plasmid constructions.E. coli BL21 (DE3) (Novagen) served as the host foroverexpressing the ResD and ResE protein. The resD genewas amplified from B. subtilis JH642 chromosomal DNA byPCR using primers FMH238 (5 0-CAT4426ATGGACCAAAC-GAACGAAAC4445-3 0) at the 5 0 end of the gene and FMH356(5 0-TTCCCGGG5146TTCAGCGCCGACCTCAA5132-3 0) at the3 0 end of the gene (the superscript numbers indicate the basepair number in GenBank accession no. L09228). The PCRproduct was cloned into pCR2.1 (Invitrogen) to constructpXH6. The resD gene was released from pXH6 by NdeI andSmaI digestion and was cloned into the NdeI and SmaI sitesof pTYB2 (New England) downstream of the T7 promoter,yielding pXH22. Thus, resD was inserted into a multiplecloning site of pTYB2 to create a gene fusion between the 3 0

bp of resD and the 5 0 bp of the gene encoding the intein. TheDNA encoding a small 5 kDa chitin binding domain fromBacillus circulans has been added to the 3 0 end of the inteingene for affinity purification of the three-part fusion. Expres-sion of the fusion construct is controlled by an IPTG-inducibleT7 promoter.

The resE gene was amplified from B. subtilis JH642chromosomal DNA by PCR using primers FMH268 (5 0-GGATCC5826GGTGAACTGGCGACC5841-3 0) at the 5 0 end ofthe gene and FMH284 (5 0-TTGAATTC6971AGTCAGCGGGTG

TTTC6954-3 0) at the 3 0 end of the gene. The PCR productwas cloned into pCR2.1 (Invitrogen) to construct pXH1. ThePCR product was digested with BamHI and EcoRI andligated to pGEX-2T (Pharmacia Biotech) and digested withthe same enzymes, yielding pXH3. All clones generated weresequenced to determine PCR accuracy.

Plasmid pES68, which contains the intercistronic ctaABsequence plus part of the 5 0 coding sequence for CtaA, wasconstructed by amplifying a 406 bp fragment, using JH642DNA as template and primers FMH254 (5 0-415TTAGCCA-TAACGTAAACCTCC385-3 0) and FMH255 (5 0-9ACAAATGTCGTCAGAACACC28-3). The amplified product was cloned intopCR2.1 and sequenced (the superscript numbers indicate thebase pair number in GenBank accession no. X54140).Plasmids containing deletions of the ctaA promoter weremade using the same method and the primers whosesequences are identified in Fig. 5A. We constructed pXH43(primers FMH853 and FMH372), pXH28 (primers FMH254and FMH370), pXH37 (primers FMH385 and FMH384),pXH38 (primers FMH383 and FMH372) and pXH41 (primersFMH254 and FMH384). Primers FMH372, FMH384 andFMH370 each contained at BamHI site added at the 5 0 end,GGATCC, which is not homologous to adjacent DNA in thectaA promoter and is not indicated in Fig. 5A.

ctaA±lacZ transcriptional promoter fusions were con-structed in pDH32, which has unique EcoRI and BamHIsites upstream of the promoterless lacZ to orientate correctlythe promoter DNA fragment. ctaA promoter fragments inpCR2.1 which were derived from PCR amplification with thecoding region proximal primer containing a BamHI site weredigested with EcoRI and BamHI to release the promoterfragment which was ligated into pDH32 digested with thesame enzymes. Thus, the EcoRI±BamHI promoter fragmentfrom pXH41 cloned into pDH32 yielded pXH42, from pXH28cloned into pDH32 yielded pXH31, from pXH38 cloned intopDH32 yielded pXH40, from pXH37 cloned into pDH32yielded pXH39, and from pXH43 cloned into pDH32 yieldedpXH44. The complete ctaA promoter in pES68 was liberatedby EcoRI digestion and cloned into the EcoRI site in pDH32to construct pXH34. The promoter orientation in pXH34 wasdetermined by restriction digestion. Strains containing singlecopy promoter±lacZ fusions were constructed by transform-ing JH642 with linearized plasmid such that the promoterfusion was integrated into the chromosome by doublecrossover homologous recombination at the amyE locuswith the amyE front and amyE back on the pDH32 plasmid.Transformants were selected on chloramphenicol and werescreened for the amyE minus phenotype. Transformation ofJH642 with pXH34 produced strain MH5909 (A1,A2,A3), withpXH42 produced MH5919 (A1,A2), with pXH31 producedMH5906 (A1), with pXH40 produced MH5916 (A3), withpXH39 produced MH5915 (A2) and with pXH44 producedMH5921 (A2,A3).

Overexpression and purification of ResD and ResE

Protein expression. E. coli BL21 (DE3) cells transformed withpXH22 or pXH3 were incubated overnight at 308C in Luria±Bertani (LB) medium containing penicillin (150 mg ml21) andwere then inoculated into 2 l of the same medium at a ratio of1:100. The cells were grown at 308C until the optical density



at 600 nm (OD600) of the culture reached 0.6. Then, 1 mMIPTG was added to the culture, and growth was continued foranother 3 h. Cells were harvested by centrifugation at 48Cand washed with cell lysis and column buffer (20 mM sodiumHEPES, 500 mM NaCl, 0.1 mM EDTA, 0.1% Triton X-100,pH 8) for ResD. The ResE-containing cells were washed withP buffer (50 mM HEPES, 50 mM KCl, 5 mM MgCl2, pH 8.0).The cell pellet was stored at 2708C.

Cell lysis. The cell pellets were suspended on ice in 40 ml oftheir respective buffer, column buffer or P buffer, containing1 mM phenylmethylsulphonyl fluoride (PMSF) and wereimmediately subjected to sonication. After a 5 min sonicationon ice at an output of 100 W, the cell lysate was centrifugedat 120 000 g for 1 h at 48C. The supernatant fraction wasfiltered through a 0.45 mm membrane.

ResD purification. The lysed cell supernatant fraction wasapplied to a 1.5 ml Chitin Bead affinity column which wasattached to a Waters 650E FPLC (fast protein liquidchromatography) system. The column was washed withcolumn buffer until the OD280 of the eluate was less than0.02, then the column was washed with three volumes ofcleavage buffer (with DTT) (20 mM HEPES, 50 mM NaCl,0.1 mM EDTA, 30 mM DTT, pH 8). The column flow wasstopped and the column was left at 48C overnight. TheResD protein was eluted using cleavage buffer (withoutDTT).

ResE purification. The supernatant fraction was applied to a2 ml glutathione±agarose (Pharmacia Biotech) affinity col-umn equilibrated with P buffer. The column was washed withP buffer until the OD280 of the eluate was less than 0.02. Thebound GST±ResE fusion protein was cleaved using P buffercontaining 20% (w/v) glycerine and 20 U of thrombin (Sigma)and was shaken at 48C overnight; the eluted ResE wascollected in the flow-through by centrifugation. The proteinwas aliquoted and stored at 2708C.

Phosphorylation of ResD by ResE

Phosphorylation assays were carried out at room tempera-ture in the presence of 5 mCi of [g-32P]-ATP (specific activity6000 Ci mmol21; Amersham Biotech) in P-buffer for 20 min.The reactions were initiated by the addition of [g-32P]-ATPinto a 20 ml reaction mixture, terminated by the addition of anequal volume of 6� sodium dodecyl sulphate (SDS) samplebuffer, and subjected to SDS±PAGE on 10% polyacrylamidegels (Laemmli, 1970). Labelled proteins were detected byautoradiography of dried gels with X-ray film (Fuji MedicalX-Ray Film).

Determination of oligomeric state of ResD and ResD,P

by gel filtration

For these studies, ResD was phosphorylated by GST±ResEbound to glutathione beads and separated from the GST±ResE by a procedure described previously for phosphory-lation of PhoP by GST±PhoR (Liu and Hulett, 1997). ResDor ResD,P were individually loaded onto an HR10/30

Superdex 75 column (Pharmacia), pre-equilibrated with50 mM tris-HCl (pH 7.5), 100 mM KCl and 4 mM DTT, andthen the protein was eluted at 0.5 ml min21 and collected in400 ml fractions. The protein molecular weight standards(MW-GF-200 kit; Sigma) were applied to the column underthe same conditions, and the elution volume and molecularweight of each standard were used to generate a standardcurve to determine the relative molecular mass of the ResDprotein.

Gel shift assays

To make probes of the ctaAB promoter for gel shift assays,we used pES68, which contains the ctaAB intercistronicregion and the 5 0 coding region of ctaA. pES68 was digestedwith HindIII and end-labelled with Klenow fragment in thepresence of [a-32P]-dATP. The inserts were released by XhoIdigestion. The probes were isolated as described by Manzaraet al. (1991). The DNA probe was further purified usingElutip-d column (Schleicher and Schuell), as instructed by themanufacturer. In each reaction, ResE (1.6 mM) was incu-bated with 74 nM, 148 nM, 296 nM, 593 nM, 1.18 mM,2.37 mM and 4.74 mM of ResD in the presence of 5 mMATP at room temperature for 15 min in binding buffer [50 mMpiperazine-N,N 0-bis(2-ethane-sulphonic acid (PIPES; pH 6.1),50 mM NaCl, 1 mM EDTA, 10% glycerol, 4 mM DTT, 4 mMMgCl2]. The probe (10 000 c.p.m.) was added to the reaction,and the incubation was continued for another 30 min. Thesamples were loaded onto a 5% native polyacrylamide gelmade in 1� Tris-borate-EDTA (Sambrook, 1989). The gelwas run at 48C for 2 h, dried by vacuum and exposed to X-rayfilm.

DNase I footprinting of the ctaAB promoter

The coding strand of each promoter fragment was labelled asdescribed above as the complete ctaA promoter cloned inpCR2.1 (pES68), and the deleted ctaA promoters cloned inpCR2.1 (plasmids pXH43, pXH28, pXH37, pXH38 andpXH41) are in same orientation. To label the non-codingstrand, pES68 was digested by XhoI and was labelled withKlenow fragment in the presence of [a-32P]-dATP and[a-32P]-dCTP. The plasmid was then digested by HindIII torelease the insert. Purification of each probe was carried outas described above. The DNA-binding reaction describedabove was used but with 20 000 c.p.m. of probe. Theamounts of ResD used in the reactions with the completepromoter probes from pES68 varied (0, 9 nM, 18 nM, 37 nM,74 nM, 148 nM, 296 nM and 594 nM). Reactions containingdeleted ctaA promoter probes contained ResD concentra-tions of 0, 37 nM, 74 nM, 296 nM, 1.18 mM or 2.37 mM. Afterthe 30 min incubation at room temperature, 3 ml of DNase I(0.9 U) in 5 mM MgCl2 and 5 mM CaCl2 was added to eachreaction mixture, and digestion was conducted for 40 s for theResD-containing samples and 20 s for ResD-free samples.The reactions were stopped, and the samples were preparedas described by Manzara et al. (1991). The samples were runon a 6% polyacrylamide gel containing 7 M urea in 1� Tris-borate-EDTA for 2 h at 65 mA. The gel was then dried andexposed to an X-ray film with an intensifying screen at 2708C.



Analysis of DNase I footprint data for calculation of the

ResD and ResD,P dissociation constant at each binding

site

The protocol of Brenowitz et al. (1996) for quantification ofprotein-binding equilibria was followed. The autoradiogram(data shown in Fig. 3) was scanned using a HP Scanjet 3Cscanner with a transparency adapter and the ResD orResD,P protection quantified using Bio-Rad QUANTITY ONE

4.1.0 graphics software. All data were fitted using the non-linear least squares fitting routine in the program TABLE-CURVE

2D (SPSS Scientific).

Growth conditions and b -galactosidase activity

The cells were grown in modified Schaeffer's sporulationmedium (SSM) (Leighton and Doi, 1971) and supplementedwith 0.5% glucose (2� SSG). A single colony of strainscontaining ctaA±lacZ promoter fusions was used to inoculate5 ml LB starter cultures which were grown for several hoursand used to inoculate 200 ml 2� SSG and incubatedovernight. The overnight cultures were used to inoculate400 ml SSG to approximately 0.2 OD540. Growth wasmeasured at OD540 and b-galactosidase specific activitywas assayed every hour for 11 h. b-Galactosidase activitywas detected using the method of Ferrari et al. (1986). Oneunit was defined as 0.33 mM o-nitrophenol produced min21 at378C. The specific activity was calculated as activity permilligram of protein.

Acknowledgements

This work was supported by NIH grant GM33471 to F.M.H.We thank Andrew Mesecar for assistance with KD determina-tions. We thank Wei Liu and Elena Sarkova for the gelfiltration data.

References

Baikalov, I., Schroder, I., Kaczor-Grzeskowiak, M.,Grzeskowiak, K., Gunsalus, R.P., and Dickerson, R.E.(1996) Structure of the Escherichia coli response regulatorNarL. Biochemistry 35: 11053±11061.

Baikalov, I., Schroder, I., Kaczor-Grzeskowiak, M., Cascio,D., Gunsalus, R.P., and Dickerson, R.E. (1998) NarLdimerization? Suggestive evidence from a new crystalform. Biochemistry 37: 3665±3676.

Birkey, S.M., Liu, W., Zhang, X., Duggan, M.F., and Hulett,F.M. (1998) Pho signal transduction network reveals directtranscriptional regulation of one two-component system byanother two-component regulator: Bacillus subtilis PhoPdirectly regulates production of ResD. Mol Microbiol 30:943±953.

Boucher, P.E., Murakami, K., Ishihama, A., and Stibitz, S.(1997) Nature of DNA binding and RNA polymeraseinteraction of the Bordetella pertussis BvgA transcrip-tional activator at the fha promoter. J Bacteriol 179:1755±1763.

Brenowitz, M., Senear, D., and Kinston, R. (eds). (1996)

DNase I footprint analysis of protein-DNA binding. NewYork: John Wiley and Sons.

Dahl, J., Wei, B., and Kadner, R. (1997) Protein phosphory-lation affects binding of the Escherichia coli transcriptionactivator UhpA to the uhpT promoter. J Biol Chem 2:1910±1919.

Eder, S., Liu, W., and Hulett, F.M. (1999) Mutational analysisof the phoD promoter in Bacillus subtilis: implications forPhoP binding and promoter activation of Pho regulonpromoters. J Bacteriol 181: 2017±2025.

Fabret, C., and Hoch, J.A. (1998) A two-component signaltransduction system essential for growth of Bacillussubtilis: implications for anti-infective therapy. J Bacteriol180: 6375±6383.

Fabret, C., Feher, V.A., and Hoch, J.A. (1999) Two-component signal transduction in Bacillus subtilis: howone organism sees its world. J Bacteriol 181: 1975±1983.

Ferrari, E., Howard, S., and Hoch, J.A. (1986) Effect of stage0 sporulation mutants on subtilisin expression. J Baceriol170: 173±179.

Fiedler, U., and Weiss, V. (1995) A common switch inactivation of the response regulators NtrC and PhoB:phosphorylation induces dimerization of the receivermodules. EMBO J 14: 3696±3705.

Harlocker, S.L., Bergstrom, L., and Inouye, M. (1995)Tandem binding of six OmpR proteins to the ompFupstream regulatory sequence of Escherichia coli. J BiolChem 270: 26849±26856.

Harrison-McMonagle, P., Denissova, N., Martinez-Hackert,E., Ebright, R.H., and Stock, A.M. (1999) Orientation ofOmpR monomers within an OmpR: DNA complex deter-mined by DNA affinity cleaving. J Mol Biol 285: 555±566.

Huang, K.J., and Igo, M.M. (1996) Identification of the basesin the ompF regulatory region, which interact with thetranscription factor OmpR. J Mol Biol 262: 615±628.

Kasahara, M., Makino, K., Amemura, M., Nakata, A., andShinagawa, H. (1991) Dual regulation of the ugp operon byphosphate and carbon starvation at two interspacedpromoters. J Bacteriol 173: 549±558.

Kim, S.K., Makino, K., Amemura, M., Shinagawa, H., andNakata, A. (1993) Molecular analysis of the phoH gene,belonging to the phosphate regulon in Escherichia coli.J Bacteriol 175: 1316±1324.

Laemmli, U.K. (1970) Cleavage of structural proteins duringthe assembly of the head of bacteriophage T4. Nature 227:680±685.

Leighton, T.J., and Doi, R.H. (1971) The stability ofmessenger ribonucleic acid during sporulation in Bacillussubtilis. J Biol Chem 246: 3189±3195.

Liu, W., and Hulett, F.M. (1997) Bacillus subtilis PhoP bindsto the phoB tandem promoter exclusively within thephosphate starvation-inducible promoter. J Bacteriol 179:6302±6310.

Liu, W., and Hulett, F.M. (1998) Comparison of PhoP bindingto the tuaA promoter with PhoP binding to other Pho-regulon promoters establishes a Bacillus subtilis Pho corebinding site. Microbiology 144: 1443±1450.

Liu, X., and Taber, H. (1998) Catabolite regulation of theBacillus subtilis ctaBCDEF gene cluster. J Bacteriol 180:6154±6163.

Makino, K., Kim, S.K., Shinagawa, H., Amemura, M., and



Nakata, A. (1991) Molecular analysis of the cryptic andfunctional phn operons for phosphonate use in Escherichiacoli K-12. J Bacteriol 173: 2665±2612.

Makino, K., Amemura, M., Kawamoto, T., Kimura, S.,Shinagawa, H., Nakata, A., and Suzuki, M. (1996) DNAbinding of PhoB and its interaction with RNA polymerase.J Mol Biol 259: 15±26.

Manzara, T., Carrasco, P., and Gruissem, W. (1991)Developmental and organ-specific changes in promoterDNA±protein interactions in the tomato rbcS gene family.Plant Cell 3: 1305±1316.

Martinez-Hackert, E., and Stock, A.M. (1997) The DNA-binding domain of OmpR: crystal structures of a wingedhelix transcription factor. Structure 5: 109±124.

Nakano, M.M., and Hulett, F.M. (1997) Adaptation ofBacillus subtilis to oxygen limitation. FEMS MicrobiolLett 157: 1±7.

Nakano, M.M., Zuber, P., Glaser, P., Danchin, A., and Hulett,F.M. (1996) Two-component regulatory proteins ResD±ResE are required for transcriptional activation of fnr uponoxygen limitation in Bacillus subtilis. J Bacteriol 178: 3796±3802.

Nakano, M.M., Hoffmann, T., Zhu, Y., and Jahn, D. (1998)Nitrogen and oxygen regulation of Bacillus subtilis nasDEFencoding NADH-dependent nitrite reductase by TnrA andResDE. J Bacteriol 180: 5344±5350.

Nakano, M.M., Zhu, Y., LaCelle, M., Zhang, X., and Hulett,F.M. (2000) Interaction of ResD with regulatory regions ofanaerobically induced genes in Bacillus subtilis. MolMicrobiol 37: 1198±1207.

Okamura, H., Hanaoka, S., Nagadoi, A., Makino, K., andNishimura, Y. (2000) Structural comparison of the PhoB

and OmpR DNA-binding/transactivation domains and thearrangement of PhoB Molecules on the phosphate box.J Mol Biol 295: 1225±1236.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989).Molecular Cloning. Cold Spring Harbor, NY: Cold SpringHarbor Laboratory Press.

Shimotsu, H., and Henner, D.J. (1986) Construction of asingle-copy integration vector and its use in analysis ofregulation of the trp operon of Bacillus subtilis. Gene 43:85±94.

Sola, M., Gomis-Ruth, F.X., Serrano, L., Gonzalez, A., andColl, M. (1999) Three-dimensional crystal structure of thetranscription factor PhoB receiver domain. J Mol Biol 285:675±687.

Sorokin, A., Zumstein, E., Azevedo, V., Ehrlich, S.D., andSerror, P. (1993) The organization of the Bacillus subtilis168 chromosome region between the spoVA and serAgenetic loci, based on sequence data. Mol Microbiol 10:385±395.

Sun, G., Sharkova, E., Chesnut, R., Birkey, S., Duggan, M.F.,Sorokin, A., et al. (1996) Regulators of aerobic andanaerobic respiration in Bacillus subtilis. J Bacteriol 178:1374±1385.

Svensson, B., and Hederstedt, L. (1994) Bacillus subtilisCtaA is a heme-containing membrane protein involved inheme A biosynthesis. J Bacteriol 176: 6663±6671.

Svensson, B., Lubben, M., and Hederstedt, L. (1993) Bacillussubtilis CtaA and CtaB function in haem A biosynthesis.Mol Microbiol 10: 193±201.

Svensson, B., Andersson, K.K., and Hederstedt, L. (1996)Low-spin heme A in the heme A biosynthetic protein CtaAfrom Bacillus subtilis. Eur J Biochem 238: 287±295.



resd signal transduction regulator of aerobic respiration in bacillus subtilis: ctaa promoter...

Documents