Review Cell Stress & Chaperones (2003) 8 (3), 207217

Q Cell Stress Society International 2003

Article no. csac. 2003.435

The Bacillus subtilis heat

shock stimulon
Wolfgang Schumann

Institute of Genetics, University of Bayreuth, D-95440 Bayreuth, Germany

Abstract All organisms respond to a sudden increase in temperature by the so-called heat shock response. This
response results in the induction of a subset of genes, designated heat shock genes coding for heat shock proteins,
which allow the cell to cope with the stress regimen. Research carried out during the last 10 years with eubacteria
has revealed that the heat shock genes of a given species fall into different classes (regulons), where each class is
regulated by a different transcriptional regulator, which could be an alternative sigma factor, a transcriptional activator,
or a transcriptional repressor. All regulons of a single species constitute the heat shock stimulon. In Bacillus subtilis,
more than 200 genes representing over 7% of the transcriptionally active genes are induced at least 3-fold in response
to a heat shock. This response becomes apparent within the first minute after exposure to heat stress, is transient,
and is coordinated by at least 5 transcriptional regulator proteins, including 2 repressors, an alternate sigma-factor,
and a 2-component signal transduction system. A detailed analysis of the regulation of all known heat shock genes
has shown that they belong to at least 6 regulons that together comprise the B subtilis heat shock stimulon. Potential
thermosensors are discussed in this article.

INTRODUCTION ognize and bind nonnative polypeptide chains in collab-

When cells of bacteria or any other organism are sub-
jected to a sudden temperature upshock, a group of genes
is transiently induced, resulting in the synthesis of so-
called heat shock proteins (Hsps). The primary structure
of most of these Hsps appears to be highly conserved
during evolution, suggesting that they exert similar func-
tions in all organisms. Most Hsps belong to either of 2
classes, namely molecular chaperones and adenosine tri-
phosphate (ATP)dependent proteases. Whereas chaper-
ones ensure that polypeptides will fold or assemble prop-
erly in the cell (Georgopoulos and Welch 1993), proteases
will degrade those misfolded proteins, which are unable
to refold into their native 3-dimensional structure (Got-
tesman 1996). In many cases, chaperones and proteases
work together in the degradation of nonnative proteins
(Gottesman et al 1997). The major problem caused for any
cell by a sudden heat shock is the immediate appearance
of denatured and misfolded proteins, often collectively
designated as nonnative proteins, which tend to aggre-
gate. This will be prevented by chaperones, which rec-
Received 28 January 2003; Revised 14 May 2003; Accepted 14 May 2003.
Correspondence to: Wolfgang Schumann, Tel: 49 921-552708; Fax: 49
921-552710; E-mail: wschumann@uni-bayreuth.de.
oration with ATP-dependent proteases that degrade those
denatured proteins unable to adopt their native confor-
mation.
All heat shock genes are expressed at ambient temper-
atures, most of them at a low level, and they are tran-
siently induced after heat challenge. Induction of heat
shock genes occurs primarily at the level of transcription,
and extensive studies carried out during the past 10 years
has shown that the heat shock genes of a single eubac-
terial species are organized into several regulons, consti-
tuting the heat shock stimulon. The Escherichia coli heat
shock stimulon is composed of 2 regulons, the s32 and
the sE regulons, and the psp operon, which is under the
positive control of PspF and s54 (for a recent review, see
Yura et al 2000). In both Streptomyces sp. and Bradyrhizo-
bium japonicum, 3 regulons have been described so far
(Narberhaus 1999; Servant and Mazodier 2001). In Bacillus
subtilis, the genetic model organism of the gram-positive
eubacteria, the heat shock stimulon comprises at least 6
classes of heat shock genes, where 2 classes are controlled
by 2 transcriptional repressors, 1 class by an alternative
sigma factor, a fourth class by a still unknown transcrip-
tional activator, the fifth by a 2-component signal trans-
duction system, and there are approximately 80 addition-

207
208 Schumann
Fig 1. The HrcA regulon. This regulon consists of the 2 operons
groE and dnaK. Both operons are expressed from sA-dependent pro-
moters. Although the groE operon is transcribed into 1 bicistronic
messenger ribonucleic acid the amount of which increases after heat
shock, the heptacistronic dnaK operon is under the control of 2 pro-
moters. Low level of transcription initiates at the 2 promoters PA1 and
PA2 in the absence of heat stress, resulting in a full-length transcript
and a second transcript encompassing the 4 distal genes. After a
temperature upshift, only PA1 becomes activated. Both operons are
under the negative control of the HrcA repressor binding to 2 in-
verted repeat structures indicated by the arrows. Hairpin structures
indicate 1 processing site (between hrcA and grpE), 3 transcription
termination sites (between dnaK and dnaJ, after orf50 and after
groEL), and 3 potential degradation sites (within the 3 distal genes
of the dnaK operon).
al heat shock genes constituting Class VI where we do
not yet know how they are regulated. All B subtilis heat
shock genes are expressed at a low level at 378C and tran-
siently induced under heat shock conditions (sudden tem-
perature increase to 48508C). Any regulation mechanism
has to explain 3 facts: (1) the low level of transcription in
the absence of heat stress, (2) the rapid increase in the
amount of heat shock proteins that normally occurs with-
in 23 minutes, and (3) the slow shut-off after about 10
minutes, even if the cells are further incubated at the high
temperature. The purpose of this article will be to review
our current understanding on the regulation of the 5
known classes of heat shock genes and to discuss the con-
sequences of this fragmentation.
CLASS I HEAT SHOCK GENES: THE HRCA
REGULON
Class I heat shock genes consist of 2 operons only, the
heptacistronic dnaK and the bicistronic groE operon (Li
and Wong 1992; Schmidt et al 1992; Homuth et al 1997)
shown in Figure 1. Both operons are preceded by a sA-
type promoter (sA is the housekeeping sigma factor) and
a perfect inverted repeat of 9 bp separated by a 9-bp spac-
er with the deoxyribonucleic acid (DNA) sequence TTA-
GCACTC-N9-GAGTGCTAA. In the case of the groE pro-
moter, it was confirmed that sequences within the 210
and 235 regions of this promoter constitute the key ele-
ments recognized by the sA-containing ribonucleic acid
(RNA) polymerase, when 1 invariant position in each re-
gion was mutated independently. These mutations could
be suppressed by mutant forms of the sA factor, illus-
Cell Stress & Chaperones (2003) 8 (3), 207217
trating that the groE promoter is transcribed by the sA-
containing RNA polymerase at both 378C and 488C (Yuan
and Wong 1995a). The heptacistronic dnaK operon con-
sists of the genes hrcA (coding for the transcriptional re-
pressor of both operons, see below), followed by the
genes grpE, dnaK, and dnaJ constituting the DnaK chap-
erone machine, and the 3 open reading frames orf35
(yqeT), orf28 (yqeU), and orf50 (yqeV). The deduced amino
acid sequence of orf35 exhibits significant homology to
the ribosomal protein L11 methyltransferase coded for by
the prmA gene of E coli (Vanet et al 1993). Although the
function of orf28 remains elusive, the deduced amino acid
sequence of orf50 reveals homology to a lipoate synthase
involved in lipoate biosynthesis. Transcriptional analyses
of both operons have shown that whereas the groE operon
is transcribed into a bicistronic transcript, expression of
the dnaK operon starts from 2 promoters, one (PA1) pre-
cedes the whole operon, whereas the other (PA2) is located
between dnaK and dnaJ (Fig 1). Both primary transcripts
are subject to posttranscriptional regulation at a process-
ing site located between hrcA and grpE, and degradation
sites are present within the 3 distal genes (Homuth et al
1999). After heat stress, only the amount of the 2 tran-
scripts initiated at PA1 is transiently enhanced (Homuth
et al 1999).
How are these 2 operons regulated in response to heat
stress? The perfect inverted repeats located in both op-
erons between the transcriptional and translational start
sites suggested a function as a binding site for a regula-
tory protein. To prove this assumption, point mutations
were introduced separately into the left, the right, and
both arms, and these 3 mutated repeat sequences were
used to replace the wild-type sequence preceding the
dnaK operon. All 3 mutated sequences led to a high con-
stitutive expression of the dnaK but not of the groE op-
eron, already under nonheat shock conditions (Zuber and
Schumann 1994). It was concluded from this result that
this inverted repeat indeed serves as a cis-active binding
site for a negative regulator. Meanwhile, the inverted re-
peat has been identified in an increasing number of bac-
terial species, and in almost all cases, it preceded either
the dnaK or the groE operon. Because both operons code
for molecular chaperones, the acronym CIRCE was
coined, which stands for controlling inverted repeat of
chaperone expression.
The gene coding for the predicted transcriptional re-
pressor has been identified by 2 groups using different
approaches. In 1 case, cells carrying a transcriptional fu-
sion between the bgaB reporter gene encoding a heat-sta-
ble b-galactosidase and the controlling region of the groE
operon were mutagenized, plated at low temperature,
and screened for blue colonies. These experiments led to
the detection of the first gene of the dnaK operon as the
gene coding for the transcriptional repressor (Yuan and

Wong 1995b). Following another strategy, construction
and analysis of knockouts of the dnaK operon, it could be
first shown that expression of the groE operon was en-
hanced significantly in an hrcA knockout, suggesting that
either hrcA itself or a gene downstream of it codes for the
repressor (Schulz et al 1995). When hrcA was completely
deleted, this knockout caused a high constitutive expres-
sion of both class I operons in the absence of heat shock
(Schulz and Schumann 1996). These 2 experiments clearly
identified the first gene of the dnaK operon as the one
coding for the repressor protein binding to the CIRCE
element, and it was subsequently named hrcA (heat shock
regulation at CIRCE; Roberts et al 1996; Schulz and Schu-
mann 1996).
How is the activity of the repressor modulated after a
heat shock? It has to be assumed that the repressor will
dissociate immediately after heat challenge to allow for
the rapid increase in transcription of the 2 operons. Fur-
thermore, because the heat shock response is shut off 5
to 10 minutes later, the HrcA repressor must bind again
to its 2 operators. Four lines of experimental evidence
suggest that the activity of the repressor is modulated by
the GroE chaperonin system (Mogk et al 1997). (1) When
the hrcA gene was transferred to E coli, it could be tran-
siently inactivated by a heat shock comparable with the
situation in B subtilis, suggesting that all components re-
quired for its transient inactivation are present within E
coli cells. When its activity was tested in different E coli
groES and groEL temperature-sensitive mutants, the basal
activity of a reporter enzyme under control of hrcA was
significantly increased in the absence of a heat shock, the
first hint that the GroE chaperonin system might be in-
volved in modulating the activity of HrcA. (2) Depletion
of the GroES and GroEL proteins in B subtilis resulted in
a high constitutive expression of the genes of the dnaK
operon at low temperature. Under these conditions, high
amounts of seemingly inactive repressor protein are pre-
sent within the cells. (3) Overproduction of just GroEL in
B subtilis lowered both the basal level of dnaK operon
expression and its expression after heat stress. (4) Purified
HrcA repressor was almost unable to bind to its operator
site as shown in a bandshift assay. Addition of purified
GroEL significantly enhanced the amount of bound DNA.
From all these results, we developed the following work-
ing hypothesis presented in Figure 2. The HrcA repressor
is released from the ribosomes as an inactive protein (de-
fined by its inability to bind to its operator). To become
active, it has to interact with the GroEL chaperonin sys-
tem, and active HrcA is able to bind to its operators.
Upon dissociation from its DNA-binding sites, HrcA is
again present in its inactive form and needs another in-
teraction with the GroE system to become converted into
its active form. Proteins that frequently need to interact
with GroEL have been described recently in E coli (Houry
B subtilis heat shock stimulon 209
et al 1999). In the absence of heat stress, sufficient
amounts of free GroE proteins are available to keep most
HrcA molecules in their active conformation. Immediate-
ly after exposure to heat stress, GroE is titrated by non-
native proteins arising as a result of the heat shock. As a
consequence, de novosynthesized HrcA proteins and re-
pressor molecules dissociated from their operators will
stay in their inactive form, allowing the RNA polymerase
to initiate transcription at a higher rate. The more non-
native proteins are removed from the cytoplasm, the
more the GroE chaperonins will be free to convert inac-
tive HrcA into its active form, resulting in a gradual turn
off of the HrcA regulon until the default state has been
reached. An open question is whether titration of GroE
immediately after a heat shock is sufficient to explain turn
on of the HrcA regulon or whether an additional mech-
anism accounts for the rapid increase in transcription of
the dnaK and groE operon. Such a mechanism could in-
volve, for example, direct interaction of nonnative pro-
teins with HrcA. Biochemical analyses of purified HrcA
protein have been hampered by its strong tendency to
aggregate. Using a genetic screen, a mutant form of the
HrcA protein (HrcA114) with increased activity was iso-
lated under conditions of decreased GroE function
(Reischl et al 2002). DNase Ifootprinting experiments ex-
hibited full protection of the CIRCE element and neigh-
boring nucleotides. Furthermore, an in vitro binding as-
say using affinity chromatography showed direct and
specific interaction between HrcA114 and GroEL. All
these data are in full agreement with our model, claiming
that HrcA needs the GroE chaperonin system for activa-
tion. Open questions concern the domain structure of
HrcA and the definition of its active and inactive form at
the molecular level.
How widespread is the HrcA-CIRCE system among
eubacteria? So far, the HrcA gene has been detected in
52 bacterial species and the CIRCE element more than
100 times (T. Wiegert and W. Schumann, in preparation).
From all the published data, it can be concluded that this
heat shock regulon is the most widespread among all
those described to date.
CLASS II HEAT SHOCK GENES: THE sB
REGULON
This class of heat shock genes is by far the largest, with
127 members (Price et al 2001). All these genes are under
the positive control of the alternative sigma factor, sB,
discovered in 1979 as the first alternate sigma factor in
bacteria (Haldenwang and Losick 1979). These class II
heat shock genes are not only induced by the classical
heat shock regimen, such as the heat shock itself and eth-
anol, but also by salt; oxidation; desiccation or acid stress;
and starvation for oxygen, glucose, or phosphate. There-
Cell Stress & Chaperones (2003) 8 (3), 207217
210 Schumann

Fig 2. The HrcA-GroE reaction cycle. Within the cells, the HrcA repressor is present in an active and an inactive form, where the latter is
unable to bind to its operator, the CIRCE element symbolized by the 2 inverted arrows. The equilibrium between these 2 forms is influenced
by the GroE chaperonin system. In the absence of heat stress, the GroE system converts most of the HrcA molecules into the active form.
After a sudden temperature upshift, nonnative proteins titrate the GroE chaperonins, leading to an increase in the amount of inactive HrcA
repressor, leading to enhanced transcription of the groE and dnaK operons.

fore, the sB regulon codes for general stress proteins
(Hecker et al 1996) in contrast to specific heat shock pro-
teins affected only by heat stress. Although the explicit
role of most of the proteins in this regulon has not yet
been elucidated, their coordinate induction in response to
these different stress stimuli suggests that they serve a
general and nonspecific protective function allowing ad-
aptation of the cells to stress and starvation. How do these
127 proteins cope with the different stressful situations
to allow survival of B subtilis cells and even growth?
Based on their known or predicted function, these genes
have been classified into 6 groups (Price 2002), with just
a few examples mentioned here: (1) direct protection
(proteases, catalases, thioredoxin, arsenate reductase), (2)
modulation of sB-activity (antisigma and anti-antisigma
factors, phosphatases; see below), (3) regulation down-
stream of sB (responses of SOS response and class II heat
shock genes, members of the AraC, DeoR, and EbsC fam-
ily), (4) influx and efflux (many permeases, antiporters,
and symporters), (5) metabolism (glucosidase, dehydro-
genases, and pyruvate oxidase), and (6) turnover (cyste-
ine protease, ribonuclease R).
The regulation of class II heat shock genes occurs in a
rather sophisticated manner and is only partly under-
stood. An important regulatory mechanism effective in
the modulation of sB activity is the presence of 2 partner
switching modules. Each one consists of 3 partner pro-
teins in which the particular binding decision is dictated
by the phosphorylation status of one of the partners (see
below). Such a mechanism was coined originally to ac-
count for the regulation of the sporulation-specific sigma
factor sF (Alper et al 1994). As already mentioned, all
these genes are under positive control by the sB sigma
factor encoded by the seventh gene of an 8-gene operon
(Fig 3). The whole operon is preceded by a sA-type pro-
moter, ensuring the basal expression of all 8 genes (Wise
and Price 1995). In addition, expression of the last 4 genes
can become initiated at a sB-dependent promoter provid-

Cell Stress & Chaperones (2003) 8 (3), 207217


Fig 3. The sB operon. The whole octacistronic sigB operon is tran-
scribed at a basal level from the sA-type promoter PA, whereas under
stress conditions sB greatly enhances transcription of the 4 distal
genes at PB. The function of the different genes is described within
the text.
Fig 4. Model of the signal transduction network leading to the ac-
tivation of sB. Under physiological conditions, most sB molecules are
sequestered by the antisigma factor RsbW, which also acts as a
kinase to phosphorylate the anti-antisigma factor RsbV to keep it
inactive. Energy stress leads to the dephosphorylation of the anti-
antisigma factor RsbV by the RsbP phosphatase, whereas environ-
mental (physical) stress activates the RsbU phosphatase. Dephos-
phorylated RsbV will attack the RsbW-sB complex, leading to the
release of the sigma factor, which is now able to transcribe all the
genes of the sB regulon. For further details, see text.
ed active sB is present, resulting in autoregulation of sB
(Kalman et al 1990). In the absence of any stress factor
(under physiological conditions), sB is sequestered by an
antisigma factor encoded by the rsbW gene (for regulation
of sigma B), thereby preventing its interaction with the
RNA polymerase core enzyme (Fig 4). Besides binding to
sB, the RsbW protein also possesses a serine kinase ac-
tivity that is responsible for the phosphorylation of the
RsbV anti-antisigma factor, thereby keeping this protein
in an inactive state. If B subtilis cells are exposed to one
of the stress factors mentioned above, the phosphate is
removed from RsbV;P by one of 2 phosphatases (RsbP,
RsbU). Dephosphorylated RsbV attacks the sB-RsbW
complex, and partner switching of RsbW from sB to RsbV
leads to the release of sB. These 2 phosphatases respond
B subtilis heat shock stimulon 211
to 2 classes of stress: (1) energy stress caused, eg, by car-
bon, oxygen, or phosphorous starvation or the addition
of oxidative uncouplers to the growth medium leads to
the activation of the energy-responsive phosphatase RsbP
and (2) environmental stresses, including ethanol, heat,
acid, and salt stress, activate the environmental-respon-
sive phosphatase RsbU.
Next, we have to explain how the 2 phosphatases be-
come activated. The energy-responsive RsbP serine phos-
phatase is composed of at least 2 domains, a C-terminal
PP2C phosphatase domain and an N-terminal Per-Arnt-
Sim (PAS) domain (Fig 4). PAS domains have been shown
to control either the interaction with other proteins or to
sense changes in the energy level of cells in general (Tay-
lor and Zhulin 1999). RsbP seems to be a flavoprotein,
and it has been speculated that the redox state of the
chromophore bound to RsbP might be involved in mod-
ulating the activity of the PP2C-phosphatase domain (Vi-
jay et al 2000). Furthermore, it has been shown that this
PAS domain is indeed essential for the energy stress re-
sponse (mentioned in Price 2001). RsbP is complexed
with RsbQ, a positive regulator, coding for an a/btype
hydrolase, but the exact function of RsbQ during signal
transduction remains to be discovered (Brody et al 2001).
It has been concluded that the RsbP phosphatase is syn-
thesized in an inactive form and converted to an active
enzyme by RsbQ. Therefore, RsbQ directly senses energy
stress, or, alternatively, one or more as yet unknown pro-
teins transmit a signal to RsbQ upon energy stress chal-
lenge. What molecule(s) is sensed by RsbQ? Because en-
ergy stress is often accompanied by a drop in the cyto-
plasmic ATP level and a subsequent increase in adenosine
diphosphate (ADP), RsbQ could measure the amount of
ADP molecules either directly or indirectly.
Activation of the environmental-responsive PP2C-type
serine phosphatase RsbU is more complex, involving at
least 10 proteins, of which only 4 are shown in Figure 4.
In the absence of environmental stress, RsbS forms a com-
plex with RsbT, thereby preventing its activation of RsbU.
Environmental stress is sensed by RsbR (and most prob-
ably by 6 paralogs), which in turn interacts with RsbT to
activate its latent kinase activity (Gaidenko et al 1999).
Next, RsbT phosphorylates RsbS, thereby causing disso-
ciation of the RsbT-RsbS complex, followed by binding of
RsbT to RsbU, the second partner-switching module
(Yang et al 1996). Return to the default state is ensured
by the RsbX protein, another phosphatase specific for
RsbS. Dephosphorylation of RsbS favors sequestering of
RsbT by RsbS and thereby shuts off the signal transduc-
tion pathway. Although RsbR modulates the sB response
after heat shock and salt stress (Akbar et al 1997), 6 RsbR
paralogs have been identified (YkoB, YojH, YqhA, YteA,
YetI, and YezB), displaying 50% sequence identity in their
C-terminal halves, but the exact function of these proteins
Cell Stress & Chaperones (2003) 8 (3), 207217
212 Schumann
has yet to be elucidated. It is tempting to speculate that
these 7 stress transmitters might sense different types of
stresses, but overlapping activities are to be expected. It
is also tempting to speculate whether one or more of the
stress transmitters interact with the feedback phospha-
tase RsbX to activate or inactivate its enzymatic activity.
In summary, activation of the energy-responsive RsbP
phosphatase is regulated in a straightforward manner,
whereas that of the environmental-responsive RsbU
seems to involve several input regulators and a feedback
loop. Activation of sB after imposition of environmental
stress is always transient, reaching a maximum 1040
minutes after the shift, but then sB activity decreases rap-
idly to the preshift level (Boylan et al 1993; Maul et al
1995; Voelker et al 1995). The RsbX phosphatase that spe-
cifically dephosphorylates RsbS;P is part of the device
ensuring the transient nature of the induction (Voelker et
al 1997; Smirnova et al 1998).
How widespread is the sB network among eubacteria?
The same 8-gene sB operon has been described in B lich-
eniformis (Brody and Price 1998) and Listeria monocytogenes
(Becker et al 1998; Wiedmann et al 1998). Smaller sB op-
erons have been described in Staphylococcus aureus (Wu et
al 1996; Kullik and Giachino 2002), Staphylococcus epider-
mis (Knobloch et al 2001), Mycobacterium tuberculosis
(DeMaio et al 1996, 1997), and Streptomyces coelicolor (Kor-
manec 2000). Unfortunately, in the 2 latter species, the
presumed sB orthologs have been called sF and sH, re-
spectively.
CLASS III HEAT SHOCK GENES: THE CTSR
REGULON
In contrast to the sB-regulon, the CtsR regulon coding for
class III heat shock genes is composed of only 6 members,
the tetracistronic clpC operon (Kruger et al 1996, 1997)
Cell Stress & Chaperones (2003) 8 (3), 207217
Fig 5. The CtsR regulon. This regulon
consists of 3 transcriptional units, the
tetracistronic clpC operon and the 2
monocistronic clpP and clpE operons.
All 3 operons are under the negative
control of the CtsR repressor binding to
2 or more direct repeats indicated by
the arrows.
and the 2 monocistronic clpP and clpE operons (Fig 5).
Three of these genes encode Clp proteins: the negative
regulator of this regulon, the CtsR repressor (for class III
stress repressor), and the mcsA and mcsB (for modulator
of CtsR repression), involved in regulating the activity of
CtsR (see below). The genes clpC and clpE code for
ATPase subunits, and clpP codes for a proteolytic subunit
of the 2 ClpCP and potential ClpEP ATP-dependent pro-
teases. All 3 operons are preceded by 2 promoters each,
the clpC and clpP operons by an upstream sB- and a
downstream sA-dependent promoter and clpE by 2 sA-
type promoters (Fig 5). All sA-dependent promoters are
under the negative control of the CtsR repressor, where
the CtsR-binding sites usually overlap the transcriptional
start site or the 235 and 210 regions of the promoter,
suggesting that the repressor prevents binding of the
RNA polymerase EsA holoenzyme to these promoters.
CtsR has been predicted to code for a repressor protein
because it contains a classical helix-turn-helix DNA-bind-
ing motif. Inactivation of this gene led to a 10- to 20-fold
increase in the expression of the clpC operon and the 2
monocistronic clpP and clpE genes (Kruger and Hecker
1998; Derre et al 1999a). The operator sequence, which is
recognized by CtsR, has been determined as a highly con-
served heptanucleotide direct repeat located upstream of
all 3 transcriptional units: A/GGTCAAA NAN A/
GGTCAAA (Derre et al 1999a, 1999b). In summary, the
CtsR repressor regulates expression of 6 genes organized
in 3 transcriptional units.
What is known about the mechanism of heat shock reg-
ulation of class III genes? CtsR is composed of at least 3
functional domains, a dimerization domain (within the
first 24 amino acids), a helix-turn-helix domain (amino
acids 6467) responsible for DNA binding, and the central
glycin-rich region (amino acids 6467), which could be

involved in heat sensing (Derre et al 2000). The 8 genes
of the CtsR regulon are expressed at a low level at 378C
and are strongly derepressed after heat challenge. This
regulatory mechanism seems to be based on maintaining
a certain steady-state level of CtsR at 378C, followed by a
transient inactivation of the repressor upon exposure to
heat stress. The steady-state level of CtsR is controlled by
the ClpXP protease degrading superfluous molecules
(Derre et al 2000). The 2 proteins encoded by the mcsA
and mcsB genes act as modulators of the CtsR repressor.
McsA reveals similarities to zinc finger proteins that func-
tion as DNA-binding proteins or participate in the me-
diation of protein-protein interactions. McsA may assist
CtsR in adopting or stabilizing its active conformation (or
both). McsB shares similarities with arginine kinases and
is involved in inactivation of CtsR, most probably by
phosphorylation, thereby targeting the repressor for
ClpCP-mediated proteolysis (Kruger et al 2001). A data-
base search revealed that the CtsR regulatory system is
highly conserved among low G 1 C gram-positive bac-
teria, with CtsR-binding sites found mostly upstream
from clp genes (Derre et al 1999a).
CLASS IV HEAT SHOCK GENE: THE HTPG
OPERON
Only 1 class IV heat shock gene has been identified so
far, the htpG gene. This gene, which is assumed to code
for a molecular chaperone, is induced about 10-fold after
a heat shock from 378C to 488C, both at the level of tran-
scription and translation (Schulz et al 1997). htpG is dis-
pensable for growth between 378C and 488C (Versteeg et
al 1999), in agreement with data published for E coli htpG
(Bardwell and Craig 1988). The regulatory site for htpG
(GAAAGG) has been identified immediately downstream
of its sA-dependent promoter (Versteeg et al 2003). Upon
deletion of this sequence, the promoter lost heat induc-
ibility. Fusion of the regulatory sequence to the lepA pro-
moter, which is not subject to activation by heat stress,
conferred heat inducibility. We concluded from all these
results that htpG is under positive control by a transcrip-
tional activator binding downstream of the 210 region
(Versteeg et al 2003). Experiments are in progress to iden-
tify this transcriptional activator.
CLASS V HEAT SHOCK GENES: THE CSSRS
REGULON
Class V heat shock genes have been described quite re-
cently and consist of 2 members so far, htrA and htrB
(Darmon et al 2002), encoding putative membrane-an-
chored proteases. Both genes are preceded by a 210 re-
gion of sA-type promoters but lack an obvious 235 re-
gion. Instead, the control regions have a 4-foldrepeated
B subtilis heat shock stimulon 213
octameric consensus sequence (TTTTCATA) positioned
close to the 235 regions (Noone et al 2001). Both genes
are under the positive control of the CssRS 2-component
signal transduction system (for control of secretion stress
regulator and sensor), which responds not only to heat
but also to secretion stress exerted, eg, by a-amylase over-
production (Darmon et al 2002). In both cases, it can be
assumed that nonnative proteins present within or on the
outer face of the cytoplasmic membrane will lead to the
activation of the sensor kinase CssS.
The CssRS 2-component system is assumed to detect
secretion stress by sensing the accumulation of misfolded
proteins at the membranecell wall interface (Hyyrylai-
nen et al 2001). In accordance with this assumption, the
CssRS regulatory system does not respond to nonnative
proteins within the cytosol because neither htrA nor htrB
is induced by the addition of puromycin to the medium
(Noone et al 2000; Darmon et al 2002). Because puro-
mycin causes premature translation termination, we can
assume that the truncated proteins synthesized under
these conditions are degraded within the cytoplasm and
not secreted to induce class V heat shock genes. It has
been reported that the cssRS operon itself becomes heat
inducible in an htrA knockout strain but not in a wild-
type background. This finding can be explained by as-
suming that the level of nonnative proteins generated by
heat stress in a wild-type background is sufficient to stim-
ulate the expression of htrA and htrB but insufficient to
stimulate expression of cssRS (Darmon et al 2002).
CLASS VI HEAT SHOCK GENES
Class VI comprises a group of genes whose expression is
also responsive to stress, but the mechanism of induction
is not affected by any one of the previously mentioned
regulators. Therefore, these genes are controlled by one
or more unknown mechanisms. Ten class VI heat shock
genes have been reported previously, where 4 are ar-
ranged in monocistronic transcriptional units (ftsH, clpX,
ykoZ, and sacB) (Gerth et al 1996; Deuerling et al 1997;
Schumann et al 2002; Zuber et al 2001) and 6 in 2 bicis-
tronic units (lonA-orfX, ahpC-ahpF, and nfrA-ywcH) (Rieth-
dorf et al 1994; Antelmann et al 1996; Moch et al 2000).
In a recent global transcriptional analysis using the DNA
microarray technique, the group of J. D. Helmann has
considerably extended class VI heat shock genes by an-
other approximately 80 members (Helmann et al 2001),
making this the second largest class. Class VI heat shock
genes can be distinguished by 3 criteria: (1) the promoter-
type; (2) the heat-induction factor; and most importantly
(3) the stress factors besides heat that cause their induc-
tion, suggesting that more than 1 mechanism is respon-
sible for their regulation.
Cell Stress & Chaperones (2003) 8 (3), 207217
214 Schumann
CELLULAR THERMOMETERS AND
THERMOSENSORS
The study of the induction mechanisms of heat shock
genes has revealed 2 fundamentally different regulatory
principles with different outcomes, direct and indirect
heat sensing. Direct heat sensors are either RNA or pro-
tein molecules that change their conformation depending
on the temperature, whereas indirect heat sensors iden-
tified so far are molecular chaperones involved in mod-
ulating the activity of a transcriptional regulator and ti-
trated by nonnative proteins. Although expression of heat
shock genes controlled by direct sensors continues at a
high rate as long as the cells are exposed to the high
temperature (these genes are not turned off), those reg-
ulated by indirect sensors are induced transiently, imply-
ing that the regulatory mechanism must ensure a return
to the prestimulus state even if the cells are kept longer
at the high temperature. To distinguish these 2 opposing
mechanisms, I suggest designating direct sensors as ther-
mosensors and indirect sensors, which involve several
components, as cellular thermometers. With a few excep-
tions, heat shock genes are transiently induced, and these
are under the control of cellular thermometers. Ribo-
somes were suggested first in 1990 to act as such a ther-
mometer (VanBogelen and Neidhardt 1990), and in the
following year, the DnaK chaperone machine of E coli
cells was claimed to act as a thermometer (Craig and
Gross 1991; McCarty and Walker 1991). Examples for heat
shock genes being under the control of a thermosensor
are the rpoH mRNA coding for s32 (Morita et al 1999), the
ROSE elements present at the beginning of the mRNA of
some heat shock genes in Rhizobia (Nocker et al 2001),
and the RheA transcriptional repressor protein (Servant
et al 2001).
The best-studied example of indirect sensing is the
modulation of the s32 activity by the DnaK chaperone ma-
chine. In the absence of heat stress, the DnaK system will
bind s32 molecules and present them to an ATP-depen-
dent protease such as FtsH (Herman et al 1995). Follow-
ing a heat shock, the DnaK chaperone machine is titrated
by nonnative proteins, allowing s32 to escape sequestra-
tion and hydrolysis. The more nonnative proteins are re-
moved by molecular chaperones and ATP-dependent pro-
teases, the more s32 molecules will be bound by the DnaK
team and presented for proteolysis, resulting in a gradual
turn off of the heat shock response. Thermosensing in the
case of the transcripts involves secondary structures that
sequester the Shine-Dalgarno sequence and the start co-
don, impairing binding of the 30S ribosomal subunit. In-
creasing temperatures will lead to the melting of the sec-
ondary structures, thereby enhancing translation of the
transcript. In the case of the Rhe repressor protein, high
temperatures induce a conformational change and sub-
Cell Stress & Chaperones (2003) 8 (3), 207217
sequent dissociation from its operator. It has to be
stressed again that direct-sensing mechanisms do not al-
low a turnoff of the affected heat shock genes.
What is the situation in B subtilis? Are we already able
to define cellular thermometers and thermosensors? At
least, we can postulate that there is more than 1 heat-
sensing device. In the case of the HrcA regulon, it has
been suggested that the GroE chaperonin system acts as
a thermometer of heat stress (Mogk et al 1997). As out-
lined above, this system is responsible for modulating the
activity of the HrcA repressor, senses nonnative proteins,
and ensures autoregulation. Based on our current evi-
dence, only HrcA and GroE are the components of the
cellular thermometer of class I heat shock genes. The sim-
plicity of this thermometer could be the reason for the
widespread use of this device by more than 50 bacterial
species.
In the case of the sB regulon, RsbR seems to be part of
a complicated second cellular thermometer. This protein
somehow senses a temperature increase and then inter-
acts with the RsbT protein to stimulate its kinase activity.
Return to the default state involves inactivation of RsbR
by phosphorylation (the quick response?) and RsbX phos-
phatase, which demodifies RsbT (the slow response?). Be-
cause puromycin fails to induce the sB regulon (Mogk et
al 1998), this observation strongly argues against nonna-
tive proteins as inducers. Therefore, we have to assume
that RsbR, present in an inactive state in the absence of
heat stress, is able to sense some as yet unknown com-
ponent by direct interaction to be converted into its active
form. But it remains completely mystifying how RsbX
becomes activated and how RsbR and RsbX might sense
molecules different from nonnative proteins such as the
ATP or ADP concentration within the cytoplasm. In vitro
experiments should be able to answer this question.
Regulation of the CtsR regulon seems to be less com-
plicated. Here, as in the case of HrcA, the activity of a
repressor protein is modulated by heat stress. This mod-
ulation seems to involve the 2 proteins, McsA and McsB,
where McsA is required for the stabilization of the CtsR
repressor. Stabilization could involve chaperoning of CtsR
into its active conformation or preventing its proteolytic
degradation. McsB has been suggested to phosphorylate
CtsR, thereby making it prone to degradation by ClpCP
(Kruger et al 2001). If this picture is correct, McsB should
be the heat sensor that leads to its activation, again by
interaction with nonnative proteins? Here, too, it is large-
ly unknown how the return to the default state is regu-
lated. But it should involve a return of the McsB kinase
from its active to its inactive conformation.
The cellular thermometer involved in the regulation of
the htpG gene, the only member of class IV heat shock
genes so far, is completely unknown. We can only state
that nonnative proteins within the cytoplasm do not play

Table 1 The heat shock stimulon of B subtilis
Name of the
Class regulon Regulator
I HrcA HrcA repressor
II sB Sigma factor
III CtsR CtsR repressor
IV ? ?
V CssRS CssS kinase, CssR response regulator
VI ? ?
a
ATP, adenosine triphosphate.
any role. We concluded that the signal that is sensed by a
still unknown protein must arise either within the mem-
brane or outside the cell (Versteeg et al 2003). A similar
conclusion has been reached for class V heat shock genes.
Here, too, nonnative proteins within the cytoplasm are not
sensed by the CssRS 2-component signal transduction sys-
tem. Instead, it has been suggested that nonnative proteins
either stuck within the membrane or accumulating on the
outside are sensed by the CssR sensor kinase, leading to
autophosphorylation followed by transfer of the phosphate
group to the response regulator CssS, which in turn acti-
vates the gene htrA and htrB (Darmon et al 2002). Because
both genes code for proteases, it is tempting to speculate
that increased synthesis of these 2 proteases results in re-
moval of the nonnative proteins and thereby turnoff of the
expression of class V heat shock genes.
CONCLUSIONS AND PERSPECTIVES
Work carried out during the past 10 years has revealed
that the B subtilis heat shock stimulon is by far the most
complicated ever described in bacteria. It codes for more
than 200 heat shock genes that are regulated by at least
6 mechanisms summarized in Table 1. Why has B subtilis
developed such a sophisticated network of heat stress re-
gulons? One answer that comes immediately to ones
mind is that whereas all these regulons respond to heat
stress, they might react to other stress factors in a specific
manner. In favor of this assumption is the observation that
class II genes are also induced by starvation of carbon
and phosphate and by acid stress, to mention just a few,
and that class V genes are also induced by secretion
stress. B subtilis as a typical soil bacterium might be ex-
posed to more stress regimen in its natural habitat as
compared with the E coli residing in the gut, where most
of the heat shock genes are members of just 1 class, the
s32 regulon (Yura et al 2000). Although, in the case of
class I and class III, only heat has been identified as an
inducing stress factor, it can be expected that additional
stress factors specific for one of these 2 classes will be
identified in future.
Is there a cross talk between these different regulons?
A weak overlap involving class II and III has been re-
B subtilis heat shock stimulon 215
Function of major heat
shock genes
Molecular chaperones
General stress proteins
Atp-dependent proteasesa
Molecular chaperone
Membrane-anchored proteases
Diverse functions
ported. The clpC and clpP operons are both preceded,
besides their major sA-type promoter, by a sB-type (see
Fig 5). But clpC is poorly induced by oxygen and phos-
phate starvation (Kruger et al 1996), and in the case of
clpP, transcription initiated at sB is very weak, accounting
for only 25% of the total expression under stress condi-
tions (Gerth et al 1998). To conclude, a cross talk between
the different regulons does not seem to exist in B subtilis,
and there are no experimental data in favor of a master
regulator. All the regulons seem to be regulated com-
pletely independent of each other.
Work in progress deals with 2 important aspects: (1)
identification of the transcriptional regulators of class IV
and class VI heat shock genes and (2) a detailed analysis
of the different cellular thermometers. There is no indi-
cation for thermosensors because all heat shock genes
studied are turned off. The ultimate goal of the study of
the heat shock response in B subtilis will be a complete
and in-depth understanding of all the regulons, which
should allow a convincing answer to the question What
is the biological significance of splitting the heat shock
genes into so many different regulons?
ACKNOWLEDGMENTS
Financial support for this work was provided by the
Deutsche Forschungsgemeinschaft and the Fonds der
Deutschen Chemie.
REFERENCES
Akbar S, Kang CM, Gaidenko TA, Price CW. 1997 Modulator protein
RsbR regulates environmental signaling in the general stress
pathway of Bacillus subtilis. Mol Microbiol 24: 567578.
Alper S, Duncan L, Losick R. 1994. An adenosine nucleotide switch
controlling the activity of a cell type-specific transcription factor
in B. subtilis. Cell 77: 195205.
Antelmann H, Engelmann S, Schmid R, Hecker M. 1996. General
and oxidative stress responses in Bacillus subtilis: cloning, ex-
pression, and mutation of the alkyl hydroperoxide reductase
operon. J Bacteriol 178: 65716578.
Bardwell JCA, Craig EA. 1988. Ancient heat shock gene is dispens-
able. J Bacteriol 170: 29772983.
Becker LA, Cetin MS, Hutkins RW, Benson AK. 1998. Identification
of the gene encoding the alternative sigma factor sB from Lis-
Cell Stress & Chaperones (2003) 8 (3), 207217
216 Schumann
teria monocytogenes and its role in osmotolerance. J Bacteriol 182:
70837087.
Boylan SA, Redfield AR, Brody MS, Price CW. 1993. Stress-induced
activation of the sB transcription factor of Bacillus subtilis. J Bac-
teriol 175: 79317937.
Brody MS, Price CW. 1998. Bacillus licheniformis sigB operon encod-
ing the general stress transcription factor sB. Gene 212: 111118.
Brody MS, Vijay K, Price CW. 2001. Catalytic function of an a/b
hydrolase is required for energy stress activation of the sB tran-
scription factor in Bacillus subtilis. J Bacteriol 183: 64226428.
Craig EA, Gross CA. 1991. Is hsp70 the cellular thermometer? Trends
Biochem Sci 16: 135140.
Darmon E, Noone D, Masson A, Bron S, Kuipers OP, Devine KM,
Van Dijl JM. 2002. A novel class of heat and secretion stress-
responsive genes is controlled by the autoregulated CssRS two-
component system of Bacillus subtilis. J Bacteriol 184: 56615671.
DeMaio J, Zhang Y, Ko C, Bishai WR. 1997. Mycobacterium tubercu-
losis sigF is part of a gene cluster with similarities to the Bacillus
subtilis sigF and sigB operons. Tuber Lung Dis 78: 312.
DeMaio J, Zhang Y, Ko C, Young DB, Bishai WR. 1996. A stationary-
phase stress-response sigma factor from Mycobacterium tuber-
culosis. Proc Natl Acad Sci U S A 93: 27902794.
Derre I, Rapoport G, Devine K, Rose M, Msadek T. 1999b. ClpE, a
novel type of HSP100 ATPase, is part of the CtsR heat shock
regulon of Bacillus subtilis. Mol Microbiol 32: 581593.
Derre I, Rapoport G, Msadek T. 1999a. CtsR, a novel regulator of
stress and heat shock response, controls clp and molecular chap-
erone gene expression in gram-positive bacteria. Mol Microbiol
31: 117131.
Derre I, Rapoport G, Msadek T. 2000. The CtsR regulator of stress
response is active as a dimer and specifically degraded in vivo
at 378C. Mol Microbiol 38: 335347.
Deuerling E, Mogk A, Richter C, Purucker M, Schumann W. 1997.
The ftsH gene of Bacillus subtilis is involved in major cellular
processes such as sporulation, stress adaptation and secretion.
Mol Microbiol 23: 921933.
Gaidenko TA, Yang XF, Lee YM, Price CW. 1999. Threonine phos-
phorylation of modulator protein RsbR governs its ability to
regulate a serine kinase in the environmental stress signaling
pathway of Bacillus subtilis. J Mol Biol 288: 2939.
Georgopoulos C, Welch WJ. 1993. Role of the major heat shock pro-
teins as molecular chaperones. Annu Rev Cell Biol 9: 601634.
Gerth U, Kruger E, Derre I, Msadek T, Hecker M. 1998. Stress in-
duction of the Bacillus subtilis clpP gene encoding a homologue
of the proteolytic component of the Clp protease and the in-
volvement of ClpP and ClpX in stress tolerance. Mol Microbiol
28: 787802.
Gerth U, Wipat A, Harwood CR, Carter N, Emmerson PT, Hecker
M. 1996. Sequence and transcriptional analysis of clpX, a class-
III heat-shock gene of Bacillus subtilis. Gene 181: 7783.
Gottesman S. 1996. Proteases and their targets in Escherichia coli.
Annu Rev Genet 30: 465506.
Gottesman S, Wickner S, Maurizi MR. 1997. Protein quality control:
triage by chaperones and proteases. Genes Dev 11: 815823.
Haldenwang WG, Losick R. 1979. A modified RNA polymerase
transcribes a cloned gene under sporulation control in Bacillus
subtilis. Nature 282: 256260.
Hecker M, Schumann W, Volker U. 1996. Heat-shock and general
stress response in Bacillus subtilis. Mol Microbiol 19: 417428.
Helmann JD, Wu MF, Kobel PA, Gamo FJ, Wilson M, Morshedi MM,
Navre M, Paddon C. 2001. Global transcriptional response of
Bacillus subtilis to heat shock. J Bacteriol 183: 73187328.
Herman C, Thevenet D, DAri R, Bouloc P. 1995. Degradation of s32,
Cell Stress & Chaperones (2003) 8 (3), 207217
the heat shock regulator in Escherichia coli, is governed by HflB.
Proc Natl Acad Sci U S A 92: 35163520.
Homuth G, Masuda S, Mogk A, Kobayashi Y, Schumann W. 1997.
The dnaK operon of Bacillus subtilis is heptacistronic. J Bacteriol
179: 11531164.
Homuth G, Mogk A, Schumann W. 1999. Post-transcriptional reg-
ulation of the Bacillus subtilis dnaK operon. Mol Microbiol 32:
11831197.
Houry WA, Frishman D, Eckerskorn C, Lottspeich F, Hartl FU. 1999.
Identification of in vivo substrates of the chaperonin GroEL. Na-
ture 402: 147154.
Hyyrylainen HL, Bolhuis A, Darmon E, et al. 2001. A novel two-
component regulatory system in Bacillus subtilis for the survival
of severe secretion stress. Mol Microbiol 41: 11591172.
Kalman S, Duncan ML, Thomas SM, Price CW. 1990. Similar orga-
nization of the sigB and spoIIA operons encoding alternate sig-
ma factors of Bacillus subtilis RNA polymerase. J Bacteriol 172:
55755585.
Knobloch J, Bartscht K, Sabottke A, Rohde H, Fuecht H-H, Mack D.
2001. Biofilm formation of Staphylococcus epidermis depends on
functional RsbU, an activator of the sigB operon: differential
activation mechanisms due to ethanol and salt stress. J Bacteriol
183: 26242633.
Kormanec J, Sevckova B, Halgasova N, Knirschova R, Rezuchova B.
2000. Identification and transcriptional characterization of the
gene encoding the stress-response s factor sH in Streptomyces
coelicolor A3(2). FEMS Microbiol Lett 189: 3138.
Kruger E, Hecker M. 1998. The first gene of the Bacillus subtilis clpC
operon, ctsR, encodes a negative regulator of its own operon
and other class III heat shock genes. J Bacteriol 180: 66816688.
Kruger E, Msadek T, Hecker M. 1996. Alternate promoters direct
stress-induced transcription of the Bacillus subtilis clpC operon.
Mol Microbiol 20: 713724.
Kruger E, Msadek T, Ohlmeier S, Hecker M. 1997. The Bacillus sub-
tilis clpC operon encodes DNA repair and competence proteins.
Microbiology 143: 13091316.
Kruger E, Zuhlke D, Witt E, Ludwig H, Hecker M. 2001. Clp-me-
diated proteolysis in gram-positive bacteria is autoregulated by
the stability of a repressor. EMBO J 20: 852863.
Kullik I, Giachino P. 2002. The alternative sigma factor sB in Staph-
ylococcus aureus: regulation of the sigB operon in response to
growth phase and heat shock. Arch Microbiol 167: 151159.
Li M, Wong S-L. 1992. Cloning and characterization of the groESL
operon from Bacillus subtilis. J Bacteriol 174: 39813992.
Maul B, Volker U, Riethdorf S, Engelmann S, Hecker M. 1995. sB-
dependent regulation of gsiB in response to multiple stimuli in
Bacillus subtilis. Mol Gen Genet 248: 114120.
McCarty JS, Walker GC. 1991. DnaK as a thermometer: threonine-
199 is site of autophosphorylation and is critical for ATPase ac-
tivity. Proc Natl Acad Sci U S A 88: 95139517.
Moch C, Schrogel O, Allmansberger R. 2000. Transcription of the
nfrA-ywcH operon from Bacillus subtilis is specifically induced
in response to heat. J Bacteriol 182: 43844393.
Mogk A, Homuth G, Scholz C, Kim L, Schmid FX, Schumann W.
1997. The GroE chaperonin machine is a major modulator of
the CIRCE heat shock regulon of Bacillus subtilis. EMBO J 16:
45794590.
Mogk A, Volker A, Engelmann S, Hecker M, Schumann W, Volker
U. 1998. Nonnative proteins induce expression of the Bacillus
subtilis CIRCE regulon. J Bacteriol 180: 28952900.
Morita MT, Tanaka Y, Kodama TS, Kyogoku Y, Yanagi H, Yura T.
1999. Translational induction of heat shock transcription factor
s32: evidence for a built-in RNA thermosensor. Genes Dev 13:
655665.

Narberhaus F. 1999. Negative regulation of bacterial heat shock
genes. Mol Microbiol 31: 18.
Nocker A, Hausherr T, Balsiger S, Krstulovic NP, Hennecke H, Nar-
berhaus F. 2001. A mRNA-based thermosensor controls expres-
sion of rhizobial heat shock genes. Nucleic Acids Res 29: 4800
4807.
Noone D, Howell A, Collery R, Devine KM. 2001. YkdA and YvtA,
HtrA-like serine proteases in Bacillus subtilis, engage in negative
autoregulation and reciprocal cross-regulation of ykdA and yvtA
gene expression. J Bacteriol 183: 654663.
Noone D, Howell A, Devine KM. 2000. Expression of ykdA, encoding
a Bacillus subtilis homologue of HrcA, is heat shock inducible
and negatively autoregulated. J Bacteriol 182: 15921599.
Price CW. 2002. General stress response. In: Bacillus subtilis and its
Closest Relatives: From Genes to Cells, ed Sonenshein AL, Hoch
JA, Losick R. American Society for Microbiology, Washington,
DC, 369384.
Price CW, Fawcett P, Ceremonie H, Su N, Murphy CK, Youngman
P. 2001. Genome-wide analysis of the general stress response in
Bacillus subtilis. Mol Microbiol 41: 757774.
Reischl S, Wiegert T, Schumann W. 2002. Isolation and analysis of
mutant alleles of the Bacillus subtilis HrcA repressor with re-
duced dependency on GroE function. J Biol Chem 277: 32659
32667.
Riethdorf S, Volker U, Gerth U, Winkler A, Engelmann S, Hecker
M. 1994. Cloning, nucleotide sequence, and expression of the
Bacillus subtilis lon gene. J Bacteriol 176: 65186527.
Roberts RC, Toochinda C, Avedissian M, Baldini RL, Gomes SL,
Shapiro L. 1996. Identification of a Caulobacter crescentus operon
encoding hrcA, involved in negatively regulating heat-inducible
transcription, and the chaperone gene grpE. J Bacteriol 178:
18291841.
Schmidt A, Schiesswohl M, Volker U, Hecker M, Schumann W. 1992.
Cloning, sequencing, mapping, and transcriptional analysis of
the groESL operon from Bacillus subtilis. J Bacteriol 174: 3993
3999.
Schulz A, Schumann W. 1996. hrcA, the first gene of the Bacillus
subtilis dnaK operon encodes a negative regulator of class I heat-
shock genes. J Bacteriol 178: 10881093.
Schulz A, Schwab S, Versteeg S, Schumann W. 1997. The htpG gene
of Bacillus subtilis belongs to class III heat shock genes and is
under negative control. J Bacteriol 10: 31033109.
Schulz A, Tzschaschel B, Schumann W. 1995. Isolation and analysis
of mutants of the dnaK operon of Bacillus subtilis. Mol Microbiol
15: 421429.
Schumann W, Hecker M, Msadek T. 2002. Regulation and function
of heat-inducible genes in Bacillus subtilis. In: Bacillus subtilis and
its Closest Relatives: From Genes to Cells, ed Sonenshein AL, Hoch
JA, Losick R. American Society for Microbiology, Washington,
DC, 359368.
Servant P, Mazodier P. 2001. Negative regulation of the heat shock
response in Streptomyces. Arch Microbiol 176: 237242.
Smirnova N, Scott J, Voelker U, Haldenwang WG. 1998. Isolation and
characterization of Bacillus subtilis sigB operon mutations that
suppress the loss of the negative regulator RsbX. J Bacteriol 180:
36713680.
B subtilis heat shock stimulon 217
Taylor BL, Zhulin IB. 1999. PAS domains: internal sensors of oxygen,
redox potential, and light. Microbiol Mol Biol Rev 63: 479506.
VanBogelen RA, Neidhardt FC. 1990. Ribosomes as sensors of heat
and cold shock in Escherichia coli. Proc Natl Acad Sci U S A 87:
55895593.
Vanet A, Plumbridge JA, Alix J-H. 1993. Cotranscription of two
genes necessary for ribosomal protein L11 methylation (prmA)
and pantothenate transport (panF) in Escherichia coli K-12. J Bac-
teriol 175: 71787188.
Versteeg S, Escher A, Wende A, Wiegert T, Schumann W. 2003. Reg-
ulation of the Bacillus subtilis heat shock gene htpG is under
positive control. J Bacteriol 185: 466474.
Versteeg S, Mogk A, Schumann W. 1999. The Bacillus subtilis htpG
gene is not involved in thermal stress management. Mol Gen
Genet 261: 582588.
Vijay K, Brody MS, Fredlund E, Price CW. 2000. A PP2C phospha-
tase containing a PAS domain is required to convey signals of
energy stress to the sB transcription factor of Bacillus subtilis.
Mol Microbiol 35: 180188.
Voelker U, Luo T, Smirnova N, Haldenwang WG. 1997. Stress acti-
vation of Bacillus subtilis sB can occur in the absence of the sB
negative regulator RsbX. J Bacteriol 179: 19801984.
Voelker U, Voelker A, Maul B, Hecker M, Dufour A, Haldenwang
WG. 1995. Separate mechanisms activate sB of Bacillus subtilis
in response to environmental and metabolic stresses. J Bacteriol
177: 37713780.
Wiedmann M, Arvik TJ, Hurley RJ, Boor KJ. 1998. General stress
transcription factor sB and its role in acid tolerance and viru-
lence of Listeria monocytogenes. J Bacteriol 180: 36503656.
Wise AA, Price CW. 1995. Four additional genes in the sigB operon
of Bacillus subtilis that control activity of the general stress factor
sB in response to environmental signals. J Bacteriol 177: 123
133.
Wu S, de Lencastre H, Tomasz A. 1996. Sigma-B, a putative operon
encoding alternate sigma factor of Staphylococcus aureus RNA
polymerase: molecular cloning and DNA sequencing. J Bacteriol
178: 60366042.
Yang X, Kang CM, Brody MS, Price CW. 1996. Opposing pairs of
serine protein kinases and phosphatases transmit signals of en-
vironmental stress to activate a bacterial transcription factor.
Genes Dev 10: 22652275.
Yuan G, Wong S-L. 1995a. Regulation of groE expression in Bacillus
subtilis: the involvement of the sA-like promoter and the roles
of the inverted repeat sequence (CIRCE). J Bacteriol 177: 5427
5433.
Yuan G, Wong S-L. 1995b. Isolation and characterization of Bacillus
subtilis regulatory mutants: evidence for orf39 in the dnaK op-
eron as a repressor gene in regulating the expression of both
groE and dnaK. J Bacteriol 177: 64626468.
Yura T, Kanemori M, Morita M. 2000. The heat shock response: reg-
ulation and function. In: Bacterial Stress Response, ed Storz G,
Hengge-Aronis R. American Society for Microbiology, Washing-
ton, DC, 318.
Zuber U, Drzewiecki K, Hecker M. 2001. Putative sigma factor SigI
(YkoZ) of Bacillus subtilis is induced by heat shock. J Bacteriol
183: 14721475.
Zuber U, Schumann W. 1994. CIRCE, a novel heat shock element
involved in regulation of heat shock operon dnaK of Bacillus
subtilis. J Bacteriol 176: 13591363.
Cell Stress & Chaperones (2003) 8 (3), 207217