Molecular Microbiology (2000) 38(5), 1093±1103
The HspR regulon of Streptomyces coelicolor: a role for
the DnaK chaperone as a transcriptional co-repressor²
Giselda Bucca,1 Anna M. E. Brassington,1
Hans-Joachim SchoÈnfeld2 and Colin P. Smith1*
1
Department of Biomolecular Sciences, UMIST, PO Box
88, Manchester, M60 1QD, UK. 2Hoffmann-La Roche Ltd,
Pharmaceutical Preclinical Research, CH-4070, Basel,
Switzerland.
Summary
The dnaK operon of Streptomyces coelicolor
encodes the DnaK chaperone machine and HspR,
the transcriptional repressor of the operon; HspR
confers repression by binding to several inverted
repeat sequences in the promoter region, dnaKp.
Here, we demonstrate that HspR specifically requires
the presence of DnaK protein to retard a dnaKp
fragment in gel-shift assays. This requirement is
independent of the co-chaperones, DnaJ and GrpE,
and it is ATP independent. Furthermore the retarded
protein±DNA complex can be `supershifted' by antiDnaK monoclonal antibody, demonstrating that DnaK
forms an integral component of the complex. It was
shown in DNase I footprinting experiments that
refolding and specific binding of HspR to its DNA
target does not require DnaK. We conclude that the
formation of the stable DnaK±HspR±DNA ternary
complex does not depend on the chaperoning activity
of DnaK. In affinity chromatography experiments
using whole-cell extracts, DnaK was shown to copurify with HspR, providing additional evidence that
the two proteins interact in vivo; it was not possible
to purify HspR away from DnaK in any experiments
unless a powerful denaturant was used. The level of
heat shock induction of chromosomal DnaK could be
partially suppressed by expressing dnaK extrachromosomally from a heterologous promoter. In addition, it is shown that DnaK confers enhanced HspRmediated repression of transcription in vitro. Taken
together, these results suggest that DnaK functions
as a transcriptional co-repressor by binding to HspR
at its operator sites. In this model, the DnaK±HspR
system would represent a novel example of feedback regulation of gene expression by a molecular
chaperone, in which DnaK directly activates a
Accepted 14 September, 2000. *For correspondence. E-mail colin.smith
@umist.ac.uk; Tel. (144) 161 200 4183; Fax (144) 161 236 0409. ²This
paper is dedicated to the memory of Gilberto Hintermann.
Q 2000 Blackwell Science Ltd
repressor, rather than inactivates an activator (as is
the case in the DnaK±s32 and Hsp70±HSF systems of
other organisms).
Introduction
Many heat shock proteins (Hsps) are universally conserved across the archaeal, bacterial and eukaryal
domains. Two well-studied families of Hsps are the
Hsp70/DnaK and Hsp60/GroE molecular chaperones.
These proteins fulfil crucial roles in cellular metabolism
under both normal and environmentally stressful growth
conditions by assisting in the folding of newly synthesized
and denatured proteins and in the assembly, transport
and degradation of other proteins (Gething and Sambrook, 1992; Hartl, 1996; Bukau and Horwich, 1998).
Although these proteins rank as perhaps the most highly
conserved proteins in nature, diverse regulatory mechanisms have evolved in different bacteria for controlling their
synthesis. Bacteria regulate heat shock gene transcription
with both positive and negative mechanisms. Two of
these mechanisms have been characterized in detail: the
s32 system, which operates in Escherichia coli and other
g-purple bacteria (Georgopoulos et al., 1994; Gross,
1996); and the HrcA/CIRCE repressor±operator system,
which has been identified in over 40 bacterial species,
including Gram-positives, proteobacteria and cyanobacteria (Narberhaus, 1999). The heat shock regulon of
E. coli is under the positive control of two alternative
sigma factors, which activate transcription of heat shock
genes in response to stimuli generated in the cytoplasm
(s32) (Bukau, 1993; Yura et al., 1993; Georgopoulos et al.,
1994; Yura and Nakahigashi, 1999) or in the periplasm
(sE) (Raina et al., 1995; RouvieÁre et al., 1995). The level
and the activity of s32 is negatively modulated by the
DnaK chaperone system, primarily through direct association of DnaK with s32, leading to inactivation and
degradation of s32 by the membrane-associated FtsH
(HflB) protease (Straus et al., 1989; Tilly et al., 1989;
Gamer et al., 1992; 1996; Liberek and Georgopoulos,
1993; Tatsuta et al., 1998). In the majority of bacteria,
heat shock gene expression is negatively regulated at the
transcriptional level (Narberhaus, 1999). The most widespread negative regulatory mechanism consists of a wellconserved inverted repeat of 9 bp called CIRCE, which is
associated in most cases with dnaK and groE chaperone
expression and is positioned in the 5 0 untranslated region
1094 G. Bucca, A. M. E. Brassington, H.-J. SchoÈnfeld and C. P. Smith
of mRNA (Zuber and Schumann, 1994). This operator site
is associated with `vegetative'-type promoters and is
recognized by the HrcA repressor protein (Yuan and
Wong, 1995; Schulz and Schumann, 1996). In Bacillus
subtilis and Bradyrhizobium japonicum, GroEL (rather
than DnaK) functions as a specific negative modulator of
the CIRCE regulons, as cells depleted of GroEL (or
GroEL4) show enhanced expression of either dnaK or
groEL4 respectively (Babst et al., 1996; Mogk et al.,
1997). In B. subtilis and other bacteria, groEL and dnaK
operons are under the same genetic control. In contrast,
different mechanisms of regulation govern the expression
of groESL and dnaK operons in bacteria such as
Caulobacter crescentus, B. japonicum, Agrobacterium
tumefaciens and Streptomyces spp. In B. japonicum, a
complex regulatory network comprising three control
systems ensures finely tuned Hsp expression according
to the prevailing environmental conditions: `ROSE' and its
unidentified repressor (Narberhaus et al., 1998a); three
alternative RpoH-like sigma factors (Narberhaus et al.,
1998b); the CIRCE/HrcA system. There is cross-talk
between the three circuits, and five differentially regulated
groEL operons ensure a supply of these proteins under
particular environmental conditions (Narberhaus, 1999).
In C. crescentus, the CIRCE/HrcA and the s32 regulatory
systems both operate to control groESL expression,
although the former system is developmentally, rather
than heat shock, regulated (Baldini et al., 1998). The two
groE operons of Streptomyces bacteria are under the
control of CIRCE/HrcA (DucheÃne et al., 1994). However,
the streptomycete dnaK operon is regulated by a different
class of autoregulatory repressor protein, designated
HspR (Bucca et al., 1995; 1997; Grandvalet et al.,
1997). HspR binds to at least three inverted repeats,
designated IR1, IR2 and IR3, in the dnaK promoter
region, centred at 275, 249 and 14 (Bucca et al., 1995).
Recently, the clpB gene, encoding the ClpB protease, was
shown to be part of the HspR regulon in Streptomyces
(Grandvalet et al., 1999). HspR binding sites that
resemble IR2 and IR3 have recently been designated
HAIR (HspR associated inverted repeat) sequences by
Grandvalet et al. (1999). Until recently, it was considered
that the HspR system was restricted to the ancient
actinomycete group of bacteria. HspR is now known to
control expression of the major molecular chaperoneencoding operons of Helicobacter pylori, although their
transcription is not influenced by heat shock in this case
(Spohn and Scarlato, 1999).
The objective of the present study was to determine
how the activity of HspR is modulated by heat shock in
Streptomyces. We present evidence that Streptomyces
has evolved a different strategy for controlling synthesis of
the DnaK chaperone machine. We demonstrate that
DnaK forms a specific complex with HspR and show
that this interaction is independent of the chaperoning
activity of DnaK. The results suggest that DnaK feedback
regulates its own synthesis by forming part of a ternary
complex with HspR bound to its DNA target. A simple
DnaK titration model is presented to account for HspRmediated control of heat shock gene expression.
Results
DnaK is required for HspR to form a stable complex with
its DNA target
Early attempts to isolate native HspR protein, overproduced in E. coli, were severely hampered by its
tendency to aggregate in the course of purification by
conventional chromatography. An amino-terminally histidine-tagged version of HspR (N-his HspR) was therefore
produced in E. coli, and the protein was purified by
immobilized metal affinity chromatography (IMAC) under
denaturing conditions. (The N-his HspR protein does not
bind to the Ni-NTA column unless it is first denatured; it is
thought that the his-tag is buried within the tertiary
structure of the protein.) As procedures for subsequent
refolding of N-his HspR by dialysis proved to be unreliable
because of variable levels of precipitation of the protein,
we assessed the ability of the purified E. coli DnaK and
GroEL chaperone machines to refold (or prevent aggregation of) the protein. The restoration of HspR activity was
first assessed by its ability to bind the dnaK promoter
region in gel-shift assays (Fig. 1). Under standard folding
conditions (see Experimental procedures), the DnaK
machine restored full binding activity, resulting in retardation of the dnaKp fragment as a discrete complex. GroE,
on the other hand, failed to confer clear binding activity,
although some degree of HspR binding was evident as a
retarded `smear' of the dnaKp fragment (Fig. 1A); this
could be explained by dissociation of the HspR±DNA
complex in the course of electrophoresis. It is well
established that DnaK functions together with the cochaperones DnaJ and GrpE to assist the refolding of
proteins and that it absolutely requires ATP for its function
(e.g. Hartl, 1996). Therefore, it was initially surprising to
observe that DNA-binding activity of HspR was restored in
the presence of DnaK alone (Fig. 1B). Furthermore, the
DnaK-mediated restoration of HspR activity was independent of ATP (Fig. 1C). These unexpected observations
indicated that the `reactivation' of HspR was not attributable to classical chaperone-mediated refolding of the
protein (see also footprinting experiments below). In
control experiments, the respective chaperones incubated
in the absence of HspR had no influence on the gel
mobility of the dnaKp fragment (data not shown).
The binding of `DnaK-activated' N-his HspR to DNA
was investigated in more detail by DNase I footprinting to
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Heat shock regulation in Streptomyces 1095
Fig. 1. Influence of molecular chaperones on the ability of N-his HspR to retard the dnaKp fragment in gel-shift assays. Denatured N-his HspR
was refolded as described in Experimental procedures. Lanes X, 4 mg of induced whole-cell extract from E. coli (pET15b::N-his-hspR); lanes
C, 3.2 pmol of HspR refolded in the absence of other proteins; F, position of free DNA fragment.
A. KJE, 1.6 and 3.2 pmol of HspR, respectively, incubated with DnaK±DnaJ±GrpE (10:1:5 ratio) in the presence of ATP; ESL, 0.4 and
1.6 pmol of HspR, respectively, incubated with GroEL±ES (14:14 ratio) (subunit molar ratios relative to HspR monomer) in the presence of
ATP.
B. Influence of individual components of the DnaK chaperone machine on HspR activity. Lanes labelled K, J and E, 1.6 pmol and 3.2 pmol of
HspR, respectively, incubated with DnaK (10:1), with DnaJ (1:1) and with GrpE (5:1) in presence of ATP.
C. Influence of ATP. Lanes labelled K contained 3.2 pmol of HspR and 32 pmol of DnaK, refolded in the absence and presence of ATP
respectively.
assess whether the extent of protection of the dnaKp
region was the same as that observed previously with the
native HspR protein (in cell extracts; Bucca et al., 1995).
The results (Fig. 2) demonstrated that N-his HspR
incubated with DnaK specifically protected the same
DNA sequences, from 285 to 117 with respect to the
transcription start site. The extent of protection was
independent of ATP. In this study, a heparin fraction of
partially purified native HspR was used as a control. It is
important to note that denatured N-his HspR was also
able specifically to protect the same DNA sequences
when it was preincubated instead with the GroE machine
or with bovine serum albumin (BSA; the latter having a
size and charge similar to that of DnaK) (Fig. 2). However,
the N-his HspR from the GroE or BSA incubations failed
to produce a discrete gel shift of the dnaKp fragment
(Fig. 1; data not shown). The immediate inference from
the above experiments is that denatured N-his HspR does
not require preincubation with DnaK in order to interact
with dnaKp (as detected by DNase I footprinting), as BSA
can substitute. Instead, HspR has a specific requirement
for DnaK in order to bind dnaKp with an affinity that is
sufficiently high to be resolved as a discrete complex in
the gel-shift assay. These footprinting observations are
critical in distinguishing chaperone-dependent refolding
from spontaneous refolding. It is clear from the BSA result
that N-his HspR can be refolded and regain specific DNAbinding activity independently of chaperones. It is known
that some DNA-binding proteins can produce a clear
footprint on a template yet not retard the same template in
a gel-shift assay; this is because the latter technique
requires the formation of a higher affinity protein±DNA
complex. Incubation of N-his HspR with GroE, DnaJ or
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
BSA produced comparable, faint, retarded smears in
the gel-shift assay, suggesting that these proteins can
promote the formation of a (loose) binary complex
between HspR and its target DNA.
In the above type of gel-shift/footprinting experiments, it
would be preferable to separate the refolding step from
the DNA-binding step by first purifying the refolded HspR
away from the DnaK chaperone. However, attempts to do
this, by size exclusion chromatography, failed because
DnaK co-purified with the HspR-containing fractions. The
only case in which we have been able to purify HspR
away from DnaK is by denaturing (8 M urea) Ni-NTA
chromatography (see below).
DnaK forms a complex with HspR bound to its DNA target
The ability of DnaK to restore DNA-binding activity to
HspR in gel-shift assays in the absence of co-chaperones
and ATP suggested that DnaK might form a ternary
complex with HspR and its DNA target. To verify this
hypothesis, a monoclonal anti-DnaK antibody was
incorporated in further gel-shift assays. N-his HspR was
denatured in 8 M urea and refolded in the presence of
DnaK (excluding ATP) at an equimolar ratio (see
Experimental procedures). After the refolding reaction
was completed, stoichiometric amounts of anti-DnaK
monoclonal antibody were added to the reactions, and
samples were subjected to the gel-shift assay. The results
(Fig. 3) clearly demonstrate that the presence of antiDnaK antibody `supershifts' the DNA±protein complex. A
comparable supershift was also obtained when a wholecell extract from E. coli overexpressing hspR was
subjected to the gel-shift DNA-binding assay in the
1096 G. Bucca, A. M. E. Brassington, H.-J. SchoÈnfeld and C. P. Smith
Fig. 2. DNase I footprinting: influence of molecular chaperones and
BSA on DNA-binding activity of refolded N-his HspR. The dnaK
operon promoter region was used as template (Bucca et al., 1995),
and HspR was refolded as described in Experimental procedures.
Additional proteins added to the respective samples are indicated
above the autoradiographs: K, DnaK; KJ, DnaK 1 DnaJ; KJE,
DnaK 1 DnaJ 1 GrpE; G, GroES 1 GroEL, B, BSA; the inclusion
or omission of ATP is also indicated. The region protected by HspR
is bracketed (numbered relative to the transcription start site), and
six DNase I-hypersensitive sites are arrowed. The sample indicated
by an asterisk was incubated with 12 pmol of native partially
purified HspR (heparin fraction). Dideoxy sequencing ladder of the
probe fragment is labelled GATC.
SDS±PAGE and Western blotting (Fig. 4). At least two
proteins of < 70 kDa were found to co-purify with HspR,
and one of these was shown to cross-react with anti-DnaK
antibodies. A second (<23 kDa) co-purifying protein has
a size similar to GrpE, but it failed to cross-react with antiGrpE antibody. This protein was observed to elute in
equivalent fractions from control cell extracts containing
the vector only (data not shown), and therefore its
binding to the Ni-NTA column was independent of HspR;
subsequently, the protein was identified as SlyD, a
histidine-rich nickel-binding peptidyl-prolyl cis/trans isomerase (Hottenrott et al., 1997), by mass spectrometric
analysis of tryptic peptides (A. M. E. Brassington, D.
Smith and J. Thomas-Oates, personal communication).
Importantly, none of the Ni-NTA-purified fractions from the
control extract cross-reacted with DnaK antibodies (data
not shown). In separate experiments, all native (non-histagged) HspR-containing protein fractions that were
obtained by conventional heparin chromatography were
also shown to contain DnaK by Western analysis (data
not shown). Thus, the interaction of DnaK with HspR is
not restricted to experiments in which HspR is previously
unfolded. These affinity chromatography results are
consistent with the proposal that HspR and DnaK interact
specifically in vivo. The purified C-his HspR-containing
fractions alone were shown to form a discrete complex
with dnaKp in gel-shift assays (data not shown), an
observation that would be predicted because DnaK
co-elutes with HspR.
presence of anti-DnaK antibody (Fig. 3). These results
confirmed that DnaK forms an ATP-independent complex
with HspR.
DnaK co-purifies with HspR
If DnaK and native HspR do interact physically, then they
should co-purify. This prediction was confirmed by
investigating the co-elution of DnaK and his-tagged
HspR in IMAC experiments. It was necessary to use a
C-his-tagged HspR for these experiments because the Nhis HspR protein does not bind to a Ni-NTA column under
non-denaturing conditions. Total protein extract from
E. coli overproducing C-his HspR was loaded onto a NiNTA column, and bound proteins were eluted with a
0±500 mM linear imidazole gradient in a buffer containing
200 mM KCl. The resulting fractions were analysed by
Fig. 3. Anti-DnaK monoclonal antibody `supershifts' the DnaK±
HspR±DNA complex. See Experimental procedures for details.
Samples indicated by an asterisk were incubated with 4 mg of
induced - extract from E. coli (pET15b::N-his-hspR), instead of
purified HspR and DnaK. m-Ab, anti-DnaK monoclonal antibody;
IgG, anti-mouse IgG (non-specific negative control).
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Heat shock regulation in Streptomyces 1097
Fig. 4. Co-purification of DnaK and C-his
HspR on a metal affinity (Ni-NTA) column.
A. Coomassie-stained gel of protein fractions
from an induced culture of E. coli (pET22b::Chis-hspR) eluted from a Ni-NTA column. M,
molecular weight markers (sizes are indicated
in kDa); K 1 J 1 E; mixture of 200 ng each
of purified DnaK, DnaJ and GrpE; HspR,
200 ng of N-his HspR. The respective
positions of DnaK and HspR in the gel are
arrowed.
B. Western analysis of protein fractions using
anti-HspR antibody; 50 ng of N-his HspR was
loaded as a control.
C. Western analysis of the same protein
fractions using anti-DnaK polyclonal antibody;
50 ng of purified DnaK was loaded as a
control; N, 10 mg of cell extract from E. coli
BB1553(DdnaK52) (negative control).
DnaK enhances HspR-mediated repression of
transcription in vitro
The biochemical studies reported above indicate that
DnaK is required for HspR to form a stable complex with
its DNA target. In vitro transcription experiments were
conducted to assess whether this interaction is necessary
for HspR to repress transcription from dnaKp efficiently.
The results in Fig. 5 demonstrate that HspR-mediated
repression of transcription is significantly increased in the
presence of stoichiometric amounts of DnaK. DnaK alone
had no inhibitory effect on transcription from dnaKp (data
not shown). The GroE chaperone machine was used as a
further control and caused no enhancement of repression;
in fact, it appeared to interfere with HspR-mediated
repression (Fig. 5). These in vitro results are consistent
with the notion that DnaK functions as a transcriptional
co-repressor of the DnaK operon.
vector pMT3206 and analysing chromosomal dnaK
transcription at 308C (normal growth temperature) and
after a heat shock at 428C.
Transcript levels were quantified by S1 nuclease mapping. dnaK transcripts originating from the chromosomal
dnaK operon could be distinguished unambiguously from
Expression of dnaK from a heterologous promoter
partially suppresses heat shock induction of the
chromosomal dnaK operon
If DnaK does play a role in negatively regulating its own
synthesis in Streptomyces, then it would be predicted that
`artificial' production of DnaK in vivo, by expressing dnaK
from a heterologous promoter, would lead to partial
suppression of the normal level of heat shock induction
of the chromosomal dnaK operon. The reasoning is that a
significant proportion of the cellular pool of DnaK is likely
to be derived from the vector-borne dnaK, placing less
reliance on heat shock induction of the native dnaK
operon as a source of DnaK. This prediction was tested
by placing dnaK under the control of the Streptomyces
coelicolor gyl promoters in the multicopy expression
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Fig. 5. DnaK enhances HspR-mediated repression of transcription
from dnaKp: in vitro `run-off' transcription assays with linear dnaKp
template DNA. Lane 1, RNA polymerase only; lane 2, plus N-his
HspR (0.56 mM); lanes 3 and 4, N-his HspR incubated with DnaK
at 1:1 and 5:1 molar ratio respectively; lanes 5 and 6, N-his HspR
incubated with GroES and GroEL at 14:14:1 and 28:28:1 subunit
molar ratios respectively; M, pBR322 MspI size markers (selected
sizes are indicated in nucleotides). The run-off transcript from
dnaKp is indicated by an arrow; E, strong signal from artifactual
end-to-end transcription of the linear template.
1098 G. Bucca, A. M. E. Brassington, H.-J. SchoÈnfeld and C. P. Smith
Fig. 6. Partial suppression of heat shock induction of the
chromosomal dnaK operon by expressing dnaK from a
heterologous plasmid-borne promoter. The level of dnaK transcript
(<70 bp hybrid) originating from the chromosomal dnaK operon
promoter under normal (308C) and heat-shocked (428C for 15 min)
conditions was quantified by S1 nuclease mapping and related to
the level of the (constitutive) glkA transcript (210 bp hybrid) from
the same respective RNA samples. V and KEJ indicate that the
RNA samples were isolated from Streptomyces cultures containing
the vector pMT3206 and pMT3206::KEJ respectively. The low level
of dnaK transcript in cultures grown at 308C is not visible under
these experimental conditions.
vector-derived dnaK transcripts on the basis of the
respective sizes of the S1 nuclease-resistant hybrids.
The glucose kinase-encoding glkA gene was used as a
constitutive internal reference for RNA integrity and
quantity in the respective RNA preparations. In these
experiments, the levels of dnaK and glkA hybrids were
compared within a particular RNA sample when deducing
the level of heat shock induction, rather than between
samples (as the latter offers less reliable comparison). In
strains carrying pMT3206::K or pMT3206::KEJ, the level
of heat shock induction of the chromosomal dnaK operon
was reproducibly lower (mean < twofold) than in strains
containing the vector alone (e.g. Fig. 6). The level of
suppression of induction was comparable in strains
expressing dnaK alone or dnaK±grpE±dnaJ from the
vector (data not shown). Although not unequivocal, these
in vivo results are consistent with the proposal that DnaK
plays a negative feedback role in modulating its own
synthesis (via interaction with HspR).
Discussion
The principal conclusions of this study are (i) the DnaK
protein forms a specific ATP-independent complex with
the Streptomyces HspR repressor; and (ii) this interaction
is necessary for HspR to retard the dnaKp fragment in
gel-shift assays.
The observed DnaK-dependent formation of a stable
complex between HspR and its DNA target cannot be
attributable per se to the chaperoning function of DnaK,
as neither GrpE nor DnaJ is required, and the DnaKmediated `activation' of HspR is independent of ATP.
DnaK is not actually required for the refolding of ureadenatured HspR, as judged from the DNase I footprinting
experiments; it was shown that BSA could substitute for
DnaK or GroES/EL in the refolding reactions and restore
specific DNA-binding capacity to HspR, indistinguishable
from that observed in the presence of DnaK. We conclude
that HspR has the capacity to refold spontaneously under
these experimental conditions and that BSA prevents the
aggregation of the purified HspR (which occurs when its
refolding is conducted in the absence of additional
proteins). However, DnaK was the only protein that
could produce a stable HspR±DNA complex that is
resolvable by gel electrophoresis. This suggests that the
binding of DnaK confers a higher affinity binding of HspR
to its DNA target. Furthermore, it was demonstrated by
`supershift' assays with monoclonal anti-DnaK antibody
that DnaK forms an integral part of this protein±DNA
complex. It is not necessary to unfold HspR in order to
detect an interaction with DnaK. Independent in vitro
evidence for the stable interaction between HspR and
DnaK was obtained by showing that the two proteins copurified in every attempt made to obtain pure HspR,
despite the high salt concentration used in some cases
(up to 300 mM NaCl was present during C-his-HspR
purification). DnaK was shown to interact with native
(untagged) HspR and both the N- and C-his HspR
derivatives in cell extracts. In our hands, it has not been
possible to purify non-denatured HspR away from DnaK,
and therefore it has not been possible to investigate the
interaction of purified native HspR with DnaK without first
denaturing the HspR and separating it by affinity
chromatography. When HspR and DnaK are mixed
together, they invariably co-purify. It is considered that,
under normal steady-state physiological conditions, HspR
is always complexed with DnaK in vivo.
Further experiments described in this study provided
evidence that DnaK exerts a negative effect on transcription from its own promoter, dnaKp. In the in vitro
transcription study, purified HspR and DnaK, in combination, were sufficient to confer substantial repression; GroE
or BSA, on the other hand, had no negative effect on
transcription from dnaKp. It is likely that the steady-state
level of the DnaK±HspR±dnaKp ternary complex is
principally responsible for determining whether the dnaK
operon is expressed or not in vivo. When dnaK was
expressed from a heterologous plasmid-borne promoter
in wild-type Streptomyces, it was shown that the
chromosomal dnaK operon was heat induced at a lower
level compared with control cultures that contained the
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Heat shock regulation in Streptomyces 1099
Fig. 7. A model for DnaK-mediated feedback regulation of the
Streptomyces coelicolor dnaK operon. See the Discussion.
vector only. This partial suppression would be expected if
DnaK fulfils a homeostatic role in feedback regulating its
own synthesis, as a significant proportion of the DnaK
pool in the cell would be contributed from the vector-borne
dnaK gene, and therefore induction of the chromosomal
dnaK operon at a wild-type level would be superfluous.
The gyl promoters on pMT3206 are glycerol inducible.
However, it was found that the gyl system is not inducible
at 428C (A. M. E. Brassington, unpublished data). Therefore, these expression experiments relied on basal
expression of dnaK from the pMT3206 derivatives, and
this may explain why the observed differences in
chromosomal dnaK induction levels are relatively modest
(as there would be a significant cellular demand for DnaK
from the latter). Western analysis of cell extracts from
Streptomyces revealed comparable levels of DnaK in
cultures containing either the vector only or pMT3206::K
(or pMT3206::KEJ) (data not shown). Again, this would be
predicted if DnaK negatively regulates its own synthesis;
expression of dnaK from pMT3206::K is insensitive to the
HspR±DnaK system because the plasmid lacks the HspR
binding sites and, therefore, the additional synthesis of
DnaK from the plasmid-derived dnaK gene would be
compensated by reduced expression from the chromosomal HspR-regulated dnaK operon. If DnaK does negatively regulate its own synthesis in Streptomyces, it would
be predicted that a dnaK null mutant would exhibit an
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
enhanced basal level of transcription of the dnaK operon.
Repeated attempts to delete dnaK (and dnaK±grpE±
dnaJ) in Streptomyces have failed (G. Bucca and A. M. E.
Brassington, unpublished data) and, consequently, we
believe that dnaK is essential in this organism. Experiments are in progress to delete the chromosomal dnaK±
grpE±dnaJ genes (leaving the dnaK operon promoter and
hspR intact) in a strain that contains dnaK±grpE±dnaJ
integrated elsewhere in the chromosome under the
transcriptional control of the gyl promoters. It will be
possible in this deletion strain to conduct DnaK depletion
experiments (by removing glycerol from the growth
medium) to test the proposal that DnaK negatively
regulates its own synthesis.
In some bacteria and in eukaryotic cells, the DnaK/
Hsp70 chaperone machine has been shown to function as
a negative regulator of heat shock gene expression by
physically associating with and inactivating specific heat
shock transcriptional activators, for example, s32 in E. coli
(Liberek and Georgopoulos, 1993) and HSF in vertebrate
cells (Morimoto, 1998; Shi et al., 1998); in these cases,
the cycle of binding and release of DnaK/Hsp70 to the
substrate is ATP dependent. In other bacteria such as
B. subtilis, it has been shown that the GroE chaperone
machine (rather than DnaK) functions as a negative
regulator of heat shock gene expression (Mogk et al.,
1997; Narberhaus, 1999). GroE is necessary for the HrcA
repressor protein to function efficiently as a repressor of
the dnaK operon in B. subtilis; in this case, it is considered
that GroE maintains HrcA in a properly folded state. We
suggest that the HspR/DnaK system in Streptomyces
differs from previously studied systems in that DnaK
functions as a transcriptional co-repressor by binding to
HspR although it is complexed with DNA. Thus, DnaK
`activates' HspR repressor, rather than inactivating an
activator (such as s32 and HSF). Despite this reversal of
function, the end-result is the same in the respective heat
shock systems ± the molecular chaperones negatively
regulate their own synthesis by modulating the activity
of their respective transcription factors. Thus, although
mechanisms controlling molecular chaperone gene
expression vary widely, homeostatic feedback control by
a key chaperone is emerging as a common theme. The
results from the present study suggest a simple model
for DnaK-mediated regulation of the S. coelicolor dnaK
operon (Fig. 7); under normal growth conditions, native
HspR binds DnaK, and this complex binds avidly to the
dnaK promoter region to repress transcription of the
operon efficiently. We speculate that in the (transient)
absence of DnaK, HspR does not bind with high affinity
to its DNA target. Thus, during heat shock, when DnaK
is sequestered by denatured or partially unfolded
proteins, the HspR protein is unable efficiently to repress
transcription from dnaKp and the operon is induced.
1100 G. Bucca, A. M. E. Brassington, H.-J. SchoÈnfeld and C. P. Smith
A major future objective will be to determine which
sequences of DnaK and HspR interact physically and to
elucidate how this interaction alters the activity of HspR. It
will also be important to investigate the oligomeric nature
of the ternary complex under physiological and heat shock
conditions.
Experimental procedures
Bacterial strains and plasmids
S. coelicolor A3(2) strain MT1110 is a SCP12, SCP22
derivative of the wild-type strain 1147 (Hopwood et al., 1985).
Streptomyces lividans 1326 is the wild-type strain (Hopwood
et al., 1985). [S. lividans is essentially the same species as
S. coelicolor, but has the advantage that it lacks the methylDNA-directed restriction system of the latter (Flett et al.,
1997); their respective dnaK operons have an identical
restriction map, and heat shock regulation appears to be
identical in both strains.] E. coli XL1-Blue (Stratagene) was
used as the general cloning host, and E. coli BL21(lDE3,
pLysS) [F2ompT hsdSB (rB2mB2) gal dcm (lDE3) pLysS
(CmR)] (Novagen) was used as the host for overproduction of
HspR. E. coli plasmid pGRP17, which contains the 3 0 end of
grpE and the complete dnaJ and hspR genes (Bucca et al.,
1997), was used in the construction of C-his hspR, and
pG480 (Bucca et al., 1997) was used for the isolation of a
200 bp SacII±NcoI fragment containing the HspR binding
sites; alternatively, it was used as a template in polymerase
chain reactions (PCRs) to produce a uniquely end-labelled
probe of the dnaKp region. E. coli plasmids pET22b and
pET11a (Novagen) were used for the construction and
overexpression of C-his hspR and native hspR respectively.
The construction of N-his hspR has been described
previously (Bucca et al., 1997).
The Streptomyces gyl-based expression vector pMT3206,
a derivative of the multicopy plasmid pIJ680 containing
the gylR±gylP1/P2 fragment `RP±S1' (Paget et al., 1994;
F. Amini, M. S. B. Paget, G. Bucca and C. P. Smith,
manuscript in preparation), was used for expressing dnaK
independently of the dnaK promoter region in Streptomyces.
In plasmid pMT3206::K, the complete S. coelicolor dnaK
gene [including its ribosome binding site (RBS)] corresponding to nucleotide co-ordinates 176±2056 in EMBL accession
number L46700, was cloned downstream from the gyl
promoters in pMT3206; this construction involved the use of
exonuclease III to remove downstream grpE-coding DNA and
required the addition of an oligonucleotide adapter to the 5 0
end of dnaK to rebuild the RBS. Plasmid pMT3206::KEJ
contains the identical dnaK sequences to pMT3206::K, but
extends further downstream to include the complete coding
sequences of grpE and dnaJ, ending at nucleotide position
4004 (as numbered in entry L46700). No PCR steps were
used in either construction, and the structures of the
respective expression plasmids were validated by DNA
sequencing across all vector±insert junctions. Further details
of the construction of each plasmid are available from the
authors on request.
Molecular chaperones
DnaK, the co-chaperones GrpE and DnaJ, and GroES/EL
were each purified from E. coli strains overexpressing the
respective genes as described previously (SchoÈnfeld et al.,
1995a, b; Behlke et al., 1997) All purified proteins were free
of ATP and stored as aliquots at 2808C in 100 mM NaCl,
50 mM Tris-HCl, pH 7.7.
Streptomyces cultivation, transformation, heat shock and
RNA isolation
Surface-grown S. lividans cultures, transformed with
pMT3206 and dnaK-containing plasmid derivatives, were
grown for 24±28 h on cellophane-coated SMMS agar
medium containing 50 mg ml21 thiostrepton and either 0.5%
mannitol or 0.5% glycerol as an additional carbon source.
The growth, heat shock conditions and RNA isolation
procedure have been described previously (Bucca et al.,
1995). All other methods and media for Streptomyces growth
and transformation are described by Hopwood et al. (1985).
Overproduction and purification of HspR in E. coli
The construction of N-his hspR in pET15b and overproduction of the protein has been reported previously (Bucca et al.,
1997). To construct a C-terminally his-tagged HspR (C-his
HspR), a BamHI±BstXI fragment containing the hspR coding
sequence was isolated from pGRP17 and ligated to a BstXI±
HindIII oligonucleotide adapter that contains the end of the
hspR coding region, excluding the stop codon. The fragment
was ligated to the E. coli pET22b expression vector and
digested with BamHI and HindIII; this construction incorporates six histidine residues at the C-terminus of the protein.
The construct was digested with NdeI (in the vector
polylinker) and SunI (codon 8 of the hspR gene), and the
purified backbone fragment was ligated to an NdeI±SunI
adapter described previously (Bucca et al., 1995) to fuse the
hspR coding region to the start codon within the expression
vector. The conditions for overproduction were as described
by Bucca et al. (1997). For the construction of native hspR,
an NdeI±BlpI fragment containing the hspR coding sequence
was isolated from pET15b::N-his hspR and cloned into
pET11a digested with the same enzymes.
N-his HspR was purified under denaturing conditions as
described in the Qiagen handbook (QIAexpressionist, 3rd
edn). The purified protein was eluted with a 0±500 mM
imidazole linear gradient in the following buffer: 10 mM TrisHCl (pH 8), 100 mM Na-phosphate buffer (pH 8) and 8 M
urea. C-his HspR was purified under native conditions, as
recommended by the above handbook, with the exception
that 200 mM KCl (instead of 300 mM NaCl) was used
throughout the purification protocol. C-His HspR was eluted
with a 0±500 mM imidazole gradient. Native HspR was
partially purified from an E. coli (pET11a::hspR) protein
extract in 100 mM Tris-HCl (pH 8), 50 mM Na-phosphate
(pH 8), 100 mM NaCl and 10% glycerol by chromatography
on a heparin column (Amersham Pharmacia); fractions were
eluted with a linear 0±1 M NaCl gradient in 100 mM Tris-HCl
(pH 8), 0.5 mM EDTA and 15% glycerol.
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Heat shock regulation in Streptomyces 1101
Refolding of N-his HspR
Unless stated otherwise elsewhere, in the refolding reaction,
40 pmol of HspR, denatured in 8 M urea, was diluted 1:100
(to 0.4 mM) in folding buffer [100 mM Tris-HCl, pH 7.7, 5 mM
dithiothreitol (DTT), 50 mM KCl, 5 mM MgCl2, 2 mM ATP]
containing either DnaK (at 10:1 or 1:1), DnaJ (1:1) and GrpE
(5:1) or GroES (14:1) and GroEL (14:1) (all expressed as
subunit molar ratios relative to HspR monomer). After 1 h
incubation at 258C, HspR activity in the samples was
analysed by either the gel-shift assay or DNase I footprinting.
As controls, HspR was incubated under the conditions
described above with or without BSA at an equimolar ratio,
but omitting the molecular chaperones. For the in vitro
transcription experiments, denatured HspR (1.9 mM) was
preincubated with DnaK alone (at 1:1 and 5:1 subunit molar
ratio) or with GroES/EL (at 14:14:1 or 28:28:1 subunit molar
ratio) relative to HspR, under the above conditions.
Gel mobility shift assays and DNase I footprinting
Procedures and labelled DNA fragments used for gel
mobility shifts and DNase I footprinting have been described
previously (Bucca et al., 1995); refolded N-his HspR (0.4±
3.2 pmol) was incubated with 1.3 fmol of end-labelled dnaKp
fragment in gel-shift assays, and 6.4 pmol of N-his HspR was
incubated with 40 fmol of the same DNA template for DNase I
footprinting. The antibody `supershift' assay was performed
with a DnaK-specific monoclonal antibody (Stressgen Biotechnologies): 1.2 pmol of anti-DnaK monoclonal antibody
was incubated for 15 min at 258C with an aliquot of refolded
HspR (containing 1.6 pmol of both DnaK and HspR), and
samples were subjected to the standard gel-shift assay. Antimouse IgG was used as a negative control. Samples
were analysed on 4% native polyacrylamide gels in standard
Tris-borate±EDTA electrophoresis buffer at 48C.
SDS±PAGE and Western analysis
Standard procedures were followed for SDS±PAGE (described by Bucca et al., 1995). Western blot analysis of HspR,
DnaK and GrpE was performed with the respective polyclonal
antibodies: anti-HspR antibodies were raised in a rabbit
using IMAC-purified N-his HspR and TiterMax Gold as the
adjuvant; anti-DnaK antibody from rabbit was kindly provided
by B. Bukau; and anti-GrpE antibody was raised in mouse
using purified recombinant GrpE. Alkaline phosphataseconjugated secondary antibodies were used, and the blots
were developed using Sigma Fast NBT/BCIP tablets. Protein
size markers were obtained from Bio-Rad.
In vitro transcription
In vitro run-off transcription assays were carried out as
described by Bucca et al. (1997), except that BSA was
omitted from the transcription buffer. After refolding the ureadenatured N-his HspR in the presence or absence of molecular
chaperones (as above, except that ATP was omitted), the
samples were incubated with the dnaKp template for 30 min at
258C before adding Streptomyces RNA polymerase. HspR was
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
present at 0.56 mM in the transcription reactions. After the
reactions were terminated, samples were phenol±chloroform
extracted before ethanol precipitation and electrophoresis on
a 7 M urea26% polyacrylamide gel.
Quantification of in vivo dnaK transcript levels
S1 nuclease mapping was used to quantify transcript levels
and conducted as described by Smith (1991). Probe KC,
used for detecting transcription from the chromosomal dnaK
operon, was a 233 bp PCR product generated using the
primer pair, oligoKUS (5 0 -CGGGCGGTGCGTCCTTCC-3 0 )
and oligoKDS (5 0 -TTAGTCGTGCCCAGGTCG; end labelled)
and pG480 as template; after nuclease treatment, a hybrid of
70 bp is generated. Probe KV used for detecting vector-borne
dnaK transcription (from pMT3206::K and pMT3206::KEJ) was
a 303 bp PCR product encompassing the tandem gylP1/P2
promoters and part of dnaK in pMT3206::K, generated using
primer pair AMBgylRP (5 0 -TTCCGGCCTTCGACGTTCC) and
oligoKDS (end labelled) and pMT3206::K as template; S1
resistant hybrids of 120 bp (from gylP2) and 170 bp (from
gylP1) were generated. Probe GLK (265 bp) was used to detect
the 5 0 end (the first 210 nucleotides) of the glucose kinaseencoding glkA (orf3) mRNA (which represented a `constitutive'
internal control; Angell et al., 1992); primer pair Abglk5p (5 0 CAGCGCATCGACCTGGACTG) and Abglk3p (5 0 -ATGCCCAC
TGCGACGATCTC; end labelled) was used with S. coelicolor
MT1110 chromosomal DNA as template.
Acknowledgements
We thank Bernd Bukau and Costa Georgopoulos for helpful
discussions, Duncan Smith and Jane Thomas-Oates for
assistance with mass spectrometric analysis, and we are
grateful to Bernd Bukau for providing anti-DnaK polyclonal
antibodies. A.M.E.B. was the recipient of a PhD studentship
from the BBSRC, UK. This work was supported by project
grant number 050565 from the Wellcome Trust to C.P.S.
References
Angell, S., Schwarz, E., and Bibb, M.J. (1992) The glucose
kinase gene of Streptomyces coelicolor A3(2): its nucleotide sequence, transcriptional analysis and role in glucose
repression. Mol Microbiol 6: 2833±2844.
Babst, M., Hennecke, H., and Fischer, H.-M. (1996) Two
different mechanisms are involved in the heat-shock
regulation of chaperonin gene expression in Bradyrhizobium japonicum. Mol Microbiol 19: 827±839.
Baldini, R., Avedissian, M., and Gomes, S.L. (1998) The
CIRCE element and its putative repressor control cell cycle
expression of the Caulobacter crescentus groESL operon.
J Bacteriol 180: 1632±1641.
Behlke, J., Ristau, O., and SchoÈnfeld, H.J. (1997) Nucleotidedependent complex formation between the Escherichia coli
chaperonins GroEL and GroES studied under equilibrium
conditions. Biochemistry 36: 5149±5156.
Bucca, G., Ferina, G., Puglia, A.M., and Smith, C.P. (1995)
The dnaK operon of Streptomyces coelicolor encodes a
1102 G. Bucca, A. M. E. Brassington, H.-J. SchoÈnfeld and C. P. Smith
novel heat-shock protein which binds to the promoter
region of the operon. Mol Microbiol 17: 663±674.
Bucca, G., Hindle, Z., and Smith, C.P. (1997) Regulation of
the dnaK operon of Streptomyces coelicolor A3(2) is
governed by HspR, an autoregulatory repressor protein.
J Bacteriol 179: 5999±6004.
Bukau, B. (1993) Regulation of the Escherichia coli heat
shock response. Mol Microbiol 9: 671±680.
Bukau, B., and Horwich, A.L. (1998) The Hsp70 and Hsp60
chaperone machines. Cell 92: 351±366.
DucheÃne, A.M., Thompson, C.J., and Mazodier, P. (1994)
Transcriptional analysis of groEL genes in Streptomyces
coelicolor A3(2). Mol Gen Genet 245: 61±68.
Flett, F., Mersinias, V., and Smith, C.P. (1997) High efficiency
intergenic conjugal transfer of plasmid DNA from Escherichia coli to methyl DNA-restricting streptomycetes. FEMS
Microbiol Lett 155: 223±229.
Gamer, J., Bujard, H., and Bukau, B. (1992) Physical
interaction between heat shock proteins DnaK, DnaJ, and
GrpE and the bacterial heat shock transcription factor
sigma 32. Cell 69: 833±842.
Gamer, J., Multhaup, G., Tomoyasu, T., McCarty, J.S.,
Rudiger, S., Schonfeld, H.J., et al. (1996) A cycle of binding
and release of DnaK, DnaJ and GrpE chaperones
regulates activity of the E. coli heat shock transcription
factor sigma 32. EMBO J 15: 607±617.
Georgopoulos, C., Liberek, K., Zylicz, M., and Ang, D. (1994)
Properties of the heat shock proteins of Escherichia coli
and the autoregulation of the heat shock response. In
Properties of the Heat-Shock Proteins of Escherichia coli
and the Autoregulation of the Heat-Shock Response.
Morimoto, R.I., TissieÁres, A., and Georgopoulos, C. (eds).
Cold Spring Harbor, NY: Cold Spring Harbor Laboratory
Press, pp. 209±249.
Gething, M.-J., and Sambrook, J. (1992) Protein folding in the
cell. Nature 355: 33±45.
Grandvalet, C., Servant, P., and Mazodier, P. (1997)
Disruption of hspR the repressor gene of the dnaK operon
in Streptomyces albus G. Mol Microbiol 23: 77±84.
Grandvalet, C., de Crecy-Lagard, V., and Mazodier, P. (1999)
The ClpB ATPase of Streptomyces albus G belongs to the
HspR heat shock regulon. Mol Microbiol 31: 521±532.
Gross (1996) Function and regulation of the heat shock
proteins. In Escherichia coli and Salmonella: Cellular and
Molecular Biology. Neidhardt, F.C., Curtiss, R., III, Ingraham, J.L., Lin, E.C.C., Low, K.B., Magasanik, B., et al.
(eds). Washington, DC: American Society for Microbiology
Press, pp. 1382±1399.
Hartl, F.U. (1996) Molecular chaperones in cellular protein
folding. Nature 381: 571±580.
Hopwood, D.A., Bibb, M.J., Chater, K.F., Kieser, T., Bruton,
C.J., Kieser, H.M., et al. (1985) Genetic Manipulation of
Streptomyces: A Laboratory Manual. Norwich: The John
Innes Foundation.
Hottenrott, S., Schumann, T., Pluckthun, A., Fischer, G., and
Rahfeld, J.U. (1997) The Escherichia coli SlyD is a metal
ion-regulated peptidyl-prolyl cis/trans-isomerase. J Biol
Chem 272: 15697±15701.
Liberek, K., and Georgopoulos, C. (1993) Autoregulation of
the Escherichia coli heat shock response by the DnaK and
DnaJ heat shock proteins. Proc Natl Acad Sci USA 90:
11019±11023.
Mogk, A., Homuth, G., Scholz, C., Kim, L., Schmid, F.X., and
Schumann, W. (1997) The GroE chaperon machine is a
major modulator of the CIRCE heat shock regulon of
Bacillus subtilis. EMBO J 16: 4579±4590.
Morimoto, R.I. (1998) Regulation of the heat shock transcriptional response: cross talk between a family of heat shock
factors, molecular chaperones, and negative regulators.
Genes Dev 12: 3788±3796.
Narberhaus, F. (1999) Negative regulation of bacterial heat
shock genes. Mol Microbiol 31: 1±8.
Narberhaus, F., Kaser, R., Nocker, A., and Hennecke, H.
(1998a) A novel DNA element that controls bacterial heat
shock gene expression. Mol Microbiol 28: 315±323.
Narberhaus, F., Kowarik, M., Beck, C., and Hennecke, H.
(1998b) Promoter selectivity of the Bradyrhizobium japonicum RpoH transcription factors in vivo and in vitro. J
Bacteriol 180: 2395±2401.
Paget, M.S.B., Hintermann, G., and Smith, C.P. (1994)
Construction and application of streptomycete promoter
probe vectors which employ the Streptomyces glaucescens tyrosinase-encoding gene as reporter. Gene 146:
105±110.
Raina, S., Missiakas, D., and Georgopoulos, C. (1995) The
rpoE gene encoding the sE (s24) heat shock sigma factor of
Escherichia coli. EMBO J 14: 1043±1055.
RouvieÁre, P., Penas, A.D.L., Mecsas, J., Lu, C.Z., Rudd,
K.E., and Gross, C.A. (1995) RpoE, the gene encoding the
second heat shock sigma factor, sE, in Escherichia coli.
EMBO J 14: 1032±1042.
SchoÈnfeld, H.J., Schmidt, D., SchroÈder, H., and Bukau, B.
(1995a) The DnaK chaperone system of Escherichia coli ±
quaternary structures and interactions of the DnaK and
GrpE components. J Biol Chem 270: 2183±2189.
SchoÈnfeld, H.-J., Schmidt, D., and Zulauf, M. (1995b)
Investigation of the molecular chaperone DnaJ by
analytical ultracentrifugation. Progr Colloid Polym Sci 99:
7±10.
Schulz, A., and 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:
1088±1093.
Shi, Y., Mosser, D., and Morimoto, R. (1998) Molecular
chaperones as HSF1-specific transcriptional repressors.
Genes Dev 12: 654±666.
Smith, C.P. (1991) Methods for mapping transcribed DNA
sequences. In Essential Molecular Biology, Vol. II. Brown,
T.A. (ed.). Oxford: IRL Press, pp. 237±252.
Spohn, G., and Scarlato, V. (1999) The autoregulatory HspR
repressor protein governs chaperone gene transcription in
Helicobacter pylori. Mol Microbiol 34: 663±674.
Straus, D., Walter, W.A., and Gross, C.A. (1989) The activity
of s32 is reduced under conditions of excess heat shock
protein production in Escherichia coli. Genes Dev 3: 2003±
2010.
Tatsuta, T., Tomoyasu, T., Bukau, B., Kitagawa, M., Mori, H.,
Karata, K., and Ogura, T. (1998) Heat shock regulation in
the ftsH null mutant of Escherichia coli: dissection of
stability and activity control mechanisms of s32 in vivo. Mol
Microbiol 30: 583±593.
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Heat shock regulation in Streptomyces 1103
Tilly, K., Spence, J., and Georgopoulos, C. (1989) Modulation
of stability of the Escherichia coli heat shock regulatory
factor s32. J Bacteriol 171: 1585±1589.
Yuan, G., and Wong, S.-L. (1995) Isolation and characterization
of Bacillus subtilis groE regulatory mutants: evidence for
orf39 in the dnaK operon as a repressor gene in regulating
the expression of both groE and dnaK. J Bacteriol 177:
6462±6468.
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 1093±1103
Yura, T., and Nakahigashi, K. (1999) Regulation of the heat
shock response. Curr Opin Microbiol 2: 153±158.
Yura, T., Nagai, H., and Mori, H. (1993) Regulation of the
heat shock response in bacteria. Annu Rev Microbiol 47:
321±350.
Zuber, U., and Schumann, W. (1994) CIRCE, a novel heat
shock element involved in regulation of heat shock operon
dnaK of Bacillus subtilis. J Bacteriol 176: 1359±1363.