ClpE, a novel type of HSP100 ATPase, is part of the CtsR heat shock regulon of Bacillus subtilis

Authors

  • Isabelle Derré,

    1. Unité de Biochimie Microbienne, URA 1300 du Centre National de la Recherche Scientifique, Institut Pasteur, 25, rue du Docteur Roux, 75724 Paris Cedex 15, France.,
    Search for more papers by this author
  • Georges Rapoport,

    1. Unité de Biochimie Microbienne, URA 1300 du Centre National de la Recherche Scientifique, Institut Pasteur, 25, rue du Docteur Roux, 75724 Paris Cedex 15, France.,
    Search for more papers by this author
  • Kevin Devine,

    1. Department of Genetics, Trinity College, Lincoln Place Gate, Dublin 2, Ireland.,
    Search for more papers by this author
  • Matthias Rose,

    1. Institute für Mikrobiologie, J. W. Goethe Universität, Marie Curie Strasse 9, Frankfurt am Main 60439, Germany.
    Search for more papers by this author
  • Tarek Msadek

    1. Unité de Biochimie Microbienne, URA 1300 du Centre National de la Recherche Scientifique, Institut Pasteur, 25, rue du Docteur Roux, 75724 Paris Cedex 15, France.,
    Search for more papers by this author

Isabelle Derré. E-mail iderre@pasteur.fr; Tel. (+33) 1 45 68 88 48; Fax (+33) 1 45 68 89 38.

Abstract

Clp ATPases, which include the ubiquitous HSP100 family, are classified according to their structural features and sequence similarities. During the course of the Bacillus subtilis genome sequencing project, we identified a gene encoding a new member of the HSP100 family. We designated this protein ClpE, as it is the prototype of a novel subfamily among the Clp ATPases, and have identified homologues in several bacteria, including Listeria monocytogenes, Enterococcus faecalis, Streptococcus pyogenes, Streptococcus pneumoniae, Lactobacillus sakei and Clostridium acetobutylicum. A unique feature of these Hsp100-type Clp ATPases is their amino-terminal zinc finger motif. Unlike the other class III genes of B. subtilis (clpC and clpP ), clpE does not appear to be required for stress tolerance. Transcriptional analysis revealed two σA-type promoters, expression from which was shown to be inducible by heat shock and puromycin treatment. Investigation of the regulatory mechanism controlling clpE expression indicates that this gene is controlled by CtsR and is thus a member of the class III heat shock genes of B. subtilis. CtsR negatively regulates clpE expression by binding to the promoter region, in which five CtsR binding sites were identified through DNase I footprinting and sequence analysis.

Introduction

Clp ATPases are highly conserved among prokaryotes and eukaryotes and have been classified according to their structural features and sequence similarities. Proteins of the first class, also known as the HSP100 family (ClpA, ClpB, ClpC and ClpD), contain two nucleotide-binding domains (NBDs), whereas the second class of smaller sized proteins have a single NBD (ClpM, ClpN, ClpX and ClpY) (Schirmer et al., 1996). Each class is further subdivided according to specific signature sequence motifs or the length of the interdomain region separating the two NBDs (Schirmer et al., 1996). Many of the Clp proteins (ClpA, ClpX and ClpC) act as the ATPase subunit of the ATP-dependent Clp protease by associating with the ClpP proteolytic subunit to which they confer substrate specificity (Gottesman et al., 1997a,b; Turgayet al., 1998).

In Escherichia coli, the clpB, clpX and clpP genes are σ32-dependent heat shock genes (Kroh and Simon, 1990; Squires et al., 1991; Gottesman et al., 1993; Wojtkowiak et al., 1993). ClpA and ClpX of E. coli have both been shown to act independently of ClpP as molecular chaperones (Wawrzynow et al., 1996), and the Clp protease is involved in the degradation of several phage replication proteins, the RpoS stationary-phase sigma factor as well as SsrA-tagged proteins (Schweder et al., 1996; Gottesman et al., 1997a,b; 1998). The ClpY ATPase associates with a distinct proteolytic subunit, ClpQ, to form the ATP-dependent ClpYQ protease, which plays a role in degrading misfolded proteins (Szabo et al., 1996; Polissi et al., 1998).

In the Gram-positive soil bacterium Bacillus subtilis, the genes encoding ClpC, ClpY, ClpQ, ClpX and ClpP have been characterized (Msadek et al., 1994; 1998; Slack et al., 1995; Gerth et al., 1996). ClpC, ClpX and ClpP are all heat shock proteins that play essential roles in stationary-phase-adaptive responses, such as competence development, motility and sporulation, and are also required for growth at high temperature and stress tolerance (exposure to salt, ethanol or puromycin) (Krüger et al., 1994; Msadek et al., 1994; 1998; Gerth et al., 1996; 1998).

Four classes of heat shock genes have been identified in B. subtilis. Class I genes, which encode classical chaperones such as DnaK, GroES and GroEL, are controlled by the HrcA repressor, which recognizes the highly conserved CIRCE operator sequence (Hecker et al., 1996). Class II genes encode general stress proteins, and their expression requires the σB sigma factor whose synthesis and activity are increased during heat shock, exposure to salt or ethanol, or starvation for glucose, phosphate or oxygen (Hecker and Volker, 1998). Class III genes, which include clpP and the six genes of the clpC operon, are negatively regulated by CtsR, which recognizes a directly repeated heptanucleotide operator sequence (A/GGTCAAA NAN A/GGTCAAA) (Derréet al., 1999). Class IV genes are defined as those devoid of the CIRCE or CtsR operator sequence and whose induction by heat shock or general stress conditions is σB independent. These genes encode clpX (Gerth et al., 1996), ftsH (Deuerling et al., 1997), lonA (Riethdorf et al., 1994), htpG (Schulz et al., 1997), ahpC (Antelmann et al., 1996; Bsat et al., 1996) and trxA (Scharf et al., 1998).

During the course of the B. subtilis genome sequencing project, a gene encoding a new Clp ATPase was identified. We show here that this gene encodes a novel type of HSP100 protein, designated ClpE, with two NBDs and an N-terminal zinc finger motif. We have identified similar proteins in many Gram-positive bacteria including several pathogens. The B. subtilis clpE gene is a class III heat shock gene controlled by CtsR, and its expression is induced by heat shock and puromycin, suggesting that it may play a role either as a chaperone or in the degradation of misfolded proteins.

Results

ClpE defines a novel subclass of HSP100

Within the framework of the European Bacillus subtilis genome sequencing project, the nucleotide sequence of the region upstream from the motA gene was determined, revealing an open reading frame (ORF) designated ykvH. The ATG initiation codon is preceded by a typical ribosome binding site (AAGGAGGT) at an appropriate distance, and the TAA stop codon is followed by a probable rho-independent transcription terminator sequence (AGCGGTTTCCT TTT AGGAAGCCGCT, ΔG = −75 kJ mol−1), followed by a poly(T) stretch. The gene immediately upstream from ykvH is in the opposite orientation, suggesting that ykvH is organized as a monocistronic transcriptional unit.

ykvH encodes a predicted protein of 699 amino acid residues with a calculated molecular mass of 77.8 kDa. Comparison of the deduced amino acid sequence with those of GenBank indicated significant similarities to class I Clp ATPases (ClpA, ClpB, ClpC and ClpD) belonging to the HSP100 family. Nevertheless, it appears to be unique among the HSP100 proteins, as it has an amino-terminal zinc finger motif (C-X2-C-X22-C-X2-C). Although the ClpX ATPases also have an amino-terminal zinc finger motif, they are smaller sized proteins with a single NBD and are not part of the HSP100 family. As this novel Clp ATPase appears to constitute a separate subclass among the HSP100 family, we have designated it ClpE according to established nomenclature (Schirmer et al., 1996).

We identified ClpE homologues in several Gram-positive bacteria by searching unfinished microbial genome databases. Indeed, ClpE of B. subtilis shares strong amino acid sequence identity with ClpE of Listeria monocytogenes (56%), Enterococcus faecalis (54%), Lactococcus lactis (54%), Streptococcus pyogenes (53%), Streptococcus pneumoniae (52%), Clostridium acetobutylicum (47%) and Lactobacillus sakei. Alignment of these proteins (Fig. 1) shows high conservation of the two NBDs with the Walker motifs (A box and B box) (Walker et al., 1982), but also of the zinc finger motif. These features, two NBDs and an amino-terminal zinc finger motif, are characteristic of this new HSP100 subclass. Partially sequenced homologues were also identified in the Streptococcus gordonii and Staphylococcus aureus incomplete genome sequences. The ClpE proteins are slightly smaller than the other HSP100-type ATPases, with a spacer region between the two NBDs that is shorter than that of the closely related ClpC subfamily, 52–54 amino acids in length rather than 61–69.

Figure 1.

. Alignment of the B. subtilis ClpE amino acid sequence (Bsu) (Z99111) with those of Listeria monocytogenes (Lmo) (AF076664), Enterococcus faecalis (Efa), Lactococcus lactis (Lla) (AF023421), Streptococcus pyogenes (Spy), Streptococcus pneumoniae (Spn), Clostridium acetobutylicum (Cac) and Lactobacillus sakei (Lsa) (U82366). Numbers indicate the positions in the amino acid sequence. Identical residues are shaded. The conserved zinc finger motif and nucleotide-binding regions are boxed. Nucleotide-binding (A box, B box) and predicted coiled-coiled motifs are overlined.

A recent report has described the presence of a predicted coiled-coil structural motif in ClpB ATPases (Celerin et al., 1998). Using the Coils2 program (Lupas et al., 1991; Lupas, 1996), we identified a coiled-coil motif within the B. subtilis ClpE interdomain spacer region (positions 330–364) (Fig. 1). This motif is also present in the ClpE proteins of E. faecalis, L. lactis and C. acetobutylicum.

Construction and analysis of a ΔclpE mutant

Mutations in the clpP, clpC and clpX genes have been shown to have pleiotropic phenotypes, affecting competence development, motility, degradative enzyme synthesis, growth at high temperature and sporulation (Msadek et al., 1994; 1998; Gerth et al., 1998). In order to examine the role of ClpE in B. subtilis, a ΔclpE mutant (QB8023) was constructed using plasmid pΔES by chromosomal replacement of the entire clpE coding sequence with the spc spectinomycin resistance gene through a double cross-over event. In contrast to the other clp mutants of B. subtilis, no obvious phenotypes were detected for the ΔclpE mutant.

Indeed, the strain was motile, and competence development and sporulation were not affected. When exponentially growing cells were transferred from 37°C to 54°C, the ΔclpE mutant displayed a significant lag (2–4 h) before resuming growth. However, apart from this lag, the strain grew at the same rate as the otherwise isogenic 168 reference strain. Stationary-phase survival was not affected, nor did the ΔclpE mutant exhibit any stress sensitivity (exposure to 0.8 M NaCl, 5% ethanol or 15 μg ml−1 puromycin). No difference in thermotolerance of the ΔclpE mutant was observed, i.e. the ability to survive otherwise lethal growth temperatures (55°C) after pre-exposing cells to an intermediate temperature that effectively induces the heat shock response (48°C). Although ClpE might play a specific role in B. subtilis, this effect could be masked in the ΔclpE strain by elevated levels of the ClpC or ClpX ATPases under stress conditions.

clpE expression is induced by heat shock and puromycin treatment

clp genes of B. subtilis are known to be stress inducible. Indeed, the expression of clpP, clpC and clpX is induced by heat shock but also by many stress conditions, such as exposure to salt, ethanol or puromycin (Krüger et al., 1994; Gerth et al., 1996; 1998). The expression of clpE was tested under different conditions. Transcriptional fusions were constructed between the upstream region of clpE and either the bgaB gene (encoding the thermostable β-galactosidase of B. stearothermophilus) or the lacZ gene of E. coli. These fusions were introduced into B. subtilis 168 by transformation giving strains QB8012 (clpE′–bgaB) and QB8016 (clpE′–lacZ). Expression of the clpE′–bgaB fusion was monitored under heat shock conditions and that of the clpE′–lacZ fusion during exposure to salt, ethanol or puromycin.

As shown in Fig. 2, clpE was weakly expressed (≈ 80 units mg−1 protein) at 37°C. When the exponentially growing culture was shifted to 48°C, clpE expression was increased 10-fold in 15 min (850 units mg−1 protein), slowly decreasing thereafter to reach ≈ 460 units mg−1 protein 2 h after heat shock. When 15 μg ml−1 puromycin was added to the culture at 37°C, a fourfold induction was observed in 30 min (≈ 330 units mg−1 protein), and clpE expression decreased to about 170 units mg−1 protein 2 h later. No difference in clpE expression was observed when 0.8 M NaCl or 5% ethanol was added to the culture at 37°C (data not shown). These results suggest that clpE is induced by heat shock and puromycin, but not by exposure to NaCl or ethanol.

Figure 2.

. clpE is induced by heat and puromycin. Expression of clpE′–bgaB (strain QB8012) at 37°C (□, ▪) or 48°C (○, •) and clpE′–lacZ (strain QB8016) at 37°C after the addition of 15 μg ml−1 puromycin (▵, ▴). Cells were grown in LB medium at 37°C until approximately OD600 = 1. The culture was then divided, and cells were transferred to 48°C, or puromycin was added. Open symbols indicate optical density at 600 nm, and solid symbols indicate β-galactosidase activity expressed as Miller units mg−1 protein as a function of time before or after heat shock or puromycin treatment.

clpE is controlled by dual σA-dependent heat-inducible promoters

Primer extension experiments were performed to determine the clpE transcription initiation site(s) using total RNA isolated 6 min after heat shock. Three signals were observed located 200 (S1), 81 (S2) and 56 (S3) bp upstream from the translation initiation codon of clpE (Fig. 3A). The nucleotide sequence of the region preceding positions −200 (S1) and −81 (S2) revealed potential −10 and −35 regions sharing similarities with the consensus sequence of promoters recognized by the vegetative form of RNA polymerase, EσA (Fig. 3B). No promoter-type sequences could be identified upstream from position −56 (S3). However, the region immediately upstream from this position contains an inverted repeat sequence (AAGAAGGTC N10 GACCTTTTT), which could form a hairpin loop secondary structure, suggesting that the S3 signal is caused by pausing of reverse transcriptase.

Figure 3.

. A. Primer extension analysis of clpE mRNA. Total RNA (10 μg) isolated from B. subtilis 168 6 min after heat shock was used as a template for reverse transcriptase. The corresponding DNA sequence of plasmid is shown on the left. Potential transcription start sites S1, S2 and S3 and their corresponding positions are indicated by arrows. B. Sequence of the clpE promoter region. Potential −35 and −10 sequences are boxed; transcriptional start points are indicated by S1 and S2; the Shine–Dalgarno sequence and the translational initiation codon are underlined. Positions are numbered relative to the ATG codon. C and D. Expression of clpE′–bgaB: strain QB8014 (amyE::clpEP1′–bgaB) (□, ▪) and QB8015 (amyE::clpEP2′–bgaB) (○, •). Open and solid symbols indicate β-galactosidase activity expressed as Miller units mg−1 protein when cells were grown at 37°C or transferred to 48°C, respectively, at an OD600 of ≈ 1.

To verify that the signals obtained by primer extension correspond to in vivo transcription initiation sites, the upstream region corresponding to each signal was transcriptionally fused to the bgaB gene. Integration of these fusions, designated P1, P2 and P3, at the amyE locus of B. subtilis 168 led to strains QB8014, QB8015 and QB8024 respectively. Expression of the different fusions was monitored during growth at 37°C or 48°C. No expression was observed for the P3 fusion, indicating that the S3 signal is very probably the result of pausing of the reverse transcriptase. The expression pattern obtained with P1 and P2 bgaB fusions (Fig. 3C and D) was the same as for the clpE′–bgaB fusion containing both promoters (Fig. 2), but the two promoters were differentially expressed. Expression from the P1 promoter gives a basal level of ≈ 70 units mg−1 protein at 37°C and is increased ≈ 13-fold by heat shock (940 units mg−1 protein), whereas the basal expression level of the P2 promoter was 12 units mg−1 protein at 37°C and increased sixfold (70 units mg−1 protein) after heat shock (Fig. 3C and D).

The expression of clpE thus involves dual heat-inducible σA-dependent promoters, which are differentially expressed.

clpE is expressed transiently at 48°C

Expression of clpE′–bgaB after heat shock suggested transient expression at 48°C (Fig. 2). Primer extension experiments were performed on total RNA isolated before and at different times after heat shock. As shown in Fig. 4, clpE mRNA levels were not detectable at 37°C but increased greatly during the first 6 min after transfer to 48°C and then decreased rapidly, with no signal remaining after 12 min.

Figure 4.

. clpE is transiently expressed. Total RNA (10 μg) isolated from B. subtilis 168 grown at 37°C (0 min) and at the times indicated after transfer to 48°C (2, 4, 6, 8, 10 and 12 min) was used as a template for reverse transcriptase. The transcriptional start sites S1 and S2 are indicated by arrows.

These results confirm that, although the expression of clpE from both σA-type promoters is induced by heat shock, this expression is transient and rapidly shut off during prolonged exposure to elevated temperatures.

CtsR regulates clpE expression negatively

Four classes of heat shock genes have been distinguished in B. subtilis, with the first three defined as those whose expression is controlled by HrcA, σB or CtsR respectively. Genes in the fourth class are not controlled by these three mechanisms, but by one or several as yet uncharacterized regulatory modes.

Sequence analysis indicates that the clpE promoter region does not contain the CIRCE operator, which is the binding site for the HrcA repressor. No σB-type promoter recognition sequences were detected, and a sigB deletion had no effect on expression from the P1 or P2 promoters at 37°C or 48°C, as determined by bgaB transcriptional fusions and primer extension experiments (data not shown). However, several potential CtsR binding sites were identified in the sequence of the clpE upstream region, suggesting that ClpE synthesis may be controlled by CtsR, as is that of ClpC and ClpP (Derréet al., 1999)

Expression of the clpE′–bgaB transcriptional fusion was compared in the ΔctsR mutant (strain QB8013) and the otherwise isogenic 168 derivative reference strain, QB8012, during growth at 37°C or 48°C. As shown in Fig. 5, clpE expression was increased 125-fold (≈ 10 000 units mg−1 protein) in the ΔctsR mutant at 37°C. Furthermore, an additional 2.4-fold increase was observed when the corresponding ΔctsR mutant was grown at 48°C. Both the P1 and P2 promoters are controlled by CtsR, as similar effects were seen when the transcriptional bgaB fusions with each separate promoter were placed in the ΔctsR mutant (strains QB8014 and QB8017, QB8015 and QB8018). Expression from the P1 promoter was increased 140-fold at 37°C in the ΔctsR mutant, with a further 2.6-fold induction at 48°C (data not shown). Expression from the P2 promoter was enhanced 50-fold at 37°C in the absence of CtsR, with an additional 3.6-fold increase at 48°C (data not shown).

Figure 5.

. clpE is negatively regulated by CtsR. Expression of clpE′–bgaB: strain QB8012 (amyE::clpE′–bgaB) at 37°C (□, ▪), QB8013 (amyE::clpE′–bgaBΔctsR) at 37°C (○, •) and after transfer to 48°C (▵, ▴). Open symbols indicate optical density at 600 nm, and solid symbols indicate β-galactosidase activity expressed as Miller units mg−1 protein as a function of time before or after heat shock.

These results indicate that CtsR negatively regulates clpE, leading to heat-inducible expression from both σA-type promoters.

CtsR binds to the clpE promoter region

Previous experiments have shown that CtsR represses the expression of clpC and clpP by binding directly to the highly conserved direct repeat A/GGTCAAA NAN A/GGTCAAA (derréet al., 1999). CtsR binding to the clpE promoter region was tested by gel mobility shift DNA binding assays. A radiolabelled, PCR-generated DNA fragment corresponding to positions −279 to +64 relative to the translation initiation codon of clpE was incubated with increasing amounts of purified CtsR. All DNA-binding assays were performed in the presence of an excess of non-specific competitor DNA [1 μg of poly-(dI–dC)]. As shown in Fig. 6, CtsR bound specifically to the radiolabelled fragment, progressively forming five complexes with a complete displacement obtained at the highest CtsR concentration. Cold competitor chase experiments were performed by adding increasing amounts of the unlabelled clpE DNA fragment, preventing DNA binding of the radiolabelled probe (data not shown). These results suggest at least five binding sites for CtsR in the clpE promoter region.

Figure 6.

. CtsR binds specifically to the clpE promoter region. DNA binding reactions were performed with radiolabelled DNA fragments corresponding to the clpE promoter region. Lanes 1–8, 10 000 c.p.m. of clpE labelled fragment and increasing amounts of purified CtsR. Lane 1, no protein; lane 2, 0.015 pmol; lane 3, 0.03 pmol; lane 4, 0.06 pmol; lane 5, 0.12 pmol; lane 6, 0.24 pmol; lane 7, 0.48 pmol; lane 8, 0.96 pmol.

DNase I footprinting assays were performed on a DNA fragment carrying the clpE promoter region to determine the location of the CtsR binding sites precisely. When the non-template strand was end-labelled, CtsR protected five distinct regions extending from positions −232 to −209, −205 to −181, −161 to −138, −86 to −63 and −40 to −16 (Fig. 7A and C). The template strand was protected from positions −237 to −213, −210 to −186, −165 to −142, −90 to −67 and −45 to −21 (Fig. 7B and C). Positions are given relative to the translation initiation codon.

Figure 7.

. DNase I footprinting analysis of CtsR binding to the clpE promoter region. A and B. Lanes contain 0.154 pmol of labelled non-template strand (A) and 0.124 pmol of template strand (B) of clpE. Fragments were incubated with increasing amounts of purified CtsR: lane 1, no protein; lane 2, 0.3 pmol; lane 3, 0.6 pmol; lane 4, 1.2 pmol; lane 5, 2.4 pmol; lane 6, 4.8 pmol. A + G Maxam and Gilbert reactions of the appropriate DNA fragments were loaded. Regions protected by CtsR are indicated by brackets. C. Extent of the DNase I protected regions. The DNase I protected areas are shaded, and the CtsR direct repeat recognition sequences are indicated by arrows. Positions are numbered relative to the translation initiation codon.

Five sites within the clpE upstream region are protected by CtsR from DNase I cleavage (Fig. 7A and B), in agreement with the five distinct protein–DNA complexes observed in the gel mobility shift assay (Fig. 6). Three of these overlap the −35 and −10 sequences of the P1 promoter and the transcription start sites of the P1 and P2 promoters, while the remaining two are located between the P1 and P2 promoters and downstream from the P2 promoter (Fig. 7C). Each of the protected regions contains a CtsR direct repeat recognition sequence (Fig. 7C). These data indicate that CtsR binds to five target sites within the clpE upstream region, effectively repressing expression from both the P1 and P2 promoters. The clpE gene is thus a member of the B. subtilis class III heat shock genes, controlled by CtsR, which also include clpP and the genes of the clpC operon.

Discussion

Clp ATPases, which include the HSP100 family, are distinguished by their structural organization and consensus sequence features into two classes and eight subclasses (Schirmer et al., 1996). During the B. subtilis genome sequencing project, a new clp gene was identified. Analysis of the encoded protein allowed us to describe a novel subclass of HSP100 ATPase characterized by two NBDs and an amino-terminal zinc finger motif. We designated this new type of HSP100 ClpE according to the previously established nomenclature (Schirmer et al., 1996).

ClpE homologues were also identified in several Gram-positive bacteria: L. monocytogenes, E. faecalis, L. lactis (H. Ingmer, personal communication), S. pyogenes, S. pneumoniae, C. acetobutylicum and L. sakei. No ClpE-type HSP100 was identified in Gram-negative bacteria, suggesting that the ClpE subfamily might be specific to Gram-positive bacteria.

Transcription analysis indicates that clpE has two tandemly located σA-dependent promoters and that expression of the gene is strongly induced by heat shock as well as by treatment with puromycin. clpE is part of the CtsR regulon, as its expression is greatly increased in the ΔctsR mutant, and purified CtsR binds to five target sites in the clpE promoter region. Regulation by CtsR seems to be conserved for the clpE genes, as CtsR binding sites were identified in their promoter region (Derréet al., 1999) in every case, except for clpE of C. acetobutylicum. In the case of S. pneumoniae, the available nucleotide sequence of the clpE gene is incomplete, lacking the beginning of the gene as well as the upstream region.

Among the class III genes of B. subtilis, clpE is the most tightly controlled by CtsR. Indeed, the promoter region contains five CtsR binding sites, most of which overlap the −35 and −10 sequences or the transcription start sites of the two σA-dependent promoters. This could explain the very low basal level of expression at 37°C and the transient induction by heat shock. Indeed, low levels of CtsR at 37°C are very probably sufficient to extinguish clpE expression completely given the number of binding sites, which presumably favour co-operative binding of CtsR. As ctsR is the first gene of the clpC operon, levels of the repressor will rise rapidly under heat shock conditions. Even if most of the repressor is inactivated and/or degraded at 48°C, the remaining levels could be sufficient to repress clpE expression after the initial heat shock, thus explaining the transient expression pattern observed at 48°C. This could also explain why no obvious phenotype was detected for the ΔclpE mutant, as the gene is presumably repressed most of the time and only induced under extreme conditions.

Expression of clpE is greatly increased at 37°C in the ΔctsR mutant. However, as shown previously for the clpP and clpC genes (derréet al., 1999), a significant residual induction of clpE expression was observed in the ΔctsR mutant at 48°C (Fig. 5). This could suggest that some additional mechanism, besides CtsR-mediated repression, could be involved in controlling clp gene expression in response to heat shock. Expression of clpP, clpE and the clpC operon were not modified in the ΔclpE mutant at 37°C or 48°C (data not shown).

Two HSP100 ATPases of the ClpE family have been shown to play a role in bacterial virulence. In S. pneumoniae, the gene was identified during a large-scale identification of virulence genes using the signature-tagged mutagenesis technique (Polissi et al., 1998). Virulence of the S. pneumoniae mutant in which the clpE gene is inactivated was significantly affected, as shown using a mouse septicaemia model (Polissi et al., 1998). In L. monocytogenes, in which the ClpC ATPase has previously been shown to play a role in virulence (Rouquette et al., 1996; 1998), a clpE gene was identified recently (Nair et al., 1999). Expression of clpE in L. monocytogenes was greatly increased in the absence of ClpC, and ClpE was shown to act synergistically with ClpC in controlling cell division and virulence (Nair et al., 1999). The negative role of ClpC in controlling clpE expression in L. monocytogenes could act through CtsR. Indeed, we have shown previously that a mutation inactivating the clpC gene of B. subtilis leads to an increase in the expression of clpP, which is CtsR dependent, suggesting a chaperone role for ClpC in controlling CtsR activity (Derréet al., 1999). However, the expression of clpE in B. subtilis was not modified in the ΔclpC mutant (data not shown). This could result from the large number of CtsR binding sites in the clpE regulatory region, as small amounts of active CtsR in the ΔclpC mutant would be sufficient to repress clpE expression.

In contrast to the ΔclpC, ΔclpX and ΔclpP mutants of B. subtilis, which are defective in cell division, stress tolerance and many stationary-phase-adaptive responses such as competence and sporulation (Krüger et al., 1994; Msadek et al., 1994; 1998; Gerth et al., 1996; 1998), the ΔclpE mutant has no obvious phenotype. Although this does not exclude ClpE interacting with ClpP, it could suggest that ClpE, like ClpB, might also act independently of ClpP, perhaps as a molecular chaperone. Indeed, as for ClpB (Celerin et al., 1998), a coiled-coil motif was identified in ClpE, and it has been suggested that this domain might prevent protein–protein interaction with ClpP (Celerin et al., 1998). However, we have also identified a coiled-coil motif in the B. subtilis ClpC protein, which has recently been shown to interact with ClpP (Turgay et al., 1998). Synthesis of ClpE is induced by both heat shock and treatment with puromycin, conditions that favour the accumulation of truncated and misfolded proteins, which supports the idea that ClpE may act as a chaperone. The amino-terminal zinc finger domain of ClpE could be involved in binding to misfolded proteins. Indeed, although initially characterized as a domain associated with DNA binding, there is increasing evidence that zinc finger motifs also play an important role in protein–protein interaction (Berg, 1990; Mackay and Crossley, 1998). In the case of the DnaJ chaperone, the zinc finger domain has been suggested as playing two roles: binding to denatured protein substrates (Szabo et al., 1996) and disulphide bond formation, reduction and isomerization (de Crouy-Chanel et al., 1995).

ClpE is thus part of a novel HSP100 subfamily, which appears to be specific to Gram-positive bacteria. Two members of this family play an important role in bacterial virulence, and we speculate that the B. subtilis ClpE ATPase may act as a molecular chaperone, as it does not appear to play a role similar to that of ClpP or the ClpX and ClpC ATPases. We have shown that synthesis of ClpE is CtsR dependent in B. subtilis and identified likely CtsR binding sites in front of most of the clpE genes. An understanding of the regulatory mechanism(s) controlling clp gene expression will probably prove essential in studying the virulence of Gram-positive pathogens.

Experimental procedures

Bacterial strains, growth conditions and transformation

The B. subtilis strains used in this work are listed in Table 1. E. coli K-12 strain TG1 (Δ(lac proAB) supE thi hsdΔ5 ( F′traD36 proAB lacI q lacZΔM15)) (Gibson, 1984) was used for cloning experiments.

Table 1. . B. subtilis strains used in this study.Thumbnail image of

E. coli was grown in Luria–Bertani (LB) medium. Electroporation procedures were used for E. coli transformation with selection on LB plates supplemented with ampicillin (100 μg ml−1) or spectinomycin (100 μg ml−1).

B. subtilis was grown in LB medium and transformed as described previously using plasmid or chromosomal DNA (Msadek et al., 1998). Transformants were selected on SP plates supplemented with chloramphenicol (5 μg ml−1), erythromycin (25 μg ml−1) or spectinomycin (100 μg ml−1).

β-Galactosidase activity was estimated on plates by 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (Xgal) hydrolysis (100 μg ml−1). β-Galactosidase specific activities were determined as described previously (Miller, 1972; Msadek et al., 1990; 1998) and expressed as Miller units mg−1 protein.

DNA manipulations

Standard procedures were used to extract plasmids from E. coli (Amory et al., 1987; Sambrook et al., 1989). Restriction enzymes were used as recommended by the manufacturers. DNA fragments were amplified by the polymerase chain reaction (PCR) technique (Mullis and Faloona, 1987; Saiki et al., 1988) as described previously (Msadek et al., 1991).

DNA sequences were determined by the dideoxy chain termination method (Sanger et al., 1977) using modified T7 DNA polymerase (Tabor and Richardson, 1987) (Pharmacia).

Plasmids and plasmid constructions

Plasmids used in this study are listed in Table 2. Plasmids pDL (Yuan and Wong, 1995) and pDH32 (Perego, 1993) were used for constructing transcriptional fusions with the B. stearothermophilus bgaB or E. coli lacZ genes, respectively, with subsequent integration at the amyE locus. The S. aureus spc gene (spectinomycin resistance) (Murphy, 1985) and plasmid pHT181, a derivative of pUC18 carrying the Tn1545 erythromycin resistance gene (Lereclus and Arantes, 1992), were used for gene deletion/replacement.

Table 2. . Plasmids used in this study.Thumbnail image of

clpE′–bgaB transcriptional fusions were constructed using EcoRI/BamHI DNA fragments corresponding to clpE upstream regions generated by PCR using the following oligonucleotide pairs: TM-212 (5′-GAAGAATTCCCTCCCAGTCGCAAATCC-3′) and TM-213 (5′-GGAGGATCCTTTGACCAATGAAAGTG-3′) (−451 to −31); TM-276 (5′-GAAGAATTCATGTT-

TTACTTTGTAACAAATC-3′) and TM-274 (5′-GGAGGATCCGGCAAACTTAACGGCATTCAAACC-3′) (−279 to −94); TM-273 (5′-GAAGAATTCAGTCTGCTTTGAGCATATTGG-3′) and TM-275 (5′-GGAGGATCCGGAATTTATTTGCATGTTAAGGCG-3′) (−141 to +64); ID40 (5′-GAAGAATTCGTTTGCCGTATACTAATAGTC-3′) and TM-275 (5′-GGAGGAT-

CCGGAATTTATTTGCATGTTAAGGCG-3′) (−101 to +64). Positions are given relative to the translation initiation codon. These fragments were cloned between the EcoRI/BamHI sites of plasmid pDL to give plasmids pClp11/4, pDL274/276, pDL273/275 and pDL40/275 respectively. The fragment generated with oligonucleotides TM-212 and TM-213 was also cloned between the EcoRI/BamHI sites of plasmid pDH32 to give plasmid pDH212/213. Linearization of these plasmids at the unique PstI site and transformation of B. subtilis 168 with selection for chloramphenicol resistance yielded strains QB8012, QB8014, QB8015, QB8024 and QB8016, in which the clpE′–bgaB or the clpE–lacZ fusion was integrated as a single copy at the amyE locus. Strains QB8013, QB8017 and QB8018 were constructed by introduction of the linearized plasmids into strain QB4991 (ΔctsR).

The pMapE plasmid was constructed by cloning a 343 bp DNA fragment corresponding to the clpE upstream region generated by PCR using oligonucleotides TM-276 and TM-275 between the EcoRI and BamHI sites of plasmid pHT181.

A deletion/replacement mutant of clpE was constructed by cloning a 334 bp EcoRI/BamHI DNA fragment, a 1.3 kb BamHI/KpnI DNA fragment carrying the spc spectinomycin resistance gene and a 303 bp KpnI/HindIII DNA fragment between the EcoRI and HindIII sites of plasmid pHT181 to give plasmid pΔES. The EcoRI/BamHI DNA fragment and the BamHI/KpnI DNA fragments correspond to the chromosomal DNA regions immediately upstream and downstream, respectively, from the clpE gene and were generated by PCR using oligonucleotide pairs TM-256 (5′-GAAGAATTCTTCCCTCCGAACGAACCGC-3′) and TM-213 (5′-GGAGGATCCTTTGACCAATGAAAGTG-3′); and TM-214 (5′-GGTG-

GTACCGTCCCTTCTCCTGAAG-3′) and TM-270 (5′-AAGAAGCTTTTTTACTAGCTTGTCTATGG-3′). Plasmid pΔES was linearized at the unique ScaI site and used to transform B. subtilis 168. SpectinomycinR/erythromycinS integrants arose through a double cross-over event in which the entire clpE coding sequence was deleted and replaced by the spc spectinomycin resistance gene. The chromosomal replacement was verified by PCR, and the corresponding ΔclpE::spc mutant was designated QB8023.

RNA extraction

Overnight cultures of B. subtilis strains were diluted 100-fold in LB medium and grown at 37°C or 48°C with vigorous aeration. Aliquots were collected (10 ml), and cells were pelleted and frozen immediately. Frozen cells were resuspended in 0.5 ml of water and disrupted with a FastPrep disintegrator (30 s at 4°C) using 0.5 g of glass beads (106 μm; Sigma) in the presence of 0.4 ml of 4% Macaloid and 0.5 ml of phenol–chloroform–isoamyl alcohol, pH 8.0 (Amresco). After centrifugation for 2 min at 14 000 r.p.m., supernatants were extracted with phenol–chloroform (1:1, v/v) and then with chloroform–isoamyl alcohol (24:1, v/v). RNA was precipitated with isopropanol in the presence of 0.2 M NaCl and resuspended in 20 μl of water. RNA concentrations were determined by measuring absorbance at 260 nm, and samples were stored at −20°C.

Primer extension

A synthetic oligodeoxynucleotide (ID28: 5′-CGCATTTGCCAAAACCTCCTTAATAATTTGATC-3′) was 5′ end-labelled with [γ-32P]-ATP (110 TBq mmol−1) using T4 polynucleotide kinase. RNA (40 μg) and 1 pmol of labelled oligonucleotide were annealed in a total volume of 18 μl of reverse transcriptase buffer (50 mM Tris-HCl, 8 mM MgCl2, 30 mM KCl, 1 mM dithiothreitol, pH 8.5). The mixture was incubated for 3 min at 65°C and then cooled slowly to room temperature. Aliquots of 1 μl (25 units) of avian myeloblastosis virus (AMV; Boehringer Mannheim) reverse transcriptase and 1 μl of a solution of all four dNTPs (20 mM each) were then added. After 30 min of incubation at 42°C, reactions were stopped by adding 5 μl of a solution containing 97.5% deionized formamide, 10 mM EDTA, 0.3% xylene cyanol and 3.3% bromophenol blue.

Gel mobility shift DNA-binding assays

An EcoRI/BamHI DNA fragment, corresponding to the promoter region of the clpE gene (−279 to +64) was generated by PCR using oligonucleotides TM-276 and TM-275. Radiolabelling, DNA binding and gel shift experiments were performed as described previously (Derréet al., 1999).

DNase I footprinting

DNA fragments used for DNase I footprinting (−288 to +64) were prepared by PCR using oligonucleotides ID-34 (5′-CGAAGAAAACATGTTTTACTTTGTAAC-3′) and ID-37 (5′-CGGAATTTATTTGCATGTTAAGGCG-3′). Labelling and DNase I treatment were performed as described previously (derréet al., 1999).

Database comparisons and sequence analysis

Computer analyses were performed using the GCG sequence analysis software package (version 9.1, Genetics Computer Group, Madison, WI, USA) and multiple alignments with the CLUSTAL V program (Higgins et al., 1992). Sequence comparisons with the GenBank database were accomplished using the National Center for Biotechnology Information BLAST2 (Altschul et al., 1997) network service with the default parameter values provided. Unfinished bacterial genome sequences were from the following sources: S. pneumoniae, E. faecalis, S. aureus, S. gordonii: http://www.tigr.org; S. pyogenes http://www.genome.ou.edu; C. acetobutylicum: http://www.cric.com.

Nucleotide sequence accession number

The nucleotide sequence data reported in this paper has been deposited in the GenBank database (accession no. Z99111).

Acknowledgements

We thank H. Agaisse for helpful discussions, S. Nair and P. Berche for providing the L. monocytogenes clpE nucleotide sequence, H. Ingmer for the L. lactis clpE nucleotide and protein sequences, and J. Bignon for excellent technical assistance. This work was supported by research funds from the Institut Pasteur, Centre National de Recherche Scientifique, Université Paris 7 and the Biotechnology programme of the European Commission (contract number BIO4-CT96-0655). I.D. was the recipient of a fellowship from the Ministère de l'Education Nationale de la Recherche et de la Technologie.

Ancillary

Advertisement