Introduction

Chaperone proteins form a multiprotein network that governs protein folding, refolding, transport, degradation and regulation (Ellis 1999; Hartl and Hayer-Hartl 2002). Proper functioning of the network is particularly important under stress conditions, such as heat shock, but chaperones are also essential for normal viability of cells under physiological conditions. In the model organism Escherichia coli, the Hsp70 chaperone machinery, one of the major elements of the network, is composed of the DnaK(Hsp70)[K], DnaJ(Hsp40)[J] and GrpE(Hsp23)[E] proteins (reviewed by Hartl and Hayer-Hartl 2002; Bukau and Horwich 1998; Houry 2001). DnaK consists of an approximately 44-kDa NH2-terminal ATPase domain and a 27-kDa COOH-terminal substrate-binding domain. The latter is divided into a subdomain with a peptide-binding cleft and an α-helical latch-like segment (Zhu et al. 1996). DnaK has weak ATPase activity, and it cycles through ATP- and ADP-bound stages, with its affinity for substrate (e.g., a partially denatured protein or a nascent peptide) being lower in the ATP-bound stage than in the ADP one (Żylicz et al. 1983; Liberek et al. 1991; Palleros et al. 1993; Buchberger et al. 1994; McCarty et al. 1995). In the former, the α-helical latch over the peptide-binding cleft is in an open conformation and rapid peptide binding occurs. Stable holding of peptide involves closing of the latch, a conformational change that is achieved by hydrolysis of bound ATP to ADP. The cycling of DnaK between these states is regulated by the co-chaperones DnaJ and GrpE (the latter functions as a homodimer). The NH2-terminal J domain binds to DnaK and accelerates hydrolysis of ATP by DnaK, thus facilitating peptide capture (Liberek et al. 1991; Karzai and McMacken 1996; Laufen et al. 1999; Mayer et al. 2000). The COOH-terminal domain of DnaJ functions as a chaperone in recognizing hydrophobic peptides and thus can recruit DnaK to substrates (Langer et al. 1992; Rudiger et al. 1999; Sha et al. 2000). GrpE induces release of ADP from DnaK, and upon rebinding of ATP the DnaK-peptide complex dissociates, completing the reaction cycle (Liberek et al. 1991; McCarty et al. 1995; Szabo et al. 1994; Jordan and McMacken 1995).

The KJE chaperone proteins are very well conserved in evolution in terms of structure and general mechanism of function (Gupta et al. 1997; Gribaldo et al. 1999; Hartl and Hayer-Hartl 2002), as documented by an impressive amount of experimental evidence. By contrast, little is known about the functional differences among the homologous chaperones from different species—especially regarding their species specificity. It has been shown that purified Hsp70 (DnaK) proteins other than DnaK of E. coli possess different affinities for model peptides, which might cause differences in substrate specificity in a cell (Levy et al. 1995; Jamsen et al. 1997; Silberg et al. 1998; Kluck et al. 2002). Despite these differences, most of the studied Hsp70 systems, representing all major groups of organisms, are able to refold thermally or chemically denatured firefly luciferase, which thus serves as a model protein to assay chaperone activity in vitro (Schröder et al. 1993; Terada et al. 1997; Lu and Cyr 1998; Żmijewski et al. 2004).

Vibrio harveyi is a free-living gram-negative γ purple bacterium found in diverse marine environments, in various geographic locations, at depths between surface and about 1000 m (Ruby et al. 1980; Baumann and Schubert 1984). Its use as a bioindicator of environmental pollution is under current investigation, especially for monitoring the levels of mutagenic substances. It has been shown that V. harveyi cells are highly sensitive to various mutagens and hypersensitive strains have been constructed for use in a modified Ames’ test (Czyż et al. 2002; Wegrzyn and Czyż 2003). However, non-mutagenic toxic substances would not be detected in this type of a test. Heat-shock proteins have been proposed as general markers of cellular aggression and their use for environmental monitoring has often been suggested (Sanders and Martin 1993; Van Dyk et al. 1994; Ait-Aissa et al. 2000; Czyż et al. 2002), due to the fact that the heat-shock genes are up-regulated while cells are subjected to various stressors, such as heavy metals, oxidizing agents, organic solvents or UV irradiation. But in order to use V. harveyi heat-shock proteins as stress markers, a thorough knowledge of the heat-shock response in this bacterium is necessary.

Although V. harveyi is not pathogenic for humans, its close relatives, Vibrio cholerae, Vibrio parahaemolyticus and Vibrio vulniformis, cause devastating diseases; thus, V. harveyi is an important, very good and safe model for studying the biology of the Vibrio genus. It was shown for V. cholerae and V parahaemolyticus that heat-shock proteins may regulate synthesis of a pathogenic toxin (Chakrabarti et al. 1999; Wong et al. 2002); hence, studies on Vibrio heat-shock response seem highly pertinent. Moreover, V. harveyi is a model organism to study luminescence, DNA repair and bacterial quorum sensing (Czyż et al. 2002; Mok et al. 2003). Participation of chaperone proteins in regulation of both processes has been suggested, which again points to the necessity of heat-shock investigations in V. harveyi.

Based on the reasons outlined above, we studied the heat-shock response of V. harveyi, which led to identification of the major heat-shock proteins DnaK, DnaJ, GroEL, GroES, and IbpA/B (Klein et al. 1995, 2001), followed by cloning and characterization of their genes, including dnaK–dnaJ operon (Klein et al. 1998; Kuchanny et al. 1998; Kuchanny-Ardigo and Lipinska 2003). We found that the V. harveyi dnaK–dnaJ operon restored the thermoresistant phenotype of a E. coli dnaJ mutant but did not suppress the thermosensitivity of E. coli dnaK at 42°C. These results suggested that the DnaK chaperone may be more species-specific than the DnaJ co-chaperone.

In this work we provide further information regarding the DnaK–DnaJ–GrpE chaperone machinery of V. harveyi with respect to the species specificity of the DnaK chaperone. We presented sequence of the entire grpE–gltP–dnaK–dnaJ region and showed that V. harveyi DnaK, DnaJ and GrpE efficiently cooperate with the members of the E. coli DnaK–DnaJ–GrpE system in vitro. In addition, we have analyzed the suppression of mutations in dnaK, dnaJ and grpE of E. coli using homologous genes of V. harveyi .

Materials and methods

Bacterial strains, phage, plasmids, media, and reagents

The bacterial strains, bacteriophage, and plasmids used in this study are listed in Table 1. Luria–Bertani (L–B) medium and Luria agar (LA) were prepared according to the method of Sambrook et al. (1989) and were supplemented with appropriate antibiotics (when necessary): 100 μg ampicillin/ml , 68 μg chloramphenicol/ml , 10 μg tetracycline/ml and 30 μg kanamycin/ml. T-agar plates and T-top agar for phage titration were prepared according to a standard procedure (Sambrook et al. 1989). Boss medium and agar were prepared as described (Klein et al. 1998). Primers for DNA amplification and sequencing were purchased from Interactiva and all enzymes for DNA manipulations were from Fermentas. All other reagents were purchased from Sigma or as indicated in the text. Molecular biology methods were done according to Sambrook et al. (1989).

Table 1 Bacterial strains, bacteriophages and plasmids used in this study
Full size table

Subcloning of dnaK Vh and dnaJ Vh

The 1,250-bp XapI DNA fragment of pGK1 dnaK Vh dnaJ Vh plasmid, containing dnaK Vh , was cloned into the EcoRI site of the pGEM Zf3 + vector. Subsequently, two plasmids, pMZ20 and pMZ33, which differed in orientation of the insert, were isolated. Orientation of the insert was determined by restriction analysis and sequencing of DNA. pMZ20 contained dnaK Vh under the control of the phage T7 ϕ10 promoter, while pMZ33 contained a lacZ’–dnaK Vh fusion. Both plasmids were used in in-vivo complementation experiments, and plasmid pMZ20 was used for overproduction of DnaK Vh , according to Tabor and Richardson’s system (1985). V. harveyi dnaJ coding sequence was amplified by PCR using V. harveyi chromosomal DNA as a template and primers: DnaJVhfor 5′-GGAATTCGATACAAACACCAAGCGG-3′ and DnaJVhrev 5′-CCTGCAGGTCGACTGGATTCAACTAACCAGC-3′. The primers were designed based on the known sequence of the dnaKdnaJ operon of V. harveyi (Klein et al. 1998). Both primers had artificially designed EcoRI or HindIII restriction enzyme digestion sites (underlined). The PCR fragment was digested with EcoRI and HindIII and ligated into corresponding sites of the pT7-5 vector plasmid. The quality of the resulting construct (pMZ1) was confirmed by restriction analysis and comparison to the known DNA sequence [AF055368]. pMZ1 was subsequently used for overproduction of DnaJ Vh using Tabor and Richardson’s system (1985).

Cloning and sequencing of grpE Vh

The gene coding GrpE Vh was amplified using PCR with chromosomal DNA of V. harveyi as template and the primers grpEVhfor 5′-TGAATTCTGGAGATATCATGAGCAACCAA-3′, containing the Shine-Dalgarno sequence and a start (ATG) codon, and grpEVhrev 5′-TTCTGCAGAATTATTTGGCAACCATCACCA-3′, containing a stop (TAA) codon and part of the intergenic (grpE–gltP) fragment. Sequences of the primers were designed based on the sequences of grpE genes of Vibrio proteolyticus [AF218211] and Vibrio cholerae [AF019558]. Additionally, sequences recognized by restrictions enzymes EcoRI or PstI were introduced into the primers (underlined). Restriction analysis of the PCR product (∼650 bp) revealed that the coding sequence of the V. harveyi grpE, in contrast to the grpE genes of V. proteolyticus and V. cholerae, contained additional sites for both restriction enzymes (EcoRI and PstI), so that these sites could not be used. Finally, grpE of V. harveyi was cloned into the SmaI site of pUC18 after creating blunt ends of the PCR product by Klenow polymerase, resulting in the plasmid pMZ40. DNA of grpE Vh was sequenced; the sequence was deposited in GenBank [AY193779], and was later resubmitted as a part of the grpE–gltP–dnaK–dnaJ gene cluster [AY639008].

Analysis of the grpE–gltP–dnaK–dnaJ gene cluster composition and the gltP sequence

In order to determine the position of grpE with respect to the dnaK–dnaJ operon on the V. harveyi chromosome, a DNA fragment containing grpE and the 5′-region of gltP was amplified by PCR using the chromosomal DNA of V. harveyi as template and the primers: grpEVhfor (as above) and gltPVhI 5′-CCTTAAGCTTCTCAGTTAGCGTGC-3′. The resulting PCR fragment of 2,300 bp was sequenced using primers: gltPVhI (as above), gltPVhII 5′-CAGAGCAAATACACCGTAAGG-3′ and gltPVhIII 5′-CCAGATCACGAATCAAACACAG-3′. Combining the resulting sequence with those of grpE Vh [AY193779] and the dnaK Vh –dnaJ Vh operon [AF055368] revealed a putative gltP gene between grpE Vh and dnaK Vh . A new database entry was created for the entire grpE Vh –gltP Vh –dnaK Vh –dnaJ Vh gene cluster [AY639008]. The sequences of grpE and gltP of V. harveyi were analyzed using the Internet programs: blastn, blastp, genome blast and blast 2 sequences on the NCBI server (www.ncbi.nih.gov). Multisequence comparisons were done using the ClustalW program on EBI server (www.ebi.ac.uk).

Suppression of the E. coli mutations by cloned V. harveyi genes

Mutants of E. coli: BM271 ΔdnaK, CG1000 ΔdnaJ, WKG190 ΔcbpAdnaJ and DA259 ΔgrpE were transformed with plasmids carrying cloned V. harveyi: dnaK, dnaJ, dnaKdnaJ, or grpE genes and suppression of thermosensitive phenotype and bacteriophage λ growth were tested as described for the archaeon Methanosarcina mazei genes (Żmijewski et al. 2004). The dnaK Vh and dnaJ Vh genes from pMZ20 and pMZ1, respectively, were expressed with the help of pGP1-2 plasmid, encoding bacteriophage T7 polymerase (Tabor and Richardson 1985). Briefly, for the suppression of the thermosensitive phenotype of E. coli mutants, bacterial suspensions in 10 mM MgSO4 (Sambrook et al. 1989) were prepared from appropriate transformants; 2 μl of serial dilutions from 10−1 to 10−6 of the suspensions were spotted onto LA plates and incubated at 30, 37, 42 and 45°C for 18 h. For the bacteriophage λ-tests, 3 ml of T-top agar (Sambrook et al. 1989) containing 100 μl of bacterial suspension in 10 mM MgSO4, prepared as described above, was plated on a T-agar plate and allowed to dry. Two microliters of serial dilutions (10−1 to 10−9 in TM buffer) of bacteriophage λ cIb2 [phage titer was at least 109 plaque-forming units (pfu) per ml] were spotted on the plates. Plates were incubated overnight at 30°C, plaques were counted, and the efficiency of plating (e.o.p.) was calculated as the ratio of pfu on the tested strain to the pfu on the wild-type reference strain (E. coli B178). All in vivo experiments were repeated at least three times.

Determination of DnaK Vh overproduction in E. coli

E. coli B178 (wt) and E. coli BM271 (ΔdnaK) were transformed with plasmids pGK1 dnaK Vh dnaJ Vh ; pMZ33 lacZ′–dnaK Vh ; pMZ20 dnaK Vh and pGP1-2. In order to estimate the amount of DnaK Vh overproduction in transformed cells, overnight cultures were diluted (1:50) and incubated in L–B medium supplemented with proper antibiotic(s) at temperature of 30°C. At an optical density (OD) of 0.6, each culture was split: one part remained at 30°C while the other was subjected to heat shock at 42°C. After 1 h of incubation, 200 μl of each culture was harvested, resuspended in lysing buffer for SDS-PAGE (bromophenol blue was omitted) (Sambrook et al. 1989) and protein concentration was measured by the method of Bradford (1976). Cell lysates equivalent to 10 μg of protein were supplemented with bromophenol blue, analyzed by SDS-PAGE and visualized by Coomassie brilliant blue staining (Sambrook et al. 1989).

Overproduction, purification, and analysis of recombinant DnaK, DnaJ and GrpE

DnaK Vh and DnaJ Vh were overproduced using the Tabor and Richardson’s (1985) system from plasmids pMZ20 dnaK Vh and pMZ1 dnaJ Vh , in E. coli BM271 ΔdnaK and CG1000 ΔdnaJ cells, respectively. Both genes were under the control of the bacteriophage T7 ϕ10 promoter and their expression was supported by the presence of pGP1-2, encoding heat-inducible T7 phage RNA polymerase (Tabor and Richardson 1985). Cultures of E. coli cells overproducing DnaK Vh and DnaJ Vh were incubated oxically at 30°C to OD595=0.6 and then heat-shocked at 42°C for 2 h. GrpE Vh was expressed from the chromosomal gene in V. harveyi cells grown at 30°C to OD595 =1 and then subjected to heat shock for 2 h at 39°C. DnaK Ec , DnaJ Ec and GrpE Ec proteins were overproduced from plasmids pMOB45 dnaK Ec , pDW19 dnaJ Ec and pRLM156 grpE Ec in an E. coli B178 background. DnaK Ec and DnaJ Ec were overproduced by induction of the genes with 1 mM IPTG at OD595=0.6, followed by incubation for an additional 2 h at 37°C. Cultures of the E. coli cells overproducing GrpE Ec protein were incubated oxically at 30°C to OD595=0.6 and then heat-shocked at 42°C for 2 h. All cultures were harvested by centrifugation (9,000g for 10 min at 4°C) and then resuspended in 50 mM Tris–HCl (pH 7.5), 10% (w/v) sucrose, 1 mM EDTA (pH 7.7) (1 ml per 1 g of cells). The suspensions were then frozen in liquid nitrogen and stored at −70°C until required. All chaperone proteins of V. harveyi and E. coli were purified as described previously (Żmijewski et al. 2004). For preparation of DnaK free of the nucleoside-diphosphate kinase (Ndk), the purified protein was concentrated on a Millipore filter (cut-off 30 kDa), and then purified by HPLC chromatograph, using a Zorbax GF-250 column (Agilent Biotechnologies).

Protein concentration was measured by the method of Bradford (1976) and confirmed by densitometry of Coomassie-stained SDS-polyacrylamide gels. All proteins used in this study were purified to more than 95% as estimated by SDS-PAGE and were free of contamination with other chaperones as judged by Western blotting. In Western blotting experiments, rabbit polyclonal antibodies were used; anti-DnaK Ec , anti-DnaJ Ec , anti-GrpE Ec were prepared as described previously by Szewczyk and Summers (1998). The N-terminal amino acid sequence of the purified proteins was analyzed on a gas-phase sequencer (Model 491, Perkin Elmer-Applied Biosystems, Foster City, Calif., USA) at BioCenter (Jagiellonian University, Krakow, Poland).

Luciferase refolding assay

Luciferase refolding assays were performed as published previously (Żmijewski et al. 2004). Briefly, firefly luciferase was pre-incubated for 10 min at 25°C in the presence of various combinations of the chaperone proteins and subsequently denatured for 10 min at 42°C. Renaturation was started by addition of 5 mM ATP followed by 15–90 min incubation at 25°C. All experiments were done in luciferase-refolding buffer L (40 mM Tris–HCl, pH 7.4, 50 mM KCl, 5 mM DTT, 10 mM MgCl2, 10% glycerol), with 1 μM DnaK, 0.2 μM DnaJ, 0.5 μM GrpE, and 80 nM luciferase. Control assays contained 0.5 mg bovine serum albumin (BSA)/ml. Modifications of this standard procedure are indicated in the figure legends, when pertinent. Luciferase activity was determined by using the Luciferase Assay System (Promega) and a Berthold luminometer. Initial activity of the native luciferase was set as 100%.

ATPase activity of DnaK

ATPase activity of the DnaK proteins was measured in the presence of various combinations of co-chaperones, as stated when pertinent. All reactions were performed at 25°C as described previously (Żmijewski et al. 2004) and free phosphate was assayed by the malachite green method (Lanzetta et al. 1979). The standard curve was obtained using H3PO4 or phosphate buffer, pH 6.8 (Sambrook et al. 1989). The enzymatic activity was calculated as pmol ATP/pmol DnaK/min.

All in vitro and in vivo tests were repeated at least three times, and mean values and standard deviations were calculated.

Results

Cloning of V. harveyi dnaK, dnaJ and grpE

dnaK

The coding sequence of dnaK of V. harveyi was cloned from the pGK1 dnaK Vh dnaJ Vh plasmid into expression vector pGEM Zf3+, as described in Materials and methods. Two plasmids were obtained: pMZ20 (orientation consistent with pT7 Φ10 promoter) and pMZ33 (orientation consistent with plac promoter). SDS-PAGE of the lysates of E. coli BM271 ΔdnaK cells transformed with plasmid pMZ33 revealed that DnaK Vh migrated with lower mobility than expected (Figs. 1a, 3b). Theoretical examination of the pMZ33 nucleotide sequence showed a possible in-frame fusion of the lacZ fragment (120 nucleotides encoding 40 N-terminal amino acids) with dnaK Vh (both genes in the same frame). This was confirmed by sequencing and, accordingly, the construct is referred to as pMZ33 lacZ′-dnaK Vh . DnaK Vh expressed from the pMZ20 plasmid migrated in SDS-PAGE as predicted from the nucleotide sequence of dnaK Vh [theoretical molecular mass of DnaK Vh is 69,076.6 Da, see Klein et al. 1998) [(Fig. 1a)]. Both proteins, DnaK Vh and LacZ′-DnaK Vh , reacted with polyclonal antibodies raised against E. coli DnaK (Fig. 1b).

Fig. 1
figure 1

Overproduction of Vibrio harveyi DnaK and DnaJ in Escherichia coli cells. a, c Electrophoretic analysis and b, d immunodetection of recombinant V. harveyi proteins synthesized in E. coli cells. a Purified DnaK Ec (lane 1); DnaK Vh overproduction from pMZ20 plasmid (in the presence of pGP1-2 plasmid) (lane 2) or from pMZ33 plasmid (lane 3) in E. coliΔdnaK bacteria; E. coliΔdnaK control cells (lane 4). b Western blotting analysis of DnaK Vh using anti-DnaK Ec polyclonal antibodies [lanes as in (a)]. c Purified DnaJ Ec (lane 1); DnaJ Vh overproduction from pMZ1 plasmid (in the presence of pGP1-2 plasmid) in E. coliΔdnaJ bacteria (lane 2); E. coliΔdnaJ control cells (lane 3). d Western blotting of DnaJ Vh using anti-DnaJ Ec polyclonal antibodies [lanes as in (c)]

Full size image

dnaJ

The coding sequence of dnaJ of V. harveyi was cloned by PCR using the chromosomal copy of the gene as template, as described in Materials and methods , and a plasmid, pMZ1 dnaJ Vh , was obtained. The dnaJ Vh protein product migrated in SDS-PAGE with lower mobility than DnaJ of E. coli, as predicted from the nucleotide sequence [calculated molecular mass of DnaJ Vh is 41,618.8 Da (Klein et al. 1998) and of DnaJ Ec is 41,100.3 Da (Bardwell et al. 1986) (Fig. 1c). The DnaJ Vh protein reacted with polyclonal antibodies raised against DnaJ Ec (Fig. 1d).

grpE

We previously reported the sequence of the dnaKdnaJ operon of V. harveyi (Klein et al. 1998) but a sequence of the third gene of the chaperone triad—grpE—remained unknown. In order to fill that gap, the gene coding V. harveyi GrpE was cloned, sequenced (see Materials and methods for details) and the sequence of the grpE Vh was deposited in GenBank (accession number AY193779). grpE Vh consists of 597 base pairs and encodes a hypothetical protein of 198 amino acids with a calculated molecular mass of 22,307 Da and isoelectric point pI=4.61; a model GrpE of E. coli is a 21,812 Da protein with pI=4.68.

Analysis of the grpE Vh DNA sequence and of hypothetical GrpE Vh revealed a high level of identity (80–90%) to genes and proteins of other members of Vibrio genus (Table 2). Interestingly, the identity of grpE genes was higher among the marine Vibrios (V. harveyi, V. parahaemolyticus, V. vulnificus and V. proteolyticus) than to the human pathogen, V. cholerae. Protein identity decreased very rapidly when the sequence of GrpE Vh was compared to other bacterial GrpE sequences, for example that of E. coli or Bacillus subtilis (Table 2). Comparison of the sequence of GrpE Vh to those from Archaea and Eukarya resulted in 29 and 27% identity, respectively (Table 2). Our results confirm that in the DnaK–DnaJ–GrpE system GrpE is the least conserved in evolution (Gupta et al. 1997; Cheetham and Caplan 1998; Gribaldo et al. 1999).

Table 2 Comparison of the sequences of V. harveyi grpE and gltP, and their products to selected bacterial, eukaryotic and archaeal sequences. It should be noted that all organisms presented here possess more than one gene encoding proton/glutamate or sodium/dicarboxylate symporter (gltP). The sequence of the protein/gene with the highest blastp/blastn score was chosen for the final comparison. The human hmge sequence was used for comparison
Full size table

Cluster of the V. harveyi heat-shock genes

We have previously reported that directly upstream of the dnaK Vh coding sequence there is an open reading frame, orf136 (Klein et al. 1998). Further investigation (sequencing of V. harveyi chromosomal DNA and database searching) resulted in extending the known sequence to 317 nucleotides (not shown), which revealed that this fragment shared about 80% of identity with the gltP genes of other Vibrios, encoding the glutamate transport proteins. The presence of the gltP homolog upstream of the V. harveyi dnaK gene (in the same orientation) suggested a similar arrangement of the V. harveyi genes to that of V. parahaemolyticus (accession number AP005075) and V. proteolyticus (accession number AF218211), namely grpE–gltP–dnaK–dnaJ, in contrast to V. cholerae, in which gltP is situated separately on the chromosome (accession number AE004170). The order of the genes on V. harveyi chromosome was confirmed by PCR and the primers: grpEVhfor and gltPVhI, followed by sequencing of the 2,300-bp PCR product (see Materials and methods for details). A sequence of the region between grpE and orf136 was obtained, which enabled us to compose a contiguous sequence of the grpE–gltP–dnaK–dnaJ region of V. harveyi. The entire sequence was deposited in GenBank under accession number AY639008. The sequence of the putative gltP gene of V. harveyi shares a very high degree of identity with the gltP genes of other Vibrio species and very low identity with the gltP genes of other Bacteria, Archaea or Eukarya (Table 2). The existence of the additional gene in the grpE–dnaK–dnaJ gene cluster in marine Vibrio species is very unique and suggests an involvement of gltP in their heat-shock response. However, we were able to find only a putative vegetative (sigma 70) promoter sequence before the gltP coding sequence—no typical heat-shock promoter could be identified (data not shown).

Cooperation of the KJE chaperone proteins of V. harveyi and E. coli in vivo

Mutants of E. coli carrying a deletion of dnaK, or dnaJ, or grpE are thermosensitive and resistant to bacteriophage λ infection (Georgopoulos et al. 1990). The E. coli mutant strains BM271 ΔdnaK, CG1000 ΔdnaJ, WKG190 ΔcbpAdnaJ and DA259 ΔgrpE were transformed with plasmids carrying the relevant V. harveyi gene, and the transformants were tested for growth at high temperatures (42 and 45°C), as well as for their ability to support growth of λ phage (see Materials and methods ).

DnaK of V. harveyi was not able to suppress thermosensitivity of E. coli mutant (Figs. 2a, 3a) when the protein was expressed from high-copy-number plasmids pMZ33 lacZ′–dnaK or pMZ20 dnaK (both pGEM3zf + derivatives). The lack of rescue by the V. harveyi dnaK was not due to the absence of dnaK Vh expression, because DnaK Vh was present at high levels in the transformed E. coli dnaK mutant cells at 42°C (Fig. 3b). The expression of the lacZ′–dnaK Vh gene from pMZ33 was high even without IPTG induction (Fig. 3b).

Fig. 2
figure 2

Complementation experiments of E. coli mutations with V. harveyi genes. E. coli strains: a ΔdnaK, b ΔdnaJ, c ΔcbpAdnaJ, d ΔgrpE were transformed with plasmids carrying relevant V. harveyi genes or vector DNA, as a control, and the ability of transformants to grow at different temperatures was tested as indicated. Filled diamonds: a, b, c, d E. coli B178 wild-type; a, b, c filled triangles: plasmid pGK1dnaK Vh dnaJ Vh (pBR322 derivative); a, b, c open triangles: pBR322 vector; filled squares: plasmid carrying the relevant V. harveyi gene: a pMZ20dnaK Vh ; b, c pMZ1dnaJ Vh and d pMZ40grpE Vh ; a filled circles: pMZ33lacZ′-dnaK Vh ; open squares: vector control: b, c pT7-5; d pGEM3zf+ (see Materials and methods for details). Additionally, cells bearing pMZ1, pMZ20 and corresponding vector controls were co-transformed with pGP1-2 plasmid. Results are presented as a number of colony forming units (CFU)

Full size image
Fig. 3
figure 3

Overexpression of V. harveyi DnaK is insufficient for growth of E. coliΔdnaK bacteria under heat-shock conditions (42°C) and is not toxic for E. coli cells. a The E. coli MC4100 (wild-type) [(left side of a plate)] or E. coliΔdnaK (right side of a plate) cells transformed with various plasmids carrying dnaK Vh were spread on LA plates and incubated at 30°C (left plate) or 42°C (right plate) for 18 h. b Electrophoretic analysis of the whole-cell lysates of E. coli MC4100 (wild-type) and E. coliΔdnaK bacteria transformed with plasmids carrying dnaK Vh (as in Fig. 2a). The following plasmids were used: pMZ20 dnaK Vh , pMZ33 lacZ′-dnaK Vh , pGK1 dnaK Vh dnaJ Vh . Cells were grown to optical density OD595=0.6 at 30°C and then shifted to 42°C or incubated at 30°C for an additional hour. See Materials and methods for details. *E. coli cells bearing pMZ20 also carried pGP1-2

Full size image

It has been previously reported that overexpression of DnaK Ec might be toxic for E. coli (Rockabrand and Blum 1995). To rule out the possibility that non-physiologically high levels of DnaK Vh had a lethal effect on E. coli growth, the growth of E. coli MC4100 dnaK+ strain transformed with pMZ20, pMZ33, or pGK1 plasmids was monitored; all transformants were able to grow at both 30 and 42°C (Fig. 3a).

According to Mogk et al. (1999), lack of suppression of the E. coli dnaK mutation by a foreign dnaK gene expressed from a plasmid might be due to an insufficient level of DnaJ Ec in the E. coliΔdnaK strain (Mogk et al. 1999). It has been previously reported that, in the case of plasmid pGK1 carrying a complete dnaK Vh dnaJ Vh operon, both DnaK and DnaJ were expressed in E. coli cells (Klein et al. 1998). However, pGK1 did not support growth of E. coli ΔdnaK, similar to plasmids expressing only dnaK Vh (Figs. 2a, 3a and Klein et al. 1998); thus, we concluded that V. harveyi DnaK is unable to function efficiently in E. coli cells. In order to gain additional information concerning in vivo activity of DnaK Vh , the ability of pMZ33 carrying lacZ′–dnaK Vh fusion to support bacteriophage λ growth in E. coli BM271 ΔdnaK cells was tested. To some extent, V. harveyi DnaK was able to support bacteriophage λ growth: the e.o.p. for mutant cells transformed with pMZ33 was 3.8×10−4 compared to an e.o.p. <10−8 obtained for mutant cells transformed with vector plasmids only. These results were similar to those reported previously for dnaK Vh dnaJ Vh cloned in pGK1 (Klein et al. 1998).

Cloned dnaJ and grpE of V. harveyi, in contrast to dnaK Vh , rescued the growth defect of the E. coli mutants at 42 and 45°C (Fig. 2b,d). Complementation by V. harveyi dnaJ was also very efficient in the double knock-out mutant ΔcbpAdnaJ, deprived not only of DnaJ but also of its homolog, CbpA, as well (Fig. 2c). The V. harveyi DnaJ is a type I DnaJ according to the classification of Cheetham and Caplan (1998), similarly as E. coli DnaJ, while CbpA belongs to class II. A single cbpA mutation does not cause thermosensitivity but the double mutant ΔcbpAdnaJ is more thermosensitive than E. coli dnaJ bacteria and commonly used for complementation tests (Ueguchi et al. 1995; Kelley and Georgopoulos 1997). Moreover, both DnaJ and GrpE of V. harveyi were able to support bacteriophage λ growth in E. coli mutant cells. The complementation efficiency was significantly higher than in the case of DnaK Vh (e.o.p was 0.05 for E. coliΔdnaJ cells transformed with pMZ1dnaJ Vh , 1.0 for E. coliΔcbpAdnaJ and 0.8 for E. coliΔgrpE transformed with pMZ40grpE Vh ). These results show that V. harveyi DnaJ and GrpE are able to function efficiently in E. coli cells. It should be noted that phage plaques formed on E. coliΔdnaK bacteria transformed with pMZ33 or pGK1 coding dnaK Vh were significantly smaller than those formed on the E. coliΔdnaJ or ΔgrpE bacteria transformed with relevant V. harveyi genes (data not shown). Results obtained for the single dnaJ of V. harveyi expressed from pMZ1 were similar to those previously reported for pGK1 dnaK Vh dnaJ Vh (Klein et al. 1998). Since the expression of dnaJ Vh from pMZ1 was significantly higher than from pGK1 (data not shown), it could be concluded that the level of DnaJ Vh had a minor effect on the results of the in vivo tests. Although the level of GrpE Vh expressed from pMZ40 was very low (data not shown), GrpE Vh was able to function efficiently in the DnaK Ec –DnaJ Ec –GrpE Vh system in vivo.

These results indicate that DnaK is more species-specific in vivo than DnaJ and GrpE, despite the fact that DnaK is the best conserved in evolution among the DnaK–DnaJ–GrpE chaperone proteins.

Purification of V. harveyi DnaK, DnaJ and GrpE

The positive complementation tests in vivo indicated that functional DnaJ Vh and GrpE Vh were expressed in E. coli cells from the cloned genes. In addition, DnaK Vh and DnaJ Vh expressed from pMZ20 and pMZ1, respectively, could be detected in bacterial lysates analyzed by SDS-PAGE as additional bands with the expected mobility, similar to those of the homologous E. coli molecules (Figs. 1a,c, 3b), and could be detected using antibodies against DnaK and DnaJ of E. coli (Fig. 1b,d). DnaK Vh and DnaJ Vh were overexpressed from pMZ20 dnaK Vh and pMZ1 dnaJ Vh , respectively, in relevant E. coli mutants, using the Tabor and Richardson’s expression system (see Materials and methods for details). While the E. coliΔdnaK mutant transformed with pMZ20 and pGP1-2 plasmids was not able to form colonies on LA plates at 42°C (Figs. 1a, 2a) and did not survive prolonged incubation at 42°C in L–B medium (data not shown), efficient overproduction of DnaK Vh was achieved after 2 h of incubation under heat-shock conditions. Although the fusion protein LacZ′–DnaK encoded by pMZ33 was highly overproduced in E. coli MC4100 (wild-type) and BM271 ΔdnaK, and seemed to possess similar activity as DnaK Vh , judging from the results of in vivo tests (Fig. 2a), it was not used for the in vitro experiments.

GrpE of V. harveyi did not cross-react with antibodies raised against the E. coli GrpE homolog and, consequently, it was not possible to detect its presence in V. harveyi wild-type cells or E. coliΔgrpE cells transformed with DNA of pMZ40 grpE Vh (data not shown). Since the expression of the cloned V. harveyi grpE in E. coli was not evident on Coomassie-stained gels (data not shown), the GrpE of V. harveyi was purified directly from V. harveyi (wild-type) cells subjected to heat shock, as described in Materials and methods .

The identity of GrpE Vh was confirmed, after purification of the candidate protein on the DnaK Ec -Sepharose affinity column, by sequencing of its N-terminus. The N-terminal amino acids of GrpE Vh (SNEENKVTEEELDQI) were as predicted from gene sequencing, except that purified protein lacked the initial methionine.

DnaK, DnaJ and GrpE of V. harveyi were purified to homogeneity using a procedure similar to that previously described for the E. coli homologues (Żmijewski et al. 2004); the proteins migrated in SDS-PAGE as 70, 41 and 24 kDa proteins, respectively (Fig. 4). The observed mobilities were consistent with the predicted molecular mass of DnaK Vh (69,076.6 Da), DnaJ Vh (41,618.8 Da) and GrpE Vh (22,307 Da).

Fig. 4
figure 4

Electrophoretic analysis of purified chaperone proteins of V. harveyi and E. coli. Purified chaperones were resolved by SDS-PAGE performed in a 12.5% gel and stained with Coomassie blue. Lane M Ladder of molecular mass standards

Full size image

Renaturation of luciferase by the KJE system of V. harveyi

Chaperone activity of the V. harveyi DnaK–DnaJ–GrpE system was studied by measuring the ability of the system to reactivate thermally denatured firefly luciferase. As a reference model of the KJE chaperone machinery, the chaperone proteins of E. coli were used. The level of luciferase activity after reactivation by the E. coli system was about 80% (Fig. 5a) and was similar to that previously published (Laufen et al. 1999).

Fig. 5
figure 5

Luciferase refolding by the V. harveyi DnaK chaperone system. Thermally denatured luciferase (80 nM) was renatured in the presence of ATP by various combinations of chaperones, and luciferase activity was then measured. a Reactivation (expressed as percentage of the initial activity; absolute value was 5×106 relative light units) after 60 min of incubation with chaperones as indicated and 5 mM ATP, at 25°C. V. harveyi chaperones were used in all experiments except that of the reference, in which E. coli proteins were employed (bar marked Ec). b Kinetics of luciferase refolding in the presence of: DnaK Ec , DnaJ Ec and GrpE Ec (filled triangles); DnaK Vh , DnaJ Vh and GrpE Vh (filled squares). In these experiments, ATP (l mM) was present, both during the denaturation and the renaturation steps. As a control, (filled circles) luciferase, without ATP, was denatured and renatured in the presence of DnaK Vh , DnaJ Vh , and GrpE Vh . In standard experiments, the protein concentrations were: DnaK 1 μM, DnaJ Ec 0.2 μM, and GrpE 0.5 μM. In one case marked with an asterisk, DnaK Vh was 2 μM. Error bars Standard deviation for at least three determinations. For the sake of clarity in b, the standard variations, which did not exceed 10%, have not been drawn

Full size image

The K–J–E chaperones from V. harveyi reactivated luciferase less efficiently than the E. coli system under standard conditions (DnaK 1 μM) (Fig. 5a). Increasing the concentration of DnaK Vh to 2 μM improved the process and reactivation of luciferase to 65% was achieved (Fig. 5a); a further increase of the DnaK Vh concentration did not result in higher reactivation (not shown). Similar results were achieved when DnaK was purified directly from V. harveyi cells (data not shown), which excluded the possibility that the recombinant DnaK differed from the cellular DnaK. There results indicate that the V. harveyi DnaK system has a lower ability to refold luciferase than the E. coli chaperone system, under conditions optimal for the E. coli system. Since V. harveyi is a marine bacterium, the optimal conditions for its chaperone proteins’ function might be different from those for the E. coli chaperones. In order to explore this possibility, various experimental conditions (modifications of the reaction buffer) were tested as follows: (1) Mops-based buffer (pH 7.2) instead of Tris–HCl (pH 7.4) was used (Minami et al. 1996); (2) since optimal growth of V. harveyi demands 3% (510 mM) NaCl in the medium, increasing concentrations of NaCl (50, 100, 200, 300 mM) instead of KCl in the reaction buffer were used; (3) magnesium acetate was substituted for magnesium chloride, as published for the eukaryal Hsp70 machineries (Terada et al. 1997); (4) various concentrations of ATP were used since we found that the ATPase activity of DnaK Vh was two-fold higher than that of DnaK Ec (see below). It has been shown that DnaJ loses its activity when Zn atom is removed from the protein (during extensive dialysis) but addition of ZnCl2 reverses the inactivation (Banecki et al. 1996). To ensure high-quality V. harveyi DnaJ, the protein was incubated in a reaction buffer supplemented with 5 mM ZnCl2 for 30 min prior to the refolding reaction. None of the presented modifications of the luciferase refolding reaction yielded higher luciferase activity (data not shown). Additionally, it was found that at increased concentrations of NaCl (200 mM and higher) reactivation of luciferase by the V. harveyi K–J–E system was inhibited.

As shown in Fig. 5a, refolding of luciferase by V. harveyi DnaK was strictly dependent on the presence of GrpE, which is typical for bacterial systems, and it was ATP-dependent, like all Hsp70 systems (Fig. 5b) (for recent review see Hartl and Hayer-Hartl 2002). Kinetics of reactivation of the heat-shock-denatured luciferase was measured and it was found that reactivation in both systems (V. harveyi and E. coli) reached maximum after 30 min (Fig. 5b).

Cooperation of DnaK, DnaJ and GrpE of E. coli and V. harveyi

The observed lack of DnaK Vh function in E. coli cells in vivo might have been caused by its inability to cooperate with endogenous co-chaperones. We therefore tested reactivation of luciferase using different combinations of E. coli and V. harveyi DnaK, DnaJ and GrpE. Reactivation was efficient in all tested variants, although a significant decrease in refolding efficiency was observed when DnaK from V. harveyi was incubated with two cochaperones of E. coli and vice versa (15–25% decrease of luciferase reactivity) [(Fig. 6)]. Our results showed that the yield of active luciferase depended mainly on DnaK activity but not on DnaJ or GrpE activity. Firstly, the KJE chaperone machinery of E. coli exhibited higher refolding activity (80%) than that of V. harveyi proteins (65%) (Fig. 5). Secondly, replacing of one co-chaperone (DnaJ or GrpE) or both DnaJ and GrpE always resulted in a decrease of refolding efficiency, compared to the homologous system from which a DnaK was derived (Fig. 6). If the activity of the heterologous system was limited by DnaJ or GrpE activities, cooperation of DnaK Vh with DnaJ and/or GrpE of E. coli should elevate the amount of active luciferase.

Fig. 6
figure 6

Cooperation of DnaK–DnaJ–GrpE chaperones from V. harveyi and E. coli in luciferase refolding. Thermally denatured luciferase (80 nM) was renatured in the presence of 5 μM ATP by various combinations of chaperones at 25°C, and luciferase activity was then measured after 60 min. For the sake of clarity, luciferase reactivation for each homologous system (all E. coli and all V. harveyi) was set as 100% and data obtained in each experiment were recalculated. Mean values of at least three experiments are shown; the standard deviations, which did not exceed 10%, have not been drawn. Absolute value of the initial activity of luciferase was 5×106 relative light units. Protein concentrations were as in standard experiments presented in Fig. 5a. Reactivation in all E. coli and all V. harveyi was 80% and 55%, respectively

Full size image

Stimulation of ATPase activity of DnaK proteins

To further understand the functional similarities and differences of the bacterial DnaK chaperone machineries, the ATPase activity of DnaK Vh was compared to that of DnaK Ec , alone and in the presence of various co-chaperones. The basic level of ATPase activity of DnaK Vh was more than two times higher than that of DnaK Ec (0.05±0.02 compared to 0.11±0.4) when DnaK proteins were purified using an additional gel filtration step, removing exogenous ATPase activity (NDP kinase), which usually co-purifies with DnaK (Barthel and Walker 1999). It should be noted that the additional purification step had no influence on the chaperone activity of either DnaK in the luciferase refolding assay (data not shown) but significantly decreased the ATPase activity of the DnaK preparations (almost three-fold for DnaK Vh ).

The ATPase activity of DnaK Vh could be stimulated by the presence of the V. harveyi and also E. coli co-chaperones, although stimulation by both DnaJ and GrpE co-chaperones of theE. coli origin was less efficient—approximately 50% of that caused by the V. harveyi co-chaperones (Fig. 7a,c). The presence of GrpE was essential for efficient stimulation of ATPase activity of V. harveyi DnaK by DnaJ proteins (Fig. 7a,c), similar to the case of the model DnaK of E. coli (Fig. 7b,d). Stimulation of the ATPase activity of DnaK Vh by co-chaperones (DnaJ and GrpE) was at least 10 times weaker than stimulation of DnaK Ec (compare Fig. 7a,c with b). The stimulation of DnaK Vh ATPase activity was more efficient at high, non-physiological concentrations of DnaJ (1 μM and higher) [(Fig. 7a,c)], similar to the case of DnaK Ec (Fig. 7b); under such conditions, DnaJ probably serves both as co-chaperone and substrate for DnaK (Laufen et al. 1999).

Fig. 7
figure 7

DnaJ-dependent stimulation of V. harveyi DnaK ATPase activity in the presence of GrpE. a, c ATPase activity of DnaK Vh and b, d DnaK Ec was measured in the presence of increasing concentrations of DnaJ (0–5 μM), either with GrpE (0.5 μM), or without GrpE. a Filled diamonds DnaK Vh + DnaJ Ec , filled squares DnaK Vh + DnaJ Ec + GrpE Ec ; b filled diamonds DnaK Ec + DnaJ Ec , filled squares DnaK Ec + DnaJ Ec + GrpE Ec ; c filled diamonds DnaK Vh + DnaJ Vh , filled squares DnaK Vh + DnaJ Vh + GrpE Ec , filled triangles DnaK Vh + DnaJ Vh + GrpE Vh ; d filled diamonds DnaK Ec + DnaJ Vh , filled squares DnaK Ec + DnaJ Vh + GrpE Ec . All values are expressed in units of increased activity (UIA in the ordinate), and were calculated as the ratio of the DnaK Vh ATPase activity in the presence of the co-chaperones to the activity without co-chaperones (ATPase activity was calculated as described in Materials and methods ). Mean values of at least three independent experiments are presented. The standard deviations did not exceed 10%

Full size image

Discussion

In this work, the V. harveyi chaperone system, composed of the DnaK–DnaJ–GrpE proteins, was characterized and compared to the model DnaK system of E. coli both in vitro and in vivo. Purified DnaK of V. harveyi was able to cooperate well with DnaJ and GrpE co-chaperones of E. coli in protein refolding (in reactivation of thermally denatured luciferase), and also in stimulation of DnaK ATPase activity. Protein refolding by the V. harveyi DnaK system was, similar to E. coli (and other bacteria), GrpE-dependent and occurred with similar kinetics, although with lower efficiency. These findings indicate that the mechanisms of action of the DnaK chaperone machineries of V. harveyi and E. coli are similar. This is in agreement with the high identity of the participating proteins, 81, 71 and 46% for DnaK, DnaJ and GrpE, respectively, and with the fact that V. harveyi and E. coli belong to the same systematic group (γ-proteobacteria). Despite the described similarities, DnaK of V. harveyi was not able to suppress the thermosensitivity of the E. coli mutant while cloned dnaJ and grpE of V. harveyi rescued the growth defect of the E. coli mutants at 42 and 45°C. Moreover, both DnaJ and GrpE of V. harveyi were able to support bacteriophage λ growth in E. coli mutant cells. The in vivo results suggest that—notwithstanding the higher sequence conservation of DnaK as compared with those of DnaJ and GrpE—the DnaK chaperone is more species-specific than are its co-chaperones.

Detailed comparison of separate domains of DnaK proteins of V. harveyi and E. coli revealed that the degree of conservation differs among the domains. The identity and similarity of the ATPase domains are 86 and 92%, and of the substrate-binding domains (SBD) 81 and 91%. Moreover, modeling of the DnaK Vh SBD domain and analysis of its sequence revealed that all the amino acids shown to be involved in the model peptide-binding by DnaK Ec , as well as those responsible for the salt bridges and H-bond formation between the SBD and the latch (Zhu et al. 1996), were preserved in DnaK Vh (not shown). The most variable part, by contrast, is located at the COOH-terminus, in the latch region (residues 537–638 of DnaK Ec ), with 60% identity and 75% similarity. It is possible that this least conserved region might be involved in creating DnaK species specificity.

Similar results, pointing to DnaK species specificity in vivo were obtained for the DnaK system of the archaeon M. mazei (Żmijewski et al. 2004) and for the Hsp70 chaperone machinery of the yeast Saccharomyces cerevisiae, consisting of the mitochondrial proteins Ssc1p–Mdj1p–Mge1p (Deloche et al. 1997). In these cases, the inability of the DnaK(Hsp70) to function in E. coli cells might be explained by the fact that E. coli, M. mazei and S. cerevisiae represent different kingdoms and their chaperones are less similar than those of V. harveyi and E. coli. Interestingly, in both studies mentioned above exogenous DnaJ and GrpE were able to complement the corresponding E. coli mutations. Lack of complementation of E. coli dnaK mutations by the cloned DnaK genes has been shown previously for Bacillus megaterium (Sussman and Setlow 1987), Borrelia burgdorferi (Tilly et al. 1993), Bradyrhizobium japonicum (Minder et al. 1997) and Mycobacterium tuberculosis (Mehlert and Young 1989). However, Mogk et al. (1999) reported that DnaK of B. subtilis was able to substitute the lack of DnaK in an E. coli mutant (BM271) at 40°C when DnaJ of E. coli was co-expressed, restoring the physiological level of that protein, and it has been shown that dnaK genes of Brucella ovis (Cellier et al. 1992), Francisella tularensis (Zuber et al. 1995) and Zymomonas mobilis (Michel 1993) rescued growth of E. coli dnaK mutants. Interestingly, there is no simple correlation between the identity level of these DnaK proteins and their ability to function in E. coli cells. It should be admitted that the results of these studies are rather difficult to compare, because in the cited experiments at least four different mutants of E. coli were used and experimental conditions were not identical. A general conclusion, however, might be drawn that exogenous DnaK is much less likely to function in E. coli cell than the DnaJ and GrpE chaperones.

Our in vitro experiments showing efficient cooperation of DnaK Vh with E. coli DnaJ and GrpE suggest that it was not the inability of a foreign DnaK to interact with endogenous co-chaperones that caused the lack of DnaK Vh function in E. coli cells.

Nonetheless, we have shown that V. harveyi DnaK was less active as a chaperone and its ATPase activity was significantly less stimulated by co-chaperones than E. coli DnaK. This indicates that, despite a similar mechanism of action, the DnaK chaperones of closely related V. harveyi and E. coli bacteria differ functionally. It is possible that these differences of quantitative nature are responsible, at least partially, for the observed species specificity of DnaK.

There are at least two more possible scenarios that could explain DnaK specificity. (1) It is well established that DnaK(Hsp70) proteins from different organisms recognize model peptides with slightly different affinities (Levy et al. 1995), which can be explained by the discrete amino acid substitutions within the DnaK substrate-binding cavity (Mayer et al. 2000). It could be speculated that chaperone proteins evolved such that they better recognized unfolded proteins of the same origin. This would cause lower affinity/efficiency of a chaperone system with heterologous DnaK. Since the KJE system participates in multiple cellular processes, even a minor impairment of its function would have a cumulative final effect of serious consequences for the cell as a whole, particularly under stress conditions (heat shock, phage infection). However, while this problem should be overcome by overexpression of a heterologous DnaK, this did not occur in the experiments described in this work or in experiments with overexpression of M. mazei DnaK in E. coli cells (Żmijewski et al. 2004). (2) It is also possible that some of the chaperone substrates in E. coli cells are not recognized by DnaK Vh . If these were key proteins, probably native, essential for cell survival and their proper function under stress conditions depended on efficient chaperone actions, the effects of the lack of recognition could be lethal. The best known example of the native E. coli protein recognized by DnaK is the heat-shock transcription factor σ32. It is well established that σ32 is negatively regulated by DnaK–DnaJ–GrpE chaperone machinery (Morita et al. 2000); thus, the different affinity of DnaK for σ32 could destabilize its turnover. Bukau and coworkers have shown that DnaK of Bacillus subtilis was able to suppress mutations of dnaK of E. coli under certain conditions but failed to substitute DnaK Ec in negative regulation of σ32 (Mogk et al. 1999). An increase in σ32 level is not likely to be lethal; however, there may be other proteins whose deregulation would be fatal for the cell. Species specificity of DnaK/Hsp70 proteins is still far from being understood but we believe that the presented work is a good starting point for further investigations of this phenomenon.