Search tips
Search criteria 


Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2012 September; 194(17): 4688–4697.
PMCID: PMC3415474

DnaK Dependence of the Mycobacterial Stress-Responsive Regulator HspR Is Mediated through Its Hydrophobic C-Terminal Tail


HspR is a repressor known to control expression of heat shock operons in a number of Eubacteria. In mycobacteria and in several other actinobacteria, this protein is synthesized from the dnaKJE-hspR operon. Previous investigations revealed that HspR binds to the operon promoter, thereby controlling its expression in an autoregulatory manner. DnaK, which is a product of the same operon, further aids this autoregulatory process by stimulating the operator binding activity of HspR. The molecular mechanism by which DnaK assists HspR in executing its function is not clearly understood. In this study, it has been shown that DnaK can augment DNA binding activity of HspR by two mechanisms: (i) DnaK can restore the activity of completely denatured HspR by forming a complex with it, and (ii) DnaK can prevent thermal instability of HspR renatured by other means. Unlike the first mechanism, the latter function does not involve complex formation. The C-terminal hydrophobic tail of HspR was found to play a significant role in determining its thermal stability and DnaK dependence properties. A deletion mutant in which this region is removed does not respond to thermal stress and functions independent of DnaK. The hydrophobic C-terminal tails of HspRs of Mycobacterium tuberculosis and related Actinomycetales therefore may have evolved to make these HspRs more sensitive to thermal stress and, at the same time, subject to regulation by DnaK.


The heat shock response is a tightly regulated process which is induced under stress conditions (17). Induction of the heat shock response results in the elevated synthesis of heat shock proteins (HSPs), which include molecular chaperones that are required for the correct folding of the newly synthesized proteins as well as denatured proteins (11). Although induction of HSP synthesis is a universal process, the regulatory mechanisms that control their synthesis differ widely between organisms. One such mechanism found in Escherichia coli involves the specific sigma factor σ32, which is required for the recognition of heat shock promoters by RNA polymerase (3). Under normal physiological conditions, this sigma factor remains associated with DnaK and therefore cannot bind to the promoter. Under heat shock conditions, the denatured proteins which accumulate within the cell titrate the DnaK away from σ32, leaving it free to function (32).

The sigma factor-dependent regulation of heat shock promoters is not conserved in prokaryotes. Indeed, in many Eubacteria, expression of heat shock genes and operons is controlled by specific repressors (14). In Bacillus subtilis, both the groEL and dnaK operons are negatively controlled by a repressor known as HrcA. It binds to an inverted repeat named CIRCE (controlling inverted repeat of chaperone expression) that is located in the regulatory regions of these operons. The repressor activity of HrcA is positively modulated by GroEL. A titration model, similar to that in the case of the DnaK/σ32 pair, has been proposed for GroEL/HrcA. In this model, under heat shock conditions, denatured proteins titrate GroEL away from HrcA, thereby inactivating it. There is, however, a basic difference in the manners in which DnaK regulates the activities of σ32 and HrcA. In the case of σ32, it acts negatively by preventing σ32 from functioning, whereas in the case of the repressor HrcA, it acts positively by assisting HrcA's ability to bind to CIRCE.

In several high-GC-content Gram-positive bacteria and also in some Gram-negative bacteria, an additional repressor known as HspR is present (14). Unlike B. subtilis, in which HrcA alone controls expression from the groEL and dnaK operons, in these organisms, there is a clear demarcation—HspR represses dnaK and HrcA represses groEL. Interestingly, the activity of HspR, like that of HrcA, is chaperone dependent. However, the chaperone involved in this case is not GroEL but DnaK (2). HspR proteins of Mycobacterium tuberculosis and several other organisms belonging to the phylum Actinobacteria are synthesized from the last gene of the dnaKJE-hspR operon. The binding site of HspR, known as HAIR (HspR-associated inverted repeats), is present upstream of several genes and operons, including the one from which it is synthesized. It has been proposed that expression from the dnaKJE-hspR operon is controlled by a homeostatic feedback mechanism involving HspR and DnaK (2). According to the proposed mechanism, the operon is induced under heat shock conditions, resulting in enhanced synthesis of DnaK as well as HspR. When the amount of DnaK and HspR within the cell reaches a critical level, they together bind to the operon promoter, thereby repressing it.

In this study, the focus is on HspR from Mycobacterium tuberculosis. This organism is the causative agent of tuberculosis (TB). A major problem with TB is chronic infection. This happens due to the ability of M. tuberculosis to persist in an infected individual for a prolonged period of time (33). The persistent form of TB is difficult to cure (20). Previous studies indicated that HspR plays a key role in determining the fate of M. tuberculosis in the infected individual. It was found that an HspR-deficient mutant of M. tuberculosis failed to survive in the chronic phase (25). The reason behind this was that the level of DnaK had increased in the mutant and this boosted an immune response in the host against the pathogen. Apart from the dnaKJE-hspR operon, the expression of several other genes of M. tuberculosis is controlled by HspR. One such gene which expresses the α-crystallin family protein Acr2 is required for pathogenesis (24). HspR therefore appears to play a key role in controlling the expression of several M. tuberculosis genes linked to pathogenesis.

The mechanism by which DnaK functions as a corepressor has been investigated in a previous study using HspR of Streptomyces coelicolor, which, like M. tuberculosis, belongs to the phylum Actinobacteria. In this study (2), it was found that DnaK can enhance the binding activity of HspR in an ATP-independent manner. Hence, it was proposed that the chaperone function of DnaK is not involved in the process. Contrary to these observations, investigations done in this laboratory (8) with the mycobacterial counterpart revealed that the DnaK-dependent DNA binding activity of HspR can indeed be stimulated in the presence of ATP, and hence, the chaperone function of DnaK may have a role to play.

Although it is generally agreed upon that DnaK plays an important role in stimulating HspR's DNA binding activity (2, 8, 14), the mechanism by which it does so is not clearly understood. A major problem regarding the study of HspR is that DnaK copurifies with the native protein (2), and therefore, it is difficult to ascertain the individual contributions made by the two toward the feedback regulatory process. Purified preparations can be obtained if HspR is isolated under denaturing conditions, but its renaturation becomes a problem, as the protein is aggregation prone. In this study, by isolating the M. tuberculosis HspR in the denatured state and renaturing it using a gel filtration-based method, it was possible to obtain a purified and active preparation free from any DnaK contamination. With the help of this renatured protein, it has been shown that DnaK dependence of HspR's DNA binding activity is due primarily to the presence of a hydrophobic C-terminal tail. Removal of this tail results in a more stable version of HspR which no longer requires DnaK for its activity.


Bacterial strains and plasmids.

Mycobacterial DnaK, N-terminally tagged with 6× His residues, was purified after expression from plasmid pMRLB.6 dnak/RV0350 (acquired through TB Research Materials & Vaccine Testing, NIH, NIAID, based at Colorado State University). The M. tuberculosis HspR expression plasmid, pTDR30, was constructed earlier (8).


Nickel-nitrilotriacetic acid (Ni2+-NTA) agarose was purchased from Qiagen (Valencia). For microtiter plate-based protein-protein interaction assays, Ni-NTA HisSorb plates (Qiagen, Valencia) were used. Other chemicals for protein expression, purification, and analysis were of the highest purity grade, obtained from SRL, India. [γ-32P]ATP (12 × 1013 Bq mmol−1) was purchased from BRIT (Mumbai, India).

Sequence alignment and phylogenetic analysis.

ClustalW (28)-based alignments of multiple sequences were performed using the software MEGA 4.0 (27). Multiple-alignment penalties of 10 and 0.1 were used for gap opening and extension, respectively, or as stated. The weight matrices chosen were either PAM or BLOSUM (10). The alignments created with MEGA 4.0 were subsequently processed using the BioEdit sequence alignment editor for convenient schematic representation. Phylogenetic trees were constructed with these alignments using MEGA 4.0. Bootstrap analysis was performed using 500 replicates.

Site-directed mutagenesis.

To create a C-terminally truncated version of HspR (HspR-ΔC), a stop codon was introduced into the hspR gene in the expression vector pTDR30 at position 119 of HspR using the QuikChange site-directed mutagenesis kit (Stratagene, Canada). The mutagenic primers were 5′-CCG AAG AGC ACC GCC TAG GTC GTC TGG AAA CCG-3′ and 5′-CGG TTT CCA GAC GAC CTA GGC GGT GCT CTT CGG-3′ (the mutation is underlined). The mutation was confirmed by DNA sequencing. The resulting mutant plasmid was designated pTDR30-1.

Purification of recombinant proteins.

Recombinant N-terminal 6× His-tagged proteins were isolated by Ni2+-NTA agarose affinity chromatography under either native or denaturing conditions using standard protocols given by the suppliers of the affinity column (Qiagen) and as described earlier (8). For DnaK-HspR interaction studies, an expression construct from which an untagged version of DnaK could be synthesized was made. The M. tuberculosis DnaK coding region (without the 6× His codons) was amplified from the plasmid pMRLB.6 dnak/RV0350 using the primers 5′-TTCGGATCCATATGGCTCGTGCGGTC-3′ (forward) and 5′-CCCCAAGCTT TCACTTGGCCTCCCG-3′ (reverse) (NdeI and HindIII sites underlined, respectively) and cloned into the multicloning site of vector pT7-7 (26). The recombinant protein was isolated by ammonium sulfate fractionation and ion-exchange chromatography essentially as described earlier (13). The purity was more than 95%, as evident from an SDS-PAGE analysis of the eluted protein.

Renaturation of unfolded HspR.

HspR isolated in the presence of 8 M urea was renatured by a stepwise reduction of the urea concentration by dialysis (2 M at each step). After dialysis, the protein samples were centrifuged at 11,400 × g for 15 min to remove aggregated proteins if any. Removal of urea was also done by size exclusion chromatography (SEC). One milliliter HspR (1 mg/ml) in buffer containing 8 M urea was layered on top of a Sephacryl S-200 size exclusion column (bed volume, 40 ml) equilibrated in buffer containing Tris-HCl (50 mM), KCl (200 mM), EDTA (5 mM), MgCl2 (5 mM), and 10% (vol/vol) glycerol. The column was run at a flow rate of 0.50 ml/min. The fractions were then analyzed using UV absorption spectrometry and SDS-PAGE analysis.

For rapid refolding in the presence or absence of chaperones, HspR isolated under denaturing conditions was refolded in a final volume of 200 μl by dilution into renaturation buffer containing 20 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol (DTT), 10 mM KCl, and 1 mM MgCl2. Renaturation was done for 1 h at room temperature. In the case of DnaK-assisted renaturation, ATP was added at a final concentration of 0.5 mM. After this time period, the refolding mixture was centrifuged and the supernatant was taken out. From this supernatant, aliquots (25 μl) were removed and used for electrophoretic mobility shift assays (EMSA).

Electrophoretic mobility shift assay.

An electrophoretic mobility shift assay was performed using a 32P-labeled, 110-bp PCR-amplified DNA fragment derived from the dnaKJE operon promoter-operator region (−128 to −38) encompassing the two inverted HAIR repeats (25). The primers used (K-128F and K-38R) and the protocols for performing EMSA and antibody supershift assays were the same as those reported in the earlier study from this laboratory (8). For competition experiments, a molar excess of either self unlabeled probe or an unrelated unlabeled probe of similar size derived from the plasmid pAL5000 origin (1) was incorporated in the preincubation phase.

Circular dichroism.

Circular dichroism (CD) measurements were done on a Jasco-815 spectropolarimeter using a 1-mm-pathway quartz cell at 25°C. The far-UV CD spectra were recorded in the range of 200 to 250 nm with a scan rate of 120 nm/min and a response time of 4 ns. Three scans were accumulated for each spectrum. The concentration of HspR was 5 μM.

Promoter assays.

The dnaKJE-hspR promoter-operator region used for EMSA was cloned in the lacZ-based promoter probe vector pSD5B (12) at the XbaI site, resulting in the promoter construct pSDTD2. IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible plasmids expressing either the wild type (WT) (pTDR30) or the C-terminally truncated mutant (ΔC) (pTDR30-1) hspR were cotransformed into E. coli along with pSDTD2. In control experiments, pSD5S30B (9), a pSD5B-based construct in which lacZ is expressed from S30, a randomly isolated Mycobacterium smegmatis promoter that functions in both E. coli and M. smegmatis(7), was used in place of pSDTD2. Cotransformed cells were grown to mid-log phase in the presence of ampicillin (100 μg/ml, for pTDR30) and kanamycin (50 μg/ml, for pSDTD2). Aliquots (1 ml) were taken out from each culture, followed by induction of HspR synthesis by the addition of 1 mM IPTG. Induction was allowed for 90 min at 37°C, after which the cells were either left alone or heat shocked at 42°C for 10 min. Recovery was allowed for 40 min at 37°C, after which β-galactosidase assays were performed using the fluorescent substrate 4-methylumbelliferyl-β-d-galactopyranoside (MUG). Fluorescence readings were taken in the microtiter plate reader POLARstar Optima (BMG Labtech). The results were expressed as arbitrary units of β-galactosidase activity (MUG units), calculated using the equation (30) F/t × A, where F, t, and A are sample fluorescence intensity (excitation and emission at 360 and 460 nm, respectively) at the end of the reaction, time of the reaction in minutes, and absorbance of the cell suspension measured at 620 nm, respectively.

Raising antibody in rabbits.

Polyclonal antibodies were raised in rabbit using the affinity-purified proteins DnaK and HspR. Preimmune and immune sera were drawn and processed as reported in an earlier study (1). The specificities of the antisera thus obtained were verified using Western blotting.

Protein-protein interaction studies.

The interaction between DnaK and HspR was assessed by an enzyme-linked immunosorbent assay (ELISA)-based method, which was used earlier to study the interaction between DnaK and a plasmid replication protein (16), with some modifications. Urea (8 M)-denatured 6× His-tagged HspR (75 μM) was diluted 100 times in 100 μl of renaturation buffer containing untagged DnaK (1 μM), resulting in a final concentration of 0.75 μM. In the case of HspR renatured by SEC or dialysis, the native protein was added in such a way that the final concentration reached 0.75 μM, the same as in the case of the denatured protein. As a negative control, a mock-interaction assay was set up by omitting HspR but retaining DnaK. Reagent blanks were set up by omitting both. Incubation was done for 1 h at 25°C. The reaction mixtures were split into two halves and added to the wells of a Ni-NTA microtiter plate (Ni-NTA HisSorb). One half was processed for detecting the amount of bound DnaK, and the other was processed for detecting the amount of HspR immobilized. Following adsorption, the wells were blocked with phosphate-buffered saline containing 0.2% bovine serum albumin (PBS-BSA) for 1 h under mild shaking conditions, followed by several washings with PBS containing 0.05% Tween 20. One well of each duplicate set was probed with anti-HspR sera, and the other was probed with anti-DnaK (1:250 dilutions). After washing and addition of alkaline phosphatase-conjugated secondary antibody, color development was done using the chromogenic substrate para-nitrophenyl phosphate (PNPP) per standard procedure. Reactions were stopped with 3 N NaOH. Color development was monitored by measuring the absorbance at 405 nm (A405) using a microtiter plate reader. An A405 corresponding to the reagent blank was considered to be zero.


Sequence alignment and phylogenetic analysis of HspRs.

Alignment of HspR sequences from various Eubacteria using ClustalW revealed that the N-terminal half of the protein, which includes the winged helix-turn-helix motif DNA binding domain (HTH-W1-W2), is highly conserved (Fig. 1A). In contrast, the sequence in the C-terminal region is variable. Phylogenetic analysis based on the conserved sequences (aligned positions 50 to 130) revealed that there are several distinct clades of HspR (Fig. 1B). The HspRs of Eubacteria belonging to the phylum Thermus-Deinococcus form a distinct and possibly ancient clade. These HspRs possess a duplication of their HspR domains, designated 1 and 2. HspRs of more-modern Eubacteria, the Campylobacterales, Campylobacter jejuni and Helicobacter pylori, form a different clade. Interestingly, these HspRs appear to be closely related to the HspR of Aquifex aeolicus, a hyperthermophilic bacterium belonging to the ancient phylum Aquificae. This observation suggests that the Campylobacterales may have acquired their hspR orthologs from ancient thermophilic bacteria through horizontal transfer. The third clade comprises HspRs derived from bacteria belonging to the phylum Actinobacteria. Within this clade, a subgroup of HspRs derived from Actinomycetales form a distinct branch. The HspRs of this group are characterized by the presence of a C-terminally located conserved motif comprising the hydrophobic sequence LVVW flanked by the positively charged amino acid residues lysine and arginine (Fig. 1A, boxed region). HspR of Propionibacterium is, however, an exception, in that although it is derived from an actinomycete, it clusters with the bifidobacterial HspR (Fig. 1B). It has been established in earlier studies that DnaK preferentially binds hydrophobic cores of four or five residues in length and that positively charged residues in the flanking regions increase binding (18, 19). The LVVW core is hydrophobic and also flanked by positively charged amino acid residues. Hence, it was hypothesized that this motif may be involved in the interaction of HspR with DnaK. To test this hypothesis, a C-terminal deletion mutant (HspR-ΔC) was constructed to exclude the LVVW motif.

Fig 1
Sequence comparison of HspRs. (A) ClustalW alignment (gap opening and extension penalties of 10 and 0.1, respectively) of representative HspRs derived from a broad spectrum of Eubacteria using MEGA 4.0. The similar/identical residues (PAM matrix) were ...

Renaturation of HspR-WT and HspR-ΔC.

The LVVW sequence is highly hydrophobic and includes a tryptophanyl residue, considered to be the most hydrophobic of all amino acids (15). Hydrophobic regions of proteins usually tend to remain buried (6). However, if they are in the exposed state, they may trigger aggregation. It was suspected that the LVVW motif may be responsible for the tendency of HspRs derived from Actinomycetales to form aggregates, as reported earlier (2, 8). To test this possibility, attempts were made to renature HspR-WT and HspR-ΔC from the denatured state by stepwise removal of the denaturant, urea, through dialysis. HspR-WT renatured by this procedure aggregated, resulting in poor recovery of the soluble form, with most of it (80%) being present in the insoluble fraction (Fig. 2A, lanes 3 and 5 compared to lane 1). HspR-ΔC, on the other hand, remained soluble. Most of the protein (~87.5%) was recovered in the soluble supernatant, whereas a small amount (~12.5%) was found to be present in the pellet fraction (Fig. 2A, lanes 4 and 6 compared to lane 2). Since HspR-WT could not be renatured efficiently through dialysis, an alternative strategy based on SEC (31) was attempted. HspR isolated under denaturing conditions through affinity purification was loaded onto a SEC column (exclusion limit, 200 kDa), followed by elution. The major part of the protein eluted as a single peak corresponding to the HspR monomer (Fig. 2B and andC).C). The secondary structure of HspR present in the pooled peak fractions was investigated using CD spectroscopy. From the spectra (Fig. 2D and andE),E), the helical contents of both of the renatured proteins were calculated to be about 33%. The results indicate that the presence of the hydrophobic C-terminal tail prevents efficient renaturation of HspR and that its removal (as in HspR-ΔC) leads to significant improvement in its solubility and ability to refold spontaneously. The results also indicate that the WT protein can be renatured effectively by performing SEC.

Fig 2
Renaturation efficiencies of HspR-WT and HspR-ΔC. (A) Renaturation by dialysis. A 500-μl aliquot of the urea-denatured sample was dialyzed serially against buffers containing decreasing concentrations of urea. After complete removal of ...

DNA binding activity of renatured HspR.

DNA binding assays were performed using HspR-WT and HspR-ΔC, renatured by SEC (WTSE) and dialysis (ΔCD), respectively. The probe used was a 32P-labeled, 110-bp DNA segment encompassing the HspR binding sites located upstream of the M. tuberculosis dnaKJE-hspR operon (−128 to −38). A concentration-dependent increase in binding of HspR-WT and HspR-ΔC to the probe was observed. At the lower concentration, C1 is formed, but as the concentration increased, a second complex, C2, appeared (Fig. 3A). To obtain an idea about specificities, competition binding experiments were performed using a molar excess of either self (specific) and or an unrelated (nonspecific) competitor, a 150-bp fragment derived from the mycobacterial plasmid pAL5000 origin of replication (1) (Fig. 3B and andC).C). The resulting competition experiments revealed that the complexes were competed out in a dose-dependent manner by only the specific and not the nonspecific competitor. This indicates that the complexes were specific. Moreover, the rates of competition by the self competitor were almost the same for WT and ΔC HspR. This indicates that the affinity of HspR for its target site did not change significantly following removal of the C-terminal region.

Fig 3
DNA binding activity of HspR. The WT protein renatured by SEC (WTSE) and the ΔC protein renatured by the dialysis method (ΔCD) were used in EMSA experiments using the labeled HspR binding site as a probe. (A) EMSA was performed with increasing ...

The C-terminal tail determines HspR's thermal stability and DnaK-dependent binding activities.

The results presented in the previous section indicate that HspR can bind to the operator without DnaK assistance. In order to test whether DnaK assistance is necessary under heat shock conditions, EMSA experiments were performed using the heat shock proteins, either WT or ΔC, in the presence or absence of either DnaK or Hsp16.3, a chaperone known to be able to protect against thermal denaturation (4). The results show that in the absence of heat shock, HspR formed the specific complexes C1 and C2 (Fig. 4, lane 2). In the presence of the chaperones, either DnaK or Hsp16.3, no substantial change in the binding pattern (Fig. 4, lanes 3 and 4) except the appearance of a low-intensity band (C3) in the presence of DnaK (Fig. 4, lane 3) was observed. Following heat shock, complex formation was abolished (Fig. 4, lane 5). Upon addition of DnaK, the DNA binding activity of HspR was restored, as indicated by the reappearance of the major HspR DNA complexes C1 and C2 (Fig. 4, lane 6). The minor complex C3 was also formed in addition to C1 and C2. Although not as efficient as DnaK, Hsp16.3 was found to be capable of preventing loss of HspR's DNA binding activity. Neither DnaK nor Hsp16.3 possesses any DNA binding activities of its own (Fig. 4, lanes 14 and 15), and hence, none of the bands are due to direct binding of these chaperones to the probe. The results obtained suggest that the presence of DnaK or, to a lesser extent, any other chaperone can assist HspR by preventing its denaturation under heat shock conditions. In sharp contrast, HspR-ΔC was stable toward heat shock. The presence or absence of DnaK or Hsp16.3 made no difference (Fig. 4, lanes 8 to 13). These results clearly indicate that removal of the C-terminal tail leads to stabilization of the protein on one hand and DnaK independence on the other.

Fig 4
DNA binding activity of HspR under heat shock conditions. EMSA experiments were performed using the WTSE or ΔCD protein (1 μM) exposed to either room (25°C) or heat shock (42°C) temperature in the presence or absence of ...

Renaturation of urea-denatured HspR.

While heat shock at 42°C is expected to mildly denature HspR, 8 M urea is expected to either completely or substantially denature it. The question that was raised was whether DnaK can help in the renaturation of HspR once it has been completely denatured. The ability of HspR to renature to the active form, either spontaneously or with the assistance of DnaK, was functionally investigated by performing EMSA experiments with the denatured samples that were renatured by dilution into denaturant-free buffer in the presence or absence of DnaK.

The results show that when denatured HspR-WT was diluted, no binding activity could be recovered, indicating that HspR-WT was unable to regain activity (Fig. 5A, lane 2). On the other hand, when denatured HspR-ΔC was similarly diluted, formation of complexes C1 and C2 was observed (Fig. 5A, lane 4). This indicated that HspR-ΔC, but not HspR-WT, could be renatured by dilution into denaturant-free buffer without the assistance of DnaK. When denatured HspR-WT was diluted into DnaK containing renaturation buffer, DNA binding activity was observed, but the mobility of the complex formed (C4) did not match that of either C1-C2 or C3. It migrated more slowly than the rest (Fig. 5A, lane 3). Unlike HspR-WT, denatured HspR-ΔC did not form C4 in the presence of DnaK. The predominant complexes were C1 and C2 (Fig. 5, lane 5). However, the minor complex C3 was also found to be present.

Fig 5
DNA binding activity of HspR renatured with DnaK assistance. (A) WT or ΔC HspRs denatured using 8 M urea were diluted 100 times into renaturation buffer containing 0.5 mM ATP in the presence or absence of 2 μM DnaK. Following preincubation, ...

To test whether the C4 complex was specific, competition binding assays were performed using self and nonspecific competitors. The results showed that a 750-fold molar excess of the specific competitor abolished complex formation, whereas at the same molar excess, the nonspecific competitor failed to compete (Fig. 5B). This indicates that the complex C4, like C1 and C2, is also specific in nature. Since C4 was formed only in the presence of DnaK, it was necessary to examine whether the added DnaK became a part of complex C4. A supershift assay was performed using antisera against DnaK. The corresponding preimmune sera were used as a control. The results show that in the presence of anti-DnaK sera, a supershift was observed (Fig. 5C). The preimmune sera had no effect. This indicates that DnaK is a part of complex C4. The important conclusions from these experiments are as follows: (i) the C-terminal tail of HspR plays a key role in preventing spontaneous renaturation of denatured HspR, and (ii) this region appears to be involved primarily in the interaction with DnaK.

DnaK-HspR interactions.

In order to assess the affinity of DnaK for HspR-WT and HspR-ΔC, direct binding assays were performed using a microtiter plate-based method. In this assay, 6× His-tagged HspR, either WT or ΔC, was allowed to bind in solution to a tag-free version of mycobacterial DnaK. The DnaK-HspR complexes formed were immobilized onto the wells of a Ni-NTA HisSorb plate in duplicate. One well of each duplicate set was probed with anti-HspR sera, and the other was probed with anti-DnaK. The results show that DnaK binding was the maximum in the case of denatured HspR-WT (Fig. 6A). In the case of renatured HspR-WT, binding was only marginally higher than background levels. That this difference is not due to variations in the amount of HspR present in the assay system is evident from the observation that there were only minor differences in the corresponding A405 (HspR) values (Fig. 6B). As in the case of the renatured WT protein, the level of DnaK binding to renatured HspR-ΔC was low, but the interesting observation was that unlike in the case of the WT, denatured HspR-ΔC did not show any significant binding to DnaK. The observed differences became more pronounced if the minor fluctuations in the amount of HspR bound to the support were taken into account and the data were normalized accordingly (Fig. 6C). Such normalization was done by finding the ratios (absorbance ratios) between the respective A405 values for DnaK and HspR. The major conclusions from this experiment are as follows. (i) The hydrophobic C-terminal tail plays an important role in the ability of HspR to interact with the DnaK. (ii) Presence of the tail is not enough; it has to be present in the exposed (denatured) state for productive interaction with DnaK.

Fig 6
ELISA to study the interaction between DnaK and either WT or ΔC HspR. Mean A405 values derived from five replicate ELISAs using antisera against DnaK (A) or HspR (B) are shown. (C) Means of the absorbance ratios between wells probed with anti-DnaK ...

In vivo implications of deleting the C-terminal tail.

In order to test whether deletion of the C-terminal tail affects the ability of HspR to act as a repressor, an experiment was designed using the surrogate host E. coli. Reporter activity from pSDTD2 or the control plasmid pSD5S30B, which expresses the reporter gene lacZ from the unrelated S30 promoter, not known to be heat inducible, was assessed in the absence or presence of HspR synthesized from the cotransformed plasmids pTDR30 (for HspR-WT) and pTDR30-1 (for HspR-ΔC). Following induction of HspR synthesis, the samples were divided into two parts; one part received no heat shock, while the other did (Fig. 7A and andB,B, respectively). The results showed that the reporter activities from both pSDTD2 and pSD5S30B were higher than that of the promoterless vector pSD5B. Hence, both promoters were active in E. coli. The activity of pSDTD2 was, however, more than that of pSD5S30B. When HspR was supplied in trans (pSDTD2+WT), partial repression of lacZ expression was observed in the absence of heat shock (Fig. 7A), but for HspR-ΔC (pSDTD2+ΔC), repression was near complete. Upon heat shock (Fig. 7B), derepression from the dnaKJE-hspR promoter was clearly evident in the case in which repression was mediated by the WT repressor (pSDTD2+WT), but no such effect was observed in the case of the ΔC mutant (pSDTD2+ΔC). What emerges from this experiment is that HspR-ΔC functions as a stronger repressor in vivo than the WT. Moreover, unlike in the case of the WT, heat shock could not relieve the repression mediated by HspR-ΔC. These effects are clearly specific to the dnaKJE-hspR operon promoter, as expression from the control plasmid pSD5S30B remained almost unaffected in the presence of either WT or ΔC HspR. Small differences, if any, were found to be statistically insignificant.

Fig 7
In vivo effect of C-terminal truncation. Fluorescence-based β galactosidase reporter assays of E. coli cells transformed with the indicated promoter constructs either alone or in combination with plasmids expressing either wild type (+WT) or mutant ...


Mycobacterium tuberculosis HspR is a repressor that controls the expression from the heat shock operon dnaKJE-hspR. It was demonstrated earlier that as in the case of Streptomyces coelicolor (2), the activity of M. tuberculosis HspR is modulated by DnaK (8). HspRs from Actinomycetales, such as M. tuberculosis and S. coelicolor, are insoluble in nature and tend to aggregate unless DnaK is present (2, 8). A sequence comparison of HspRs from diverse groups of bacteria revealed that although HspR is a highly conserved protein, there are significant sequence variations in the C-terminal region. Phylogenetic analysis revealed that HspRs can be divided into two broad groups. One group comprises the HspRs from ancient thermophilic bacteria, whereas the other group comprises the Actinomycetales HspRs. Surprisingly, the HspRs of the Campylobacterales H. pylori and C. jejuni cluster with those derived from thermophilic bacteria. In particular, their HspRs were found to be closely related to that of the ancient hyperthermophile Aquifex aeolicus. It appears that the Campylobacterales may have acquired their hspRs from ancient bacteria through horizontal transfer.

The actinobacterial HspRs form a distinct clade, and most of them possess the conserved hydrophobic motif LVVW in the C-terminal region which potentially constitutes a high-affinity DnaK binding site (18). To investigate the role of this motif in HspR's DnaK-dependent DNA binding activity, the C-terminal tail region of the M. tuberculosis HspR was deleted so as to eliminate this motif. The solubility of the mutant protein (HspR-ΔC) was found to be significantly more than that of the WT. It could be easily renatured by dialysis against denaturant-free buffer or by simple dilution into renaturation buffer. The amount of protein lost due to aggregation was significantly less for HspR-ΔC than for the WT. As HspR-WT could not be renatured by these methods, an alternative strategy in which removal of the denaturant was done using size exclusion chromatography was used (31). This method gave better results, possibly because in this procedure, the denaturant gets removed in infinitesimal steps and, therefore, the protein gets more opportunity to achieve the biologically active structure.

EMSA experiments with the refolded proteins revealed that both HspR-WT and HspR-ΔC were capable of binding to the target site unaided. The binding patterns were the same for both. Two complexes, C1 and C2, were formed in a concentration-dependent manner. At the lower concentration, C1 was formed, and at the higher, C2 was formed. C2 migrated more slowly than C1, and hence, it appears to be a higher-order complex than C1. Considering that there are two HspR binding sites in the dnaKJE-hspR operon operator region (25), it is most likely that C1 is monomeric and C2 is dimeric. The binding was found to be specific, as the complexes formed were competed out by the self competitor but not by an unrelated DNA fragment. Marginal batch-to-batch differences in binding efficiencies were encountered, probably due to differences in renaturation efficiencies. This might explain why, in the concentration-dependent DNA binding experiment, the observed binding efficiency of HspR-ΔC was somewhat less than that of the WT. The proportion of active molecules in the HspR-ΔC preparation used in this experiment was probably less than that of the WT. An alternative possibility is that HspR-ΔC has lower DNA binding affinity than the WT. The results of the self competition DNA binding assays, however, do not support this. In both cases, increasing concentrations of the self competitor abolished complex formation at almost the same rate. The affinities of the two proteins for the target site are thus unlikely to be significantly different.

HspR renatured using the SEC method could bind to the target independent of DnaK, resulting in the complexes C1 and C2, and hence, contrary to previous claims, presence of DnaK is not obligatorily required (2). However, DnaK certainly has a role to play, particularly in the context of heat shock. It was revealed that HspR is thermolabile, and it loses activity at the mild heat shock temperature of 42°C. It is in this context that the presence of DnaK becomes relevant. It protects against thermal denaturation of renatured HspR. At a lower level, the chaperone Hsp16.3, known for its ability to prevent thermal denaturation of proteins (4), also gave protection. This indicates that as far as protection against thermal denaturation is concerned, other chaperones may at least partially substitute for DnaK. The introduction of DnaK did not result in further retardation in the movement of C1 and C2. Hence, the protective action does not appear to involve complex formation between HspR and DnaK. In the presence of DnaK, however, the minor complex C3 was formed in addition to C1 and C2 and migrated more slowly than both. The nature of this minor complex is not clear at present. No evidence for the presence of DnaK was obtained using antibody supershift assays (data not shown). It is likely to represent a complex involving HspR multimers. Even if it is assumed that DnaK is a part of C3, the basic conclusion that DnaK gives protection against the thermal instability of HspR without forming a complex remains unchanged, as the mobilities of the DnaK-independent complexes C1 and C2 remain unaffected. It has been demonstrated in this study that DnaK has high affinity for wild-type protein, particularly when it is present in the denatured state. It is possible that a small amount of incompletely renatured protein present in the sample may have interacted with DnaK, resulting in a C3-type complex.

Unlike protection, renaturation of the denatured protein obligatorily requires formation of a stable complex between DnaK and HspR. Under this condition, a single retarded complex (C4) which migrates more slowly than all the other complexes is formed. Supershift assays convincingly demonstrate that DnaK is a part of the complex. The results of protein-protein interaction experiments also indicate that if HspR-WT is initially in the denatured state, then the formation of a DnaK-HspR complex is facilitated. In contrast, when HspR-ΔC was renatured in the same way, C4 did not form and the DnaK-independent complexes C1 and C2 were the predominant species. This indicates that DnaK did not interact with HspR-ΔC, even when it was present in the denatured state. This observation finds support from the results of the protein-protein interaction experiment, which shows that, indeed, HspR-ΔC has little or no DnaK binding activity, even when it is present in the denatured state. The results thus obtained clearly indicate that the C-terminal hydrophobic tail of HspR which includes the LVVW motif is the primary site where DnaK binds. However, considering that only the denatured form of the WT protein gave a high level of DnaK binding activity, it may be concluded that the presence of the hydrophobic tail is not enough and that it must be in the exposed state so that interaction with DnaK is possible. A minor complex of the C3 type was nevertheless seen in the case of denatured HspR-ΔC. Hence, although the LVVW motif constitutes the primary DnaK binding site, there may be others which are secondary in nature.

To understand how the C-terminal deletion affects the repressor activity of HspR, an in vivo experiment was performed using E. coli as a surrogate host. E. coli does not have an ortholog of HspR. Hence, there is no possibility of background interference. The ability of HspR to repress the activity of the dnaKJE-hspR operon promoter in trans was demonstrated using reporter assays. Derepression of the operon promoter was found to occur under heat shock conditions. HspR-WT, however, appeared to be a weak repressor, as it repressed only partially. In contrast, HspR-ΔC proved to be a stronger repressor which reduced the activity of the promoter to the level observed in the case of the promoterless vector. When subjected to heat shock, no evidence of derepression was observed in the case of HspR-ΔC. This observation gives a valuable clue regarding the physiological significance of the presence of the C-terminal hydrophobic motif. It is due to the presence of this motif that the protein becomes unstable and, therefore, the system responds to heat efficiently. It is interesting that although HspR-WT and HspR-ΔC were demonstrated to have nearly identical DNA binding affinities in vitro, in vivo, HspR-ΔC was found to be more efficient. This can be explained on the basis of the differences in the thermal stabilities of the WT and mutant proteins. While the in vitro DNA binding experiments were all performed at 0°C, the in vivo experiments required that the temperature be maintained at 37°C so as to allow optimal bacterial growth. The WT protein, being thermally unstable, may be only partially active at the relatively high temperature of 37°C, and therefore, repression was leaky. In the case of HspR-ΔC, which is thermally stable, there is no such constraint, and hence, it functions as a better repressor.

HspRs can be divided into two types depending on their solubility—the relatively insoluble type found in the Actinomycetales (2, 8, 29) and the more-soluble types found in Campylobacterales (23) and bacteria belonging to the phylum Thermus-Deinococcus (22). What makes one particular HspR more insoluble and, hence, more dependent on feedback control is an intriguing question. The LVVW motif located in the C-terminal region of HspR, which is highly conserved in the Actinomycetales but not in Campylobacterales or Thermus-Deinococcus bacteria, may play a deciding role. In the case of HrcA, too, similar regulation occurs at least in some cases. Thus, the C-terminal tail of chlamydial HrcA was found to inhibit its DNA binding activity, and in this case, GroEL can relieve this inhibition, apparently by interacting with the tail (5). Negative regulation of heat shock operon repressors by their own C-terminal tails may constitute a universal strategy by which Eubacteria control their responses to heat shock.


We thank CSIR and DST (Government of India) for financial support.


Published ahead of print 29 June 2012


1. Basu A, Chawla-Sarkar M, Chakrabarti S, Das Gupta SK. 2002. Origin binding activity of the mycobacterial plasmid pAL5000 replication protein RepB is stimulated through interactions with host factors and coupled expression of repA. J. Bacteriol. 184:2204–2214 [PMC free article] [PubMed]
2. Bucca G, Brassington AM, Schonfeld HJ, Smith CP. 2000. The HspR regulon of Streptomyces coelicolor: a role for the DnaK chaperone as a transcriptional co-repressor. Mol. Microbiol. 38:1093–1103 [PubMed]
3. Bukau B. 1993. Regulation of the Escherichia coli heat-shock response. Mol. Microbiol. 9:671–680 [PubMed]
4. Chang Z, et al. 1996. Mycobacterium tuberculosis 16-kDa antigen (Hsp16.3) functions as an oligomeric structure in vitro to suppress thermal aggregation. J. Biol. Chem. 271:7218–7223 [PubMed]
5. Chen AL, Wilson AC, Tan M. 2011. A Chlamydia-specific C-terminal region of the stress response regulator HrcA modulates its repressor activity. J. Bacteriol. 193:6733–6741 [PMC free article] [PubMed]
6. Creighton TE. 1990. Protein folding. Biochem. J. 270:1–16 [PubMed]
7. Das Gupta SK, Bashyam MD, Tyagi AK. 1993. Cloning and assessment of mycobacterial promoters by using a plasmid shuttle vector. J. Bacteriol. 175:5186–5192 [PMC free article] [PubMed]
8. Das Gupta T, Bandyopadhyay B, Das Gupta SK. 2008. Modulation of DNA-binding activity of Mycobacterium tuberculosis HspR by chaperones. Microbiology 154:484–490 [PubMed]
9. Das Gupta SK, Jain S, Kaushal D, Tyagi AK. 1998. Expression systems for study of mycobacterial gene regulation and development of recombinant BCG vaccines. Biochem. Biophys. Res. Commun. 246:797–804 [PubMed]
10. Dayhoff MO, Barker WC, Hunt LT. 1983. Establishing homologies in protein sequences. Methods Enzymol. 91:524–545 [PubMed]
11. Hendrick JP, Hartl FU. 1993. Molecular chaperone functions of heat-shock proteins. Annu. Rev. Biochem. 62:349–384 [PubMed]
12. Jain S, Kaushal D, DasGupta SK, Tyagi AK. 1997. Construction of shuttle vectors for genetic manipulation and molecular analysis of mycobacteria. Gene 190:37–44 [PubMed]
13. Klostermeier D, Seidel R, Reinstein J. 1998. Functional properties of the molecular chaperone DnaK from Thermus thermophilus. J. Mol. Biol. 279:841–853 [PubMed]
14. Narberhaus F. 1999. Negative regulation of bacterial heat shock genes. Mol. Microbiol. 31:1–8 [PubMed]
15. Nozaki Y, Tanford C. 1971. The solubility of amino acids and two glycine peptides in aqueous ethanol and dioxane solutions. Establishment of a hydrophobicity scale. J. Biol. Chem. 246:2211–2217 [PubMed]
16. Ratnakar PV, Mohanty BK, Lobert M, Bastia D. 1996. The replication initiator protein pi of the plasmid R6K specifically interacts with the host-encoded helicase DnaB. Proc. Natl. Acad. Sci. U. S. A. 93:5522–5526 [PubMed]
17. Richter K, Haslbeck M, Buchner J. 2010. The heat shock response: life on the verge of death. Mol. Cell 40:253–266 [PubMed]
18. Rudiger S, Mayer MP, Schneider-Mergener J, Bukau B. 2000. Modulation of substrate specificity of the DnaK chaperone by alteration of a hydrophobic arch. J. Mol. Biol. 304:245–251 [PubMed]
19. Rudiger S, Schneider-Mergener J, Bukau B. 2001. Its substrate specificity characterizes the DnaJ co-chaperone as a scanning factor for the DnaK chaperone. EMBO J. 20:1042–1050 [PubMed]
20. Sacchettini JC, Rubin EJ, Freundlich JS. 2008. Drugs versus bugs: in pursuit of the persistent predator Mycobacterium tuberculosis. Nat. Rev. Microbiol. 6:41–52 [PubMed]
21. Saitou N, Nei M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425 [PubMed]
22. Schmid AK, Howell HA, Battista JR, Peterson SN, Lidstrom ME. 2005. HspR is a global negative regulator of heat shock gene expression in Deinococcus radiodurans. Mol. Microbiol. 55:1579–1590 [PubMed]
23. Spohn G, Scarlato V. 1999. The autoregulatory HspR repressor protein governs chaperone gene transcription in Helicobacter pylori. Mol. Microbiol. 34:663–674 [PubMed]
24. Stewart GR, et al. 2005. The stress-responsive chaperone alpha-crystallin 2 is required for pathogenesis of Mycobacterium tuberculosis. Mol. Microbiol. 55:1127–1137 [PubMed]
25. Stewart GR, et al. 2001. Overexpression of heat-shock proteins reduces survival of Mycobacterium tuberculosis in the chronic phase of infection. Nat. Med. 7:732–737 [PubMed]
26. Tabor S, Richardson CC. 1985. A bacteriophage T7 RNA polymerase/promoter system for controlled exclusive expression of specific genes. Proc. Natl. Acad. Sci. U. S. A. 82:1074–1078 [PubMed]
27. Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24:1596–1599 [PubMed]
28. Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673–4680 [PMC free article] [PubMed]
29. Ventura M, Kenny JG, Zhang Z, Fitzgerald GF, van Sinderen D. 2005. The clpB gene of Bifidobacterium breve UCC 2003: transcriptional analysis and first insights into stress induction. Microbiology 151:2861–2872 [PubMed]
30. Vidal-Aroca F, et al. 2006. One-step high-throughput assay for quantitative detection of beta-galactosidase activity in intact gram-negative bacteria, yeast, and mammalian cells. Biotechniques 40:433–438 [PubMed]
31. Werner MH, Clore GM, Gronenborn AM, Kondoh A, Fisher RJ. 1994. Refolding proteins by gel filtration chromatography. FEBS Lett. 345:125–130 [PubMed]
32. Yura T, Nagai H, Mori H. 1993. Regulation of the heat-shock response in bacteria. Annu. Rev. Microbiol. 47:321–350 [PubMed]
33. Zhang Y. 2004. Persistent and dormant tubercle bacilli and latent tuberculosis. Front. Biosci. 9:1136–1156 [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)