Dissection of the heat-shock response in Mycobacterium tuberculosis using mutants and microarraysa

Graham R. Stewart1, Lorenz Wernisch2, Richard Stabler3, Joseph A. Mangan3, Jason Hinds3, Ken G. Laing3, Douglas B. Young1 and Philip D. Butcher3

Department of Infectious Diseases and Microbiology, Centre for Molecular Microbiology and Infection, Imperial College of Science Technology and Medicine, London SW7 2AZ, UK1
School of Crystallography, Birkbeck College, University of London, Malet Street, London WC1E 7HX, UK2
Department of Medical Microbiology, St George’s Hospital Medical School, Cranmer Terrace, London SW17 0RE, UK3

Author for correspondence: Graham Stewart. Tel: +44 207 594 3090. Fax: +44 207 594 3095. e-mail: g.stewart{at}ic.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Regulation of the expression of heat-shock proteins plays an important role in the pathogenesis of Mycobacterium tuberculosis. The heat-shock response of bacteria involves genome-wide changes in gene expression. A combination of targeted mutagenesis and whole-genome expression profiling was used to characterize transcription factors responsible for control of genes encoding the major heat-shock proteins of M. tuberculosis. Two heat-shock regulons were identified. HspR acts as a transcriptional repressor for the members of the Hsp70 (DnaK) regulon, and HrcA similarly regulates the Hsp60 (GroE) response. These two specific repressor circuits overlap with broader transcriptional changes mediated by alternative sigma factors during exposure to high temperatures. Several previously undescribed heat-shock genes were identified as members of the HspR and HrcA regulons. A novel HspR-controlled operon encodes a member of the low-molecular-mass {alpha}-crystallin family. This protein is one of the most prominent features of the M. tuberculosis heat-shock response and is related to a major antigen induced in response to anaerobic stress.

Keywords: HspR, HrcA, Hsp70, Hsp60, transcriptional regulator

Abbreviations: ADC, albumin/dextrose (glucose)/catalase; CIRCE, controlling inverted repeat of chaperone expression; HAIR, HspR-associated inverted repeat; OADC, oleic acid/dextrose (glucose)/albumin/catalase

a A list of the 100 ORFs most highly induced by heat shock is provided as supplementary data with the online version of this paper (http://mic.sgmjournals.org).


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The heat-shock response is a ubiquitous adaptive pathway involved in the survival of cells exposed to a sudden increase in ambient temperature. It is characterized by global transcriptional changes including elevated expression of a set of highly conserved heat-shock proteins. Heat shock has been exploited as a model system for studying gene regulation in a wide variety of cell types, and analysis of the proteins themselves has provided fundamental insights into cell biology. The major heat-shock proteins function as molecular chaperones, ensuring appropriate folding, translocation and assembly of polypeptide structures (Hartl, 1996 ). Increased expression of heat-shock proteins is triggered by a range of stress conditions, with accumulation of unfolded polypeptides representing the likely common stimulus (Narberhaus, 1999 ). Heat-shock proteins are induced in both host and pathogen during the process of infection, helping to maintain cell integrity and contributing to immune signalling (Asea et al., 2000 ) and recognition of the pathogen (Castellino et al., 2000 ; Suto & Srivastava, 1995 ). We have recently demonstrated that partial disruption of heat-shock regulation in Mycobacterium tuberculosis has an important impact on virulence, impairing the ability of the bacteria to establish a chronic infection (Stewart et al., 2001 ). The aim of the present study was to characterize transcriptional regulation of the major heat-shock proteins in this important pathogen.

We have previously demonstrated that the regulation of the 70 kDa heat-shock protein, Hsp70 or DnaK, is mediated, as in Bacillus subtilis, by release of transcriptional repression (Stewart et al., 2001 ). In addition, positive regulation of Hsp70 via an alternative sigma factor, as in Escherichia coli (Grossman et al., 1984 ), has also recently been demonstrated (Raman et al., 2001 ). In B. subtilis, a single repressor, HrcA, coregulates the Hsp70 heat-shock genes together with those encoding the Hsp60 GroE family (Hecker et al., 1996 ). In contrast, deletion of the Hsp70 regulator, HspR, in M. tuberculosis had no detectable effect on expression of Hsp60 proteins (Stewart et al., 2001 ). This suggests that a second repressor which has similarity to HrcA may play a role in the mycobacterial heat-shock response. To test this hypothesis, we have generated mutant strains of M. tuberculosis with deletions affecting both of the repressors and characterized the resulting effects on patterns of gene expression.

In order to study transcriptional regulation at a global level, we have used microarrays prepared by spotting DNA sequences representing 3924 of the ORFs identified in the M. tuberculosis H37Rv genome sequence (Cole et al., 1998 ). Competitive hybridization to the microarray using RNA/cDNA prepared from different strains of M. tuberculosis or from bacteria exposed to different conditions provides a window on the overall pattern of transcriptional changes. By combining whole-genome analysis with targeted mutagenesis, we have been able to distinguish the contribution of two regulatory circuits to the overall heat-shock response and to reveal a number of novel heat-shock genes.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
All DNA vector construction was performed in Escherichia coli DH5{alpha}. E. coli were grown at 37 °C in Luria–Bertani broth and agar containing 150 µg hygromycin ml-1 or 50 µg kanamycin ml-1 where appropriate. M. tuberculosis H37Rv, {Delta}hspR and {Delta}hspR {Delta}hrcA were grown at 37 °C in Middlebrook 7H9 broth (Difco) containing 10% albumin/dextrose(glucose)/catalase (ADC) enrichment or on Middlebrook 7H11 agar medium (Difco) containing 10% oleic acid/dextrose (glucose)/albumin/catalase (OADC) enrichment. Hygromycin at 50 µg ml-1 and kanamycin at 15 µg ml-1 were added where appropriate. Two per cent sucrose was added to media for counterselection of sacB. Heat shock was performed by splitting 20 ml broth cultures at late exponential phase into two universal tubes and placing one tube at 37 °C and the other at 45 °C for 30 min.

Deletion of hspR, hrcA in M. tuberculosis.
The gene replacement of hspR with the hygromycin B phosphotransferase gene (hyg) from Streptomyces hygroscopicus has been described previously (Stewart et al., 2001 ). The sequential deletion of hrcA to generate a double hspR hrcA mutant strain was achieved using a similar suicide delivery strategy but replacing the target gene, hrcA, with the kanamycin-resistance gene (aph) from Tn903. Briefly, 1·5 kb regions of DNA up- and downstream of hrcA were cloned around the aph gene in the mycobacterial suicide plasmid pSMT99 to make pSMT163. This plasmid cannot replicate in mycobacteria and carries sacB for counterselection against single crossover and illegitimate integration of the plasmid. One microgram of plasmid was irradiated with 100 mJ cm-2 UV (Hinds et al., 1999 ) and electroporated into M. tuberculosis or M. tuberculosis {Delta}hspR (Wards & Collins, 1996 ). Following overnight recovery of the cells in 7H9/ADC, gene replacement transformants were directly selected on 7H11/OADC containing hygromycin, kanamycin and sucrose. Deletion of hrcA was confirmed by Southern blotting of KpnI digested genomic DNA using the 1·5 kb upstream hrcA fragment as hybridization probe.

Complementation of M. tuberculosis {Delta}hspR.
pKinta is a ColE1-based E. coli plasmid which carries the aph kanamycin-resistance gene and the int gene and attP site from the L5 mycobacteriophage (Stover et al., 1992 ). This plasmid integrates into the chromosome in a single copy by site-specific recombination at the attB site. The Hsp70 operon promoter containing the two HspR-associated inverted repeat (HAIR)-regulated promoter regions (Stewart et al., 2001 ) was amplified by PCR using the primers Hsp701 (tcggtcaagctggcggactga) and Hsp702 (agccatggtgaatcctcctg) and cloned into the SacI site of pKinta. The hspR ORF was then amplified and cloned downstream of the hsp70 promoter so as to transcriptionally fuse the ORF with its own promoter albeit without the intervening hsp70, grpE and dnaJ sequence. The resultant plasmid, pSMT168, was introduced into M. tuberculosis {Delta}hspR by electroporation.

RNA extraction and cDNA labelling.
Ten millilitres of broth culture in late exponential phase was added directly to 40 ml GTC solution containing 5 M guanidinium thiocyanate, 0·5% sodium N-lauryl sarcosine, 0·1 M ß-mercaptoethanol and 0·5% Tween 80. The bacteria were pelleted by centrifugation and resuspended in 1·2 ml Trizol (Life Technologies) with 0·5 ml of 0·1 mm silica/ceramic beads and processed at 6·5 W for 45 s in a Ribolyser (reciprocal shaker: Hybaid). The samples were centrifuged at 13000 r.p.m. for 15 min in a microfuge to remove bacterial debris and the supernatant removed to a fresh tube. The phases were separated by the addition of 0·6 ml chloroform, mixing and centrifugation. The aqueous phase was re-extracted with chloroform and the RNA precipitated with propan-2-ol, washed in 70% ethanol and dissolved in RNase-free water. The RNA was treated with amplification-grade DNase I (Life Technologies) and cleaned up by RNeasy purification (Qiagen).

cDNA was labelled by incorporation of Cy3 or Cy5 dCTP (Amersham) during reverse transcription of RNA. RNA (2–10 µg) was mixed with 3 µg random hexamer oligonucleotides in 11 µl water, heated to 95 °C and snap-cooled. In a total volume of 25 µl, the labelling reaction was initiated by the addition of 5 µl First Strand Buffer, 25 mM DTT, 1 mM each dATP, dGTP and dTTP, 0·4 mM dCTP, 2 nmol Cy3- or Cy5-dCTP and 500 U Superscript II reverse transcriptase (Life Technologies). The reaction was incubated in the dark at 25 °C for 10 min and then at 42 °C for 90 min. The relevant pairs of Cy3- (wild-type H37Rv) and Cy5- (mutant strain or heat-shocked cells) labelled cDNA were mixed and purified using a Qiagen MinElute kit, eluting in water.

Microarrays and hybridizations.
Whole genome microarrays were constructed by robotic spotting on to poly-L-lysine-coated glass microscope slides (MicroGrid II, BioRobotics, UK) of PCR amplicons (size range 60–1000 bp; mean 517 bp) derived from portions of each of the 3924 predicted ORFs of the sequenced strain of M. tuberculosis H37Rv. Primer pairs for each ORF were designed with Primer 3 software and selected by BLAST analysis to have minimal cross-homology with all other ORFs. All procedures used, including post-processing of deposited arrays, were as described by others (Wilson et al., 1999 , 2001 ).

The microarray was incubated in prehybridization solution (3·5xSSC, 0·1% SDS and 10 mg BSA ml-1) at 65 °C for 20 min. The slide was rinsed in water for 1 min and propan-2-ol for 1 min before drying by centrifugation at 1500 r.p.m. for 5 min.

The purified Cy3/Cy5-labelled cDNA was adjusted to 16 µl in 4x SSC and 0·3% SDS. This hybridization solution was heated to 95 °C for 2 min, briefly centrifuged and applied to the array under a cover slip. The slide was sealed in a humid hybridization cassette and incubated at 65 °C in the dark for 16–20 h. The slide was washed for 2 min at 65 °C in 1x SSC/0·05% SDS, for 2 min at room temperature in 0·06x SSC and then dried by centrifugation. The hybridized microarrays were scanned with an Affymetrix 428 scanner. The scanned images were analysed with ImaGene4.1 and the median spot intensities calculated.

Data processing and statistical analysis.
For each strain or condition, three or four independent RNA preparations were analysed. Background values were subtracted from signal values. In cases where this resulted in negative values, a small positive constant was assigned to prevent numerical problems when forming ratios or taking logarithms. Data were normalized to median signal intensities. All values were log2-transformed for further analysis.

Significance values were calculated for each ORF in the mutant:wild-type comparisons through an ANOVA analysis. Each of the three datasets (wild-type versus {Delta}hspR; wild-type versus {Delta}hspR pSMT168; wild-type versus {Delta}hspRhrcA) forms a balanced factorial design. Three main effects were taken into account: the array effect A for each array, the gene effect G for each gene and the variety effect V for the two varieties, mutant and control. In addition, the three pairwise interactions between the main effects, that is, interactions A:V, A:G and V:G, have been accounted for. The resulting residuals stem from the A:V:G interaction of all three main effects which were used to estimate the standard error. One problem is that the residual or error variance is much higher for low expression values, which is not unexpected considering the higher uncertainty in these values. Hence, using one standard error estimate for all genes does not seem appropriate. Instead, we resampled from the residuals, redistributed them over the expected response values and fitted new models to these bootstrap replicates. The multitude of models allowed us to calculate confidence intervals for the estimates of the effects. The final value is based on the difference in estimated V:G effects, which represents the influence of variety, that is, mutant or control, on gene expression. Confidence intervals for these effects are calculated through the resampling procedure as above. Final P-values are obtained from confidence intervals by Bonferroni correction for multiple testing, that is, all raw P-values are multiplied by the number of genes resulting in the final adjusted P-values.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Overview of the M. tuberculosis heat-shock response
Previous reports have described the induction of heat-shock proteins in cultures of M. tuberculosis exposed to temperatures ranging from 37 to 48 °C for varying lengths of time and demonstrated transcriptional regulation of selected heat-shock genes (Patel et al., 1991 ; Young & Garbe, 1991 ). These studies demonstrate a complex response, which varies with both temperature and time of exposure. To obtain an overview of the heat-shock response, we used whole-genome microarray analysis to generate a transcriptome snapshot of the changes induced by incubation at 45 °C for 30 min, conditions previously demonstrated to result in high-level expression particularly of the Hsp70 regulon. This is displayed in the scatter plot (Fig. 1a), which shows the global nature of the transcriptional changes induced by heat shock, the expression ratio of many genes lying away from the zero line demonstrating altered expression. A list of the 100 most highly induced ORFs is provided as supplementary data with the online version of this paper (http://mic.sgmjournals.org). The functional distribution of the induced genes varied from that found across the genome, with a bias towards heat induction of adaptation/detoxification and regulatory genes, and away from cell-wall-associated genes (Fig. 2). The induced genes included all the known members of the HspR regulon, as well as the groEL and groES genes and other previously identified heat-shock-inducible genes including those encoding the alternative sigma factors {sigma}B, {sigma}H and {sigma}E (Fernandes et al., 1999 ; Manganelli et al., 1999 ). This set of heat-inducible genes included five of the nine genes preceded by a {sigma}E consensus promoter sequence (Manganelli et al., 2001 ) and all seven genes identified by Raman et al. (2001) as containing {sigma}H consensus promoter regions. This is consistent with identification of these sigma factors as both heat-inducible genes and regulators of the heat-shock response. To characterize regulation of genes encoding the major heat-shock proteins, we next extended the microarray approach to analysis of mutant strains of M. tuberculosis from which predicted transcriptional repressors had been deleted.



View larger version (18K):
[in this window]
[in a new window]
 
Fig. 1. Gene expression profiles of M. tuberculosis during heat shock and of M. tuberculosis lacking the transcriptional repressor, HspR. Scatter plots show log Cy5/Cy3 signal ratios against log total signal intensity, where log ratios are centralized such that mean log Cy5 and Cy3 are equal to zero. (a) Expression of M. tuberculosis genes at 45 °C (Cy5) versus 37 °C (Cy3). (b) Expression in M. tuberculosis {Delta}hspR (Cy5) versus wild-type M. tuberculosis H37Rv (Cy3) at 37 °C. (c) Expression in M. tuberculosis {Delta}hspR complemented with a functional copy of hspR on the integrating plasmid pSMT168 (Cy5) versus wild-type M. tuberculosis H37Rv (Cy3) at 37 °C.

 


View larger version (19K):
[in this window]
[in a new window]
 
Fig. 2. Functional distribution of genes upregulated during heat shock. Frequency of genes among functional groups (http://genolist.pasteur.fr/TubercuList/) across the genome (white bars) and among heat-shock-induced genes (black bars).

 
The HspR regulon
By examining the gene expression profile at 37 °C of an M. tuberculosis strain lacking the transcriptional repressor HspR ({Delta}hspR), we aimed to isolate any de-repressed genes and identify the subset of heat-inducible genes directly under HspR control. In contrast to the heat-shocked bacteria, transcription of the majority of genes was unaltered in the mutant strain, but there were several obvious upregulated genes (Fig. 1b). ANOVA analysis also revealed the less obvious upregulated genes, exposing a set of 49 upregulated ORFs (P<0·01) in the mutant strain, including the members of the Hsp70 operon (dnaK, grpE and dnaJ) (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Upregulated genes in M. tuberculosis {Delta}hspR compared with wild-type H37Rv

 
We searched the genome for sequences that resembled the HspR binding site, HAIR CTTGAGT-N7-ACTCAAG (Grandvalet et al., 1999 ) and compared the locations of potential sites to the gene expression analysis of both heat-shocked M. tuberculosis and M. tuberculosis {Delta}hspR. In addition to the HAIR sequences already identified upstream of the Hsp70 operon and clpB (Stewart et al., 2001 ), a HAIR-like domain was present 71 bp upstream of the start codon of Rv0251c (Fig. 3a). This gene bears similarity to the {alpha}-crystallin (acr)/14 kDa antigen of M. tuberculosis (41% identity over 98 amino acids), so we have termed it acr2. It appears to be at the head of an operon preceding Rv0250c and Rv0249c, as these are also upregulated in the mutant (Table 1) and are arranged with minimal intergenic regions. There were no other HAIR-like sequences associated with any of the other upregulated genes in the {Delta}hspR strain.



View larger version (34K):
[in this window]
[in a new window]
 
Fig. 3. Heat-shock repressor binding sites within M. tuberculosis. (a) HspR associated inverted repeat or HAIR sequences. (b) HrcA binding sites or CIRCE.

 
As expected, the Hsp70 operon genes along with acr2 and Rv0250c were upregulated in response to heat shock. Under the conditions used in this study, acr2 was the most heat-inducible gene in the genome (Fig. 1a). Other {Delta}hspR-regulated ORFs demonstrated to be induced under heat shock were Rv3654c, bfrB and groES. Rv3654c encodes an 8 kDa protein of unknown function, and bfrB encodes a bacterioferritin involved in iron acquisition; neither gene has an identifiable HAIR-like sequence in its vicinity, and both may therefore be under some indirect control by HspR. Perhaps most interesting is the inclusion of the chaperone gene groES, as our previous studies had not indicated that this gene was controlled by HspR. Indeed, the level of induction is considerably less than that of the Hsp70 or Acr2 operons. The HspR associated control over groES expression may be indirect, as there is no HAIR sequence in the promoter region, but there is a weak HAIR-like sequence situated 24 bases downstream of the groES initiation codon. The remaining non-heat-induced genes upregulated in the {Delta}hspR mutant presumably reflect adaptive responses triggered by constitutive overexpression of the genes normally controlled by HspR. Notable members of this group included genes encoding the alternative sigma factor {sigma}C, the Sec-independent protein translocase, TatA, and four ribosomal proteins. Indeed, there was a general trend among nearly all the ribosomal protein genes to be upregulated in the {Delta}hspR mutant.

We have previously described unsuccessful attempts to complement the M. tuberculosis {Delta}hspR strain (Stewart et al., 2001 ). Reintroduction of the gene with a constitutive promoter or even gently induced expression from the acetamidase promoter (Parish & Stoker, 1997 ) rendered the bacteria non-viable. These findings suggest that expression of reintroduced hspR would have to be appropriately regulated so as to closely match wild-type expression dynamics. To achieve this, the hspR gene was cloned under the control of the natural promoter of the hsp70 operon, which includes two HAIR sequences. A single copy of this construct was inserted at the attB phage integration site in the chromosome of M. tuberculosis {Delta}hspR. In contrast to previous attempts at complementation, this strain was fully viable. Whole-genome expression profiling of the complemented mutant showed a pattern largely similar to the original wild-type strain (Fig. 1c). The reintroduced hspR gene was approximately twofold overexpressed, demonstrating that the complementing construct did not express hspR identically to wild-type. This may be due to the changed position of hspR in relation to its promoter, or perhaps reflects some stoichiometric relationship between hspR expression and the number of HAIR sites. However, all the genes overexpressed in the {Delta}hspR strain showed a complete or substantial reduction of overexpression in the complemented strain (Table 1). This demonstrates that the altered transcriptome of the mutant was specifically due to the absence of hspR and not to polar effects on neighbouring genes or to an inadvertently selected mutation.

The HrcA regulon
ORF Rv2374c in the M. tuberculosis genome shares sequence homology with the family of heat-shock repressors related to the hrcA gene of B. subtilis. To test whether this ORF is similarly involved in heat-shock regulation in M. tuberculosis, we undertook a deletion strategy analogous to that used to generate the {Delta}hspR mutant, replacing hrcA with a kanamycin-resistance gene. We were unable to generate {Delta}hrcA mutants in wild-type M. tuberculosis, yet were successful at introducing the mutation into M. tuberculosis {Delta}hspR (Fig. 4a). SDS–PAGE analysis of the total protein profile of the double knock-out M. tuberculosis {Delta}hspR{Delta}hrcA demonstrated constitutive overexpression of proteins consistent in size with Hsp70, Hsp60 (GroEL) and GroES, as well as an additional band at approximately 20 kDa (Fig. 4b).



View larger version (76K):
[in this window]
[in a new window]
 
Fig. 4. Deletion of hrcA and hspR results in overexpression of Hsp70 (DnaK), Hsp60 (GroEL), Hsp10 (GroES) and a protein consistent in size with Acr2. (a) Southern blot of KpnI-digested genomic DNA demonstrating deletion of hrcA in M. tuberculosis {Delta}hspR. Lane 1, HindIII-digested {lambda} DNA; lane 2, M. tuberculosis {Delta}hspR (3634 bp wild-type hrcA hybridizing fragment); lane 3, M. tuberculosis {Delta}hspR{Delta}hrcA (6526 bp hrcA-deleted fragment). (b) Protein extracts of 37 °C cultured M. tuberculosis H37Rv (lane 1) and M. tuberculosis {Delta}hspR{Delta}hrcA (lane 2) separated by SDS-PAGE and stained with Coomassie brilliant blue.

 
Whole-genome expression profiling of M. tuberculosis {Delta}hspR{Delta}hrcA at 37 °C revealed enhanced expression of a set of 48 ORFs (P<0·01) (Table 2). Twelve ORFs upregulated in the single {Delta}hspR mutant were also upregulated in the {Delta}hspR{Delta}hrcA strain. These included members of the Hsp70 and Acr2 operons as well as sigC, tatA and groES. The upregulation of groES was much greater in the {Delta}hspR{Delta}hrcA mutant than in the {Delta}hspR strain (9·60- and 1·96-fold, respectively). This indicated that although transcription of groES can be induced by an HspR-associated mechanism, the predominant mode of transcriptional control is through the HrcA repressor. HrcA also seemed the likely mechanism of control for the two M. tuberculosis groEL genes, as these were both strongly upregulated in the {Delta}hspR{Delta}hrcA strain. We searched the genome for the HrcA binding site, controlling inverted repeat of chaperone expression (CIRCE) TTAGCACTC-N9-GAGTGCTAA (Hecker et al., 1996 ) and, as for HspR, compared the putative CIRCE locations with both the heat-shock expression data and the double mutant transcriptional profile. groEL2 is preceded by two CIRCE-like elements and groES/groEL1 by one (Fig. 3b). This confirmed the hypothesis that HrcA acts as the main regulator for the GroE/Hsp60 heat-shock protein family.


View this table:
[in this window]
[in a new window]
 
Table 2. Upregulated genes in M. tuberculosis {Delta}hspR{Delta}hrcA compared with wild-type H37Rv

 
A CIRCE-like sequence was also identified 28 bp upstream of the initiation codon of Rv0991c (Fig. 3b). This ORF is predicted to encode an 11·5 kDa conserved hypothetical protein and was significantly upregulated in the {Delta}hspR{Delta}hrcA mutant (Table 2). Both Rv0991c and the immediately adjacent downstream gene Rv0990c were upregulated after heat shock for 30 min at 45 °C in the wild-type. Although no significant change was detected in transcription of Rv0990c in the mutant strain, this suggests that the two genes may be coregulated. None of the remaining {Delta}hspR{Delta}hrcA upregulated genes were associated with CIRCE-like elements nor were they induced under heat shock in the wild-type. Similarly to the single {Delta}hspR mutant, there was a trend for ORFs encoding ribosomal proteins to be upregulated, but in addition, the gene encoding ribosome recycling factor, frr, was also significantly upregulated.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
M. tuberculosis currently infects one-third of the global population (Sudre et al., 1992 ), claiming several million lives every year. There is an urgent need for improved drugs and vaccines for treatment and prevention of the disease. The availability of the complete genome sequence of M. tuberculosis (Cole et al., 1998 ) represents a vast resource of information concerning the evolution and pathobiology of the organism, but it remains a considerable challenge to translate this information into functionally useful understanding. Whole-genome expression profiling provides an important approach to rapid accumulation of information about adaptive responses of M. tuberculosis. In a recent report by Sherman et al. (2001) , this approach was used to document the set of M. tuberculosis genes induced by exposure to oxygen deprivation, allowing identification of one of the regulatory elements involved in the hypoxic response. Manganelli et al. (2001) combined targeted mutagenesis with expression profiling to characterize the genes transcribed by the heat-inducible alternative sigma factor, {sigma}E. In the present report, we have used deletion mutants to characterize the two regulatory circuits responsible for the control of the Hsp60 and Hsp70 heat-shock response, along with the expression of other newly identified heat-shock proteins.

Heat shock is both a convenient and robust laboratory model for studying transcriptional regulation as well as an important element in the pathogenesis of M. tuberculosis. Heat-shock proteins are induced following uptake of M. tuberculosis by host cells (Lee & Horwitz, 1995 ; Monahan et al., 2001 ), presumably contributing to bacterial survival in the phagocytic environment. However, they are also used by the host cell as peptide chaperones to enhance MHC class I presentation (Castellino et al., 2000 ; Suto & Srivastava, 1995 ) and as a signal to the immune system (Asea et al., 2000 ). Directly because of these immunological roles of heat-shock proteins, the constitutive overexpression of bacterial Hsp70 proteins results in reduced survival of M. tuberculosis during prolonged infection (Stewart et al., 2001 ). By understanding the regulation of the heat-shock response, we may be able to design interventions capable of enhancing resistance to persistent infection with tuberculosis.

Microarray analysis of an hspR deletion mutant of M. tuberculosis confirms and extends previous studies of Hsp70 regulation. HspR is a DNA-binding protein related to the MerR family. It recognizes either of two inverted repeat sequences (HAIR) in the promoter region of the hsp70 operon, reducing the level of transcription in unstressed conditions. The HspR protein interacts tightly with Hsp70 in vitro (Bucca et al., 2000 ; Stewart et al., 2001 ). A system where this heterodimer forms the functional repressor unit with feedback achieved by titration of Hsp70 away from the HspR complex in the presence of unfolded polypeptides represents an attractive model for regulation (Bucca et al., 2000 ; Narberhaus, 1999 ). As predicted, we show that in the absence of HspR, there is release of transcriptional repression, and the genes of the Hsp70 operon are upregulated. There were also a further 46 genes with significantly elevated transcription. Of these, only three genes (Rv0249c–Rv0251c), arranged consecutively in an apparent operon, were associated with a HAIR-like sequence. Interestingly, the lead gene Rv0251c has also been shown to be under the control of the heat-shock responsive ECF sigma factor, {sigma}E, and is also prominent in response to treatment with SDS (Manganelli et al., 2001 ). This dual control mechanism may account for the relatively modest elevation of Rv0251c transcription in the {Delta}hspR mutant compared with that observed under heat-shock conditions in the wild-type. Rv0251c encodes a 159 amino acid protein belonging to the small heat-shock protein family, termed Hsp20, or the {alpha}-crystallin family. Its predicted size is consistent with the approximately 20 kDa protein found by SDS-PAGE to be upregulated in the {Delta}hspR{Delta}hrcA mutant (Fig. 3b). The small heat-shock proteins, like the larger heat-shock protein families, are found widely in bacterial and eukaryotic cells, and appear to function as molecular chaperones at least in vitro (Chang et al., 1996 ; Yang et al., 1999 ). There are two members of this family in M. tuberculosis. The other family member was originally identified as a prominent antigen and is variously referred to as the 14 kDa antigen, 16 kDa antigen, Hsp16·3, {alpha}-crystallin (Acr) or HspX. This gene is not induced by heat shock but is upregulated in stationary-phase cultures and during the hypoxic response (Cunningham & Spreadbury, 1998 ; Sherman et al., 2001 ; Yuan et al., 1996 , 1998 ). It is possible that the different Acr homologues fulfil analogous functional roles in response to different stresses. The Acr gene is induced following phagocytosis of M. tuberculosis (Monahan et al., 2001 ) and is required for growth in macrophages (Yuan et al., 1998 ). It will be of interest to determine whether the protein encoded by Rv0251c, which we term Acr2, also plays a role during infection.

Within the {Delta}hspR-upregulated ORF set, the Hsp70 and Acr2 operon genes were upregulated during heat shock along with bfrB, groES and Rv3654c. The bacterioferritin gene, bfrB, and Rv3654c, encoding an 8 kDa protein with unknown function, are not preceded by obvious HspR binding sites, but their coregulation with HAIR-associated genes in both heat shock and the mutant suggest an indirect link to HspR. The majority of genes upregulated in the mutant were neither associated with HAIR sequences nor upregulated during heat shock. We conclude that the induction of these genes is a consequence of the physiological changes associated with overexpression of the HspR-regulated proteins and may not be directly relevant to the normal heat-shock response. An interesting example of this was the trend for upregulation of ribosomal protein expression, which was also mirrored in the {Delta}hspR{Delta}hrcA strain.

A surprising omission from the {Delta}hspR upregulated list was clpB, which encodes another probable molecular chaperone. We have previously shown the elevation of ClpB expression in the mutant by proteomic analysis (Stewart et al., 2001 ), which suggests that the clpB mRNA is of a sufficiently short half-life to preclude detection of the {Delta}hspR-associated transcriptional increase. The detection of substantially increased clpB mRNA in the wild-type after heat shock at 45 °C is explained by upregulation of clpB transcription by the heat-inducible sigma factor, {sigma}H, as well as release of HspR repression (Raman et al., 2001 ).

We were able to delete the proposed hrcA gene in the {Delta}hspR mutant, but the same approach has been unsuccessful with wild-type M. tuberculosis. This may reflect some technical problem, but it is also possible that overexpression of Hsp70 proteins compensates in some way for a deleterious effect of hrcA deletion. Upregulation of the major HspR-regulated genes was preserved in the double mutant, alongside upregulation of the HrcA regulon, which included the Hsp60 family genes, groES, groEL1 and groEL2. GroES is functionally related to GroEL, and its gene is situated immediately upstream of groEL1. While the expression of groES was enhanced in the {Delta}hspR mutant, its upregulation in the {Delta}hspR{Delta}hrcA strain was much greater. The M. tuberculosis HrcA protein has yet to be analysed for DNA binding in vitro, but it has a strong sequence similarity to B. subtilis HrcA, and analogous CIRCE-like structures are present in the groES/groEL1 and groEL2 promoter regions. Thus, we can conclude that the HrcA repressor acts as the main transcriptional controller of the Hsp60/GroE family heat-shock response, with some cross-talk between the Hsp60 and Hsp70 responses demonstrated by the induction of GroES expression in the hspR deleted strain. The mechanism for this cross-talk is unclear, although a weak match for the HspR binding site, HAIR, is present at the beginning of the GroES ORF. Interaction of HspR with this inverted repeat could conceivably have a more subtle effect on transcription than that observed with HAIR sequences that directly overlap the RNA polymerase footprint.

A good match for the CIRCE sequence is found upstream of another {Delta}hspR{Delta}hrcA upregulated gene, Rv0991c, which encodes a conserved hypothetical protein with unknown function. Expression of both Rv0991c and the adjacent downstream ORF, Rv0990c, was elevated during heat shock, but Rv0990c was not significantly upregulated in the mutant. Whether the two genes are transcribed as a bicistronic message or are separately regulated and transcribed remains to be conclusively determined. Thus, it is clear that HrcA regulates not just the Hsp60 heat-shock response but also Rv0991c and probably Rv0990c. In light of the effect of the {Delta}hspR mutation on the virulence of M. tuberculosis (Stewart et al., 2001 ), it will be of considerable interest to study the double mutant in infection models.

It is obvious from the 45 °C transcriptional snapshot that the HspR and HrcA regulons, which dominate the heat-shock proteome, constitute only a part of the overall adaptive response. Genes regulated by {sigma}H and {sigma}E are prominent in the 45 °C response, and upregulation of the {sigma}B gene suggests overlap with the general stress response. These different regulatory layers are interlinked, with hsp70 and clpB under dual HspR and {sigma}H control, and acr2 under dual HspR and {sigma}E control. Moreover, the heat-inducible expression of {sigma}B and {sigma}E is dependent on {sigma}H, which autoregulates its own expression (Raman et al., 2001 ). In addition, it is probable that the functional activity of the sigma factors is subject to post-translational control by anti-sigma factor pathways (Helmann, 1999 ). Detailed analysis of bacteria exposed to different temperatures for different time periods will be important in further dissection of this complex pattern of regulatory circuits.


   ACKNOWLEDGEMENTS
 
We thank Brian Robertson and Peadar O’Gaora for help with this study. This work was supported by a Wellcome Trust Programme Grant to D.B.Y. and by The Wellcome Trust functional genomics Development Initiative (grant number: 062511) to P.D.B. (http://www.sghms.ac.uk/depts/medmicro/bugs/).


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Asea, A., Kraeft, S. K., Kurt-Jones, E. A., Stevenson, M. A., Chen, L. B., Finberg, R. W., Koo, G. C. & Calderwood, S. K. (2000). HSP70 stimulates cytokine production through a CD14-dependent pathway, demonstrating its dual role as a chaperone and cytokine. Nat Med 6, 435-442.[Medline]

Bucca, G., Brassington, A. M., Schonfeld, H. J. & Smith, C. P. (2000). The HspR regulon of Streptomyces coelicolor: a role for the DnaK chaperone as a transcriptional co-repressor. Mol Microbiol 38, 1093-1103.[Medline]

Castellino, F., Boucher, P. E., Eichelberg, K., Mayhew, M., Rothman, J. E., Houghton, A. N. & Germain, R. N. (2000). Receptor-mediated uptake of antigen/heat shock protein complexes results in major histocompatibility complex class I antigen presentation via two distinct processing pathways. J Exp Med 191, 1957-1964.[Abstract/Free Full Text]

Chang, Z., Primm, T. P., Jakana, J., Lee, I. H., Serysheva, I., Chiu, W., Gilbert, H. F. & Quiocho, F. A. (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.[Abstract/Free Full Text]

Cole, S. T., Brosch, R., Parkhill, J. & 39 other authors (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544.[Medline]

Cunningham, A. F. & Spreadbury, C. L. (1998). Mycobacterial stationary phase induced by low oxygen tension: cell wall thickening and localization of the 16-kilodalton {alpha}-crystallin homolog. J Bacteriol 180, 801-808.[Abstract/Free Full Text]

Fernandes, N. D., Wu, Q. L., Kong, D., Puyang, X., Garg, S. & Husson, R. N. (1999). A mycobacterial extracytoplasmic sigma factor involved in survival following heat shock and oxidative stress. J Bacteriol 181, 4266-4274.[Abstract/Free Full Text]

Grandvalet, C., de Crecy-Lagard, V. & Mazodier, P. (1999). The ClpB ATPase of Streptomyces albus G belongs to the HspR heat shock regulon. Mol Microbiol 31, 521-532.[Medline]

Grossman, A. D., Erickson, J. W. & Gross, C. A. (1984). The htpR gene product of E. coli is a sigma factor for heat-shock promoters. Cell 38, 383-390.[Medline]

Hartl, F. U. (1996). Molecular chaperones in cellular protein folding. Nature 381, 571-579.[Medline]

Hecker, M., Schumann, W. & Volker, U. (1996). Heat-shock and general stress response in Bacillus subtilis. Mol Microbiol 19, 417-428.[Medline]

Helmann, J. D. (1999). Anti-sigma factors. Curr Opin Microbiol 2, 135-141.[Medline]

Hinds, J., Mahenthiralingam, E., Kempsell, K. E., Duncan, K., Stokes, R. W., Parish, T. & Stoker, N. G. (1999). Enhanced gene replacement in mycobacteria. Microbiology 145, 519-527.[Abstract]

Lee, B. Y. & Horwitz, M. A. (1995). Identification of macrophage and stress-induced proteins of Mycobacterium tuberculosis. J Clin Invest 96, 245-249.[Medline]

Manganelli, R., Dubnau, E., Tyagi, S., Kramer, F. R. & Smith, I. (1999). Differential expression of 10 sigma factor genes in Mycobacterium tuberculosis. Mol Microbiol 31, 715-724.[Medline]

Manganelli, R., Voskuil, M. I., Schoolnik, G. K. & Smith, I. (2001). The Mycobacterium tuberculosis ECF sigma factor {sigma}E: role in global gene expression and survival in macrophages. Mol Microbiol 41, 423-437.[Medline]

Monahan, I., Betts, J., Banerjee, D. & Butcher, P. (2001). Differential expression of mycobacterial proteins following phagocytosis by macrophages. Microbiology 147, 459-471.[Abstract/Free Full Text]

Narberhaus, F. (1999). Negative regulation of bacterial heat shock genes. Mol Microbiol 31, 1-8.[Medline]

Parish, T. & Stoker, N. G. (1997). Development and use of a conditional antisense mutagenesis system in mycobacteria. FEMS Microbiol Lett 154, 151-157.[Medline]

Patel, B. K., Banerjee, D. K. & Butcher, P. D. (1991). Characterization of the heat shock response in Mycobacterium bovis BCG. J Bacteriol 173, 7982-7987.[Medline]

Raman, S., Song, T., Puyang, X., Bardarov, S., Jacobs, W. R.Jr & Husson, R. N. (2001). The alternative sigma factor SigH regulates major components of oxidative and heat stress responses in Mycobacterium tuberculosis. J Bacteriol 183, 6119-6125.[Abstract/Free Full Text]

Sherman, D. R., Voskuil, M., Schnappinger, D., Liao, R., Harrell, M. I. & Schoolnik, G. K. (2001). Regulation of the Mycobacterium tuberculosis hypoxic response gene encoding alpha-crystallin. Proc Natl Acad Sci USA 98, 7534-7539.[Abstract/Free Full Text]

Stewart, G. R., Snewin, V. A., Walzl, G. & 7 other authors (2001). Overexpression of heat-shock proteins reduces survival of Mycobacterium tuberculosis in the chronic phase of infection. Nat Med 7, 732–737.[Medline]

Stover, C. K., de la Cruz, V. F., Bansal, G. P., Hanson, M. S., Fuerst, T. R., Jacobs, W. R.Jr & Bloom, B. R. (1992). Use of recombinant BCG as a vaccine delivery vehicle. Adv Exp Med Biol 327, 175-182.[Medline]

Sudre, P., ten Dam, G. & Kochi, A. (1992). Tuberculosis: a global overview of the situation today. Bull W H O 70, 149-159.[Medline]

Suto, R. & Srivastava, P. K. (1995). A mechanism for the specific immunogenicity of heat shock protein-chaperoned peptides. Science 269, 1585-1588.[Medline]

Wards, B. J. & Collins, D. M. (1996). Electroporation at elevated temperatures substantially improves transformation efficiency of slow-growing mycobacteria. FEMS Microbiol Lett 145, 101-105.[Medline]

Wilson, M., DeRisi, J., Kristensen, H. H., Imboden, P., Rane, S., Brown, P. O. & Schoolnik, G. K. (1999). Exploring drug-induced alterations in gene expression in Mycobacterium tuberculosis by microarray hybridization. Proc Natl Acad Sci USA 96, 12833-12838.[Abstract/Free Full Text]

Wilson, M., Voskuil, M., Schnappinger, D. & Schoolnik, G. K. (2001). Functional genomics of Mycobacterium tuberculosis using DNA microarrays. In Methods in Molecular Medicine, vol. 54: Mycobacterium tuberculosis Protocols , pp. 335-357. Edited by T. Parish & N. G. Stoker. Totowa, NJ:Humana Press.

Yang, H., Huang, S., Dai, H., Gong, Y., Zheng, C. & Chang, Z. (1999). The Mycobacterium tuberculosis small heat shock protein Hsp16·3 exposes hydrophobic surfaces at mild conditions: conformational flexibility and molecular chaperone activity. Protein Sci 8, 174-179.[Abstract]

Young, D. B. & Garbe, T. R. (1991). Heat shock proteins and antigens of Mycobacterium tuberculosis. Infect Immun 59, 3086-3093.[Medline]

Yuan, Y., Crane, D. D. & Barry, C. E.3rd (1996). Stationary phase-associated protein expression in Mycobacterium tuberculosis: function of the mycobacterial {alpha}-crystallin homolog. J Bacteriol 178, 4484-4492.[Abstract]

Yuan, Y., Crane, D. D., Simpson, R. M., Zhu, Y. Q., Hickey, M. J., Sherman, D. R. & Barry, C. E.3rd (1998). The 16-kDa {alpha}-crystallin (Acr) protein of Mycobacterium tuberculosis is required for growth in macrophages. Proc Natl Acad Sci USA 95, 9578-9583.[Abstract/Free Full Text]

Received 16 April 2002; revised 1 July 2002; accepted 4 July 2002.