DevR–DevS is a bona fide two-component system of Mycobacterium tuberculosis that is hypoxia-responsive in the absence of the DNA-binding domain of DevR

Deepak Kumar Saini1, Vandana Malhotra1, Deepanwita Dey1, Neha Pant1, Taposh K. Das2 and Jaya Sivaswami Tyagi1

1 Department of Biotechnology, All India Institute of Medical Sciences, New Delhi 110029, India
2 Department of Anatomy, All India Institute of Medical Sciences, New Delhi 110029, India

Correspondence
Jaya Sivaswami Tyagi
jstyagi54in{at}indiatimes.com


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Two-component systems play a central role in the adaptation of pathogenic bacteria to the environment prevailing within host tissues. The genes encoding the response regulator DevR (Rv3133c/DosR) and the cytoplasmic portion (DevS201) of the histidine kinase DevS (Rv3132c/DosS), a putative two-component system of Mycobacterium tuberculosis, were cloned and the protein products were overexpressed, purified and refolded as N-terminally His6-tagged proteins from Escherichia coli. DevS201 underwent autophosphorylation and participated in rapid phosphotransfer to DevR in a Mg2+-dependent manner. Chemical stability analysis and site-directed mutagenesis implicated the highly conserved residues His395 and Asp54 as the sites of phosphorylation in DevS and DevR, respectively. Mutations in Asp8 and Asp9 residues, postulated to form the acidic Mg2+-binding pocket, and the invariant Lys104 of DevR, abrogated phosphoryl transfer from DevS201 to DevR. DevR–DevS was thus established as a typical two-component regulatory system based on His-to-Asp phosphoryl transfer. Expression of the Rv3134c–devR–devS operon was induced at the RNA level in hypoxic cultures of M. tuberculosis H37Rv and was associated with an increase in the level of DevR protein. However, in a devR mutant strain expressing the N-terminal domain of DevR, induction was observed at the level of RNA expression but not at that of protein. DevS was translated independently of DevR and induction of devS transcripts was not associated with an increase in protein level in either wild-type or mutant strains, reflecting differential regulation of this locus during hypoxia.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Intracellular signal transfer in bacteria is dominated by phosphoryl transfer between conserved transmitter and receiver domains in regulatory proteins of two-component systems. Mycobacterium tuberculosis contains eleven complete systems and seven orphan proteins belonging to sensor kinase and response regulator protein families. They are expected to modulate gene expression and direct the adaptation of tubercle bacilli to environmental challenges. In the first step, a sensor, which is usually a membrane-associated histidine protein kinase, is autophosphorylated at a conserved histidine residue in an ATP-dependent manner in response to an environmental stimulus such as O2 tension, temperature, pH, ions, metabolites, etc. In the second step, the phosphoryl group is transferred to a highly conserved aspartic acid residue in the N-terminal domain of the response regulator (Parkinson & Kofoid, 1992; Stock et al., 1995). Two-component regulatory proteins dephosphorylate response regulators in three different ways: through the phosphatase activity of sensor kinases; through reverse phosphorylation, which involves transfer of the phosphoryl group from response regulators back to sensor kinases; or by autodephosphorylation. Regulation may therefore take place by modulation of either the kinase activity or the phosphatase activity of these proteins. A change in the phosphorylation status of the response regulator protein in turn alters its affinity for DNA motifs present in the vicinity of target genes. These unique DNA–protein interactions typically modulate gene transcription, culminating in an appropriate adaptive response (Stock et al., 1995, 2000).

The devR–devS genetic system was identified among genes differentially expressed in the virulent strain of M. tuberculosis (Kinger & Tyagi, 1993; Dasgupta et al., 2000). DevR (Rv3133c) demonstrated homology to response regulator proteins of the NarL/UhpA/FixJ subfamily, and DevS (Rv3132c) to histidine sensor kinases. devR and devS are cotranscribed and conserved in M. tuberculosis and Mycobacterium bovis BCG (Dasgupta et al., 2000); their nucleotide sequences are identical in M. tuberculosis (Cole et al., 1998) and M. bovis except for a single amino acid replacement at codon 283 (Ile to Thr in M. bovis; D. Sharma & J. S. Tyagi, unpublished observation) in the putative transmembrane domain TM2 of DevS (DosS). Recent reports from our and other laboratories have indicated devR–devS expression to be hypoxia-responsive in pathogenic and nonpathogenic mycobacteria including M. tuberculosis (Sherman et al., 2001), M. bovis BCG (Boon et al., 2001) and Mycobacterium smegmatis (Mayuri et al., 2002). DevR was recently designated DosR (dormancy survival regulator) on account of its function in dormancy survival and gene regulation in M. bovis BCG (Boon & Dick, 2002).

In vitro phosphorylation studies have established the functionality of three two-component systems of M. tuberculosis, TrcR–TrcS (Haydel et al., 1999), RegX3–SenX3 (Himpens et al., 2000) and MprA–MprB (Zahrt et al., 2003). Here we report the functional in vitro reconstitution of the DevR–DevS two-component system of M. tuberculosis and identify His395 of DevS and Asp54 of DevR as the sites of phosphorylation. We further show that the DevR–DevS system is transcriptionally induced in hypoxic cultures of M. tuberculosis expressing the N-terminal domain of DevR. RNA and protein expression analysis of devR–devS under aerobic and hypoxic conditions in wild-type and devR-disrupted strains of M. tuberculosis highlight differential regulation of this locus at the levels of transcription and translation.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Construction of recombinant expression plasmids.
Recombinant DNA techniques were performed as described by Sambrook & Russell (2001). Template DNA for PCR reactions was prepared by boiling M. tuberculosis H37Rv cells in 0·1 % Triton X-100. devR and truncated devS201 genes were amplified using Pfu DNA polymerase and gene-specific primers (underlined bases represent mutations introduced in primer sequences) containing engineered restriction enzyme sites to facilitate cloning. The devR gene was amplified using primers DevRNde (5'-GCCCATATGGTAAAGGTCTTCTTGG-3') and DevRSal (5'-CCGGCTTTTTCGTCGACGAGG-3') and the DNA segment encoding the cytosolic portion of DevS (from amino acids 378 to 578) was amplified using DevScytoF (5'-CAACGTCGGATCCGCGAACTCGACG-3') and DevSBamR (5'-GGCGCCGGGATCCTGGCACTAGG-3'). The devR PCR product was digested with NdeI and SalI, treated with T4 DNA polymerase and cloned in the blunted EcoRI site of the pPROEx-HTa plasmid expression vector (Invitrogen) to generate pDSR217. The devS201 PCR product was digested with BamHI and cloned in the BamHI site of pPROEx-HTc to generate plasmid pSCS201.

Overexpression, purification and refolding of recombinant proteins.
All recombinant plasmids were freshly transformed into Escherichia coli BL23(DE3) and grown at 37 °C in 2xYT medium containing 100 µg ampicillin ml–1 to an OD600 of 0·4–0·6. The production of recombinant proteins was induced by the addition of IPTG to a final concentration of 1 mM for 5–6 h. Cell pellets were resuspended in phosphate-buffered saline (PBS), pH 7·4, and disrupted by sonication (5 pulses of 2 min each, Branson Sonifier 450). The insoluble protein fraction containing recombinant proteins was solubilized using denaturation buffer (20 mM Tris/HCl, pH 7·9, 10 %, v/v, glycerol, 500 mM NaCl and 20 mM imidazole containing 6 M guanidinium hydrochloride or 8 M urea) and proteins refolded on a solid matrix as described by Saini et al. (2002).

Site-directed mutagenesis.
Point mutations were introduced into pDSR217 and pSCS201 expressing DevR and DevS201, respectively, using the QuikChange site-directed mutagenesis kit (Stratagene). All the wild-type and mutant protein expression constructs were verified by DNA sequence analysis using an ABI Prism 377 DNA sequencer.

In vitro phosphorylation assays.
For autophosphorylation, DevS201 (15 µM) was incubated with 5 µCi [{gamma}-32P]ATP (1·85x105 Bq mmol–1, 5000 Ci mmol–1; BRIT, India) in 10 µl reaction buffer (50 mM Tris/HCl, pH 8·0, 50 mM KCl, 25 mM MgCl2, 50 µM ATP) at 25 °C for various time periods, after which aliquots were withdrawn and stored at –70 °C until analysis. To study phosphotransfer, DevS201 was phosphorylated for 60 min as described above. Subsequently DevR protein (20 µM) was added, the reaction mixture was incubated at 25 °C, and at indicated time points, aliquots were removed and stored at –70 °C until analysis. Assays with mutant proteins were performed under identical conditions.

Experiments involving the simultaneous use of DevS201 and DevR were performed with His6-DevS201 and DevR free of His6 tag. The latter was generated by digesting His6-DevR with rTEV protease (Invitrogen) at 30 °C for 6 h followed by Ni2+-NTA affinity chromatography. To assess the stability of DevR~32P in the absence of DevS201, the latter was phosphorylated with [{gamma}-32P]ATP as described above and immobilized on Ni2+-NTA agarose beads. The beads were washed free of unbound DevS and incubated with DevR for 2 min in reaction buffer, centrifuged at low speed for 30 s and the supernatant fraction containing purified DevR~32P was incubated in aliquots for up to 30 min at 25 °C and subsequently analysed by SDS-PAGE and autoradiography.

Analysis of chemical stability of DevR~P and DevS201~P.
Phosphorylated DevR and DevS201 proteins were prepared as described above and treated in the presence of 2 % SDS with 1 M HCl, 0·1 M HCl, 1 M NaOH, 0·1 M NaOH or 0·1 M Tris/HCl, pH 8·0, for 30 min at 37 °C. The reactions were neutralized with 0·25 vol. 2 M Tris/HCl, pH 8·0, and analysed by SDS-PAGE. All reactions were terminated with 5 µl 3x stop buffer (250 mM Tris/HCl, pH 6·8, 10 %, v/v, glycerol, 1 % SDS, 280 mM {beta}-mercaptoethanol, 0·01 % bromophenol blue) and analysed by 15 % SDS-PAGE and autoradiography.

Culture conditions.
Primary cultures of M. tuberculosis H37Rv and devR mutant strains were grown in 10 ml Dubos-Tween-Albumin (DTA) medium without any antibiotics or in the presence of kanamycin (20 µg ml–1), respectively, at 37 °C in 50 ml conical tubes to an OD590 of about 0·8–1·0. For expression analysis, cultures were set up as described below.

Setup 1.
The primary culture was diluted to an OD590 of 0·025–0·05 in 300 ml (or 100 ml) fresh DTA media in a 1 litre (or 500 ml) flask and grown with shaking at 190 r.p.m. at 37 °C (the flasks were plugged with cotton to maintain aeration) to an OD590 of 0·5–0·8. A 100 ml (or 30 ml) portion of the culture was immediately harvested by centrifugation (‘aerobic culture’). For hypoxia, 10 ml or 13 ml cultures were dispensed into 15 ml screw-capped tubes and kept static for 48 h (‘hypoxic culture’), after which the culture was pelleted in the same tube. In one of the culture tubes, methylene blue, added to monitor hypoxia generation, faded at 48 h, indicating that the cultures were hypoxic (but not the extent of hypoxia).

Setup 2.
Alternatively, secondary cultures were set up at a very low initial OD590 of 0·005 in DTA medium (160 ml in 180 ml bottles) and allowed to grow under sealed static conditions for up to 40 days. RNA was isolated from 30 day and 40 day hypoxic cultures (which were anoxic/hypoxic as assessed by decolorization/fading of methylene blue in wild-type/mutant cultures). In this setup, cultures grew and adapted to hypoxia over a considerably longer period than in setup 1, where exponential-phase cultures were shifted to static conditions for 48 h, during which hypoxia was generated by consumption of oxygen.

RT-PCR analysis.
The cell pellets were either immediately processed or stored at –70 °C for a maximum of 3 days before RNA isolation. RNA was isolated using an RNAeasy mini kit (Qiagen). DNase-treated RNA and 200 ng random hexamer primers were mixed and denatured by heating at 70 °C for 10 min followed by quenching on ice. RT reactions (20 µl volume) were set up in 50 mM Tris/HCl, pH 8·3, 75 mM KCl, 3 mM MgCl2, 10 mM DTT, 0·5 mM dNTPs, 40 U ribonuclease inhibitor and 200 U Mo-MuLV RT enzyme (Invitrogen) at 25 °C for 10 min and subsequently at 42 °C for 50 min. The absence of DNA contamination in RNA was confirmed by checking for 23S rDNA in RNA preparations. RT reactions were heated at 70 °C for 15 min and stored at –70 °C until further use in PCR assays using gene-specific primers. RNA expression analysis was performed with two separate RNA preparations and two to four times with the same RNA preparation (i.e. four to eight analyses in total) in setup 1 cultures. In setup 2, analysis was performed with one RNA preparation three times. RT-PCR gels were scanned and quantified using ChemiImager densitometric software (Alpha Innotech) and plotted as arbitrary densitometric units (mean ± SD) after normalizing against 23S rRNA. Rv3134c expression was analysed once each with two RNA preparations and SD was not plotted in this case. For each target, expression level is depicted relative to the H37Rv aerobic value (fixed at 100).

M. tuberculosis protein lysate preparation.
Setup 1 cultures harvested as indicated above were gamma-irradiated and frozen at –20 °C. Cell pellets were thawed, washed with 3 vols buffer [PBS containing protease inhibitor cocktail (Roche), 1 mM DTT, 0·5 mM EDTA and 0·1 mM PMSF], resuspended in 1 ml buffer and subjected to bead beating using an equal volume of 0·5 mm zirconium-silica beads for 5x1 min in a mini-bead-beater (Biospec). The cleared lysate was obtained by centrifugation at 20 000 g for 10 min at 4 °C and stored at –20 °C for use in immunoblot analysis.

Immunoblotting.
The protein concentration in lysates was estimated by the BCA assay system (Pierce Biologicals). About 3 and 6 µg protein equivalent of lysates were subjected to Western blot analysis as described by Saini et al. (2002). Briefly, proteins were resolved by 12·5 % SDS-PAGE and transferred to nitrocellulose membrane. Anti-DevR and anti-DevS201 rabbit polyclonal sera were cleaned of anti-E. coli antibodies as described by Dasgupta et al. (2000) and used at a dilution of 1 : 4000. Anti-rabbit horseradish peroxidase conjugate (Bangalore Genie) was used at 1 : 2000 dilution. M. tuberculosis whole-cell lysate (25 mg protein ml–1; obtained from Dr John Belisle, Colorado State University, Fort Collins, CO, USA, under NIH contract AI-75320 and provided by Dr H. K. Prasad, AIIMS) was used as a positive control in all experiments. Blots were scanned and signals were quantified and plotted as arbitrary densitometric units relative to protein expression in H37Rv aerobic cultures (fixed at 100). The DevR value is the mean±SD of eight experiments from four cultures and the devs value is the mean±SD of four experiments from four cultures.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Overexpression, purification and refolding of His6-tagged proteins
A schematic representation of DevR and DevS proteins and their mutant derivatives is shown in Fig. 1(a). The C-terminal region of DevS is conserved among two-component sensors and predicted to function as a kinase; this region was overexpressed as DevS201. DevR and DevS201 were purified from inclusion bodies by Ni2+-NTA affinity chromatography as N-terminally His6-tagged proteins of approximate molecular mass 25·5 kDa and 26·5 kDa, respectively (Fig. 1b). Soluble proteins were obtained by on-column protein refolding as described by Saini et al. (2002). The yields of DevR and DevS201 were 25 mg l–1 and 40 mg l–1, respectively. The mutant proteins were also purified by the same protocol with similar purity and yields. All the proteins were judged to be >90 % pure by Coomassie staining after SDS-PAGE.



View larger version (19K):
[in this window]
[in a new window]
 
Fig. 1. (a) Features of DevR, DevS and mutant derivative proteins of M. tuberculosis. The white boxes at the N-termini of the overexpressed proteins represent the His6-tag (approx. 3 kDa). H-T-H represents the helix–turn–helix motif (predicted by PROSITE). TM 1, 2 and 3 refer to transmembrane domains (predicted by TMpred). The white boxes below DevS578 show GAF1 (residues 63–205) and GAF2 (residues 231–377) domains identified by InterProScan search. PROSITE, TMpred and InterProScan searches were performed at www.expasy.ch. The grey boxes indicate the H, N, D/G1 and G2 conserved sequence motifs of sensor histidine kinases (Parkinson & Kofoid, 1992). D8N, etc., represent mutated amino acid residues in single-letter code. DevS201 refers to the 201 amino acid cytoplasmic domain of DevS (residues 378–578 inclusive) used in the phosphorylation experiments described in this study. (b) SDS-PAGE analysis of E. coli lysates and purified proteins. Lanes 1 and 5, E. coli lysate of uninduced cells carrying recombinant expression constructs pDSR217 and pSCS201, respectively; lanes 2 and 6, E. coli lysate of IPTG-induced cells carrying these constructs; lanes 3 and 7, purified DevR and DevS201, respectively; lane 4, protein molecular mass marker.

 
Autophosphorylation of DevS201 and stability of DevS201~P
DevS201 was rapidly labelled with [{gamma}-32P]ATP in an autophosphorylation reaction; about 80 % and 98 % of the maximum radioactivity incorporated in the protein was detected at 1 h and 4 h, respectively (Fig. 2a). The protein was not labelled when [{alpha}-32P]ATP was used in place of [{gamma}-32P]ATP, indicating that phosphorylation of DevS201 occurred only through transfer of a {gamma}-phosphate group from ATP and not from non-specific binding of ATP to the protein (data not shown). The radioactivity associated with DevS201 did not decrease rapidly when DevS201~32P purified free from ATP was incubated further; the protein retained >90 % and about 30 % of the radioactivity at 4 h and 24 h, respectively (Fig. 2b). Further, the radiolabelled phosphoryl moiety incorporated in DevS201~32P was not replaced by the addition of a 100-fold molar excess of unlabelled ATP. Likewise, [{gamma}-32P]ATP could not replace the phosphoryl moiety in DevS201~P phosphorylated previously with 500 nM unlabelled ATP (Fig. 2c). These experiments suggest DevS201~P to be a stable species under the conditions tested. Autophosphorylation was optimal in the presence of 25 mM MgCl2 (50 µM to 100 mM range tested; data not shown). MnCl2 could replace MgCl2 to some extent (approx. 60 % efficiency) while CaCl2 was unsuitable as cofactor in the phosphorylation reaction (Fig. 2d). Among two-component systems, Mg2+ has been reported to be the most commonly used divalent ion in the reaction (Emmerich et al., 1999; Walker & DeMoss, 1993; Mukai et al., 1990); Mn2+ and Ca2+ have been used as optimal ions in place of Mg2+ in certain instances, such as for TrcS (Haydel et al., 1999).



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 2. Autophosphorylation of DevS201. (a) Time-dependent autophosphorylation of DevS201. DevS201 was incubated in reaction buffer as described and samples were removed at indicated time points, chilled on ice and analysed by SDS-PAGE. Gel slices corresponding to the labelled protein were excised and radioactivity incorporated was quantified by liquid scintillation counting. (b) Stability of DevS201~32P. DevS201~32P was prepared as described, unincorporated [{gamma}-32P]ATP was removed by filtration through a 10 k Nanosep device and stability of DevS~P was analysed by incubation at 25 °C for up to 24 h. (c) Stability of DevS201~P in the presence of competitor. DevS201~P was phosphorylated with either [{gamma}-32P]ATP or unlabelled ATP. An excess of competitor ATP was added, [{gamma}-32P]ATP to unlabelled DevS201~P and unlabelled ATP to DevS201~32P, respectively. Samples were removed at indicated time points and analysed by SDS-PAGE. (d) Ionic requirement for autophosphorylation. DevS201 was autophosphorylated in reaction buffer containing the divalent ions indicated (25 mM each, as chloride salts): lane 1, Mg2+; lane 2, Mn2+, lane 3, Ca2+; lane 4, 25 mM EDTA.

 
Phosphotransfer from DevS201~P to DevR
To assess phosphotransfer, DevS201 was autophosphorylated with [{gamma}-32P]ATP for 60 min before the addition of DevR. Phosphotransfer occurred rapidly and the maximum signal in DevR~32P was obtained within 2 min of DevR addition. In the reaction containing both DevR and DevS201, dephosphorylation of DevR was not accompanied by a reverse phosphotransfer to DevS. Rather, a rapid net loss of protein-bound radiolabelled phosphate was noted and dephosphorylation of both proteins was nearly complete within 10 min (Fig. 3a). Thus while DevS201~P was of intrinsically high stability, it was labile in the presence of DevR and rapidly transferred its phosphoryl group to generate DevR~P. Purified DevR~32P was stable for longer in the absence (half-life approx. 20 min) than in the presence of DevS~P (half-life approx. 3 min, earliest time point evaluated), implying the presence of phosphatase activity in DevS201 (data not shown). Furthermore, in the presence of increasing concentrations of DevR (2·5–40 µM), there was a proportionate increase in the extent of dephosphorylation of DevS201~32P within a 2·5 min period (data not shown). To examine the requirement for Mg2+ in the phosphotransfer reaction, labelled DevS201~32P was synthesized and purified by filtration through a 10 k Nanosep device (Pall Life Sciences) to remove Mg2+ and ATP. Phosphotransfer from DevS201~P to DevR occurred only in the presence of MgCl2 (Fig. 3b), not that of MnCl2 or CaCl2 (not shown). Formation of DevR~P from DevS~P was not dependent on the presence of ATP (Fig. 3b).



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 3. (a) Phosphotransfer from DevS201~P to DevR. DevS201~32P was prepared as described, DevR was added and the tubes were incubated for up to 30 min. (b) Mg2+-dependent phosphotransfer from DevS201~P to DevR. DevS201~32P was prepared as described, filtered through a 10 k Nanosep device and washed to remove MgCl2 and ATP. DevR (20 µM) and Mg2+ were added as indicated and the tubes incubated for 3 min. (c) Chemical stability of DevS201~P and DevR~P. DevS201~32P and DevR~32P were prepared as described. Lane 1, control reaction containing 2 % SDS; lane 2, plus 1 M HCl; lane 3, plus 0·1M HCl; lane 4, plus 1 M NaOH; lane 5, plus 0·1 M NaOH. (d) Phosphorylation of DevS201 mutant proteins. DevS201 protein (WT, wild-type) and its mutants were phosphorylated as described. (e) Phosphotransfer to DevR mutant proteins. DevS201~32P was prepared as described, incubated with different mutants for 3 min and analysed. DevR-D54V behaved like DevR-D54N in this assay (not shown). The lower panels in (d) and (e) show the Coomassie-stained profile of the radioactive panel.

 
Phosphorylation sites in DevS201 and DevR proteins
To analyse the nature and type of amino acids phosphorylated in DevS201 and DevR proteins, chemical stability analysis of the phosphorylated proteins was performed. Phosphoramidates, which are generated by phosphorylation of the conserved histidine residue in sensor kinases, are stable under alkaline conditions but sensitive to acidic pH, whereas the opposite is true for phosphoaspartic acid, which is generated by phosphorylation of the conserved aspartate residue in response regulator proteins (Stock et al., 1995, 2000). The phosphatidyl bond of DevS201~P was acid-labile and base-stable, while the acyl phosphate bond in DevR~P was stable in mildly acidic conditions and labile under basic conditions (Fig. 3c). The pH stabilities of the phosphorylated proteins are consistent with the existence of a histidyl-phosphate group for DevS and an aspartyl-phosphate for DevR.

By sequence homology with known proteins belonging to two-component systems, the conserved His residue at position 395 in DevS and the conserved Asp residue at position 54 in DevR were predicted to be the sites of phosphorylation (Dasgupta et al., 2000). Their involvement in the phosphorylation reactions was confirmed by the analysis of mutant proteins. His395 and Asp54 residues in DevS201 and DevR proteins, respectively, were altered by mutagenesis to create mutant proteins DevS201-H395Q and DevR-D54V. The mutant proteins were overproduced, purified as His6-tagged proteins, refolded and analysed in phosphorylation reactions. The DevS201-H395Q mutant protein was defective in autophosphorylation (Fig. 3d). In contrast, the mutation of His397 to Ala or Gln in DevS201 did not affect its phosphorylation, confirming that phosphorylation occurs only at the conserved His395 residue within the H box. Likewise, mutation of the Asp54 residue in DevR to Val rendered it unable to accept the phosphoryl moiety from DevS201~P (Fig. 3e).

The pH stability and mutagenesis experiments collectively confirmed that His395 of DevS201 and Asp54 of DevR participate in the auto- and transphosphorylation reactions, respectively, as is typical of proteins belonging to two-component systems.

Mutational analysis of other conserved residues in DevS201 and DevR
The mutation of the Asn503 residue located in the N box of DevS (Fig. 1a) to Asp also blocked the phosphorylation of DevS201 (Fig. 3d). The N box residue, together with the D/G1 and G2 boxes, is implicated in the formation of the ATP-binding region of DevS. Likewise, Asp8 and Asp9 in DevR were predicted to correspond to acidic residues at analogous positions in other response regulator proteins that together with Asp54 form the acidic pocket and coordinate Mg2+ required for the phosphorylation reaction (Stock et al., 2000). Mutant DevR proteins were created wherein these two residues were singly mutated to Asn (DevR-D8N and DevR-D9N) and the invariant Lys104 was mutated to Glu (DevR-K104E). All the mutant proteins were defective in phosphotransfer from wild-type DevS201~P protein, confirming the essentiality of the Mg2+-binding pocket and Lys104 in the phosphotransfer reaction (Fig. 3e).

Full-length DevR is necessary for normal basal expression under aerobic conditions
The effect of devR disruption on its own transcription was analysed in a M. tuberculosis devR mutant constructed by allelic exchange (Malhotra et al., 2004). Briefly, a kanamycin-resistance cassette encoding AphI was cloned into a unique PpuMI site within the devR gene. DNA sequence analysis of the devR : : kan locus in the mutant strain indicated the disruption of the devR ORF accompanied by the creation of an ORF potentially encoding a putative 406 amino acid fusion polypeptide composed of 145 amino acids corresponding to the N-terminal receiver domain of DevR and the AphI protein (designated DevRN-Aph, Fig. 4) and lacking the DNA-binding domain in the C-terminal region.



View larger version (51K):
[in this window]
[in a new window]
 
Fig. 4. RT-PCR analysis of M. tuberculosis cultures in response to hypoxia. (a) Analysis of RNA from setup 1; (b) Analysis of RNA from setup 2. Culture conditions are described in Methods. The primers used for RT-PCR analysis are indicated below the genes in the diagram at the top. Positive and negative controls were included for all PCR reactions. All PCR assays involved 35 cycles of amplification except for devS (45 cycles), devRf2/devRf3 (30 cycles) and 23S rRNA (17 cycles). devR transcripts were not detected in mutant cultures using the f2r2 primer pair that flanks the kan cassette and fails to give an amplicon owing to its large size. A and H refer to aerated and hypoxic cultures, respectively and 30dH and 40dH to hypoxic cultures grown under sealed static conditions for 30 days and 40 days, respectively.

 
Since devR and devS are cotranscribed with the upstream Rv3134c gene (encoding a protein containing two Usp domains) in the wild-type strain (Dasgupta et al., 2000), it is likely that devS is transcribed from the promoter(s) located upstream of Rv3134c and/or devR. Further, insertion of the kanamycin-resistance cassette into the devR ORF could cause premature termination and exert a polar effect on devS. Using RNA from aerobic cultures, basal levels of devR, devS and Rv3134c transcripts in the devR mutant were compared to those in the wild-type strain. The aerobic expression level was consistently lower in mutant vs wild-type strains cultured in both setup 1 and setup 2 (Fig. 4a, b). This suggested that (1) full-length DevR was necessary for basal expression of this locus, and (2) the lower level of devS transcripts in the mutant strain (Fig. 4: compare lanes 1 and 3, devS panel) was not due to polar effects as the expression of the upstream Rv3134c gene was also lower (Fig. 4a, b). Our results are not in agreement with those of Sherman et al. (2001), who reported the abrogation of transcription from this locus in an Rv3134c mutant. The possible reasons for the discrepancy in our observations are discussed later.

The DevR N-terminal domain is sufficient for hypoxic induction of RNA from the Rv3134cdevR–devS locus
Next we compared transcript levels of this locus in the devR mutant and wild-type strains in hypoxic cultures. A modest but consistent increase in the expression of the locus was noted during hypoxia in wild-type cultures (Fig. 4a, b). In the devR mutant strain, wild-type devR transcripts were obviously not detected (Fig. 4a, b, second panel), but transcripts from the flanking regions were induced during hypoxia (Fig. 4a, b). Furthermore, aphI transcripts from the kanamycin-resistance cassette were also induced, which indicated Rv3134c/devR promoter activation since the cloned cassette lacks its own promoter. In hypoxic cultures we consistently observed higher devR transcript levels 3' to the cloned cassette and of the cassette itself, for which we have no explanation. But we found 12/18 and 11/20 matches in a sequence stretch (in devR upstream of the cassette) to the DNA motif proposed by Florczyk et al. (2003) and Park et al. (2003) to bind DevR. Thus the induction of these genes occurred reproducibly in the mutant and the fold-induction matched that in the wild-type strain. Importantly, hypoxic induction occurred in the absence of the DNA-binding C-terminal domain of DevR.

Levels of DevR and DevS proteins in aerobic cultures of wild-type and devR mutant strains
The devR and devS genes overlap by 1 bp and are cotranscribed as mentioned earlier (Dasgupta et al., 2000). A classical Shine–Dalgarno sequence was not apparent in sequences immediately upstream of the devS ORF, although a TGATGGA sequence having a 4/7 match with a proposed mycobacterial consensus sequence (Dale & Patki, 1990) was present 2 bp upstream of its AUG initiation codon. Therefore DevS may be made either independently or by cotranslation with DevR. In the mutant strain, the latter possibility is discarded since stop codons were present in all three frames downstream of the AphI-coding region. So while the presence of devS and devRaphI transcripts was not in doubt it was imperative to detect DevS and DevRN–Aph proteins in the mutant strain.

Aerobic expression of DevR and DevS was confirmed by Western blotting. Using anti-DevR sera, an approximately 23 kDa protein was detected in wild-type lysates, while an approximately 45 kDa protein of the expected size was detected in mutant lysates (Fig. 5a, lanes 1 and 2, upper panels). In mutant lysates, labelling resulted from immunoreactivity of the N-terminal two-thirds of DevR in the DevRN–Aph fusion protein (Fig. 4). The DevR protein level was modestly but consistently higher in wild-type cultures as compared to mutant cultures (Fig. 5a, upper panels) and reflected the differences observed at the RNA level. However, it is possible that the difference could also be due to differences in seroreactivity of DevR vs DevRN–Aph. With anti-DevS sera, an approximately 62 kDa protein corresponding to DevS was detected in both wild-type and mutant bacteria (Fig. 5a, lanes 1 and 2, lower panels). DevS protein expression in the mutant strain confirmed that DevS was translated independently of DevR.



View larger version (44K):
[in this window]
[in a new window]
 
Fig. 5. Immunoblot analysis of DevR and DevS proteins expressed in M. tuberculosis in aerobic and anaerobic setup 1 cultures. Lanes 1 and 3, 3 µg total protein; lanes 2 and 4, 6 µg total protein. Lane 5, M. tuberculosis whole-cell lysate (see Methods). (b) Western blots were scanned and quantified as described.

 
Dichotomy of DevR and DevS protein expression in response to hypoxia
To confirm the observation of RNA induction, DevR and DevS protein levels were assessed in hypoxic cultures also. Hypoxia-adapted wild-type cultures showed a modest but consistent increase in the level of DevR protein compared to aerobic cultures (Fig. 5a, upper left panel, lanes 3 and 4). This was in agreement with earlier observations made in M. bovis BCG (Boon et al., 2001). However, the level of DevRN–Aph protein was not elevated in hypoxic mutant bacteria (Fig. 5a, upper right panel, lanes 3 and 4). The possibility that the fusion protein is inherently less stable was excluded, as discrete immunoreactive bands were detected in hypoxic mutant lysates. In contrast, the levels of DevS protein remained unchanged in hypoxia in both wild-type and mutant cultures (Fig. 5a, lower panels, lane 4 vs lane 2; Fig. 5b). Thus the transcriptional induction of devS is not associated with an increase in the level of DevS protein under hypoxic conditions.

Immunoelectron microscopic analysis of these cultures using anti-DevS and anti-DevR sera showed labelling in the entire cellular compartment, with no indication of a specific pattern of intracellular localization in either aerobic or hypoxia-adapted cultures (Fig. 6).



View larger version (66K):
[in this window]
[in a new window]
 
Fig. 6. Immunoelectron microscopy of DevR and DevS proteins. M. tuberculosis H37Rv and devR mutant bacteria (aerobic and hypoxic cultures from setup 1) were fixed with glutaraldehyde, developed with anti-DevR or anti-DevS201 antibodies and gold-conjugated anti-rabbit second antibody and assessed by immunoelectron microscopy as described by Saini et al. (2002).

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
DevR and DevS201 proteins of M. tuberculosis were overexpressed and purified from E. coli. The refolded proteins exhibited properties typical of two-component system proteins, namely autophosphorylation of a conserved histidine residue (His395) in the cytoplasmic domain of the histidine kinase DevS and phosphotransfer of phosphoryl moiety from DevS201 to a conserved aspartic acid residue (Asp54) in the response regulator DevR. DevR–DevS was thereby established to be a bona fide two-component system of M. tuberculosis.

The stability of DevS201~P was modulated by DevR; the phosphorylated species had a very short half-life (<1 min) in the presence of DevR but was very stable in its absence (90 % of radioactivity retained in DevS~P at 4 h). DevS201~P stability was in contrast to that of some other sensor kinases such as RegS of Bradyrhizobium japonicum and NarX of E. coli, the half-lives of whose phosphorylated forms were reported to be 15 min and >1 h. Dephosphorylation of DevR was enhanced in the presence of DevS201, as reported for the phosphorylated response regulators NarL and RegR in the presence of their cognate sensor kinases (Emmerich et al., 1999; Walker & DeMoss, 1993). The phosphorylated species of M. tuberculosis two-component systems TrcR–TrcS and RegX3–SenX3, when present together, appeared to be more stable than those of DevR–DevS. At 10 min (only a single point was studied; Haydel et al., 1999), TrcR~P and TrcS~P were markedly more stable than DevR~P and DevS~P; the phosphorylated species of both proteins all but disappeared in the latter system at the same time point. Likewise, the phosphorylated species of SenX3 and RegX3 were quite stable up to at least 30 min (last time point examined; Himpens et al., 2000). The phosphorylation kinetics of DevR–DevS indicate a rapid rate of phosphotransfer between DevS201 and DevR and a rapid rate of signal dissipation from DevR. In vivo, these kinetics would enable bacilli to respond quickly to hypoxia and also revert rapidly to the basal state upon hypoxia reversal. However, the in vitro phosphorylation properties of DevS201 may not truly reflect its in vivo response as it lacks the N-terminal domain; the results therefore need to be viewed with caution.

Our results suggest DevS to be the cognate sensor kinase of DevR. Histidine kinases typically sense a specific signal by their N-terminal domains, and the N-terminal domain of DevS is predicted to contain two GAF domains that are present in many proteins implicated in sensory and signalling pathways (Ho et al., 2000). The nature of the sensed signal and the precise role of DevS and its GAF domains in the hypoxia response are yet to be elucidated. The absence of a severe phenotype in a devS mutant with regard to hypoxic viability compared to a severe loss in hypoxic viability and expression of dormancy proteins such as Acr in a devR mutant of M. bovis BCG, along with a lack of phenotype in Acr expression in a devS mutant of M. tuberculosis, suggested the operation of a compensatory pathway involving signalling from an alternative sensor kinase (Boon & Dick, 2002; Sherman et al., 2001). We had earlier highlighted the similarity between the putative orphan histidine kinase Rv2027c and DevS (Dasgupta et al., 2000). So Rv2027c could compensate for the lack of DevS by generating phosphorylated DevR. DevR may also be activated by a low-molecular-mass phosphate donor such as acetyl phosphate. Another possibility is that DevR–DevS is not the primary system to sense and respond to hypoxia but part of a cascade involving other proteins. The M. tuberculosis genome does not have homologues to bacterial proteins that sense oxygen such as FixL FNR, Dos and HemAT (Patschkowski et al., 1996; Kiley & Bienert, 1998; Delgado-Nixon et al., 2000; Hou et al., 2001). Yet another possibility is that DevS does not perceive oxygen but some metabolite generated as a consequence of oxygen deprivation, as does the E. coli histidine kinase ArcB (Georgellis et al., 1999). Cytochrome c oxidase was recently suggested as the sensor for oxygen limitation and nitric oxide in M. tuberculosis (Voskuil et al., 2003).

Rv3134cdevR–devS expression at the RNA level exhibited two notable properties. First, it was expressed in the devR mutant strain, albeit at a lower level compared to the wild-type strain. Since the mutant strain expresses the N-terminal domain of DevR we conclude that basal transcription of the operon was not absolutely dependent on the presence of full-length DevR. Secondly, hypoxic induction occurred in both the wild-type and mutant strains. Our observations regarding devR–devS expression are at variance with those of Sherman et al. (2001) and Park et al. (2003) with regard to (1) the induction noted in the present study vs the loss of induction in their study in mutant bacteria, and (2) the fold-induction in hypoxia (less than twofold in this study vs tenfold in their study). These discrepancies could be ascribed to differences in mutant bacilli, the assays used (semi-quantitative RT-PCR vs microarrays) and the conditions employed to develop hypoxia. The devR mutant strain used by us expressed the N-terminal domain of DevR while their mutant was deleted of nearly all Rv3134c sequences and also some upstream sequences (Sherman et al., 2001) that possess promoter activity which is hypoxia-responsive (Bagchi et al., 2003). Our culture conditions involved the setting in of hypoxia from utilization of oxygen by exponential-phase cultures in sealed tubes during a 48 h period (setup 1) or by adaptation of cultures of low starting density to hypoxia in sealed bottles over a 40 day growth period (setup 2). It is possible that our cultures (smaller headspace and no stirring) may be in NRP2 rather than NRP1 (Wayne & Hayes, 1996) and the peak of induction falls away so that we see lower levels of induction. Such a peaking and falling of devR promoter activity was observed by us in M. smegmatis cultured under similar conditions (Bagchi et al., 2003). The discrepancy in fold-induction can be tested by generating hypoxia in cultures exposed to 0·2 % oxygen for 2 h as described by Sherman et al. (2001). Thus the differences in the rate and extent of hypoxia generation and their effect on the physiological state of the bacilli might be the root cause of the observed differences in fold-induction. However, this does not change the observation that the Rv3134c–devR–devS locus is induced during hypoxia in the mutant strain.

In the mutant strain DevS can be phosphorylated at the onset of hypoxia and then could rapidly transfer the phosphoryl moiety to DevRN–Aph in the first step of the signalling cascade. This is feasible, as phosphotransfer from kinases to independent N-terminal domains of response regulators was reported to occur in Spo0A, RegR and TrcR (Fujita & Losick, 2003; Emmerich et al., 1999; Haydel et al., 1999). However, the second step of phosphorylation-mediated activation of DNA binding would not occur, as mutant DevR protein lacks the DNA-binding gene-activating domain. In this context, the hypoxic induction of this locus in the mutant strain is striking and suggests a role for the N-terminal domain of DevR. The absence of an increase in DevS protein levels during hypoxia is also striking. Since phosphosignal transfer from DevS~P to DevR is extremely rapid and occurs at a rate that is proportional to the concentration of the latter, perhaps an enhanced amount of DevS protein is unnecessary for efficient signalling during hypoxia. In the case of DevR, however, the constitutive level of DevR protein may be insufficient for hypoxic regulation of the devR regulon and a mechanism of protein amplification may be necessary to elicit activation of the regulon. This may be very critical after infection, when the bacteria would need to implement widespread changes in gene expression in response to hypoxia.

In summary, the modest but reproducible hypoxic induction of devR–devS expression in M. tuberculosis expressing the DevR N-terminal domain (this study) but not in a devR deletion mutant (Park et al., 2003) suggests a critical role for the DevR N-terminal domain and warrants detailed studies of the receiver and activation domains in DevR. In vitro reconstitution of DevR–DevS and the availability of a mutant strain expressing the N-terminal domain of DevR provide an unique opportunity to decipher the regulatory features of this locus in M. tuberculosis. The biochemical properties of DevS and DevR proteins provide us with valuable assays to screen for potential inhibitors that interfere with their function and obstruct the initiation and/or maintenance of the hypoxia response (US patent pending). The study of the DevR–DevS signalling pathway will improve our understanding of the dormancy response in M. tuberculosis and perhaps provide novel molecules for the treatment of persistent tuberculosis.


   ACKNOWLEDGEMENTS
 
Deepak Sharma is acknowledged for readily providing able assistance in computer analysis. Illah Joshi is thanked for providing anti-DevR rabbit polyclonal sera. D. K. S. is grateful to the University Grants Commission for a Senior Research Fellowship. Financial assistance provided to J. S. T. by the Council for Scientific and Industrial Research, Government of India is acknowledged.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bagchi, G., Mayuri & Tyagi, J. S. (2003). Hypoxia-responsive expression of Mycobacterium tuberculosis Rv3134c and devR promoters in Mycobacterium smegmatis. Microbiology 149, 2303–2305.[Free Full Text]

Boon, C. & Dick, T. (2002). Mycobacterium bovis BCG response regulator essential for hypoxic dormancy. J Bacteriol 184, 6760–6767.[Abstract/Free Full Text]

Boon, C., Li, R., Qi, R. & Dick, T. (2001). Proteins of Mycobacterium bovis BCG induced in Wayne dormancy model. J Bacteriol 182, 2672–2676.

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.[CrossRef][Medline]

Dale, J. W. & Patki, A. (1990). Mycobacterial gene expression and regulation. In Molecular Biology of Mycobacteria, pp. 173–198. Edited by J. McFadden. Guildford, UK: Surrey University Press.

Dasgupta, N., Kapur, V., Singh, K. K., Das, T. K., Sachdeva, S., Jyothisri, K. & Tyagi, J. S. (2000). Characterization of a two-component system, devR-devS, of Mycobacterium tuberculosis. Tuber Lung Dis 80, 141–159.[CrossRef][Medline]

Delgado-Nixon, V. M., Gonzalez, G. & Gilles-Gonzalez, M. A. (2000). Dos, a heme-binding PAS protein from Escherichia coli, is a direct oxygen sensor. Biochemistry 39, 2685–2691.[CrossRef][Medline]

Emmerich, R., Panglungtshang, K., Strehler, P., Hennecke, H. & Fischer, H. M. (1999). Phosphorylation, dephosphorylation and DNA-binding of Bradyrhizobium japonicum RegSR two-component regulatory proteins. Eur J Biochem 263, 455–463.[Abstract/Free Full Text]

Florczyk, M. A., McCue, L. A., Purkayastha, A., Currenti, E., Wolin, M. J. & McDonough, K. A. (2003). A family of acr-coregulated Mycobacterium tuberculosis genes shares a common DNA motif and requires Rv3133c (dosR or devR) for expression. Infect Immun 71, 5332–5343.[Abstract/Free Full Text]

Fujita, M. & Losick, R. (2003). The master regulator for entry into sporulation in Bacillus subtilis becomes a cell-specific transcription factor after asymmetric division. Genes Dev 17, 1166–1174.[Abstract/Free Full Text]

Georgellis, D., Kwon, O. & Lin, E. C. (1999). Amplification of signaling activity of the arc two-component system of Escherichia coli by anaerobic metabolites. An in vitro study with different protein modules. J Biol Chem 274, 35950–35954.[Abstract/Free Full Text]

Haydel, S. E., Dunlap, N. E. & Benjamin, W. H., Jr (1999). In vitro evidence of two-component system phosphorylation between the Mycobacterium tuberculosis TrcR/TrcS proteins. Microb Pathog 26, 195–206.[CrossRef][Medline]

Himpens, S., Locht, C. & Supply, P. (2000). Molecular characterization of the mycobacterial SenX3-RegX3 two-component system: evidence for autoregulation. Microbiology 146, 3091–3098.[Abstract/Free Full Text]

Ho, Y. S., Burden, L. M. & Hurley, J. H. (2000). Structure of the GAF domain, a ubiquitous signaling motif and a new class of cyclic GMP receptor. EMBO J 19, 5288–5299.[Abstract/Free Full Text]

Hou, S., Freitas, T., Larsen, R. W. & 7 other authors (2001). Globin-coupled sensors: a class of heme-containing sensors in Archaea and Bacteria. Proc Natl Acad Sci U S A 98, 9353–9398.[Abstract/Free Full Text]

Kiley, P. J. & Beinert, H. (1998). Oxygen sensing by global regulator, FNR – the role of the iron–sulfur cluster. FEMS Microbiol Rev 22, 341–352.[CrossRef][Medline]

Kinger, A. K. & Tyagi, J. S. (1993). Identification and cloning of genes differentially expressed in the virulent strain of Mycobacterium tuberculosis. Gene 131, 113–117.[CrossRef][Medline]

Malhotra, V., Sharma, D., Ramanathan, V. D. & 8 other authors (2004). Disruption of response regulator gene, devR, leads to attenuation in virulence of Mycobacterium tuberculosis. FEMS Microbiol Lett 231, 237–245.[CrossRef][Medline]

Mayuri, Bagchi, G., Das, T. K. & Tyagi, J. S. (2002). Molecular analysis of the dormancy response in Mycobacterium smegmatis: expression analysis of genes encoding the DevR-DevS two-component system, Rv3134c and chaperone alpha-crystallin homologues. FEMS Microbiol Lett 211, 231–237.[CrossRef][Medline]

Mukai, K., Kawata, M. & Tanaka, T. (1990). Isolation and phosphorylation of the Bacillus subtilis degS and degU gene products. J Biol Chem 265, 20000–20006.[Abstract/Free Full Text]

Park, H. D., Guinn, K. M., Harrell, M. I., Liao, R., Voskuil, M. I., Tompa, M., Schoolnik, G. K. & Sherman, D. R. (2003). Rv3133c/dosR is a transcription factor that mediates the hypoxic response of Mycobacterium tuberculosis. Mol Microbiol 48, 833–843.[CrossRef][Medline]

Parkinson, J. S. & Kofoid, E. C. (1992). Communication modules in bacterial signalling proteins. Annu Rev Genet 26, 71–112.[CrossRef][Medline]

Patschkowski, T., Schluter, A. & Priefer, U. B. (1996). Rhizobium legumisosarum bv. viciae contains a second fnr/fixK-like gene and an unusual fixL homologue. Mol Microbiol 21, 267–80.[CrossRef][Medline]

Saini, D. K., Pant, N., Das, T. K. & Tyagi, J. S. (2002). Cloning, overexpression, purification and matrix-assisted refolding of DevS (Rv3132c) histidine protein kinase of Mycobacterium tuberculosis. Protein Expr Purif 25, 203–208.[CrossRef][Medline]

Sambrook, J. & Russell, D. W. (2001). Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

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

Stock, J. B., Surette, M. G., Levit, M. & Park, P. (1995). Two-component signal transduction systems: structure-function relationships and mechanisms of catalysis. In Two-Component Signal Transduction, pp. 25–51. Edited by J. A. Hoch & T. J. Silhavy. Washington, DC: American Society for Microbiology.

Stock, A. M., Robinson, V. L. & Goudreau, P. N. (2000). Two-component signal transduction. Annu Rev Biochem 69, 183–215.[CrossRef][Medline]

Voskuil, M. I., Schnappinger, D., Visconti, K. C., Harrell, M. I., Dolganov, G. M., Sherman, D. R. & Schoolnik, G. K. (2003). Inhibition of respiration by nitric oxide induces a Mycobacterium tuberculosis dormancy program. J Exp Med 198, 705–713.[Abstract/Free Full Text]

Walker, M. S. & DeMoss, J. A. (1993). Phosphorylation and dephosphorylation catalyzed in vitro by purified components of the nitrate sensing system, NarX and NarL. J Biol Chem 268, 8391–8393.[Abstract/Free Full Text]

Wayne, L. G. & Hayes, L. G. (1996). An in vitro model for sequential study of shiftdown of Mycobacterium tuberculosis through two stages of nonreplicating persistence. Infect Immun 64, 2062–2069.[Abstract]

Zahrt, T. C., Wozniak, C., Jones, D. & Trevett, A. (2003). Functional analysis of the Mycobacterium tuberculosis MprAB two-component signal transduction system. Infect Immun 71, 6962–6970.[Abstract/Free Full Text]

Received 3 January 2003; revised 16 December 2003; accepted 17 December 2003.