Mycobacterium tuberculosis is a major pathogen that has the ability to establish, and emerge from, a persistent state. Wbl family proteins are associated with developmental processes in actinomycetes, and M. tuberculosis has seven such proteins. In the present study it is shown that the M. tuberculosis H37Rv whiB1 gene is essential. The WhiB1 protein possesses a [4Fe-4S]2+ cluster that is stable in air but reacts rapidly with eight equivalents of nitric oxide to yield two dinuclear dinitrosyl-iron thiol complexes. The [4Fe-4S] form of WhiB1 did not bind whiB1 promoter DNA, but the reduced and oxidized apo-WhiB1, and nitric oxide-treated holo-WhiB1 did bind to DNA. Mycobacterium smegmatis RNA polymerase induced transcription of whiB1 in vitro; however, in the presence of apo-WhiB1, transcription was severely inhibited, irrespective of the presence or absence of the CRP (cAMP receptor protein) Rv3676, which is known to activate whiB1 expression. Footprinting suggested that autorepression of whiB1 is achieved by apo-WhiB1 binding at a region that overlaps the core promoter elements. A model incorporating regulation of whiB1 expression in response to nitric oxide and cAMP is discussed with implications for sensing two important signals in establishing M. tuberculosis infections.
- iron–sulfur cluster
- nitric oxide
- Wbl protein
Approximately one third of the world's population is infected by Mycobacterium tuberculosis . Most infected individuals are essentially asymptomatic, a state characterized by the presence of M. tuberculosis in a persistent non-replicating state. However, one in ten infected individuals will become ill with active tuberculosis during their lifetimes, contributing to ~1.8 million deaths in 2008 . Thus prevention of reactivation tuberculosis is a major healthcare challenge, and understanding the reprogramming of gene expression that facilitates entry into, and emergence from, the persistent state is an important research goal [2,3].
The CRPMt (M. tuberculosis cAMP receptor protein Rv3676) is a global regulator that is required for virulence . Among the genes regulated by Rv3676 (CRPMt) are rpfA and whiB1 [4,5]. The rpfA gene encodes a protein that is involved in reviving dormant bacteria , and whiB1 encodes a member of a family of proteins (Wbl) that are involved in developmental processes in actinomycetes . Wbl proteins possess four conserved cysteine residues that bind a redox-sensitive iron–sulfur cluster [7–9]. It is assumed that the iron–sulfur cluster will prove essential in controlling Wbl protein function. This assumption is supported by experiments showing that all four conserved cysteine residues are essential for WhiD function in Streptomyces coelicolor . Recombinant expression and aerobic isolation of all seven Wbl proteins of M. tuberculosis resulted in proteins partially occupied by [2Fe-2S] clusters, but after anaerobic reconstitution these proteins contained [4Fe-4S] clusters . The presence of a predicted helix–turn–helix, combined with phenotypic studies of mutants, led to the suggestion that Wbl proteins are DNA-binding transcription factors (, and references therein). Consistent with this suggestion, M. tuberculosis WhiB3 interacts with promoter DNA of genes known to be differentially regulated in a whiB3 mutant, and with the essential sigma factor SigA [10–12]. It has also been suggested that Wbl proteins, including M. tuberculosis WhiB1, function as protein disulfide reductases , although WhiD from S. coelicolor was shown to lack such activity .
The M. tuberculosis WhiB3 protein contributes to virulence  and whiB3 expression is induced in mouse lungs and macrophages . The WhiB3 [4Fe-4S] cluster is slowly degraded in the presence of O2, resulting in total cluster loss, in contrast with the rapid reaction with O2 of the Escherichia coli O2-sensing transcription factor FNR, which proceeds by cluster conversion from the [4Fe-4S] into the [2Fe-2S] form [16–18]. The WhiB3 and FNR [4Fe-4S] clusters also react with NO to form DNICs (dinitrosyl-iron thiol complexes) [16,19]. Thus it was suggested that WhiB3 acts as a sensor of the physiologically significant gases O2 and NO to control expression of genes involved in intermediary metabolism.
In the present study, we show that the M. tuberculosis H37Rv whiB1 gene is essential, and that the WhiB1 protein possesses a [4Fe-4S]2+ cluster that reacts rapidly with eight NO molecules per cluster. Furthermore, apo-WhiB1 and NO-treated holo-WhiB1 (but not holo-WhiB1) bind to PwhiB1 (whiB1 promoter region) to repress transcription. Therefore it is concluded that the WhiB1 [4Fe-4S] cluster is a NO sensor, and that exposure to NO converts holo-WhiB1 from a non-DNA-binding form into a DNA-binding form capable of regulating transcription.
MATERIALS AND METHODS
Bacterial strains, plasmids and growth conditions
Bacterial strains and plasmids are listed in Table 1. Escherichia coli cultures were grown in LB (Luria–Bertani) medium  in a 1:5 volume/flask ratio at 37 °C with shaking at 250 rev./min. For isolation of WhiB1, the expression plasmid (pGS2164) was selected for by addition of kanamycin (50 μg/ml). DNA was manipulated using standard methodologies . The whiB1 open reading frame was amplified by PCR using primers MS40 (5′-TTTTTTGAATTCGATTGGCGCCACAAGGCGGT-3′) and MS41 (5′-TTTTTTCTCGAGTCAGACCCCGGTACGGGCTT-3′) containing engineered EcoRI and XhoI sites respectively. The amplified DNA was ligated into the corresponding sites of pET28a. The authenticity of the resulting plasmid (pGS2164) was confirmed by DNA sequencing and encoded a His6–WhiB1 fusion protein.
Mycobacterium tuberculosis cultures (100 ml) were grown in 1 litre polycarbonate culture bottles (Techmate) in a Bellco roll-in incubator (2 rev./min) at 37 °C in Dubos broth containing 0.05% Tween 80 supplemented with 0.2% glycerol and 4% Dubos medium albumin. Middlebrook 7H11 agar was used for growth on plates. Where required, kanamycin was added at a final concentration of 25 μg/ml, hygromycin at 50 μg/ml and gentamicin at 15 μg/ml.
Creation of a M. tuberculosis H37Rv conditional whiB1 mutant
Approximately 1.5 kb of DNA sequence from each side of the whiB1 gene (Rv3219) was amplified from M. tuberculosis genomic DNA using PfuUltra (Stratagene) with the following primer pairs: for the 5′ side Myc642 (5′-GCGGATCCCGAACAGGCACAGCATCA-3′) and Myc643 (5′-GCGGATCCCGCCAATCCATTAGTCGT-3′) (bp 3594224-3595722), and for the 3′ side Myc644 (5′-CGGCGGCCGCGGTCTGACGACTCAGTTCT-3′) and Myc645 (5′-CGGCGGCCGCGACGGTGTCCTGGTGTGC-3′) (bp 3595961-3597391). The 5′ fragments were cloned into the BamHI site and 3′ fragments into the NotI site of the suicide gene delivery vector p2NIL, which is capable of replication within E. coli but lacks a mycobacterial origin of replication . A hygromycin cassette was cloned into the KpnI site of p2NIL to replace the whiB1 gene. A PacI fragment from pGOAL17 containing the lacZ and sacB genes was inserted to give the completed suicide delivery vector pwhiB1-KO (Table 1). This vector contains the sacB gene from Bacillus subtilis that causes lethality when expressed in M. tuberculosis, making it an effective counter-selectable marker, and a lacZ gene. The virulent M. tuberculosis strain H37Rv was transformed with pwhiB1-KO and selection was made for potential single crossovers as kanamycin- and hygromycin-resistant colonies expressing the lacZ gene. The veracity of the single crossover event was demonstrated by PCR with the following primers internal to the hygromycin cassette and external to the regions of homology: Myc152 (in hygromycin cassette, 5′-CGTTAGAACGCGGCTAC-3′) and Myc1786 (5′-GCCATTGACGATGTGCTG-3′); Myc153 (in hygromycin cassette, 5′-GGTCAGCGAACCAATCA-3′) and Myc1789 (5′-GATTACCTGATGTGGGTTCGC-3′); Myc1786 and Myc1787 (5′-GCACCGATTACAGACCAGT-3′); and Myc1789 and Myc1788 (5′-GCAACGCCCGCACGAAAGCC-3′). These colonies were streaked on to plates containing hygromycin to allow for double crossovers to occur and then serially diluted on to plates containing hygromycin, sucrose and X-Gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside). No colonies with a double crossover phenotype (hygromycin-resistant, sucroseS and white on X-Gal) were obtained at this stage, suggesting that the whiB1 deletion was lethal to the bacteria. Accordingly, a single-cross strain was transformed with a complementing plasmid (pDMH1) carrying the whiB1 gene plus 293 bp 5′-flanking and 14 bp 3′-flanking DNA cloned into the EcoRI site of pKP203 (gentamicin-resistant), an integrase-negative derivative of the integrating vector pMV306. This was prepared by PCR of genomic DNA using the primers Myc1777 (5′-GCGAATTC-GCAAGAAAGCGGATCTGAGC-3′) and Myc1778 (5′-GCGAATTC-AGAACTGAGTCGTCAGACC-3′). Plasmid pDMH1 was co-transformed into the whiB1 single crossover strain along with plasmid pBS-int carrying the integrase gene necessary to achieve integration of the plasmid into the chromosome. The pBS-int plasmid lacks a mycobacterial origin of replication and is therefore lost from the bacterium. Using this complemented strain for selection of double crossover events resulted in colonies with the required null phenotype (gentamicin-resistant, hygromycin-resistant, sucroseS and white on X-Gal) of which after testing by PCR as above, seven out of eight colonies were genuine double crossovers. This deletion removes all but the initial 11 5′ DNA bases and three 3′ bases of the whiB1 gene. Subsequently, the complementing plasmid in the whiB1-null mutant was switched  using pKP186 (kanamycin-resistant) plasmids with 1192 bp DNA 5′ to whiB1 [pDMH2, using primers Myc1779 (5′-GCGAATTC-GCCGCGACCTGCTGGCGCAC-3′) and Myc1778] and 1192 bp DNA 5′ and 535 bp DNA 3′ to the whiB1 gene [pDMH3, using primers Myc1779 and Myc1785 (5′-GCGAATTC-CACGATGCGTTGTCGATGTC-3′)] to give faster growing colonies (see the Results section).
Quantification of whiB1 expression by qRT-PCR (quantitative real-time PCR)
RNA isolated as previously described  was used to prepare cDNA using a QuantiTect reverse transcription kit (Qiagen). qRT-PCR on this cDNA was carried out on the ABI Prism 7700 Sequence Detection system using the Fast SYBR Green Master Mix (Applied Biosystems). The primers were designed using the Primer Express software (Applied Biosystems). The forward whiB1 primer was Myc947 (5′-TTCTTCCCGGTAGGAAACAGTG-3′), the reverse primer was Myc948 (5′-ATTACAGACCAGTTTCGCGTCA-3′). The forward primer from sigA, the normalizer gene, was Myc1790 (5′-TCGGTTCGCGCCTACCT-3′), the reverse primer was Myc1791 (5′-GGCTAGCTCGACCTCTTCCT-3′).
Overproduction and purification of WhiB1
For isolation of recombinant WhiB1, cultures of the E. coli expression strain (JRG6050) were grown at 37 °C in Lennox broth  to an attenuance (D) of ~0.6, when IPTG (isopropyl β-D-thiogalactoside; 120 μg/ml) was added and the cultures were incubated at 25 °C for a further 2 h. Cells were lysed after resuspension in 20 mM sodium phosphate buffer (pH 7.4) containing 0.5 M NaCl by two passages through a French pressure cell at 37 MPa. The lysate was cleared by centrifugation, and the resulting cell-free extract was applied to a nickel-charged Hi-Trap chelating column (GE Healthcare). The recombinant His6–WhiB1 protein was eluted using a linear imidazole gradient (0–500 mM in 20 ml). WhiB1-containing fractions were either used immediately or stored at −20 °C after addition of 1 mM DTT (dithiothreitol).
Reconstitution of WhiB1
WhiB1 was reconstituted under anaerobic conditions in a Mark3 anaerobic workstation (DW Scientific). Protein samples were typically incubated with 25 mM Tris/HCl buffer (pH 7.4), L-cysteine (1 mM), DTT (2.5 mM) and a 10-fold molar excess of ammonium ferrous sulfate and an aliquot of NifS (225 nM final concentration) as described by Crack et al. . Reactions were incubated at room temperature (20 °C) for up to 16 h before purification of the reconstituted protein by chromatography on a 1 ml HiTrap heparin column .
Total amino acid analysis, iron and reactive thiol assays
Amino acid analysis was carried out by Alta Bioscience (University of Birmingham, U.K.) by complete acid hydrolysis, for 24 h at 110 °C, of aliquots of WhiB1 that had been previously assayed for protein concentration using the Bio-Rad protein reagent with BSA as the standard. The results of the analyses were used to determine that the Bio-Rad assay overestimated the amount of protein present in the samples and a correction factor of 0.86 was required to obtain an accurate measurement of protein concentration.
Aliquots of reconstituted WhiB1 were used to determine their iron content. Samples were boiled in 1% (w/v) trichloroacetic acid for 5 min. The clarified supernatants were retained and mixed with saturated sodium acetate, bathophenanthroline sulfonic acid (0.3%) and sodium ascorbate (1.8%). After centrifugation, the absorbance at 595 nm was measured. The amount of iron present was estimated by comparison with a standard curve prepared using an iron standard solution (BDH).
Reactive thiols were measured according to a method described by Thelander  using ~4 μM apo-WhiB1 protein in each reaction. Where indicated, the apo-WhiB1 protein was denatured by pre-treatment with SDS to a final concentration of 1% (w/v).
WhiB1 samples were transferred to EPR tubes and, where indicated, a 27-fold molar excess of proline NONOate was added for the indicated times before the samples were frozen in liquid nitrogen. EPR spectra were recorded with a Bruker EMX spectrometer equipped with a TE-102 microwave cavity and an ESR-900 helium flow cryostat (Oxford Instruments). Spectra were recorded at both 20 K and 77 K. No [3Fe-4S]1+ signals were detected in any of the nitrosylated samples (results not shown). The microwave power and frequency were 2 mW and 9.68 GHz respectively, and the field modulation amplitude was 1 mT. Spectra were normalized to the same receiver gain. Spin intensities of paramagnetic samples were estimated by double integration of EPR spectra using 1 mM Cu(II) and 10 mM EDTA as the standard.
Scanning spectrometry was carried out using a Cary 50 Bio UV-Vis spectrophotometer using Hellma 10 mm cuvettes with a screw-top lid to maintain anaerobic conditions. Oxygen or NO was injected into the cuvette as air-saturated buffer or anaerobic solution of proline NONOate through the Hellma silicone seal in the lid using a Hamilton syringe. The concentration of NO released by aliquots of proline NONOate was measured using a NO electrode (ISO-NO™ Mark II; World Precision Instruments, Stevenage, U.K.) calibrated using dilutions of a saturated solution of NO gas in H2O (1.91 mM at 20 °C).
Protein disulfide reductase activity
EMSAs (electrophoretic mobility-shift assays), DNase I footprints and in vitro transcription reactions
EMSAs were performed using the indicated promoter DNA as described by Stapleton et al. , except that His6–WhiB1 was used in place of CRPMt. Apo-WhiB1, reconstituted holo-WhiB1 or holo-WhiB1 treated with a 20-fold molar excess of proline NONOate was used as indicated. Radiolabelled DNA (~2 nM) was incubated with 0–50 μM His6–WhiB1 in the presence of 40 mM Tris/HCl, pH 8.0, 1 mM EDTA, 114 mM NaCl, 1 mM DTT, 10 mM MgCl2, 0.25 mg/ml BSA and 1 μg of calf thymus DNA, for 5 min on ice. Where indicated, apo-WhiB1 was oxidized by treatment with 5 mM diamide before analysis. The resulting complexes were then separated on 6% polyacrylamide gels buffered with 0.5×TBE (45 mM Tris/borate and 1 mM EDTA). The sizes of the DNA fragments used were: whiB1, −289 to −4, obtained from pGS2060; rpfA, −531 to −349; 3616c, −1119 to −948, and −217 to −16; and ahpC, −127 to +1 (all obtained by PCR with genomic DNA as the template). Numbering is relative to the ATG start codons. For the competition assays, 100-fold excess unlabelled whiB1 or rpfA DNA was included in the EMSA with radiolabelled whiB1 as the target.
For DNase I footprinting, radiolabelled PwhiB1 DNA (~60 ng) was incubated with 20 μM apo-WhiB1 and 2.5–20 μM CRPMt in the presence of 40 mM Tris/HCl, pH 8.0, 75 mM KCl, 10 mM MgCl2, 0.5 mM EDTA, 5% (v/v) glycerol, 1 mM DTT and 250 μg/ml BSA. The complexes were then digested with 1 unit of DNase I for 15–60 s at 25 °C. Reactions were stopped by the addition of 50 mM EDTA, followed by phenol/chloroform extraction. The DNA was ethanol precipitated and resuspended in loading buffer [80% (v/v) formamide, 0.1% SDS, 10% (v/v) glycerol, 8 mM EDTA, 0.1% Bromophenol Blue and 0.1% Xylene Cyanole] for electrophoretic fractionation on 6% polyacrylamide-urea gels and autoradiographic analysis. Maxam and Gilbert G tracks of the DNA fragments were used to provide a calibration .
In vitro transcription reactions were carried out as described by Stapleton et al. , except that apo-WhiB1 was included as indicated. Apo-WhiB1 protein preparations lacked significant deoxyribonuclease and ribonuclease activities.
The M. tuberculosis whiB1 gene is essential
Attempts to create a M. tuberculosis whiB1 deletion mutant were unsuccessful. Although single crossovers were obtained, it was not possible to select for double crossover events. Therefore a conditional mutant was created by including a wild-type copy of the whiB1 gene on a plasmid (pDMH1) together with 293 bp DNA upstream of the whiB1 open reading frame. When pDMH1 was present, double crossovers that disrupted the chromosomal whiB1 gene were isolated. When the vector (pKP203) was present, double crossover events were not be selected. Therefore it was concluded that under the conditions used in the present study the whiB1 gene is essential for M. tuberculosis H37Rv. However, even in the presence of plasmid-encoded WhiB1 (pDMH1), the conditional whiB1 mutant grew slowly, forming small colonies on 7H11 medium. This growth defect was accounted for by the low level of expression of whiB1 from pDMH1 in the whiB1 mutant, which was only ~20% of that obtained from the chromosomal copy of the wild-type strain, as indicated by qRT-PCR. Complementation by a plasmid containing an extra 899 bp upstream of whiB1 (pDMH2) resulted in an improved growth rate, the formation of normal sized colonies, and an increase in expression of whiB1 (equivalent to ~65% of the wild-type). A plasmid that in addition also contained an extra 521 bp downstream of whiB1 (pDMH3) resulted in a further increase in whiB1 expression (equivalent to ~130% of wild-type). These observations suggest the presence of positive regulatory elements located upstream of −293 bp in the whiB1 promoter, and also up to 521 bp downstream of the whiB1 gene, or that the local structure of the chromatin for the chromosome and plasmid are sufficiently different to facilitate enhanced expression of chromosomal whiB1 compared with plasmid-borne whiB1. The whiB1 gene has its own promoter [5,29], and the gene downstream of whiB1 (Rv3220c) has the opposite orientation. Hence, it is unlikely that polar effects could account for either the apparent essentiality of whiB1 or the observation that DNA downstream of whiB1 is needed for full complementation of whiB1 expression. It should be noted that whiB1 has previously been reported as a non-essential gene based on TraSH (transposon site hybridization) screening , but there are several possible reasons for false negative results, including transposon insertion in non-essential regions either in or adjacent to essential genes resulting in TraSH probes that are still able to hybridize. This could be a particular problem for small genes such as whiB1.
WhiB1 contains an O2-insensitive [4Fe-4S]2+ cluster
The essential nature of M. tuberculosis whiB1 causes difficulty in assigning function. Several members of the Wbl family of proteins, including WhiB1, have been shown to possess iron–sulfur clusters . Although the properties of these iron–sulfur clusters are relatively poorly characterized, it is assumed that they are important for Wbl protein function. Therefore the M. tuberculosis H37Rv WhiB1 protein was overproduced as a soluble His6-tagged protein in E. coli BL21(λDE3) transformed with pGS2164. On the affinity column a dark brown band was formed, which eluted as pale yellow-brown fractions, with UV–visible spectra typical of a [4Fe-4S] cluster (results not shown). Upon longer-term storage under aerobic conditions, the preparations lost their colour, and in the absence of reducing agents the WhiB1 protein was prone to aggregation. Analysis by non-reducing SDS/PAGE showed the presence of different species with molecular masses and N-terminal amino acid sequences consistent with the presence of monomeric, dimeric and trimeric WhiB1 containing inter- and intra-molecular disulfide bonds (results not shown). The presence of disulfide bonds in apo-WhiB1 was confirmed by measuring the amount of reactive thiol using DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)] in the absence of reducing agents under native conditions and under denaturing conditions (after treatment with 1% SDS). This revealed the presence of two (1.88±0.27, n=9) reactive thiols in the native soluble apo-protein, implying the presence, on average, of one disulfide per apo-WhiB1. The number of reactive thiols per apo-WhiB1 decreased to 1.15±0.15 (n=9) after treatment with SDS, implying that unfolding the protein promoted additional disulfide bond formation (on average 1.5 per monomer). This propensity to form disulfide bonds accounts for the pattern of WhiB1 migration on non-reducing SDS/PAGE.
Reconstitution of the WhiB1 iron–sulfur cluster under anaerobic conditions was accompanied by an increase in absorbance at approx. 420 nm (Figure 1A). Total amino acid analysis confirmed the authenticity of the isolated protein and allowed accurate estimation of protein content such that, after removing excess reagents from reconstitution mixtures by chromatography on heparin Sepharose, the iron content of the reconstituted protein was measured as 3.57±0.06 Fe atoms per WhiB1 protein (n=3). In this form the protein was stable for several days if kept under anaerobic conditions and exhibited no tendency to aggregate, unlike apo-WhiB1 (see above). The reconstituted WhiB1 absorbance band at 420 nm was bleached by the addition of the reductant dithionite (Figure 1B). Exposure of the reconstituted protein to O2, by addition of air-saturated buffer equivalent to a final concentration of 110 μM O2 for up to 2 h, only slightly decreased absorbance at 420 nm (Figure 1C). Thus, it appears that the WhiB1 [4Fe-4S] cluster is more stable in the presence of O2 than those of other Wbl proteins (WhiD and WhiB3), which degrade significantly when exposed to air [14,16]. The anaerobic CD spectrum of WhiB1 had positive features at 427 nm and 509 nm (Figure 1D) that were similar to those of WhiD , but different from those of the O2 sensor FNR , indicating that the environment of the O2-insensitive WhiB1 iron–sulfur cluster is different from that of the O2-sensitive FNR iron–sulfur cluster.
The WhiB1 iron–sulfur cluster is highly sensitive to NO
NO is a key component of the host response to mycobacterial infection, and iron–sulfur clusters serve as sensors of NO in some bacterial regulatory proteins (e.g. NsrR and FNR) . Therefore reconstituted WhiB1 was titrated with the NO-releasing compound proline NONOate under anaerobic conditions and the response of the iron–sulfur cluster was monitored by obtaining UV–visible spectra (Figure 2A). Before addition of NO, the molar absorption coefficient of the WhiB1 iron–sulfur cluster was estimated to be ~18000 M−1·cm−1, which is similar to reported values for other [4Fe-4S]-containing proteins. NO addition caused decreased absorbance at 420 nm and increased absorbance at ~355 nm, with apparent isosbestic points at 395 nm and 480 nm, indicative of formation of DNICs . Under these conditions, eight NO (7.5–8.2) molecules reacted per WhiB1 [4Fe-4S] cluster (Figure 2B), suggesting the formation of an unusual octa-nitrosylated iron–sulfur cluster. The spectral properties of the product and the observed stoichiometry of the reaction are consistent with the formation of two dinuclear DNICs. Accordingly, using the molar absorption coefficient (~8500 M−1·cm−1) determined by Cruz-Ramos et al. , the absorbance increase at 360 nm upon completion of the NO titration was equivalent to the formation of ~2.5 dinuclear DNICs per iron–sulfur cluster. Remarkably, unlike the reaction of WhiB1 with O2, the spectral changes upon exposure to NO were extremely rapid. Using stopped flow to monitor absorbance changes at 364 nm upon mixing of a 92-fold molar excess of NO with holo-WhiB1 (7 μM iron–sulfur cluster), the reaction was complete within 30 s of mixing; the kinetic details of the reaction of WhiB1 with NO will be the subject of further study.
Reconstituted WhiB1 under anaerobic conditions was EPR silent, consistent with the presence of a diamagnetic [4Fe-4S]2+ cluster (Figure 3). However, monomeric DNICs have characteristic EPR spectra . After treatment with a 27-fold molar excess of proline NONOate under anaerobic conditions for 10 min, an axial EPR spectrum (g=2.048, 2.03 and 2.022) was detected, indicative of the presence of 6 μM monomeric DNIC arising from 17 μM iron–sulfur cluster (i.e. ~10% of total Fe was present as an EPR-detectable monomeric DNIC). Because DTT was present in the reaction buffer, some of this intensity may have arisen from a DTT-co-ordinated DNIC species. The intensity of the 2.03 signal decreased to that equivalent to 2.8 μM monomeric DNIC (equivalent to 4% total Fe as monomeric DNIC) after further incubation. Taken together with the UV–visible spectra, to which both mononuclear and dinuclear DNICs contribute, the EPR data suggest that exposure of holo-WhiB1 to NO results in the formation of a mixture of EPR active monomeric and EPR silent DNICs, the latter presumed to be the dinuclear form, with the EPR-silent form as the major species. Similar observations were made with the E. coli FNR protein after exposure to NO .
Under anaerobic conditions, the DNIC form of WhiB1 was stable for at least 2 h at 25 °C in the presence of 1 mM DTT (Figure 4, spectrum 2). Introduction of air caused the absorbance at 360 nm to decrease by ~20% after 2 h (Figure 4, spectrum 3), and only after >24 h was the DNIC form degraded to apo-WhiB1 (Figure 4, spectrum 4).
WhiB1 is an NO-responsive DNA-binding protein
It has been suggested that Wbl proteins are transcription factors and/or general protein disulfide reductases (see above). Therefore the protein disulfide reductase activity of apo-WhiB1, prepared by prolonged exposure to air, was tested in an insulin reduction assay . Upon reduction of disulfide bonds the insulin-B chain precipitates, which can be monitored at 600 nm. For comparison, E. coli thioredoxin was used as a bone fide general protein disulfide reductase. Although increased insulin reduction was observed as the concentration of apo-WhiB1 increased, the rates of reaction were slow compared with those obtained with a much lower concentration of thioredoxin and were similar to the rate obtained when both thioredoxin and WhiB1 were omitted from the reactions (Table 2). Garg et al.  reported significant WhiB1 protein disulfide reductase activity using a similar assay, and later showed that WhiB1 specifically interacted with and was able to reduce disulfide bonds in the α(1,4)-glucan branching enzyme GlgB . However, the results shown in the present study indicate that M. tuberculosis WhiB1 is not a general protein disulfide reductase, but does not exclude the possibility that apo-WhiB1 could act as a specific protein disulfide reductase.
The DNA-binding ability of WhiB1 was tested using EMSAs with the whiB1 promoter (PwhiB1) as the target. Apo-WhiB1 (0.08±0.04 Fe atoms per monomer), in the presence of the reducing agent DTT, bound PwhiB1 (Figure 5A, lanes 2–5). The smeared appearance of the EMSAs at lower protein concentrations suggests that the complex is relatively unstable and dissociates during electrophoresis (Figure 5A, lane 3). DNA-binding by WhiB1 was inhibited by the presence of the [4Fe-4S] cluster (Figure 5A, lanes 6–9), but was restored after treatment of holo-WhiB1 with NO (Figure 5B). Furthermore, DNA binding was abolished following reconstitution of an iron–sulfur cluster into NO-treated WhiB1 (results not shown). In the absence of additional reducing agent (DTT), NO-treated WhiB1 formed a second retarded species of lower mobility (Figure 5B, lane 5), which was absent following the addition of DTT (Figure 5B, lane 10). A similar species was also present when apo-WhiB1 was analysed (Figure 5A, lane 5). Oxidation of apo-WhiB1 by treatment with 5 mM diamide increased the amount of the slow migrating DTT-sensitive complex (Figure 5C, compare lanes 2 and 5). These data, and the relative stability of the DNIC-WhiB1 (Figures 3 and 4), suggests that the aerobic EMSAs shown in Figure 5(B) contained mostly dinuclear DNIC-WhiB1 mixed with some reduced and oxidized apo-WhiB1. These results are consistent with the assignments of the slower-migrating complex as that formed between PwhiB1 and oxidized (disulfide) WhiB1, and the faster-migrating complex as that formed between PwhiB1 and reduced (dithiol) apo-WhiB1 and/or DNIC-WhiB1. Specificity of the DNA binding was demonstrated by the inability of apo-WhiB1 to bind rpfA, ahpC and Rv3616c promoter DNA (results not shown). Specificity of DNA binding was further demonstrated by competition experiments in which a 100-fold excess of unlabelled PwhiB1 DNA inhibited binding of both oxidized (Figure 5C, compare lanes 2 and 3) and reduced (Figure 5D, compare lanes 2 and 3) apo-WhiB1 to radiolabelled PwhiB1, whereas 100-fold excess unlabelled rpfA promoter DNA did not (Figure 5C, compare lanes 2 and 4; Figure 5D, compare lanes 2 and 4). Thus it was concluded that DNA-binding by WhiB1 is complex and is affected by the oxidation state of the apo-protein and the chemical form ([4Fe-4S] or DNIC) of its cofactor.
DNase I footprinting of apo-WhiB1 at PwhiB1
Footprinting of apo-WhiB1 at PwhiB1 revealed a distinctive DNase I cleavage pattern characterized by a protected region (W1) located at −40 to −3 relative to the transcript start (Figure 6). The footprint also exhibited several hypersensitive sites, suggesting that apo-WhiB1 causes significant distortion of PwhiB1 (Figure 6). W1 was located downstream of the activating Rv3676-binding site (CRP1, centred at −58.5 ) and overlapped the repressing Rv3676-binding site (CRP2, centred at −37.5) and the −10 element of PwhiB1.
Apo-WhiB1 represses whiB1 transcription in vitro
Previous work has shown that whiB1 expression is activated by intermediate concentrations (2.5 μM) of the M. tuberculosis cAMP receptor protein (Rv3676), but that higher concentrations (20 μM) result in repression of whiB1 expression . In vitro transcription reactions confirmed this pattern of regulation (Figure 7, compare lanes 1, 5 and 9). In the absence of Rv3676, apo-WhiB1 significantly inhibited whiB1 transcription at concentrations as low as 2 μM (Figure 7, lanes 1–4). In the presence of activating concentrations of Rv3676 (Figure 7, lanes 5–8), or in the presence of repressing concentrations of Rv3676 (Figure 7, lanes 9–12), apo-WhiB1 still repressed whiB1 transcription. It was therefore concluded that apo-WhiB1 acts as an inhibitor of whiB1 expression, and that this inhibition is more severe than that mediated by high concentrations of Rv3676 and silences any Rv3676-dependent activation of whiB1 expression.
NO reacts with proteins that contain iron–sulfur clusters, non-haem iron, haem and thiols , and the consequent toxic effects are exploited by host immune systems in the response to pathogenic bacteria , providing the selection pressure for such bacteria to evolve mechanisms for NO detoxification . To mount an effective defence to counter the effects of NO, bacteria must reprogramme gene expression. Several NO-sensing regulators are known, including the non-haem iron protein NorR, the haem-containing DosS-R and DosT-R two-component systems, and the iron–sulfur proteins FNR, IRP1 (iron-regulatory protein-1), NsrR, SoxR, and WhiB3 [19,31,37–44]. Reaction of NO with the [4Fe-4S] cluster of the O2 sensor FNR yields a DNIC form that is incapable of DNA binding . Similarly, NO nitrosylates the [2Fe-2S] cluster of NsrR to inactivate the protein . In contrast, reaction of the SoxR [2Fe-2S] cluster with NO to form a protein-bound DNIC activates the protein . The [4Fe-4S] cluster of WhiB3 reacts with NO to form DNIC species, but it is the cluster-free oxidized form that binds DNA with high affinity . IRP1 is a post-transcriptional regulator of iron homoeostasis in higher organisms . Similar to WhiB3, the apo-form of IRP1 binds nucleic acid (in this case mRNA), and its [4Fe-4S] cluster reacts with NO to yield protein-bound DNIC [41,42]. Exposure of IRP1 to NO activates mRNA binding, probably by promoting cluster disassembly [43,44]. We have shown that the WhiB1 iron–sulfur cluster reacts specifically and very rapidly with NO to convert the protein from a non-DNA-binding form into a DNA-binding form capable of transcription regulation in vitro. Expression of whiB1 in response to NO in vivo has not been specifically tested, but the evidence obtained thus far suggests that the essential protein WhiB1 is likely to play a role in establishing M. tuberculosis infections by controlling gene expression in response to macrophage-generated or endogenous NO.
Iron–sulfur clusters are intrinsically sensitive to O2 and thus the stability of the WhiB1 iron–sulfur cluster in air is worthy of comment. The highly conserved CXXCXXXXXC motif in Wbl proteins, which contains three of the four cysteine ligands of the iron–sulfur cluster, is also present in the extremely sensitive O2 sensor FNR (Figure 8) . In the latter, the O2 sensitivity of the iron–sulfur cluster is determined by the protein fold [17,18,45,46]. A systematic study of FNR in which the amino acid (serine) located downstream of Cys 2 (labelled 2 in Figure 8) was replaced by all possible alternatives showed that a proline residue in this position enhanced the stability of the FNR iron–sulfur cluster in air . The presence of the proline residue immediately adjacent to Cys 2 may go some way to account for the relative O2 insensitivity of the iron–sulfur cluster of WhiB1 and other Wbl proteins. A second stabilizing substitution is replacement of the leucine residue located upstream of Cys 3 by histidine . In WhiB1 and WhiD, this position is occupied by a glutamate residue; however, a systematic study of the effects of amino acid replacements at this position has not been reported and so the effects of the presence of a glutamate rather than a leucine residue at this position are unknown.
In spite of its insensitivity to O2, the WhiB1 [4Fe-4S] cluster is remarkably sensitive to NO, reacting extremely rapidly to generate DNIC species. The CD spectrum of WhiB1 is different from that of the O2 sensor FNR , indicating that the environment of the O2-insensitive WhiB1 iron–sulfur cluster is different from that of the O2-sensitive FNR iron–sulfur cluster. Further detailed biochemical analysis will be necessary to determine the molecular basis underpinning the different responses of these two gas sensors.
In contrast with previous reports , the data in the present study indicate that WhiB1 is not a general protein disulfide reductase (Table 2). Nevertheless, the propensity of the apo-protein to form intra- and inter-molecular disulfide bonds suggests that such redox chemistry is possible, but might require an as yet unidentified reductase and specific target proteins for reduction . WhiB1 interacts with the α(1,4)-glucan branching enzyme GlgB, which is essential for optimal growth of M. tuberculosis , and thus WhiB1 might be bifunctional, acting as both a gene regulator and a specific GlgB reductase.
The apo-WhiB1 footprint at PwhiB1 showed a 37 bp protected region (−40 to −3, relative to the transcript start) that contained related DNA sequences (T−32GACA−28 and T−14GAAA−10), but it would be premature to propose an apo-WhiB1-binding site sequence; further footprints and/or SELEX (systematic evolution of ligands by exponential enrichment) experiments are necessary. The footprint was characterized by hypersensitive sites spanning a region that was too large to be accounted for by simple binding of an apo-WhiB1 monomer, suggesting multiple apo-WhiB1–DNA interactions (Figure 6). Based on the location of the apo-WhiB1 footprint, the simplest explanation for repression of whiB1 transcription is that apo-WhiB1 blocks RNAP (RNA polymerase) binding. Alternatively, by analogy with WhiB3 , apo-WhiB1 could interact with the sigma factor, SigA, to repress whiB1 transcription by inhibiting RNAP activity or by preventing RNAP from escaping the promoter.
It has recently been shown that upon infection of macrophages, a mycobacterial-derived cAMP burst promotes bacterial survival by interfering with host signalling pathways . Moreover, cAMP is important in M. tuberculosis gene regulation [4,5,48]. Dual regulation of whiB1 expression by the cAMP-responsive regulator CRPMt (Rv3676) and the NO-responsive WhiB1 protein might provide a mechanism to integrate the transcriptional response to two important signals associated with infection (Figure 9). Whatever the detailed mechanism proves to be, the evidence thus far suggests that in the macrophage environment (high NO, high cAMP) transcription of the essential gene whiB1 will be inhibited. This is consistent with entry into the persistent state. At low NO concentrations (immunocompromised state), holo-WhiB1 is formed, de-repressing whiB1 transcription, consistent with emergence from the dormant state. The reason for the essentiality of whiB1 in vitro is not established, but the data suggest that either sufficient NO is generated during in vitro culture of M. tuberculosis to activate WhiB1, or that the [4Fe-4S] form regulates as yet unidentified essential gene(s), or that the interaction between WhiB1 and GlgB [30,33] is essential for normal growth. Note that WhiB2 is also essential and binds DNA in the apo-form [49,50], and that the transcription factor IscR regulates distinct groups of genes in its [2Fe-2S] and apo-forms .
In conclusion, it is shown that M. tuberculosis WhiB1 is an essential [4Fe-4S] protein that acts as a specific NO-sensing DNA-binding protein. The essential nature of the whiB1 gene and switching of WhiB1 DNA-binding activity in response to NO has implications for the control of M. tuberculosis gene expression in the host environment. Although high NO concentrations kill M. tuberculosis , lower concentrations promote entry into the latent state , and thus NO sensing and gene expression reprogramming by WhiB1 could contribute to developmental adaptations in response to host-generated NO.
Laura Smith, Melanie Stapleton, Gavin Fullstone, Jason Crack, Salvatore Adinolfi, Debbie Hunt and Evelyn Harvey contributed to the experimental design, carried out the experiments, analysed the data, and helped with the drafting of the paper. Andrew Thomson, Nick Le Brun, Roger Buxton and Jeffrey Green initiated the project, contributed to the experimental design, data analysis and preparation of the paper.
This work was supported by The Wellcome Trust [grant number 078731/Z/05/Z], the Medical Research Council [grant number U117585867] and the Biotechnology and Biological Sciences Research Council [BB/G019347/1].
We thank Pearline Benjamin (National Institute for Medical Research, London, U.K.) for constructing the whiB1 knock-out plasmid, Myles Cheesman (Department of Chemistry, University of East Anglia, Norwich, U.K.) for the use of CD and EPR spectrometers, Arthur J. Moir (Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield, U.K.) for N-terminal amino acid analyses, and Robert K. Poole (Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield, U.K.) for access to the NO electrode.
Abbreviations: CRPMt, Mycobacterium tuberculosis cAMP receptor protein Rv3676; DNIC, dinitrosyl-iron thiol complex; DTT, dithiothreitol; EMSA, electrophoretic mobility-shift assay; IRP1, iron-regulatory protein-1; PwhiB1, whiB1 promoter region; qRT-PCR, quantitative real-time PCR; RNAP, RNA polymerase; TraSH, transposon site hybridization; X-Gal, 5-bromo-4-chloroindol-3-yl β-D-galactopyranoside
- © The Authors Journal compilation © 2010 Biochemical Society