By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
P. Agrawal, Institute of Microbial Technology, Sector-39A, Chandigarh 160 036, India Fax: +91 172 269 0585 Tel: +91 172 263 6680/263 6681; Ext 3264 E-mail: email@example.com
The whiB-like genes (1-7) of Mycobacterium tuberculosis are involved in cell division, nutrient starvation, pathogenesis, antibiotic resistance and stress sensing. Although the biochemical properties of WhiB1, WhiB3 and WhiB4 are known, there is no information about the other proteins. Here, we elucidate in detail the biochemical and biophysical properties of WhiB2, WhiB5, WhiB6 and WhiB7 of M. tuberculosis and present a comprehensive comparative study on the molecular properties of all WhiB proteins. UV–Vis spectroscopy has suggested the presence of a redox-sensitive [2Fe–2S] cluster in each of the WhiB proteins, which remains stably bound to the proteins in the presence of 8 m urea. The [2Fe–2S] cluster of each protein was oxidation labile but the rate of cluster loss decreased under reducing environments. The [2Fe–2S] cluster of each WhiB protein responded differently to the oxidative effect of air and oxidized glutathione. In all cases, disassembly of the [2Fe–2S] cluster was coupled with the oxidation of cysteine-thiols and the formation of two intramolecular disulfide bonds. Both CD and fluorescence spectroscopy revealed that WhiB proteins are structurally divergent members of the same family. Similar to WhiB1, WhiB3 and WhiB4, apo WhiB5, WhiB6 and WhiB7 also reduced the disulfide of insulin, a model substrate. However, the reduction efficiency varied significantly. Surprisingly, WhiB2 did not reduce the insulin disulfide, even though its basic properties were similar to those of others. The structural and functional divergence among WhiB proteins indicated that each WhiB protein is a distinguished member of the same family and together they may represent a novel redox system for M. tuberculosis.
Mycobacterium tuberculosis has a remarkable ability to survive under hostile conditions it encounters during infection . Despite extensive research directed towards understanding the physiology of M. tuberculosis and its molecular pathogenesis [1–3], many fundamental questions about the mechanisms of survival during early infection and persistence remain poorly understood. Among several intriguing questions, are: (a) what are the bacterial determinants necessary for early infection, (b) how does the bacterium counteract or evade its host’s defenses to survive the vigorous host-immune response, (c) what regulates the transition from initial growth to persistence and back to active growth, (d) are the bacteria present in a non-replicating ‘spore-like’ state or do they replicate at all during latency, and (e) how does the bacterium adapt to survive under the anaerobic and nutritionally altered environment within the granuloma? The answers to these questions are likely to provide insight into the mechanisms by which M. tuberculosis establishes infection and persists within the host and the means to eliminate latent infection, a phase of the disease that poses the most significant obstacle to the eradication of tuberculosis. To survive and establish successful infection, M. tuberculosis appears to have acquired a strong network of genes to sense and respond to stress conditions; the properties of many of these are poorly understood.
A family of genes, whiB, has received attention because of their involvement in cell division (whiB2), fatty acid metabolism and pathogenesis (whiB3), antibiotic resistance (whiB7) and in sensing a variety of stress conditions [4–9]. Seven genes, whiB1/Rv3219, whiB2/Rv3260c, whiB3/Rv3416, whiB4/Rv3681c, whiB5/Rv0022c, whiB6/Rv3862c and whiB7/Rv3197A, have been identified in M. tuberculosis [10,11] as orthologs of the whiB gene of Streptomyces coelicolor A3(2), which has been shown to be involved in sporulation .
Although, WhiB proteins are annotated as putative transcription factors , to date it has not been shown directly that these proteins work as transcription factors. We have previously reported that WhiB1/Rv3219 , WhiB3/Rv3416  and WhiB4/Rv3681c  are protein disulfide reductases. WhiB4 has been postulated to act as a sensor of oxidative stress, wherein the inactive holo protein (containing a [4Fe–4S] cluster) transformed into an active apo protein (without an iron–sulfur cluster) in oxidizing environments and gained protein disulfide reductase activity . However, to date the biochemical features of WhiB2, WhiB5, WhiB6 and WhiB7 from M. tuberculosis have not been reported. The observations that different whiB mutations impart distinct phenotypes and respond differently to stress conditions indicate importance of each member separately in mycobacterial physiology. The available information on WhiB proteins demands careful investigation of the biochemical and biophysical properties of each.
Mycobacterial WhiB proteins have 22–67% identity with WhiB protein of S. coelicolor A3(2). Sequence analysis of M. tuberculosis WhiB proteins shows the presence of four conserved cysteines arranged as ‘C-X19-36-C-X-X-C-X5-7-C’ . Notably, two cysteines are present in a conserved CXXC motif, except in WhiB5/Rv0022c where it is CXXXC (CLRRC). Proteins with the CXXC motif have been implicated in diverse functions, for example, protein disulfide oxidoreductase activity , redox sensing  and the coordination of metal cofactors . The functional importance of the conserved cysteine residues in iron–sulfur cluster coordination and protein disulfide reductase has been demonstrated in WhiB4 . Recently, cysteines of WhiB3 have also been shown to act as a ligand for the O2- and NO-responsive [4Fe–4S] cluster .
The presence of four conserved cysteines and a CXXC motif in WhiB proteins from M. tuberculosis raises several questions: are all WhiB proteins coordinated with an iron–sulfur cluster? If yes, then what are their basic properties? Are the iron–sulfur clusters equally oxidation labile? Does removal of the iron–sulfur cluster lead to disulfide bond formation? Are the structural features of mycobacterial WhiB proteins similar? Do all WhiB proteins behave like protein disulfide reductase? The objective of this study is to answer several of the questions raised above.
This is the first study to report the biochemical and biophysical properties of WhiB2, WhiB5, WhiB6 and WhiB7 of M. tuberculosis and also compare the properties of all seven WhiB proteins. We show that, similar to WhiB3 and WhiB4, other freshly purified WhiB proteins also coordinate a [2Fe–2S] cluster which respond differently to the oxidizing environment. Except WhiB2, apo WhiB5, WhiB6 and WhiB7 also reduce insulin in vitro, but the efficiency of the reduction varies. An extensive biophysical study suggested that the WhiB proteins of M. tuberculosis are structurally different. The functional relevance of their divergent molecular properties is discussed.
All seven whiB genes of M. tuberculosis encode iron–sulfur proteins
Previous work on WhiB3  and WhiB4  identified the presence of cysteine-bound iron–sulfur cluster in these proteins. We speculated that all seven WhiB proteins may also coordinate an iron–sulfur cluster. Therefore, we overexpressed the recombinant WhiB proteins (with an N-terminal S-tag and C-terminal 6 × His tag) in Escherichia coli BL21 (DE3). Overexpression at 37 °C for 3 h led to the formation of light brown inclusion bodies. However, induction at 16 °C for 20 h resulted in the expression of ∼ 10–20% of each WhiB protein in the soluble form. On SDS/PAGE, the mass of Ni2+-NTA-purified WhiB proteins corresponded to their theoretically calculated molecular mass (predicted molecular mass + ∼ 5 kDa tags) (Fig. S1). Proteins purified from a soluble fraction or after denaturation or by in-column refolding were ∼ 98% pure and were brownish (Figs S1 and S2).
The presence of four conserved cysteines and the brownish appearance of purified WhiB1, WhiB2, WhiB5, WhiB6 and WhiB7 indicated the presence of an iron–sulfur cluster. To identify and confirm the presence of the iron–sulfur cluster, the absorption spectra of the purified proteins were recorded in the range 200–700 nm. In addition to a peak at 280 nm, two additional peaks at ∼ 333–340 and ∼ 420–424 nm, along with two broad shoulders at ∼ 460 and ∼ 560–580 nm were observed (Fig. 1). The peaks were characteristic of an [2Fe–2S] cluster , therefore, it was assumed that freshly purified WhiB1, WhiB2, WhiB5, WhiB6 and WhiB7 also coordinated the [2Fe–2S] cluster. The absorption spectra of different WhiB proteins were largely indistinguishable, however, in WhiB6 and WhiB7, the shoulder at ∼ 460 nm was more prominent than in others. This subtle change in the peak pattern may be because of their differential electronic environment. The nature and type of amino acids and their side-chain orientations around iron–sulfur cluster coordination sites are the likely cause of minor variations in the electronic properties, which were reflected in their absorption spectra.
The brownish appearance of the protein purified in the presence of 8 m urea indicated that the iron–sulfur cluster of WhiB proteins had survived treatment by a denaturant, a feature very similar to WhiB3  and WhiB4 . Unlike proteins purified from the soluble fraction, which had a spectral feature typical of the [2Fe–2S] cluster, proteins in 8 m urea showed a single peak at ∼ 400–415 nm (Fig. S3). The differential peak features may be due to the solvent-induced conformational change, which is possibly because of changes in the chemical environment around the iron–sulfur cluster, the partial destruction of the cluster or its conversion to other forms. In order to investigate the probable reason(s) for the observed difference, the proteins were processed for in-column refolding. The absorption spectra of the in-column refolded proteins were similar to those of their native counterparts (Fig. S3). Interestingly, iron–sulfur cluster-specific peak intensities were similar in both conditions. These data suggest that the coordination of iron–sulfur clusters to the WhiB proteins was unaffected by 8 m urea and the differences in peak patterns were due to the presence of urea. In order to acquire firm evidence for this observation, the total iron content of proteins purified under different conditions was measured.
The total iron content of the native and in-column refolded protein varied between 0.14 and 0.20 atoms per monomer (Table 1). The sub-stoichiometric iron content of iron–sulfur proteins is generally due to the impaired incorporation of the cluster into the protein during overexpression in E. coli and/or loss during purification when conditions are not strictly anaerobic . We attempted to reconstitute the iron–sulfur cluster in WhiB proteins in vitro using FeCl3 and Na2S, but did not succeed. Therefore, incorporated l-cysteine as a sulfur source in the reconstitution assay. IscS/Rv3025c, a cysteine desulfurase  of M. tuberculosis was cloned, expressed in E. coli and purified by metal-affinity chromatography (data not shown). The WhiB proteins were incubated in the reaction mixture along with FeCl3, IscS and 35S-cysteine. We observed an IscS-dependent mobilization of sulfur from l-cysteine to the iron–sulfur cluster of WhiB proteins (Fig. 2A). In the control reactions, where IscS was excluded or the iron concentration was limited (10-fold less), we did not observe any signal (Fig. 2A). Further characterization of the iron–sulfur cluster of the reconstituted samples could not be carried out because none of the samples gave an EPR signal at 120K using liquid nitrogen (data not shown). It is possible that a further decrease in temperature (using liquid helium) would be required in order to detect the EPR signal. Nevertheless, the absorption spectra of the reconstituted proteins showed a single peak at ∼ 420 nm indicating the presence of a [4Fe–4S] cluster (Fig. 2B). The presence of a similar cluster has been reported in WhiB3 and WhiB4.
Table 1. Total iron content in WhiB proteins. Proteins under different conditions were purified as described in Experimental procedures. Data for each protein sample are expressed as means ± SD (three independent protein preparations).
Atoms of iron per monomer
0.131 ± 0.014
WhiB1 (in 8 m urea)
0.141 ± 0.052
0.138 ± 0.028
0.008 ± 0.005
0.145 ± 0.022
WhiB2 (in 8 m urea)
0.142 ± 0.020
0.142 ± 0.035
0.006 ± 0.005
0.185 ± 0.028
WhiB5 (in 8 m urea)
0.186 ± 0.036
0.188 ± 0.045
0.010 ± 0.007
0.212 ± 0.065
WhiB6 (in 8 m urea)
0.198 ± 0.050
0.208 ± 0.072
0.007 ± 0.003
0.182 ± 0.035
WhiB7 (in 8 m urea)
0.189 ± 0.066
0.175 ± 0.020
0.007 ± 0.005
The iron content of proteins purified from the soluble fraction, from inclusion bodies, under denaturing conditions and after refolding was similar (Table 1). The data clearly suggested that the protein fold responsible for holding the iron–sulfur cluster was resistant to the denaturing effect of 8 m urea. The ability of the iron–sulfur cluster to survive the effects of protein denaturants is a feature of high potential iron–sulfur proteins . It is possible that WhiB proteins also fall into the same category. However, detailed analysis would be required to establish this.
Iron–sulfur clusters of M. tuberculosis WhiB proteins are redox sensitive
Previously, we reported that the iron–sulfur cluster of WhiB4 disintegrates under an oxidizing environment, but not under reducing conditions . The rate of disintegration was directly correlated with the duration and strength of the oxidizing environment. Similarly, in this study, the intensity of the brown color and the iron–sulfur cluster-specific peaks decreased gradually as the time of exposure to air increased (Fig. S4). The results suggest that the iron–sulfur clusters were susceptible to oxidative degradation, although the rate of degradation varied significantly. WhiB1 lost ∼ 65% of its iron–sulfur clusters in the initial 6 h, whereas in WhiB6 and WhiB7 the loss was ∼ 8–10%. After 48 h of air exposure, losses were as follows: ∼ 80% in WhiB1, ∼ 75% in WhiB2 and WhiB5, ∼ 65% in WhiB3, ∼ 60% in WhiB4, and ∼ 35–40% in WhiB6 and WhiB7 (Fig. 3A). It was evident that the iron–sulfur clusters were most stable against air oxidation in WhiB6 and WhiB7, and most labile in WhiB1.
To study the sensitivity towards oxidized glutathione (GSSG), reduced glutathione (GSH) and dithiothreitol, proteins were incubated with 10 mm of each agent and the absorbance at 424 nm (A424) was recorded at different time intervals up to 42 h. All WhiB proteins showed differential sensitivity towards oxidation by GSSG, and similar to air oxidation, the iron–sulfur clusters of WhiB6 and WhiB7 were comparatively more stable (Fig. 3B). A reducing environment (in the presence of GSH or dithiothreitol) significantly lowered the rate of disintegration of the iron–sulfur cluster in each of the WhiB proteins (Fig. 3C,D). Therefore, disassembly of the iron–sulfur cluster under oxidizing conditions and its stability under reducing conditions suggested that the iron–sulfur clusters of M. tuberculosis WhiB proteins are redox sensitive. We assume that the iron–sulfur clusters of different WhiB proteins would respond differently to the oxidative stress encountered by M. tuberculosis in vivo.
Iron–sulfur clusters of WhiB proteins are differentially exposed to the external environment
The differential sensitivity of the iron–sulfur cluster towards different oxidizing agents could be attributed to their relative surface accessibility. We hypothesized that the iron–sulfur cluster of WhiB6 and WhiB7 may be sequestered in the interior of the holo protein, thereby shielding it from oxidative degradation. In order to study the surface accessibility of the iron–sulfur cluster of WhiB proteins, freshly purified proteins were incubated with 20 mm EDTA and the degree of iron chelation was monitored by recording the A424 at different intervals up to 20 h. Immediate chelation of iron was not observed in any WhiB protein. However, as time increased, the degree of iron chelation increased and the extent of chelation varied (Table 2). Almost 30% of the iron was chelated within 2 h in WhiB1 and WhiB2, whereas in WhiB6 and WhiB7 it was negligible over the same period. Even after 20 h of incubation, iron chelation was ∼ 15–20% (minimum) in WhiB6 and WhiB7, whereas it was ∼ 60% (maximum) in WhiB1 and WhiB2; in the other proteins it varied between 20% and 60% (Table 2). From the data, it appears that the surface accessibility of the iron–sulfur cluster is significantly different in different WhiB proteins.
Table 2. Stability of the iron–sulfur cluster from various WhiB proteins against EDTA. The initial reading was set to 100% and the change in A420 (residual) is expressed relative to the reading at t = 0. Data are expressed as means ± SD (three independent protein preparations).
Change in A420 (%)
62 ± 4
53 ± 4
44 ± 3
72 ± 5
44 ± 3
38 ± 6
82 ± 3
70 ± 5
58 ± 3
92 ± 2
59 ± 5
49 ± 6
80 ± 2
65 ± 3
58 ± 3
96 ± 3
89 ± 2
84 ± 2
98 ± 2
96 ± 4
80 ± 4
Cysteine residues form two intramolecular disulfide bonds after removal of the iron–sulfur cluster
It has been shown in WhiB4 that the cysteine-thiols, which are ligands of the iron–sulfur cluster, undergo oxidation and form two intramolecular disulfide bonds after disassembly of the iron–sulfur cluster . The presence of two intramolecular disulfide bonds has also been demonstrated in apo WhiB1  and apo WhiB3 . Proteins containing intramolecular disulfide bond(s) often show retarded mobility on SDS/PAGE under reducing conditions [15,23]. Both apo WhiB2 and apo WhiB5 showed significant retarded mobility on SDS/PAGE under reducing conditions, indicating the presence of intramolecular disulfide bond(s) (Fig. 4). Alkylation of cysteine by iodoacetamide (IAA) coupled with MS was used to determine the status of cysteines after removal of the iron–sulfur cluster. Reaction of IAA with a cysteine-thiol causes an increase in molecular mass of 57 Da, therefore, the total increase in the mass after reduction of the disulfide bond reflects the total number of cysteines present in the thiol and disulfide forms. In the oxidized state, both WhiB2 and WhiB5 showed a major peak corresponding to the theoretical molecular mass of the recombinant protein. However, the reduced proteins had increased molecular masses, representing alkylation of four cysteine residues in each case (Fig. 5). Although, WhiB5 and WhiB6 did not show any mobility differences under reducing conditions, a similar increase in mass was found after reduction (Fig. 5). The difference in mass between the oxidized and reduced forms suggested the presence of four cysteine-thiols in the reduced apo WhiB proteins. Because none of the cysteines was present in a thiol form in the oxidized protein (except for one in WhiB6 which has five cysteines), it was concluded that the apo form of all WhiB proteins contained two intramolecular disulfide bonds.
All apo WhiB proteins, except WhiB2, reduce the insulin disulfide
Previously, we reported that apo WhiB1  WhiB3  and WhiB4  are protein disulfide reductases. The enzymatic activity of WhiB4 was shown to be governed by the CXXC motif . Because the CXXC motif is present in all WhiB family members of M. tuberculosis, except WhiB5/Rv0022c (CXXXC), we tested the protein disulfide reductase activity of WhiB2, WhiB5, WhiB6 and WhiB7 by insulin disulfide reduction assay. This is a standard assay to asses the disulfide reductase activity of any protein in which reduction of the insulin disulfide by dithiothreitol in the presence of a test protein is monitored . Reductase activity was calculated by dividing the maximal slope of the curve (ΔA650·min−1) by the onset time of precipitation (time when A650 reached 0.05) . Except WhiB2, all WhiB proteins catalyzed the reduction of insulin disulfide (Table 3, Fig. S5). However, for WhiB2, the possibility of the presence of a natural in vivo substrate protein(s) still remains. Because M. tuberculosis RshA also has a C86XXC89 motif, and purified recombinant RshA (lab preparation) was therefore used as a control, but it did not catalyze insulin reduction.
Table 3. Protein disulfide reductase activity of apo WhiB proteins. The data for each protein sample (3 μm) are expressed as means ± SD (three independent protein preparations).
Reductase activity (× 10−3ΔA650 nm·min−2)
a Taken from Alam & Agrawal . b Taken from Alam et al. .
4.78 ± 0.25
0.56 ± 0.16
4.19 ± 0.62
42.2 ± 0.86
10.96 ± 0.55
2.79 ± 0.22
5.44 ± 0.82
0.58 ± 0.10
0.62 ± 0.15
Formation of a reversible intramolecular disulfide bond between the cysteines of the CXXC motif is essential for protein disulfide reductase activity [16,26]. Therefore, we assume that in WhiB proteins, one disulfide bond is formed between the two cysteines of the CXXC motif (CXXXC in the case of WhiB5) and the other between the remaining two cysteines. The assumption is supported by our earlier data, in which a similar arrangement of intramolecular disulfide bonds in WhiB4 was established . In WhiB6, one of the intramolecular disulfide bonds appeared to have formed between Cys53 and Cys56 but the involvement of cysteines for the second bond is little hard to predict, as it contains five cysteines (Cys12, Cys34, Cys53, Cys56 and Cys62).
Divergence in the secondary structure composition of WhiB proteins of M. tuberculosis
The multiple sequence alignment of WhiB proteins of M. tuberculosis showed 49–66% sequence homology and 31–50% identity with respect to each other (Table S1). However, because of the variation in amino acid composition, it is possible that structural variations may be an important determinant of their functional properties in vivo. Therefore, the structural organization of each M. tuberculosis WhiB protein was studied using biophysical tools. The secondary structure of each WhiB protein was analyzed by CD spectroscopy. The far-UV CD spectra of WhiB proteins were dissimilar because their molar ellipticities varied significantly (Fig. 6). The spectra showed α-helix, β-strand and random coil features. However, the proportion of each feature varied among WhiB proteins, as evident from the difference in negative molar ellipticity at specific wavelengths, i.e. 208 and 222 nm (α helix signature), 218 nm (β strand signature), 202–204 nm (random coil signature). In WhiB5 and WhiB6, the structure was dominated by α helices and β strands and the proportion of these structural elements was higher in WhiB6. WhiB1, WhiB2 and WhiB4 showed relatively increased molar ellipticity at 202–204 nm, indicating the presence of a significant proportion of random coils (Fig. 6). The proportion of all three secondary structural elements in WhiB3 and WhiB7 appeared similar. Disulfide bond formation did not affect the secondary structure, except for WhiB5 and WhiB6, and the effect was more pronounced in WhiB6 (Fig. 6).
It was observed that the secondary structure of WhiB1 , WhiB3  and WhiB4  resists thermal denaturation. Thus, we asked would other WhiB proteins also show similar features? Surprisingly, the secondary structures of WhiB5 and WhiB6 started to melt at 50 and 77 °C respectively (Fig. 7). At 70 °C, WhiB5 lost almost all its secondary structure, whereas in WhiB6 some secondary structure was maintained even at 95 °C. Neither WhiB2 nor WhiB7 showed thermal denaturation. Together, the data suggest that considerable structural differences exist among the WhiB proteins of M. tuberculosis, with WhiB5 and WhiB6 appearing to be the most structurally divergent family members.
CD spectroscopy is considered more sensitive for the recognition of α helices and less reliable for β strands, thus the data obtained from CD spectroscopy are an approximate assessment. To estimate the level of β-structures in different WhiB proteins, a thioflavin T (ThT)-binding assay was performed. ThT shows strong fluorescence in the presence of crossed β-sheet structures [27,28]. The binding of ThT to each apo WhiB protein was measured by fluorescence spectroscopy. The fluorescence intensities of each protein varied with respect to each other (Fig. 8). In WhiB5 and WhiB6 it was several fold higher than in the others, whereas for the rest it was within a similar range (Fig. 8). From the data, it appears that β-sheet structures are a major contributors in WhiB5 and WhiB6, but the proportion is relatively low and similar in other WhiB proteins. It should be noted that WhiB5 aligned only with WhiB3 and WhiB4, whereas the WhiB6 sequence aligned only with WhiB4 (NCBI; http://www.ncbi.nlm.nih.gov/blast/bl2seq/wblast2.cgi) (Table S1), suggesting that at the amino acid sequence level both WhiB5 and WhiB6 differ from the others, and the difference is clearly reflected in their secondary structure. The other important point is that although the WhiB proteins were predicted to attain a α-helical structure , WhiB5 and WhiB6 appeared to have significant amounts of β strand. In WhiB5 and WhiB6 the reduction of intramolecular disulfide bonds resulted in a slight decrease in β-sheet structure (decreased fluorescence intensity), whereas this was reversed in the other proteins.
Structural diversification at the tertiary structure level among different WhiB proteins
Both CD spectroscopy and the ThT-binding assay provide information only about the secondary structure. Therefore, to investigate differences at a tertiary level, the surface topology of each of the WhiB proteins was studied. The presence of total exposed hydrophobic patches was probed using 8-anilinonapthalene-1-sulfonate (ANS) binding by fluorimetry. The ANS fluorescence of various WhiB proteins was significantly different (Table 4). The fluorescence intensity of WhiB3 was maximal and threefold higher than that of WhiB6 (minimum). Based on ANS fluorescence, the surface hydrophobicity of various WhiB proteins decreases in the following order: WhiB3 > WhiB1 > WhiB4 > WhiB2 > WhiB5 > WhiB7 > WhiB6 (Table 4). WhiB4 showed a significant increase in ANS fluorescence after reduction of the disulfide bonds, whereas WhiB1, WhiB2, WhiB5 and WhiB7 showed a low but reproducible increase in ANS fluorescence after reduction of the disulfide bonds, indicating a minor conformational change followed by the exposure of certain buried hydrophobic patches.
Table 4. ANS binding to WhiB proteins. Surface hydrophobicities were calculated as described in Experimental procedures. The values obtained in blanks (without protein) were subtracted form the sample readings. Data are expressed as means ± SD (three independent protein preparations).
Surface hydrophobicity (AU)
5.68 ± 0.33
6.40 ± 0.38
4.85 ± 0.24
5.62 ± 0.21
7.44 ± 0.47
7.71 ± 0.46
4.44 ± 0.22
6.00 ± 0.42
4.21 ± 0.21
4.61 ± 0.70
2.30 ± 0.18
2.38 ± 0.21
3.25 ± 0.26
3.58 ± 0.61
CD and fluorescence data clearly suggested structural variations among seven WhiB proteins of M. tuberculosis. Furthermore, we observed that in western blot analysis polyclonal antibodies raised against WhiB1 did not cross-react with WhiB4 and vice versa (data not shown). This may be because of variations in the antigenic epitopes of WhiB1 and WhiB4, which is likely to be due to conformational differences between the two proteins. The relative cross-reactivity of polyclonal antibodies was used to probe the conformational variations at the tertiary level among WhiB proteins. Polyclonal antibodies against all WhiB proteins were raised in rabbit and their cross-reactivity was tested by ELISA. Dilutions of primary antibodies which gave A450 = 0.5 ± 0.05 were selected for the cross-reactivity assay and were as follows: 1 : 35 000 for WhiB1; 1 : 25 000 for WhiB2; 1 : 20 000 for WhiB3, WhiB4 and WhiB5; 1 : 15 000 for WhiB6 and 1 : 50 000 for WhiB7. The activity of the positive control (antibody raised against that particular antigen) was taken to be 100% and cross-reactivity with other WhiB proteins was represented relative to the positive control. Figure 9 clearly shows considerable differences in antibody cross-reactivity. The differential cross-reactivity suggested that the antigenic sites (mostly conformational epitopes) on different WhiB proteins are heterogeneous. We assume that the conformational difference among various WhiB proteins is the possible cause of this heterogeneity. Because WhiB proteins share significant sequence homology minor cross-reactivity may be due to the presence of antibodies which are reacting to the linear epitopes or to conserved conformational epitopes. The data showed that there are significant structural differences among WhiB proteins of M. tuberculosis.
The purified recombinant WhiB proteins had spectral resemblance to proteins coordinating the [2Fe–2S] cluster. Earlier studies also showed that recombinant WhiB3  and WhiB4  of M. tuberculosis, purified under normal conditions, coordinate a [2Fe–2S] cluster. However, based upon the in vitro reconstitution assay, both were found to be [4Fe–4S] cluster coordinating proteins. Iron–sulfur clusters are one of the most ancient and versatile cofactors of several important class of proteins and have been implicated in variety of functions [29,30]. Iron and sulfur can be combined in several ways to produce different cluster types, e.g. [2Fe–2S], [3Fe–4S], [4Fe–4S] and more complex structures . The susceptibility of iron–sulfur clusters to oxidation makes them a good sensor of redox conditions within a cell [32,33]. It has been commonly observed that [4Fe–4S] clusters are highly oxidation labile and rapidly transform into the relatively stable [2Fe–2S] cluster [34,35], therefore, it is possible that during purification under normal conditions (as strict anaerobic conditions could not be maintained), the [4Fe–4S] cluster undergoes oxidation and forms the [2Fe–2S] cluster which is relatively stable. However, in vivo, the possibility of the presence of [4Fe–4S] in WhiB proteins remains very high.
Protein ligands for the canonical clusters are typically sulfide ions of cysteines. The involvement of four conserved cysteines in the coordination of the [4Fe–4S] cluster has been confirmed in WhiB3  and WhiB4 . The brown color of the purified protein and the iron–sulfur cluster-specific peaks also disappeared when proteins were treated with IAA (Figs 1 and S2), supporting the notion that the cysteines are the probable iron–sulfur cluster coordination sites in WhiB proteins of M. tuberculosis. Because WhiB6 has five cysteine residues, a detailed investigation is needed to determine the exact coordination sites. Although the arrangement of cysteines in the primary structure is not identical, the ‘C-X2-C-X5-C’ arrangement (except for WhiB5 where it is ‘C-X3-C-X7-C’) is conserved and the position of the N-terminal cysteine appears to be flexible. Therefore, in the primary structure, although the cysteines lie far apart with respect to each other, in a 3D form they are likely to come closer to hold the iron–sulfur cluster. The formation of two intramolecular disulfide bonds upon cluster removal strengthened the argument that in the 3D form, the cysteines are in close proximity.
Conditional expression of whiB genes in response to environmental stress is well documented [8,36,37]. The differential expression of whiB genes under different growth phases  indicated that the WhiB proteins of M. tuberculosis are segregated temporally. The differential response showed by whiB genes under different stress conditions  and our observation that all seven WhiB proteins responded differently to oxidizing environments in vitro, propose that the WhiB proteins of M. tuberculosis would respond to the redox signal differently.
M. tuberculosis has evolved a remarkable ability to adapt to the changing environment during infection. In the process, it has evolved a highly efficient network of specific gene products which assures survival and multiplication under unfavorable nutritional, pH and redox conditions . Among several mechanisms for resistance to intracellular killing, the scavenging of free radicals and the reactivation of degenerated proteins during infection by a family of proteins named ‘thioredoxins’ is one of the elegant mechanisms of defense and self-sustenance seen in several organisms [38–40]. Oxidized thioredoxin (Trx) with a disulfide at its active site (CXXC) is reduced by NADPH and thioredoxin reductase (TrxR), and then functions as a general protein disulfide reductase [41,42].
Antioxidant defense in M. tuberculosis always appeared unusual in many aspects. Similar to other actinomycetes, M. tuberculosis lacks a glutathione-dependent detoxification system. Instead, it contains mycothiol  and a mycothiol reductase . In the absence of functional oxyR, fnr, etc. [10,45] it appears that M. tuberculosis has evolved several accessory signaling molecules and a protective network which remain unexplored. The presence of the WhiB family in the form of disulfide reductase is likely to be an important components of this unexplored system.
Although insulin is a non-natural substrate, the difference in activity indicates that the efficiency of reduction of target disulfides by WhiB proteins may vary in vivo. Moreover, the involvement of a particular WhiB protein in vivo would depend upon their structural features and redox potential. The redox potential of disulfide oxidoreductases has been shown to be largely determined by the amino acids present in and around the active site (CXXC motif) [46,47]. WhiB proteins of M. tuberculosis did not show any conservation in the dipeptide present between the two cysteines of CXXC motif (Fig. S6). Therefore, we argue that each protein is likely to have a different redox potential. The crystal structures of several thioredoxins have shown that the active site is located at the N-terminus of α-2 helix of thioredoxin fold and is separated by a kink from rest of the helix due to presence of a proline [25,48–50]. The kink provides proper positioning of the active site to facilitate electron flow during the thiol–disulfide exchange reactions . Interestingly, in all WhiB proteins except WhiB2, the amino acid immediately after the C-terminal cysteine of the CXXC motif is proline (Fig. S6). In WhiB2 the absence of proline and therefore of the kink might not allow the protein to attain an optimal geometry, even though there is a CXXC motif, it lacked reductase activity. Gel-filtration chromatography suggested that apo WhiB2 exists as a homodimer (data not shown) indicating that the overall surface topology is well constructed and the non-covalent interactions are properly established. However, a lack of reductase activity due to the absence of a native-like structure could not be ruled out. Interestingly, in WhiB5 despite the insertion of an extra amino acid between the two cysteines of the CXXC motif, the protein retained disulfide reductase activity. It is possible that during evolution, or due to some mutational event, an extra amino acid was inserted, however WhiB5 retained disulfide reductase activity due to its functional importance (yet to be discovered) in M. tuberculosis.
E. coli Trx has been implicated in at least 26 distinct cellular processes . The cell division protein FtsZ, rod-determining protein MreB, enzymes of fatty acid metabolism, etc. have been shown to interact with Trx [52,53]. Thioredoxin-mediated modulation in the activity of transcription factors NF-κB and AP-1 is well-documented [54,55]. Therefore, a change in gene expression after the deletion of whiB7  could be seen as the absence of a component which possibly exerts its function via a protein–protein interaction rather than a DNA–protein interaction. To date, there has not been a single report to show that the WhiB proteins work as a DNA-binding protein. It is possible that they act as an essential structural component of a multiprotein complex and/or are directly or indirectly involved in mediating the assembly of transcription factors that regulate the expression of downstream gene(s).
Iron–sulfur clusters are known to modulate the structural and functional properties of several proteins [29,56]. Because the cysteine-thiols would be ligated to iron–sulfur cluster in holo WhiB proteins, they are not free for electron flow and disulfide exchange. Therefore, there is a high possibility that removal of the cluster is essential for the reductase activity of WhiB proteins. Indeed, the iron–sulfur cluster of M. tuberculosis WhiB4 has a regulatory role . Removal of the iron–sulfur cluster under oxidizing conditions is associated with conformational change followed by a gain of disulfide reductase activity.
A fundamental question is that if one thioredoxin-related protein ‘can do it all’, why was evolution directed to choose multiple Trx-like proteins. We do not have clear answer to this question, but in general, multigene families are thought to have evolved to address questions of substrate specificity. In a scenario in which there is a difference in structural organization among WhiB proteins and because different whiB genes are involved in different cellular processes, the existence of target/substrate specificity appears conceivable. Identification of the cellular targets of WhiB proteins is our next objective. From the data described here, we assume that each of the WhiB proteins in M. tuberculosis, although a part of single family, is an entity in itself. Furthermore, because of the seven proteins, six are disulfide reductase, we propose that WhiB proteins represent a novel redox system in M. tuberculosis.
Bacterial strains and reagents
Genomic DNA from M. tuberculosis H37Rv was prepared as described previously . E. coli DH5α was used for general cloning procedures, whereas expression of recombinant proteins was carried out in BL21(DE3). Plasmid pET-29a and E. coli BL21(DE3) were purchased from Novagen (Darmstadt, Germany). T4 DNA ligase was obtained from Promega (Madison, WI, USA) and restriction/DNA-modifying enzymes were from New England Biolabs (Ipswich, MA, USA). Oligonucleotide primers were supplied by Biobasic Inc. (Markham, Canada). Ni2+-NTA agarose and other PCR/plasmid purification kits were from Qiagen (Hilden, Germany). The EDTA-free protease inhibitor cocktail was from Roche (Mannheim, Germany). All analytical grade chemicals were either from Sigma-Aldrich (Bangalore, India) or as indicated. Throughout the study, all buffers were thoroughly degassed and purged with nitrogen just before use. When required, protein samples were purged with nitrogen before storage or transfer. Standard recombinant DNA techniques were followed as described elsewhere .
Protein overexpression and purification
Gene-specific primers were used to amplify the complete ORFs (without stop codon) of whiB2/Rv3260c, whiB5/Rv0022c, whiB6/Rv3862c, whiB7/Rv3197A and iscS/Rv3025c using genomic DNA of M. tuberculosis H37Rv as a template. The nucleotide sequence of primers is listed in Table S2. The authenticity of PCR products was determined by sequencing both DNA strands in an ABI Prism automated DNA Sequencer 310. PCR-amplified ORFs were cloned at EcoRI/XhoI (whiB2 and whiB6), EcoRI/SalI (whiB5 and whiB7) or SacI/HindIII (iscS) sites of pET-29a to get pET-O (O defines the respective ORFs) under isopropyl thio-β-d-galactoside-inducible promoter. The pET-O transformed E. coli BL21 (DE3) cells were grown at 37 °C in Luria–Bertani broth containing 30 μg·mL−1 kanamycin until D600 was ∼ 0.5–0.6 and expression was induced by the addition of 0.3 mm isopropyl thio-β-d-galactoside for 18–20 h at 16 °C (for WhiB proteins) or for 3 h at 30 °C (for IscS). Cells from 1000 mL culture were re-suspended in 10 mL buffer A (50 mm NaH2PO4, 300 mm NaCl, 30 mm imidazole, protease inhibitor cocktail, pH 8.0). After one freeze–thaw cycle, cells were lyzed by sonication. The proteins were purified from the soluble fraction using Ni2+-NTA affinity column as described by the manufacturer. WhiB1, WhiB3 and WhiB4 were expressed and purified using previously made constructs as described previously [13–15]. Proteins were purified under denaturing conditions as described by the manufacturers (Qiagen), whereas for refolding the ‘matrix-assisted in-column refolding’ method was used as described previously . Purified proteins were either used directly for spectroscopic/biochemical analysis or were dialyzed as required. Throughout the study, unless otherwise mentioned, proteins purified from the soluble fraction of the lysate were used.
Effect of EDTA on iron chelation
Freshly purified proteins (50 μm) were incubated with 20 mm EDTA at 25 °C. Samples were aliquoted at different time intervals, and the extent of chelation by EDTA was studied by measuring absorption at 420 nm. In parallel, samples without EDTA (as a control) were also incubated similarly and A420 was measured. To exclude the effect of air oxidation, readings from control samples were substracted from the test samples and the values thus obtained are represented. The initial reading was set to 100% and the change in A420 (residual) is expressed relative to the reading at t = 0. Suitable baseline corrections were made before recording the absorbance.
In vitro assembly of iron–sulfur cluster by IscS using radiolabeled cysteine
The purified proteins were dialyzed against buffer B (50 mm Tris/HCl, pH 9.0, 150 mm NaCl, 10 mm dithiothreitol). Purified IscS from M. tuberculosis was used as a cysteine desulfurase. The reaction (500 μL) was set in buffer B and contained 30 μm WhiB protein, 2 μm IscS, 300 μm FeCl3, 300 μm l-cysteine and 15 μCi 35S-labeled l-cysteine (BARC, Mumbai, India). The reaction was carried out at 10 °C for 16–18 h in darkness. After incubation, the samples were extensively dialyzed against buffer B and absorption spectra were recorded. In order to study the incorporation of 35S into the iron–sulfur cluster and therefore in the WhiB proteins, 30 μL samples were spotted onto a poly(vinylidene difluoride) membrane, air dried and visualized by phosphorimaging.
Purified apo WhiB proteins (10 μm) were reduced with 1 mm dithiothreitol for 1 h at 25 °C in buffer C (50 mm Tris/HCl, pH 8.0, 200 mm NaCl). Free thiols were alkylated using 20 mm IAA for 1 h at 37 °C in darkness. The protein without dithiothreitol treatment was alkylated with IAA as a control. Protein samples were precipitated with 10% trichloroacetic acid for 15 min on ice, washed twice with chilled acetone, air dried and re-suspended in 0.1% trifluoroacetic acid. Protein (2 μL) was mixed with an equal volume of the matrix (sinnapinic acid) and the mass was measured in MALDI-TOF-MS, Voyager 4402 (Applied Biosystems, Foster City, CA, USA).
Far-UV CD spectra of apo proteins were recorded in a JASCO J810 CD spectropolarimeter (Tokyo, Japan) at a protein concentration of 0.2 mg·mL−1 in buffer C at 25 °C, with a scan speed of 20 nm·min−1 and 2 nm bandwidth. Reduction was carried out by incubating the protein with 1 mm dithiothreitol for 1 h at 25 °C in the same buffer. Thermal denaturation of the oxidized and reduced apo proteins (0.2 mg·mL−1) in buffer C were monitored by measuring the change in ellipticity at 222 nm with increasing temperature from 25 to 95 °C at a speed of 2 °C·min−1. Data were recorded at every 0.5 °C, with an 8 s averaging time and at a 1.5 nm bandwidth. For all spectral measurements, a path length of 1 mm was used. Ten spectra were averaged for each sample.
All fluorescence measurements were recorded at 25 °C in a Perkin–Elmer Luminescence Spectrometer LS50B (Waltham, MA, USA). To measure the surface hydrophobicities, 10 μm of oxidized apo proteins in buffer C was mixed with ANS (250 μm) and incubated for 2 min at 25 °C. Thereafter, the samples were excited at 388 nm and fluorescence emission spectra were recorded in the range 400–600 nm. The surface hydrophobicity (SH) was calculated using a formula:
To study the binding of ThT to different WhiB proteins, 50 μm ThT was added to 5 μm apo proteins in buffer C. The excitation was set at 430 nm and the emission was recorded between 420 and 600 nm. All spectra were recorded at a scan speed of 50 nm·min−1. The excitation and emission slit width was 3 nm for ANS fluorescence, whereas for ThT binding it was 5 and 10 nm respectively. Ten spectra were averaged for each sample. Throughout the study, the readings of solvent blanks were subtracted from the sample spectra before plotting the graphs.
Insulin disulfide reduction assay
WhiB proteins catalyzed reduction of insulin disulfides by dithiothreitol was analyzed as described previously , with some modifications. Three different concentrations 1, 2 and 3 μm of apo proteins were pre-incubated for 1 h in reaction buffer D (0.1 m phosphate buffer pH 7.5, 200 mm NaCl and 1 mm dithiothreitol). The reaction was started by the addition of insulin to a final concentration of 0.13 mm. Precipitation of the reduced insulin chains was monitored at 650 nm in a Perkin–Elmer (λ35) spectrophotometer, every 0.1 min until A650 reached to 0.8 or for 50 min.
Immunization of rabbit, collection of sera and ELISA
New Zealand White rabbits were immunized with 250 μg of purified WhiB proteins. Before immunization 5 mL blood was collected for pre-immune sera. Equal volumes (500 μL) of protein and incomplete Freund’s adjuvants were mixed and an emulsion was prepared. The rabbits were immunized subcutaneously. After 21 days, the first booster was given and blood was collected after 5 days. The serum was isolated and the titer was determined by ELISA. Use of animals for the generation of polyclonal antibodies was carried out with prior approval from the Institutional Animal Ethics Committee (institutional registration number 55/1999/CPCSEA, dated 11 November 1999). Animal handling and experimental design was performed in accordance with the approved guidelines.
Each well of a flat-bottom microtitre ELISA plate was coated with 0.5 μg of purified WhiB protein (100μl·well−1) in 0.05 m bicarbonate buffer, pH 9.6, and incubated for overnight at 4 °C. After blocking, different dilutions of test sera (1 : 5000 to 1 : 80 000) prepared in NaCl/Pi containing 0.1% skimmed milk were added (100 μL) to each well and incubated at 37 °C for 3 h. Goat anti-rabbit IgG (100 μL) conjugated to horseradish peroxidase were diluted 1 : 5000 in NaC/Pi containing 0.1% skimmed milk, added to wells and incubated at 37 °C for 1 h. For the activity assay, 100 μL of peroxidase substrate, 3,3′,5,5′-tetramethylbenzidine (Bangalore Genei, Bangalore, India) (diluted 1 : 20 in water), was added to the wells and incubated at 37 °C for 30 min. The reaction was stopped by adding 25 μL of 1 m H2SO4 and A450 was measured in an ELISA reader (PowerWave XS Bio-Tek Instruments, Inc, Winooski, VT, USA). The antibody dilution which gave A450 = 0.5 ± 0.05 was used for the cross-reactivity assay.
UV–Vis absorption spectra of WhiB proteins (50 μm) purified under native, denaturing and in-column refolding conditions were recorded in the range 200–700 nm, immediately after purification using a Perkin-Elmer (λ35) spectrophotometer at 25 °C where the elution buffer was used for the baseline correction. The total amount of non-heme iron was estimated by the o-bathophenanthroline disulfonate method [59,60] as described elsewhere .
To remove the iron–sulfur cluster, purified proteins were incubated with EDTA and potassium ferricyanide at a molar ratio of 1 : 50 : 20 (protein : EDTA : ferricyanide) at 25 °C for 30 min. The chelated iron was removed by dialysis against buffer D and was used as apo protein for spectral studies and activity assays.
The molar extinction coefficient of various WhiB proteins was determined using vector-nti software. The values were as follows: 17 550 m−1·cm−1 for WhiB1, 13 140 m−1·cm−1 for WhiB2, 18 830 m−1·cm−1 for WhiB3 and WhiB4, 16 980 m−1·cm−1 for WhiB5, 27 200 m−1·cm−1 for WhiB6 and 17 550 m−1·cm−1 for WhiB7. Throughout, the study protein concentrations were calculated using A280 reading.
MSA and SKG are thankful to the ‘Council of Scientific and Industrial Research’ (CSIR), Govt of India, for ‘Senior Research Fellowship’. Financial assistance by CSIR and ‘Department of Biotechnology’ (DBT), Govt of India, is duly acknowledged. This report is IMTECH communication no. 08/2008. MSA, SKG and PA designed the experiments. SKG carried out some of the work on WhiB1. MSA carried out the rest of the work. MSA and PA wrote the manuscript.