A single-nucleotide mutation in the −10 promoter region inactivates the narK2X promoter in Mycobacterium bovis and Mycobacterium bovis BCG and has an application in diagnosis


  • Editor: Roger Buxton

Correspondence: Jaya Sivaswami Tyagi, Department of Biotechnology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India. Tel.: +91 11 2658 8491; fax: +91 11 2658 8663; e-mail: jstyagi@aiims.ac.in


Nitrate reduction is believed to be vital for the survival of tubercle bacteria under hypoxic/anaerobic conditions that are thought to prevail within granulomas. Nitrate reductase activity is rapidly induced in Mycobacterium tuberculosis (M. tb) under hypoxic conditions and is attributed to the induced expression of the nitrate/nitrite transporter gene, narK2. By contrast, Mycobacterium bovis (M. bovis) and M. bovis BCG (BCG) do not support the hypoxic induction of either nitrate reductase activity or narK2. Here, we show that the induction defect in the narK2X operon in M. bovis and BCG is caused by a −6T/C single nucleotide polymorphism (SNP) in the −10 promoter element essential for narK2X promoter activity. Complementation of M. bovis with both narGHJI and narK2X genes from M. tb failed to restore nitrate reductase activity in M. bovis, suggesting the involvement of additional genes/regulatory mechanisms for nitrate reduction that are absent in M. bovis. The −6T/C promoter-linked SNP enabled clear differentiation of M. tb from the other members of the M. tb complex, including M. bovis, BCG, Mycobacterium africanum and Mycobacterium microti, through a PCR-RFLP assay.


Tuberculosis in humans is chiefly caused by Mycobacterium tuberculosis (M. tb). However, Mycobacterium bovis (M. bovis), the major tuberculosis pathogen in cattle, also causes disease in humans and is usually implicated in extrapulmonary tuberculosis (Wilkins et al., 1986). Other members of the M. tb complex (MTC), such as M. bovis BCG (BCG), Mycobacterium africanum and Mycobacterium microti, rarely cause disease in immunocompromised populations (Metchock et al., 1999; Niemann et al., 2000). Zoonotic transmission of these organisms to humans, especially of M. bovis from cattle and unpasteurized milk, is an important health concern (O'Reilly & Daborn, 1995; Shah et al., 2006). Because M. bovis is naturally resistant to pyrazinamide (Scorpio & Zhang, 1996), a first-line antituberculosis drug, therefore, differentiation of M. tb infection from M. bovis infection is of paramount importance for administering the appropriate treatment.

A classical assay that differentiates M. tb from M. bovis is its high aerobic nitrate reductase activity (Virtanen, 1960). Furthermore, the nitrate reductase activity of M. tb, but not M. bovis, increases drastically upon entry into the anaerobic dormant state (Virtanen, 1960; Wayne & Doubek, 1965; Weber et al., 2000). It is thought that M. tb might survive in low-oxygen microenvironments (granulomas) by reducing nitrate to nitrite, using nitrate as a terminal electron acceptor in respiration (Wayne & Hayes, 1998; Wayne & Sohaskey, 2001). Nitrate reduction was shown to be mediated by narGHJI-encoded nitrate reductase in M. tb, but the enhanced reduction of nitrate during hypoxia was attributed to upregulation of NarK2, a putative nitrate/nitrite transporter (Sohaskey & Wayne, 2003). The inability of M. bovis and BCG to efficiently reduce nitrate under both aerobic and hypoxic conditions was ascribed to inactive narGHJI and narK2X gene/gene products (Stermann et al., 2004; Honaker et al., 2008; Sohaskey & Modesti, 2009). Single nucleotide polymorphisms (SNPs) were detected in the narGHJI promoter region (−215T/C), although it was not ruled out that other SNPs within the narGHJI operon itself could also contribute to this difference in activity (Garnier et al., 2003; Stermann et al., 2004).

The response regulator DevR controls the transcription of narK2X in M. tb by binding to multiple Dev boxes (Chauhan & Tyagi, 2008a). A recent study showed that two DevR regulon genes, narK2 and narX, are inactive in M. bovis and BCG, compared with M. tb, wherein they are strongly induced (Honaker et al., 2008). In the absence of differences in their coding regions, the lack of hypoxic induction of narK2X in M. bovis and BCG was hypothesized to be caused by an SNP in the narK2 upstream region (Honaker et al., 2008) that was reported to map at −17 from the narK2 transcription start point (TSP) (Hutter & Dick, 2000). We recently reported that this SNP is located at −6 position (T→C, −6T/C) with respect to the narK2 TSP of M. tb (Chauhan & Tyagi, 2008a; Fig. 1). A conflicting report described inducible narK2 promoter activity in BCG harbouring a T→C mutation at a different position, −16 (Hutter & Dick, 2000). Thus, while a −6T/C SNP was linked to a lack of hypoxic narK2X induction (Honaker et al., 2008), a −5T/C SNP was associated with inducible promoter activity in BCG (Hutter & Dick, 2000). As both these mutations map in the −10 promoter element, we analysed the effect of these and other mutations on promoter activity. Here, we show that the −6T/C SNP is responsible for the inactivation of the narK2X promoter and hence of the narK2X operon in M. bovis. We also show that the −5T/C SNP significantly reduces, but does not abolish, inducible narK2X promoter activity. Lastly, the −6T/C promoter SNP is useful to differentiate M. tb from M. bovis, BCG and other members of the MTC by a new PCR-RFLP assay.

Figure 1.

 Schematic representation and nucleotide sequence of the Mycobacterium tuberculosis narK2X promoter. The TSP, −10/−35 promoter elements (in boxes) and Dev box D1 (grey shaded box), described here or previously (Chauhan & Tyagi, 2008a), are shown in line 1. NarK2R1 and NarK2F denote primers utilized in the PCR-RFLP assay. The position of the −6T→C SNP in Mycobacterium bovis and other MTC species (Honaker et al., 2008) is shown in bold and the resultant NheI site (GCTAGC) is underlined (line 2). The −5T→C SNP in BCG according to Hutter and Dick (2000) is indicated in line 3. GFP fluorescence of M. tb hypoxic cultures harbouring reporter plasmids bearing narK2X WT or −10 mutant promoter elements (pnarK2 series) is shown. RFU/OD, obtained after subtracting RFU/OD values of empty vector control cultures, is expressed as mean±SD of two experiments, each performed in quadruplicate.

Materials and methods

Bacterial strains, plasmids and culture conditions

Mycobacterium tuberculosis (H37Rv), M. bovis (AN5) and BCG (vaccine strain, Chennai, India) were cultured in Dubos medium containing 0.05% Tween-80 and OADC (oleic–albumin–dextrose–catalase, Difco, France) under shaking conditions (220 r.p.m.) or hypoxic conditions as described (Chauhan & Tyagi, 2008b). Escherichia coli strains were cultured as described previously (Bagchi et al., 2005). When required, kanamycin was used at a concentration of 25 μg mL−1, hygromycin at 50 μg mL−1 and gentamycin at 12.5 μg mL−1 during mycobacterial culture. The plasmids and primers used in this study are listed in Tables 1 and 2. The presence of the −6T/C SNP in M. bovis AN5 and BCG (vaccine strain, Chennai, India) in the narK2X promoter was confirmed by DNA sequencing (not shown).

Table 1.   Plasmids used in this study
  1. All promoter coordinates mentioned in parenthesis are with respect to Mycobacterium tuberculosis narK2X TSP.

pFPV27E. coli–mycobacterial shuttle plasmid with promoterless gfp, KmrValdivia et al. (1996)
pnarK2pFPV27 containing the M. tb narK2 promoter (−220 to +57), hygrChauhan & Tyagi (2008a)
pnarK2 −4A→CpnarK2 mutated at the −4 position, A→C, hygrThis study
pnarK2 −5T→CpnarK2 mutated at the −5 position, T→C, hygrThis study
pnarK2 −6T→CpnarK2 mutated at the −6 position, T→C, hygrThis study
pnarK2 −7G→CpnarK2 mutated at the −7 position, G→C, hygrThis study
pnarK2 −8G→TpnarK2 mutated at the −8 position, G→T, hygrThis study
pnark2ΔuppnarK2 deleted of −122 to −220 bp upstream sequences, hygrChauhan & Tyagi (2008a)
pnark2ΔdownpnarK2 deleted of +14 to +57 bp downstream sequences, hygrChauhan & Tyagi (2008a)
pNarG-GM1Insertional vector pUC–INT–GM carrying M. tb narGHJI with its native promoterSohaskey & Modesti (2009)
pNarK2XPlasmid pFPV27 carrying the M. tb narK2X operon with a 280-bp upstream promoterThis study
Table 2.   Primers used in this study
PrimerSequence (5′→3′)Application
  1. Mutated nucleotides are underlined.

NarK2-4acFGACAGCCATGCGCGAACCCTGCTTSite-directed mutagenesis
NarK2-8gtFGCCATGCGCTAACACTGGCCTTCGASite-directed mutagenesis

Construction of reporter plasmids and green fluorescent protein (GFP) reporter assays

The GFP reporter vector pnarK2, carrying the M. tb narK2 promoter (Chauhan & Tyagi, 2008a), was used to generate various mutants in the putative −10 promoter region by either the site-directed mutagenesis method or the mega primer mutagenesis method as described (Sambrook & Russell, 2001; Chauhan & Tyagi, 2008b). Briefly, PCR was performed with mutated primers using wild-type (WT) pnarK2 plasmid as a template and Pfu Turbo DNA polymerase (Stratagene). The amplified PCR product was digested with DpnI enzyme for 1 h and a 10-μL aliquot of this reaction was transformed in E. coli. All the mutations were confirmed by DNA sequencing. The various plasmids were electroporated into M. tb H37Rv and GFP reporter assays were performed as described (Chauhan & Tyagi, 2008b). Briefly, stock cultures of M. tb were aerobically subcultured twice to the midlogarithmic phase (A595 nm∼0.3) and then diluted to an A595 nm of 0.05. Hypoxic cultures (standing) were established by dispensing 200 μL culture aliquots into 96-well black, clear-bottom microtitre plates and incubating the plates at 37 °C. The aerobic promoter activity was measured in cultures that were simultaneously grown in 50-mL tubes (5 mL of culture). Culture aliquots of 200 μL were sampled at 48 h and the GFP fluorescence was measured in a spectrofluorimeter (Molecular Devices, Sunnyvale, CA) with an excitation wavelength of 483 nm and an emission wavelength of 515 nm.

PCR-restriction fragment length polymorphism (RFLP)

The 178-bp narK2 promoter region was amplified by PCR using NarK2R1 and NarK2F primers (Fig. 1, Table 2) and genomic DNA of the various standard or clinical strains. The PCR conditions were a 10-min initial denaturation phase at 94 °C, followed by 40 cycles of 30 s at 94 °C, 30 s at 60 °C and 30 s at 72 °C and, finally, 7 min at 72 °C. A 10-μL aliquot of the PCR product was digested with NheI for 90 min, electrophoresed on a 6% nondenaturing polyacrylamide gel and visualized using ethidium bromide.

Construction of narGHJI and narK2X complementation strains of M. bovis

Mycobacterium bovis AN5 was complemented with the integrating plasmid pNarG-GM1 expressing the M. tb narGHJI operon (Sohaskey & Modesti, 2009) or the pNarK2X plasmid expressing the M. tb narK2X operon, (see Table 1) or both pNarG-GM1 and pNarK2X. To construct pNarK2X, the region encompassing the coding regions of narK2 and narX along with a 280-bp upstream promoter was amplified by PCR using Fusion DNA polymerase (NEB, UK) and M. tb H37Rv DNA and cloned in the EcoR1 and HindIII sites of pFPV27 mycobacterial shuttle vector. The resultant plasmid was electroporated into M. bovis or M. bovis-harbouring pNarG-GM1.

Nitrate reductase assay

Nitrate reductase assay was performed with aerobic shaking and 48-h standing cultures (hypoxic). Briefly, the cultures were grown aerobically as described above in the presence of 5 mM nitrate and standing cultures (starting OD595 nm, 0.05) were maintained for 48 h in 96-well microtitre plates as described previously (Chauhan & Tyagi, 2008b). The nitrite concentration was determined using the Griess reaction as described (Wayne & Doubek, 1965). Briefly, 50 μL of sulfanilamide was added to 50 μL of cultures (both aerobic and standing) and incubated at room temperature for 5–10 min. Next, 50 μL of N-1-napthylethylenediamine dihydrochloride was added and the A595 nm was measured in a plate reader (Biorad).

Results and discussion

A single thymine to cytosine mutation renders the narK2X operon promoter inactive in M. bovis/BCG

To test the hypothesis that the lack of hypoxic induction of narK2 and narX in M. bovis/BCG is because of a −6T/C SNP in the narK2X promoter region, we mutated the M. tb narK2 promoter by changing thymine at the −6 position to cytosine (−6T→C) in the narK2 promoter plasmid, pnarK2, to mimic the observed mutation at this site in M. bovis/BCG. The effect of this mutation on promoter activity was assessed in M. tb H37Rv under hypoxic conditions using the GFP reporter assay. The −6T→C mutation completely abolished the hypoxic induction of pnarK2 (Fig. 1), proving that the −6T nucleotide is absolutely essential for narK2 promoter activity and that its mutation to the C nucleotide rendered the M. bovis/BCG narK2X promoter inactive. To confirm that no other trans-acting factor (e.g. repressor) contributed to the loss of promoter activity in M. bovis/BCG, the pnarK2 plasmid (harbouring the M. tb WT narK2 promoter) was introduced into M. bovis and BCG strains and the GFP reporter assay was performed under hypoxic conditions. The M. tb WT narK2X promoter was well induced and to the same level in M. bovis and BCG as in M. tb (Table 3), which suggests that the −6T→C mutation, and not a trans-acting negative regulator, is responsible for the absence of narK2X promoter activity in M. bovis and BCG. To further validate that the M. tb narK2X promoter behaves similarly in M. bovis/BCG and M. tb, two additional truncated narK2X promoter GFP reporter constructs pnarK2Δdown and pnarK2Δup (described in Chauhan & Tyagi, 2008a) were also assessed for GFP fluorescence in M. bovis and BCG under hypoxic conditions. In pnarK2Δdown, the so-called ‘downstream inhibitory region’ is removed (+14 to +57 with respect to M. tb narK2X TSP), whereas in pnarK2Δup, the so-called ‘upstream activating region’ (described by Hutter & Dick, 2000) is deleted (−122 to −220 relative to M. tb narK2X TSP). A similar level of hypoxic induction was observed for both promoter constructs in all three strains (Table 3), demonstrating that the M. tb narK2X promoter has similar activity in M. tb, M. bovis and BCG. These results suggest that all the trans-acting (including DevR) and cis elements that control the narK2X promoter are functionally conserved in M. tb, M. bovis and BCG, except for the −6T/C SNP.

Table 3.   GFP reporter activity of wild type and truncated narK2 promoter constructs in hypoxic cultures of Mycobacterium tuberculosis, Mycobacterium bovis and BCG.
Promoter constructM. tbM. bovisBCG
  1. GFP reporter activity is expressed as mean relative fluorescence units (RFU/OD595 nm) ± SD from two experiments performed in triplicate.

pnarK21226 ± 461322 ± 331109 ± 25
pnark2Δup1523 ± 901600 ± 541400 ± 76
pnark2Δdown2482 ± 352201 ± 822143 ± 49

The −6T/C SNP maps in the SigA −10 promoter element of the narK2X promoter

A putative −10 element was recognized upstream of the experimentally detected TSP of narK2 that was reported previously (Chauhan & Tyagi, 2008a). In the present study, it was functionally characterized by individually mutating additional nucleotides at −4, −5, −7 and −8 positions with respect to the TSP (Fig. 1). Both −4A→C and −5T→C mutations significantly or completely reduced the promoter activity and demonstrated the importance of these nucleotides in promoter function. Taken together, the results of mutation analysis indicate that ‘ATT’ nucleotides present at −4, −5 and −6 positions are essential for promoter activity and likely to be recognized by the transcriptional machinery. Note that the −5T→C mutation was reported to be present in BCG by Hutter and Dick (2000), and the introduction of this mutation significantly reduced inducible promoter activity by ∼7-fold, but did not abolish it (Fig. 1). The −7G→C or −8G→T mutations, which also altered the overlapping putative SigC −10 sequence, did not adversely affect narK2X promoter activity. Although the −10 element of the narK2X promoter showed only a modest resemblance (2/6) to the SigA consensus sequence (Unniraman et al., 2002; Agarwal & Tyagi, 2006), mutational analysis suggests that the ‘GTTAG’ sequence bears greater functional resemblance to the SigA −10 consensus sequence rather than the SigC consensus sequence we had previously suggested (Chauhan & Tyagi, 2008a). These results suggest that the SigA σ factor could be utilized by RNA polymerase for transcribing the narK2X promoter. However, further experimentation is required to confirm the possibility.

Nitrate reductase activity of M. bovis was not rescued by complementation with M. tb NarK2X and NarGHJI

The introduction of M. tb narGHJI or narK2 into M. bovis did not result in an increase in its nitrate reductase activity either under aerobic or hypoxic conditions (Sohaskey & Modesti, 2009). Therefore, it was speculated that the underlying reason for the low nitrate reductase activity in M. bovis could be the absence of functional copies of both narGHJI and narK2 genes (Sohaskey & Modesti, 2009). Hence, we complemented M. bovis with both pNarG-GM1 (integrative vector) and pNarK2X (extrachromosomal vector) carrying narGHJI genes and narK2 along with the downstream gene narX gene, respectively. The nitrate reductase activity of M. tb H37Rv was moderate under aerobic conditions and was induced ∼17-fold under hypoxic conditions as expected (Table 4). However, very low aerobic activity and no hypoxic induction of nitrate reductase activity were observed in M. bovis or strains harbouring either pNarG-GM1 or pNarK2X or both (Table 4). These results suggest the possibility that robust nitrate reduction in M. tb requires the presence of not merely functional narGHJI and narK2X operons but also some unidentified additional mechanism(s) that is defective in M. bovis. This notion is supported by the fact that even aerobic nitrate reductase activity of M. bovis was not equivalent to that in the M. tb level despite complementation with M. tb narGHJI here, or as described previously (Sohaskey & Modesti, 2009).

Table 4.   Production of nitrite in aerobic and hypoxic cultures of Mycobacterium tuberculosis, Mycobacterium bovis and various complemented strains.
StrainsNitrite, μM ± SD
  1. Nitrate reductase activity was measured as nitrite production (in μM) in 48-h shaking (aerobic) or 48-h standing cultures in triplicate. Values are mean ± SD per 2.5 × 106 cells.

M. tb (H37Rv)111 ± 81985 ± 82
M. bovis (AN5)37 ± 743 ± 4
M. bovis+pNarK2X42 ± 337 ± 9
M. bovis+pNarG-GM138 ± 261 ± 9
M. bovis+pNarG-GM1+pNarK2X50 ± 760 ± 8

The −6T/C SNP differentiates M. tb from other members of the MTC in a PCR-RFLP assay

A unique NheI restriction site (GCTAGC) is created in the 280-bp promoter region as a consequence of the −6T/C SNP in the narK2X promoter of M. bovis/BCG (Fig. 1). This SNP was exploited to design a new PCR-RFLP assay aimed at differentiating M. tb from M. bovis/BCG. After amplification of the 178-bp narK2X promoter region and NheI-mediated cleavage of the PCR products, two digestion product bands of 120 and 58 bp were observed with DNA from M. bovis AN5 and BCG (vaccine strain, Chennai, India), whereas an intact band of 178 bp was observed with DNA amplified from M. tb (Fig. 2a and b). To further extend the analysis, 36 clinical isolates including M. tb (10), M. bovis (20), BCG (two), M. microti (two) and M. africanum (two) were tested for this RFLP. Except for the M. tb strains, all other MTC member strains produced a two-band pattern and established that the −6T/C SNP is present in all of them. A representative analysis is shown in Fig. 2c. blast analysis of the sequence (http://www.sanger.ac.uk) confirmed the presence of this SNP in M. microti and M. africanum and its absence in Mycobacterium canetti.

Figure 2.

 PCR-RFLP assay for differentiating Mycobacterium tuberculosis from Mycobacterium bovis, BCG and other members of the MTC. (a) The 178-bp PCR product, amplified using primers NarK2R1 and NarK2F, was digested with the NheI restriction enzyme. Representative electrophoretic PCR-RFLP profiles with (b) standard and (c) clinical isolates of Mycobacterium microti, Mycobacterium africanum, M. bovis, BCG and M. tb are shown. M, 100-bp DNA ladder. Arrows indicate fragment sizes in base pairs.

Two PCR-RFLP methods based on SNPs in gyrB and narGHJI were previously used to differentiate M. tb from MTC members (Niemann et al., 2000; Stermann et al., 2004). Here, we present a newly devised PCR-RFLP method to differentiate M.tb from other members of the MTC on the basis of the −6T/C narK2 promoter SNP. Because none of these PCR-RFLP methods can fully differentiate the MTC members, a combination of different PCR-RFLP assays can improve the confidence of diagnosis.

Previously, the −215T/C SNP upstream of narGHJI was correlated with differential nitrate reduction in MTC (Stermann et al., 2004); M. tb having the −215T genotype could reduce nitrate, whereas other members having the −215C genotype could not reduce nitrate. However, an extension of this analysis to more strains refuted this hypothesis on the basis of high nitrate reductase activity in M. canetti (−215C genotype) and some other ancestral strains of the M. tb−215C genotype (Goh et al., 2005). Interestingly, the nitrate reductase activity of MTC strains correlates better with the −6T/C SNP present in the narK2X promoter. Both M. tb and M. canetti have the −6T genotype and both can reduce nitrate, whereas M. africanum, M. microti, M. bovis and BCG have the −6C genotype and all lack nitrate reductase activity. Because the −6C genotype results in completely abolishing narK2X promoter activity, it is likely that all these species lack functional narK2X genes and this promoter defect is one of the key factors contributing to the lack of inducible nitrate reductase activity in them.


The plasmid pNarG-GM1 was a generous gift provided by Dr C.D. Sohaskey. We sincerely acknowledge Dr H.K. Prasad, India, for providing genomic DNA of clinical isolates of the MTC. This work was financially supported by a grant to J.S.T. from the Department of Biotechnology, Government of India. S.C. is grateful to CSIR for a research associateship. We acknowledge the facilities of the Biotechnology Information Systems (BTIS), Department of Biotechnology, and Government of India.