The wag31 gene of Mycobacterium tuberculosis is positively regulated by the stringent response


  • Editor: Roger Buxton

Correspondence: John L. Dahl, Department of Biology, University of Minnesota Duluth, SSB 207, 1035 Kirby Dr., Duluth, MN 55812, USA. Tel.: +1 218 726 6614; fax: +1 218 726 8142; e-mail:


The stringent response of Mycobacterium tuberculosis is coordinated by Rel and is required for full virulence in animal models. A serological-based approach identified Wag31Mtb as a protein that is upregulated in M. tuberculosis in a rel-dependent manner. This positive regulation was confirmed by analysis of M. tuberculosis mRNA expression. Mycobacterium smegmatis was used to confirm that the expression of wag31Mtb from its native promoter is positively regulated by the stringent response. Furthermore, elevated wag31Mtb expression in M. smegmatis drastically alters the cell-surface hydrophobic properties.


The stringent response is a global regulatory network found in all bacteria, and it allows cells to adapt to amino acid or carbon source deprivation (Cashel et al., 1996). Unlike Escherichia coli, which has two different proteins that can synthesize (p)ppGpp (RelA and SpoT), mycobacteria have only one such protein that is referred to as Rel (Mittenhuber, 2001). The deletion of relMtb causes the inactivation of the stringent response in Mycobacterium tuberculosis, which does not alter bacterial survival inside macrophages (Primm et al., 2000), but does result in a 500-fold reduction in the survival of tubercle bacilli inside a mouse host (Dahl et al., 2003) or inside a guinea-pig host (Klinkenberg et al., 2010). Rel regulates a number of responses critical for pathogenicity in a number of bacteria (Hammer & Swanson, 1999; Singh et al., 2001; Taylor et al., 2002; Haralalka et al., 2003). Rel likely regulates M. tuberculosis-specific genes required for survival within a host. Mycobacterium tuberculosis cells deficient for relMtb have reduced survival when tested under in vitro conditions designed to mimic the interior environment of the granuloma, the presumed site of bacteria during persistent M. tuberculosis infections (Primm et al., 2000). Mycobacterium smegmatis cells with an inactivated stringent response are also unable to survive under prolonged exposure to nutrient deprivation and hypoxia (Dahl et al., 2005). Microarray studies reveal that relMtb is associated with the upregulation of known M. tuberculosis virulence factors and the downregulation of immunodominant M. tuberculosis proteins (Dahl et al., 2003). Numerous genes of unknown function are also differentially regulated by relMtb in M. tuberculosis. Therefore, studying the mycobacterial stringent response may provide insights into the identification of novel M. tuberculosis genes involved in pathogenesis. Our laboratory recently established M. smegmatis as a useful tool for studying rel-dependent M. tuberculosis genes. Using a strain of M. smegmatis inactivated for relMsm (mc2155Δrel), we showed that the regulation patterns of M. tuberculosis genes hspX and eis on multicopy plasmids mimicked the observed Rel-dependent regulation of these genes on chromosomes in M. tuberculosis (Dahl et al., 2005).

Direct correlations do not always exist between cellular transcriptional activity and corresponding protein expression (Anderson & Seilhamer, 1997; Gygi et al., 1999; Skiba et al., 2010). The expression of bacterial virulence factors can occur at the levels of transcriptional regulation, mRNA stability, translational frequency, and protein stability (Dorman & Smith, 2001). We have previously reported a global transcriptional difference between wild-type M. tuberculosis (strain H37Rv) and H37RvΔrelMtb (Dahl et al., 2003), and a goal of the current study is to compare relMtb-dependent differences in protein patterns between strains with and without Rel.

Materials and methods

Bacteria and culture conditions

Mycobacterium tuberculosis strains (H37Rv and H37RvΔrel) have been described previously (Dahl et al., 2003) and were grown in Middlebrook 7H9 medium supplemented with albumin, dextrose and catalase, and 0.2% glycerol+0.05% Tween 80. Cultures were grown to the stationary phase at 37 °C in rolling flasks. Mycobacterium smegmatis strains (mc2155 and mc2155Δrel; described in Dahl et al., 2005) were grown in 7H9 with 0.2% glycerol+0.05% Tween 80 at 37 °C by shaking or on 7H10 agar plates. Hygromycin (50 μL mL−1) was added to M. smegmatis cultures to ensure plasmid stability in strains.

Generation of polyclonal antibodies

To prepare lysates for antibody production, 50-mL aliquot of 3-week-old culture of M. tuberculosis H37Rv were pelleted by centrifugation and washed 3 × in phosphate-buffered saline (PBS) before suspending in 1 mL of lysis buffer [0.3% SDS, 200 mM DTT, 30 mM Tris (pH 7.5)], and breaking cells open with glass beads (0.5 mm diameter) using a FastPrep FP120 bead-beating device (ThermoSavant). Cells were shaken at a speed of 6.5 m s−1 for 45 s and then incubated on ice for 5 min. This cycle was repeated 5 × before samples were boiled for 10 min to enhance cell lysis. Samples were then bead-beaten again five more times, as described above. Lysed samples were centrifuged at 12 000 g for 10 min at 4 °C to remove cellular debris. Supernatants were filter sterilized (0.22 μm) and stored at −20 °C until being mixed with a Titermax Gold adjuvant (Sigma), as recommended by the manufacturers. This mixture was used to subcutaneously immunize white New Zealand female rabbits immunized 3 × over a period of 6 weeks. Before each immunization, marginal ear bleedings were performed to evaluate the reactivity of the antisera against the M. tuberculosis proteins by Western blot analysis. Two weeks after the final immunization, approximately 75 mL of blood was obtained from each rabbit by cardiac terminal bleed. The blood was allowed to coagulate and the sera were separated from the clots. The serum obtained from each rabbit was stored at −80 °C until use in Western blot analysis.

Western blot analysis

Proteins were visualized by Western blot analysis, as described previously (Dahl et al., 2001). Briefly, protein lysates for each strain (50 μg per lane) were separated by 12% SDS-PAGE, transferred to nitrocellulose membranes, incubated with rabbit sera for 5 h at room temperature, washed 3 × with PBS, incubated with a 1 : 2500 dilution of an alkaline phosphatase-labeled anti-rabbit immunoglobulin G antibody (Zymed) overnight at 4 °C, washed 3 × with PBS, and developed using alkaline phosphatase buffer+nitroblue tetrazolium chloride+5-bromo-4-chloro-3′-indolylphosphate p-toluidine salt.

Protein identification

A protein band of about 40 kDa was excised from a 12% polyacrylamide gel stained with Coomassie brilliant blue. The gel band was destained for 2 h in a solution of 50% methanol+5% glacial acetic acid in distilled water. The gel band was dehydrated with acetonitrile, followed by reduction and alkylation with 10 mM DTT+50 mM iodoacetamide in 100 mM NH4HCO3, dehydrated, rehydrated in 100 mM NH4HCO3, dehydrated again, and digested with trypsin (20 ng μL) in ice-cold 50 mM NH4HCO3. The sample was incubated overnight at 37 °C with 20 μL of 50 mM NH4HCO3. After this incubation, the solution containing the digested peptides was desalted and concentrated using C18 Zip-Tips (Millipore). The sample was analyzed by matrix-assisted laser desorption/ionization using the Voyager DE RP system (Applied Biosystems). In order to identify the protein, the Mascot database (Matrix Science) was searched for monoisotopic peptide masses between the ranges 700 and 4000 Da detected in the sample.

Expression of wag31Mtb in M. smegmatis

The wag31Mtb gene, including a 350-bp upstream region, was amplified by PCR from M. tuberculosis genomic DNA using the primers 5′-CTGGTTGCGTTCATCGGTAT-3′ and 5′-GAAAACTGGCGCGTGTCC-3′. The PCR product was cloned into the pDRIVE cloning vector (Qiagen). After digestion with ApaI and PstI, the DNA insert was gel purified and cloned into the mycobacterial shuttle vector pOLYG (Garbe et al., 1994), and the resulting plasmid was named pwag31Mtb.

Reverse transcription (RT)-PCR analysis

RNA was extracted from stationary-phase-grown M. tuberculosis or M. smegmatis (OD600 nm 2.8–3.0) by suspending cell pellets in TRIzol (Invitrogen), lysing cells with 0.5-mm-diameter glass beads using a FastPrep FP120 bead-beating device, and precipitating nucleic acids with isopropanol. Nucleic acids were treated with DNase I (Roche) and mRNA was cleaned using an RNeasy kit (Qiagen). cDNA was generated using Superscript III Reverse Transcriptase (Invitrogen). The resulting cDNA was used to amplify the gene Rv2145c (wag31Mtb) by PCR using the primers 5′-CTGGTTGCGTTCATCGGTAT-3′ and 5′-GAAAACTGGCGCGTGTCC-3′. The cDNA from the dnaJ1 genes was amplified as a control using the primers 5′-ARICCICCCAAIARRTCICC-3′ and 5′-CGIGARTGGGTYGARAARG-3′ (Yamada-Noda et al., 2007). All PCR reactions were performed under the following conditions: one cycle of 94 °C (2 min); 35 repeating cycles of 94 °C (30 s), 54 °C (30 s), and 72 °C (60 s); and a final cycle of 72 °C (7 min). PCR products were analyzed by 1% agarose gels and ethidium bromide staining.

Electron microscopy

Formvar carbon-coated nickel grids were used to lift individual M. smegmatis cells from 7H10 agar plates, which were then stained with 2% phosphotungstic acid, as described previously (Arora et al., 2008). Samples were viewed using a Joel TEM 1200 EX electron microscope (Joel USA Inc., Peabody, MA), and images were captured using a Mega View III camera (Lakewood, CO).

Statistical analysis of data

The results of assays for liquid-culture turbidity are expressed as means ± SDs from three independent experiments. Student's t-test was used to assess differences between various groups with a level of significance set at 0.005.


Identification of elevated levels of Wag31Mtb in M. tuberculosis with a functional relMtb gene

Previous studies have shown that RelMtb is involved in the regulation of more than 150 genes in M. tuberculosis, including virulence factors and antigens (Dahl et al., 2003). In order to identify some of these antigens potentially regulated by RelMtb, lysates of H37Rv, H37RvΔrelMtb, and the complemented mutant strain H37RvΔrelMtbattBrelMtb were compared using polyclonal antibodies raised against the wild-type H37Rv strain (Fig. 1a). Western blot analysis was conducted on bacterial whole-cell lysates of M. tuberculosis strains grown to the late stationary phase (OD600 nm 2.8). Previous studies have shown that cells in this stage of bacterial growth are activated for the stringent response (Primm et al., 2000; Dahl et al., 2003, 2005). One protein band was observed with a 4.5-fold reduction in expression level in the H37RvΔrelMtb strain, and this protein is approximately 45 kDa in size (Fig. 1a; arrow). A protein band at this position was visualized in the corresponding Coomassie brilliant blue-stained polyacrylamide gels of H37Rv protein lysates (data not shown) and was excised, destained, and subjected to trypsin digestion and analysis by matrix-assisted laser desorption. The 45-kDa protein was identified as the M. tuberculosis Rv2145c gene product Wag31Mtb (Cole et al., 1998). In M. tuberculosis, this protein is also known as DivIVA (Kang et al., 2005) and antigen 84 (Hermans et al., 1995), and it is an ortholog of MinE in E. coli (Hu et al., 2003). Previous microarray comparisons reveal that wag31Mtb is expressed 2.6-fold higher in cells that have an intact rel gene and are starved for nutrients (Dahl et al., 2003). This Western blot analysis is Fig. 1a confirms this rel-dependent expression of wag31Mtb. To determine whether RelMtb causes the differential transcription of wag31Mtb, RT-PCR analysis was performed and shows that elevated levels of mRNA for wag31Mtb correspond with functioning Rel (Fig. 1b, upper panel). Densitometry analysis of the levels of PCR products shows that wag31Mtb was expressed at a level 11-fold higher in H37Rv cells containing relMtb. As a control to ensure that equivalent amounts of cellular mRNA were subjected to reverse transcription, expression of a Rel-independent gene was examined. rRNA levels were not compared because rRNA is downregulated in the presence of Rel (Cashel et al., 1996). Instead, dnaJ-specific mRNA levels were compared between M. tuberculosis strains with and without relMtb (Fig. 1b, lower panel). dnaJ encodes for a chaperone-like protein (Yamada-Noda et al., 2007). Our previous microarray studies showed that dnaJ is not differentially regulated by RelMtb in M. tuberculosis (Dahl et al., 2003), making the gene an appropriate control to ensure that equivalent levels of mRNA were harvested from H37Rv and H37RvΔrel cells. Collectively, these Western blot and RT-PCR analyses confirm that wag31Mtb is positively regulated by the stringent response in M. tuberculosis.

Figure 1.

wag31Mtb is expressed at higher levels in Mycobacterium tuberculosis when rel is present. Both proteins and mRNA products for wag31Mtb were analyzed in M. tuberculosis cells with or without relMtb. (a) Western blot analysis using polyclonal rabbit sera. Lane 1, H37Rv; lane 2, H37RvΔrelMtb; lane 3, H37RvΔrelMtbattBrelMtb. Arrow indicates the presence of a band that has a 4.5-fold lower intensity in cells without relMtb (lane 2). (b) cDNA from M. tuberculosis H37Rv (lanes 1, 3, 5, and 7) or from M. tuberculosisΔrelMtb (lanes 2, 4, 6, and 8) was used as a template for PCR reactions. The top panel is of PCR amplifications using wag31Mtb-specific primers and the bottom panel is of PCR amplifications using dnaJ-specific primers. cDNA was diluted 1 : 50 (lanes 1 and 2), 1 : 25 (lanes 3 and 4), or 1 : 2 (lanes 5, 6, 7, and 8). PCR reactions for lanes 7 and 8 were performed on samples that had not been reverse transcribed and they served as negative controls to show all chromosomal DNA was absent from the reactions. Lane 9 is a positive control using M. tuberculosis chromosomal DNA used as a template.

Expression of wag31Mtb in M. smegmatis

The wag31Mtb gene was further examined to see whether it was differentially regulated by the stringent response in the surrogate mycobacterial host M. smegmatis. We previously inactivated the stringent response in M. smegmatis mc2155 to facilitate the study of M. tuberculosis genes suspected of being either positively or negatively regulated by Rel (Dahl et al., 2003, 2005). To determine whether the wag31Mtb gene was similarly regulated by the stringent response in M. smegmatis, the relative levels of Wag31Mtb protein and wag31Mtb mRNA were examined in M. smegmatis strains expressing the gene. Mycobacterium smegmatis strains containing either the vector pOLYG or pwag31Mtb were grown in Middlebrook 7H9 medium+hygromycin (50 μg mL−1) with shaking for 4 days with culture densities measured using a spectrophotometer. No differences were observed in the growth rates regardless of the strain or the presence of pwag31Mtb (data not shown). To examine gene expression, the levels of wag31Mtb protein and mRNA products were determined (Fig. 2). Densitometry readings of the Wag31Mtb-specific bands reveal a 1.4-fold increased level of this protein in M. smegmatis mc2155 cells compared with the isogenic ΔrelMsm strain. The anti-H37Rv polyclonal antibodies raised against M. tuberculosis cell lysates do not appear to recognize the Wag31 homolog in M. smegmatis, as evident by the absence of a corresponding 45 kDa band in cells containing the cloning vector pOLYG (Fig. 2a, lanes 1 and 2). The Wag31 proteins of M. tuberculosis and M. smegmatis share 79% identity and 87% similarity, and the essential wag31Msm gene can be successfully replaced by wag31Mtb (Mukherjee et al., 2009). However, Wag31Mtb contains two regions of amino acids (8 and 12 residue stretches within a 34-residue region) that are dissimilar or missing from Wag31Msm (data not shown). These dissimilar domains may be the epitopes recognized by the polyclonal antibodies binding to Wag31Mtb. RT-PCR was performed on cDNA prepared from M. smegmatis mc2155/pwag31Mtb and M. smegmatis mc2155Δrel/pwag31Mtb, and this showed that Rel has a positive effect on the expression of wag31Mtb (Fig. 2b, upper panel).

Figure 2.

 Expression of wag31Mtb is regulated in the same rel-dependent fashion in Mycobacterium smegmatis as in Mycobacterium tuberculosis. (a) Western blot analysis of wag31Mtb expression in M. smegmatis. Protein lysates are from the following strains: lane 1, M. smegmatis/pOLYG; lane 2, M. smegmatisΔrel/pOLYG; lane 3, M. smegmatis/pwag31Mtb; lane 4, M. smegmatisΔrel/pwag31Mtb. The position of Wag31Mtb is indicated by the arrow. (b) RT-PCR analysis of cDNA from mc2155/pwag31Mtb (lanes 1, 3, 5, and 7) or from mc2155ΔrelMsm/pwag31Mtb (lanes 2, 4, 6, and 8). Either wag31Mtb-specific primers (top panel) or dnaJ-specific primers (bottom panel) were used for PCR reactions. cDNA was diluted 1 : 50 (lanes 1 and 2), 1 : 25 (lanes 3 and 4), or 1 : 2 (lanes 5, 6, 7, and 8). PCR reactions for lanes 7 and 8 were performed on samples that had not been reverse transcribed and they served as negative controls to show all chromosomal DNA was absent from the reactions. Lane 9 is a positive control using M. tuberculosis chromosomal DNA used as a template.

Alternation in M. smegmatis cell shape and cell surface by Wag31Mtb

The presence of wag31Mtb on a multicopy plasmid in M. smegmatis alters both the cell shape and the colony morphologies in a rel-dependent fashion (Fig. 3). In the presence of the shuttle vector pOLYG, the wild-type and ΔrelMsm colonies have raised ridges with average cell lengths of 2.1 ± 0.3 μm for mc2155/pOLYG and 3.4 ± 0.9 μm for ΔrelMsm/pOLYG (Fig. 3a and b). However, the presence of pwag31Mtb in both strains leads to a flattened, smoother colony formation. There are almost no colony ridges for mc2155/pwag31 (Fig. 3c), while the ΔrelMsm/pwag31 strain has a modest amount of ridges near the circumference of colonies (Fig. 3d). For both strains, the presence of pwag31Mtb leads to shorter cells, with mc2155/pwag31 cells averaging 0.9 ± 0.4 μm in length and ΔrelMsm/pwag31 cells averaging 1.6 ± 0.3 μm in length (Fig. 3c and d insets, respectively). Smoother colony appearance suggests a change in the cell surface properties like a decrease in hydrophobicity. To test this theory, cell dispersion was compared for cells grown in the presence or in the absence of the detergent Tween 80 (0.05% v/v) (Fig. 4). The absence of Tween 80 from standing cultures led to a 60% decrease in the levels of cell suspension, except for mc2155/pwag31Mtb cells, which remained dispersed even without the detergent. This is the first reported instance of Wag31 playing a role in altering mycobacterial cell surfaces.

Figure 3.

 Colony and cell morphologies of Mycobacterium smegmatis strains expressing wag31Mtb. Mycobacterium smegmatis strains were grown in Middlebrook 7H11 agar medium with 0.2% glycerol for 10 days. (a) Mycobacterium smegmatis mc2155/pOLYG. (b) Mycobacterium smegmatis mc2155ΔrelMsm/pOLYG. (c) mc2155/pwag31Mtb. (d) mc2155ΔrelMsm/pwag31Mtb. Insets show individual cells visualized by TEM with a magnification of × 5000. Scale bars=1 μm.

Figure 4.

 Expression of wag31Mtb leads to reduced cell-surface hydrophobicity in Mycobacterium smegmatis. Liquid cultures were of M. smegmatis strains grown in 7H9+0.05% Tween 80+hygromycin (50 μg mL−1) to OD600 nm 1.0 before equilibrating cell densities and resuspending cells in 7H9 with 0.05% Tween 80 (open bars) or without Tween 80 (black bars). Cells were allowed to incubate in stationary positions for 12 h at 37°C before the turbidity of the cell suspensions was measured using a spectrophotometer. *P<0.005 compared with the mc2155 strain with pOLYG grown without Tween 80.


Wag31 was first identified as a mycobacterial antigen using serum antibodies of leprosy and tuberculosis patients (Hermans et al., 1995), which helps explain the observed humoral response of rabbits to this protein (Fig. 1a). Wag31 was eventually discovered to be a homolog of DivIVA, a protein known to regulate cell shape and cell division in gram-positive bacteria (Cha & Stewart, 1997; Cole et al., 1998; Flardh, 2003), and wag31 has been shown to be an essential gene in M. tuberculosis (Sassetti et al., 2003) and in M. smegmatis (Kang et al., 2005; Nguyen et al., 2007). Wag31 is receptive to extracellular signaling by two serine/threonine protein kinases, PknA and PknB, and the action of Wag31 as a determinant of cell shape is dependent on a protein phosphorylation domain in the protein (Kang et al., 2005). The bacterial role of Wag31 appears to be regulating cell division and cell shape as evident by the appearance of the protein at the poles of dividing cells (Nguyen et al., 2007; Kang et al., 2008), by its appearance on the cell surface of mycobacteria (He & De Buck, 2010), by wag31-dependent alterations in cell envelope proteins and lipids (Hamasha et al., 2009), and by alterations of cell shape (Kang et al., 2005, 2008; Nguyen et al., 2007). Whereas previous studies have examined wag31-dependent functions by expressing the gene with an acetamide-inducible promoter (Kang et al., 2008), a tetracycline-inducible promoter (Hamasha et al., 2009), or a heat shock promoter (Kang et al., 2005), this current study is the first to examine wag31Mtb expression using its native promoter. This promoter appears to be upregulated by the mycobacterial stringent response (Figs 1 and 2).

The stringent response is necessary for persistent M. tuberculosis infections in mammalian hosts (Dahl et al., 2003; Klinkenberg et al., 2010). Here, we report that the stringent response is needed for higher expression of wag31, suggesting a potential connection between Wag31 and virulence. Although Wag31 is involved in mycobacterial cell wall synthesis, Wag31 may be playing some alternative roles during the infection process. Cao et al. (2008) recently reported that Wag31Mtb stimulates XCL2 expression in macrophages. XCL2 is a chemokine in macrophages that serves as a chemoattractant for CD8+ and CD4+ T cells. Therefore, wag31Mtb expression may contribute to the formation of granulomas that are extremely diminished in size and in numbers in animals infected with M. tuberculosisΔrel strains (Dahl et al., 2003; Klinkenberg et al., 2010). Although traditionally thought to function as a host defense strategy, the role of the granuloma is being re-evaluated as providing a potential benefit to mycobacterial pathogens (Flynn, 2004). Also, elevated wag31 expression may enhance M. tuberculosis survival in macrophages by enhancing resistance to oxidative stress. Wag31 may do this by stabilizing penicillin-binding protein 3 (PBP3) against cleavage by the M. tuberculosis metalloprotease Rv2869c. This metalloprotease is essential for M. tuberculosis cells to infect mice lungs, and it likely acts to regulate the bacterial lipid and membrane composition necessary for survival in the host (Madinoshima & Glickman, 2005). However, without protection by Wag31 binding, PBP3 is susceptible to deleterious cleavage by Rv2869c, leading to reduced survival of M. tuberculosis within macrophages (Mukherjee et al., 2009).


We thank Christine Davitt for assistance with TEM analysis, Gerhard Munske for help with proteomic identification of Wag31, and Mike Konkel for assistance with antibody production. This research was supported by internal funds from the University of Minnesota Duluth.