Toxin–antitoxin vapBC locus participates in formation of the dormant state in Mycobacterium smegmatis

Authors

  • Oksana I. Demidenok,

    Corresponding author
    1. Laboratory of Biochemistry of Stresses in Microorganisms, A.N. Bach Institute of Biochemistry Russian Academy of Sciences, Moscow, Russia
    • Correspondence: Oksana I. Demidenok, Laboratory of Biochemistry of Stresses in Microorganisms, A.N. Bach Institute of Biochemistry Russian Academy of Sciences, Leninsky prospekt 33, build. 2, 119071 Moscow, Russia. Tel.: +7 495 954 40 47; fax: +7 495 954 27 32;

      e-mail: demidenoksana@gmail.com

    Search for more papers by this author
  • Arseny S. Kaprelyants,

    1. Laboratory of Biochemistry of Stresses in Microorganisms, A.N. Bach Institute of Biochemistry Russian Academy of Sciences, Moscow, Russia
    Search for more papers by this author
  • Anna V. Goncharenko

    1. Laboratory of Biochemistry of Stresses in Microorganisms, A.N. Bach Institute of Biochemistry Russian Academy of Sciences, Moscow, Russia
    Search for more papers by this author

Abstract

Toxin–antitoxin (TA) loci are widely spread in bacterial plasmids and chromosomes. Toxins affect important functions of bacterial cells such as translation, replication and cell-wall synthesis, whereas antitoxins are toxin inhibitors. Participation in formation of the dormant state in bacteria is suggested to be a possible function of toxins. Here we show that overexpression of VapC toxin in Mycobacterium smegmatis results in development of morphologically distinct ovoid cells. The ovoid cells were nonreplicating and revealed a low level of uracil incorporation and respiration that indicated their dormant status. To validate the role of VapBC in dormancy formation, we used a model of dormant, ‘nonculturable' (NC) M. smegmatis cells obtained in potassium-limited conditions. Overexpression of VapB antitoxin prevented transition to dormancy, presumably due to a decreased level of the free VapC protein. Indeed, this effect of the VapB was neutralized by coexpression of the cognate VapC as a part of the vapBC operon. In summary, these findings reveal participation of vapBC products in formation of the dormant state in M. smegmatis.

Introduction

The toxin–antitoxin (TA) systems were discovered in the 1980s as plasmid genes that ensure inheritance of a plasmid in daughter cells (Ogura & Hiraga, 1983; Gerdes et al., 1986). Homologues of TA plasmid system modules were later found in the chromosomes of most bacteria, and their functions in bacterial cells have been intensively investigated (Masuda et al., 1993; Gerdes, 2000; Hazan & Engelberg-Kulka, 2004; Magnuson, 2007; Saavedra De Bast et al., 2008; Van Melderen, 2010; Wang & Wood, 2011). Most TA systems are composed of two components: a toxin that affects important cellular functions such as translation, replication and cell-wall synthesis, and an antitoxin that controls toxin activity by formation of protein–protein, RNA–RNA and RNA–protein complexes (Fozo et al., 2008; Fineran et al., 2009; Makarova et al., 2009). The genome of Mycobacterium smegmatis contains three TA modules: vapBC, mazEF and phd/doc (Pandey & Gerdes, 2005). The vapBC family is the most widespread TA locus in microbial genomes and is prevalent in several organisms. VapC toxin belongs to the Pilt N-terminal domain (PIN domain) family of proteins and is configured in an RNaseH-like fold (Arcus et al., 2005). Overexpression of VapC arrests bacterial cell growth of Escherichia coli, M. smegmatis and Mycobacterium tuberculosis and inhibits the rate of translation. The growth-inhibitory effect of VapC was neutralized by expression of the cognate VapB antitoxin (Robson et al., 2009; Winther & Gerdes, 2009; Ahidjo et al., 2011; Sharp et al., 2012). Microarray data suggest that VapC targets specific mRNA transcripts, acting as a post-transcriptional regulator of metabolic flux (McKenzie et al., 2012).

One hypothetical function of TA loci is their participation in the transition of bacteria to the dormant state. The dormant state has been defined as a ‘reversible state characterized by low level of metabolic activity or its complete absence, which cell may be in for a long time without division’ (Kaprelyants et al., 1993). Dormancy of mycobacteria is often accompanied by so-called ‘non-culturability’ – a temporary loss of ability to grow on solid media (Shleeva et al., 2002; Mukamolova et al., 2003). However, mycobacteria in the dormant state that retain culturability have also been described (Anuchin et al., 2009, 2010). It has been suggested that formation of the dormant state is controlled by a particular ‘programme’ and that TA systems probably participate in this process (Arcus et al., 2005; Condon, 2006; Zhu et al., 2006). The bacteriostatic action of VapC toxin suggests it may be a participant in dormancy formation in contrast to MazF toxin, the bactericidal effect of which on mycobacteria has been shown (Frampton et al., 2012).

In present study we demonstrate the significance of the vapBC system for the formation of dormancy in M. smegmatis cells.

Materials and methods

Bacterial strains, plasmids and growth conditions

Escherichia coli BMH was grown in Luria–Bertani (LB) medium at 37 °C with agitation (200 r.p.m.) or on LB agar plates. All the M. smegmatis strains were grown in LB medium with 0.05% Tween-80 (LBT) and modified potassium-limited Hartmans-de Bont (HdB) medium at 37 °C with agitation (200 r.p.m.; Shleeva et al., 2004). The growth media contained ampicillin (Amp; 50 μg mL−1 for E. coli) or kanamycin (Km; 50 μg mL−1 for E. coli, 10 μg mL−1 for M. smegmatis) where appropriate. Overexpression of recombinant genes was induced by addition of tetracycline (Tet; 20 ng mL−1). All strains and plasmids used in this study are listed in Table 1.

Table 1. Strains and plasmids used in this study
Bacterial strain/plasmidDescriptionSource
  1. Kmr, kanamycin resistance; Hgr, hygromycin B resistance; Ampr, ampicillin resistance.

Strain
E. coli BMH 71-18thi, supE, Δ(lac-proAB), [F′, proAB, lacI(q)ZΔM15].Promega
M. smegmatis mc2 155Wild-type strain of M. smegmatisLaboratory collection
M. smegmatis – pMindStrain mc2 155 transformed with pMind plasmid; Kmr, HgrThis study
M. smegmatisvapBVapB overexpression strain M. smegmatis mc2 155; Kmr, HgrThis study
M. smegmatis – vapC VapC overexpression strain M. smegmatis mc2 155; Kmr, HgrThis study
M. smegmatisvapBCVapBC overexpression strain M. smegmatis mc2 155;Kmr, HgrThis study
M. smegmatis∆vapBCvapBC locus deletion strain M. smegmatis mc2 155; HgrThis study
M. smegmatis –∆vapBC + vapBC vapBC locus deletion strain M. smegmatis mc2 155 with vapBC overexpressionThis study
M. smegmatis∆vapBC + vapBvapBC locus deletion strain M. smegmatis mc2 155 with VapB overexpression; Kmr, HgrThis study
Plasmid
pGEME. coli TA cloning vector; AmprPromega
pMindTetracycline-inducible expression vector; Kmr, HgrBlokpoel et al. (2005)
pMind - vapBpMind harbouring vapB gene; Kmr, HgrThis study
pMind - vapCpMind harbouring vapC gene; Kmr, HgrThis study
pMind - vapBCpMind harbouring vapBC genes; Kmr, HgrThis study

RNA isolation

RNA was extracted from M. smegmatis – pMind cells in potassium-limited HdB medium and M. smegmatis – vapB in LBT medium. For each experimental point, 50-mL culture samples (108 cells mL−1) were harvested. The total amount of cells was estimated based on microscope counts of colony-forming units. Cells were harvested by centrifugation (4000 g, 10 min) and 1 mL Trizol reagent was added to the pellets. Cells were disrupted using zirconia beads (0.1 mm) in a mini Bead Beater (BioSpec Products, Inc. Bartlesville, OK). After centrifugation the supernatant was extracted once with chloroform. Nucleic acids were then precipitated with isopropanol, harvested by centrifugation, washed with 70% ethanol and redissolved in nuclease-free water containing RNAsin ribonuclease inhibitor (Promega, Madison, WI). Each RNA sample was finally treated with RNase-free Turbo DNase (Ambion, Austin, TX). RNA was then isolated using an RNeasy Mini kit (Promega).

Quantitative real-time PCR (qRT-PCR)

The initial cDNA synthesis was carried out with M–MLV RT polymerase (αFerment) at 54 °C for 30 min. For qRT-PCR, the iQ™ SYBR Green RT-PCR kit (BioRad, Hercules, CA) was employed. Each reaction contained 20 pmol μL−1 of each of the paired gene-specific primers (Supporting information, Table S1). All primers were optimized for annealing temperature to ensure that only a single product of the correct size was amplified. DNA amplification was for 40 cycles of 20 s at 94 °C, then 20 s at 55 °C and then 20 s at 72 °C. The PCR cycle at which the amplification threshold was attained was converted to copy number using standard curves prepared with M. smegmatis genomic DNA.

DNA manipulation

All primer sequences used in this study are listed in Table S1.

pMind-vapBC overexpression plasmid was constructed as follows: amplification was carried out on the genomic DNA of M. smegmatis mc2155 by using primers Up pMind-vapBC and Low pMind-vapBC. The amplification product was first cloned into pGEM-T vector (Promega) and then subcloned into pMind using BamHI and SpeI restriction enzymes.

pMind-vapB overexpression plasmid was obtained from pMind-vapBC plasmid by removing the major part of the vapC gene by BsaBI digestion. As a result, the vapC gene (total length 390 bp) was removed from the 4th to 218th nucleotides.

pMind-vapC overexpression plasmid was constructed in the same way as pMind-vapBC using primers vapCpMind F (Robson et al., 2009) and Low pMind-vapC.

Knockout mutant ΔvapBC. To create a construct for deletion of the vapBC gene we amplified upstream (L) and downstream (R) regions by using primers Up ∆vapBC/L and Low ∆vapBC/L, Up ∆vapBC/R and Low ∆vapBC/R, respectively. L was then TA-cloned into pGEM-T Easy vector (Promega), and R was TA-cloned into pGEM-T (Promega). Cleaved out with NotI and SpeI endonucleases, L was then transferred into pGEM-R. The resulting pGEM–L-R was digested with BamHI and HindIII and ligated with a hygromycin resistance (HgR) cassette (cleaved from plamsid pMind with the same endonucleases beforehand). The resulting pGEM–L-HgR-R construct was digested with NotI and XbaI to obtain fragment L-HgR-R, and the latter was then inserted into NotI- and XbaI-cleaved pPR27.

Escherichia coli and M. smegmatis were transformed with vectors by electoporation. After growth on selective medium, colonies were subjected to PCR with specific primers.

Selection of toxin-overexpressing colonies

Individual PCR-positive vapC colonies were resuspended in 1 mL LBT medium, and transferred into two wells of a 24-well cell culture plate. One of the wells was supplemented with 20 ng mL−1 Tet for toxin induction. After 48 h of cultivation with constant stirring (120 r.p.m.) cell growth was analysed. Cells from wells (−Tet) which have counterpart wells with negative growth (+ Tet) were used for further experiments. The plates were kept at room temperature statically for at least 2 weeks.

Dormant ‘non-culturable’ M. smegmatis cells

Mycobacterium smegmatis mc2155 and its recombinant variants were cultivated in 50 mL LBT medium for 30 h at 37 °C with agitation (200 r.p.m.). Then, 0.5 mL of the culture was inoculated into modified HdB medium and grown for 120 h. Bacterial cell culturability was monitored based on CFU counts by plating serial dilutions of bacteria culture onto LB agar medium. After 72 h of cultivation in modified HdB medium, CFUs decreased to zero, reflecting transition of cells to the ‘nonculturable’ (NC) state. Viability of NC cells was restored using a reactivation procedure. NC cells were washed three times in Sauton medium (Parish & Stoker, 1998) and inoculated in 100 mL medium for further reactivation under agitation (200 r.p.m.) at 37 °C. Periodically samples were withdrawn for CFU counts or OD determination.

Measurement of [5, 6-H3]uracil incorporation and respiration

For measurement of [5, 6-H3]uracil incorporation and respiration, morphologically altered, ovoid cells of M. smegmatis – vapC were used 2 weeks following toxin induction. As a control, cells of M. smegmatis with empty pMind vector in the exponential and the stationary phases were used. Cells of the control strains were grown with 20 ng mL−1 Tet. A 1-mL aliquot of the cell culture was placed in a plastic vial with 1 μL [5, 6-H3]uracil solution (370 kBq mL−1) and incubated at 37 °C at 200 r.p.m. for 4 h. Then, 0.2 mL of the sample was placed in a 15-mL test tube, containing 3 mL 10% trichloroacetic acid, and incubated at 0 °C for 15 min. The mixture was then passed through a glass filter (Whatmann) followed by washing with 3 mL 10% trichloroacetic acid and 4 mL 96% ethanol. The filters were placed in 10 mL scintillation fluid and the number of pulses was determined using a Beckman Instruments LS analyser (Porterville, CA). The level of [5, 6-H3]uracil incorporation was estimated as counts per minute (c.p.m.) per 108 viable cells. The number of viable cells was estimated using the MPN method in LBT medium (Shleeva et al., 2002).

The endogenous respiration of M. smegmatis cells in 50 mM potassium phosphate buffer, pH 7.0, was estimated by measurement of oxygen consumption with a closed Clark-type oxygen electrode.

Light microscopy

Microscopic observations were performed using an Eclipse E4000 microscope (Nikon, Tokyo, Japan) in phase contrast mode.

Results

Overexpression of VapC toxin causes changes in cell morphology

After transformation of M. smegmatis with expression plasmid pMind-vapC carrying the vapC toxin gene, two colony types were obtained on a selective medium without tetracycline: small colonies and normal-sized colonies. Small colonies consisted of cells with plasmid carrying the vapC gene while normally growing colonies consisted of cells with deleterious mutations in the plasmid vapC, as determined by sequencing. Note that continuous cultivation with agitation in liquid media for more than 48 h often resulted in a loss of VapC-expressing cells due to propagation of mutated clones. Therefore, all the experiments with M. smegmatisvapC were carried out using newly transformed cells. The overexpression of vapC was confirmed by qRT-PCR (Fig. S2).

Induction of VapC toxin overexpression in M. smegmatis resulted in cell growth arrest in liquid medium. This observation confirmed previously published results (Robson et al., 2009). However, microscopic examination revealed that VapC toxin overexpression led to formation of morphologically altered, non-replicating ovoid cells (Fig. 1a and b). The proportion of modified cells in the culture ranged from 20 to about 90% in different experiments. Cells with an altered morphology began to appear after 96 h of cultivation without agitation and reached maximum numbers after 2 weeks, remaining constant over 1 month of observation when cells were kept at room temperature statically. When agitated at 37 °C (200 r.p.m.) in liquid medium with the inducer, cells resumed their growth after 48–72 h as a result of mutations in vapC (see above). Note that morphologically altered, nonreplicating cells could be grown on solid or liquid medium without the inducer. The majority of cells grown on agar without inducer retained plasmid with an intact vapC gene and were responsive to Tet. These observations prove the culturability of ovoid cells and demonstrate the reversibility of VapC's effect.

Figure 1.

Induction of VapC toxin overexpression results in formation of morphologically altered, ovoid cells. Cells of the Mycobacterium smegmatis – vapC strain were grown in 24-well cell culture plates in LBT medium supplemented with 10 μg mL−1 Km and 20 ng mL−1 Tet. After 48 h of cultivation at constant agitation (120 r.p.m.) cells were kept at room temperature statically. (a) Cells of M. smegmatisvapC before VapC induction; (b) cells of M. smegmatisvapC 96 h after VapC induction with Tet (20 ng mL−1); (c) cells of the control strain M. smegmatis harbouring empty vector pMind before Tet addition; (d) cells of the control strain M. smegmatis harbouring empty vector pMind 96 h after Tet addition. This experiment was repeated five times with similar resulta. A typical result is shown. Scale bars = 3 μm (a,b) and 5 μm (c,d).

To confirm that this result is not due to an effect of Tet, the latter was added to the control strain harbouring empty pMind vector. No morphological changes of control strain cells were seen (Fig. 1c and d).

Morphologically altered, ovoid cells of M. smegmatis – vapC demonstrated low level of [5,6-H3]uracil incorporation and respiration

To monitor transcriptional activity of ovoid M. smegmatisvapC, the incorporation of [5, 6-H3]uracil was measured. The rate of uracil incorporation by ovoid M. smegmatisvapC cells was significantly lower (ca. 5000 c.p.m. per 108 cells) compared with control cells (2.88 × 105 c.p.m. per 108 cells for exponential cells and 2.5 × 104 c.p.m. per 108 cells for stationary cells), indicating the significant decrease of transcription in toxin-expressing cells (Fig. 2).

Figure 2.

Morphologically altered, ovoid cells of Mycobacterium smegmatis – vapC demonstrate low levels of [5,6-H3]uracil incorporation. Two-week-old morphologically altered, ovoid cells were used in the experiment. Cells of M. smegmatis harbouring empty pMind vector in the exponential and the stationary phases were used as a control. For details of cell preparation see legend to Fig. 1. [5,6-H3]uracil was added to the cell suspensions for 4 h. Levels of [5, 6-H3]uracil incorporation were estimated as c.p.m. per 108 viable cells. (1) [5,6-H3]uracil incorporation of M. smegmatis cells in exponential phase; (2) [5,6-H3]uracil incorporation of M. smegmatis cells in stationary phase; (3) [5,6-H3]uracil incorporation of M. smegmatis – vapC non-replicating ovoid cells. This experiment was repeated five times with similar results. A typical result is shown.

It was also found that in contrast to the control cells (M. smegmatis – pMind strain) in exponential and stationary phases, ovoid cells of the M. smegmatis – vapC strain did not respire (Fig. 3).

Figure 3.

Morphologically altered, ovoid cells of Mycobacterium smegmatis – vapC strain do not respire. Two-week-old morphologically altered, ovoid cells were used in the experiment. Cells of M. smegmatis – pMind harbouring empty pMind vector in the exponential and the stationary phases were used as a control. For details of cell preparation see legend to Fig. 1. The number of cells was equalized (c. 108 mL−1) before measurement of oxygen consumption. (image_n/fml12380-gra-0001.png) Oxygen consumption of M. smegmatis cells in the exponential phase; (image_n/fml12380-gra-0002.png) oxygen consumption of M. smegmatis cells in stationary phase; (image_n/fml12380-gra-0003.png) oxygen consumption of two-week-old morphologically altered, ovoid cells of M. smegmatis – vapC. This experiment was repeated five times with similar results. A typical result is shown.

Overexpression of VapB antitoxin causes a loss of ability to develop the NC state

Mycobacterium smegmatisvapB strain, transformed with overexpression plasmid pMind-vapB carrying the vapB antitoxin gene, was tested for its ability to form NC cells when cells were grown under potassium limitation in modified HdB medium. Normally, under such conditions wild-type cells lose their culturability on solid medium after 72 h of cultivation (Fig. 4), transitioning to the NC dormant (but reversible) state. Despite zero colonies on plates, NC cells are potentially viable as they reactivate in an appropriate liquid medium (Shleeva et al., 2004). We found that overexpression of VapB antitoxin caused a complete loss of the ability of the cells to produce NC forms and allowed them to maintain their culturability for a long time in post-stationary phase. This effect was observed both with and without inducer. As shown with qRT-PCR, thei was because vapB expression increased in overexpressed strain M. smegmatis - vapB in comparison with the wild-type for both uninduced and induced variants (Fig. S1). The maintenance of culturability in M. smegmatis – vapB was presumably due to the binding of free VapC toxin by an abundant ectopic antitoxin VapB. Indeed, coexpression of VapB and VapC of the entire vapBC locus in M. smegmatis – vapBC with pMind-vapBC plasmid revealed a transition to the NC state under potassium limitation with similar kinetics to the wild-type strain (Fig. 4).

Figure 4.

The overexpression of VapB antitoxin prevented transition to the NC state. Cells of Mycobacterium smegmatis – pMind (image_n/fml12380-gra-0004.png), VapB and VapC co-expressing M. smegmatis – vapBC (image_n/fml12380-gra-0005.png) and M. smegmatisvapB overexpressing VapB antitoxin (image_n/fml12380-gra-0006.png) strains were initially grown in LBT medium for 30 h and then were used as an inoculum for growth in potassium-limited HdB medium. Transition to the NC state was noted by CFU counts. This experiment was repeated five times with similar results. A typical result is shown.

M. smegmatisΔvapBC strain loses viability under potassium limitation

The M. smegmatis – ΔvapBC knockout strain was tested for its ability to form NC cells when cells were grown under potassium limitation. We found that the strain lost viability under this condition following 64 h of cultivation (6–8 h earlier than for the wild-type strain; Fig. S3). In contrast to the wild-type (M. smegmatis – pMind) these cells were unable to recover, as judged from CFU counts and uracil incorporation. This result demonstrated M. smegmatis – ΔvapBC cell death under the conditions described (Fig. S4). The complemented M. smegmatis – ΔvapBC vapBC strain revealed transition to the NC state with similar kinetics to the wild-type strain (Fig. S5), retaining the ability to recover from the NC state similarly to the wild-type strain (Fig. S6). It is interesting that the knockout strain with oversexpressed antitoxin M. smegmatis – ΔvapBC + vapB did not loss viability and did not develop an NC state under potassium limitation exhibiting the same phenotype as the M. smegmatisvapB strain (Fig. S5).

Transition to the NC state is accompanied by an increase of expression level of vapC toxin

Alteration of vapC expression level in M. smegmatis cells during transition to the NC state was examined with qRT-PCR. Expression of vapC was increased (c. 20-fold) before formation of the NC state (Fig. 5). Levels of vapC expression in NC cells were decreased (c. six-fold) in comparison with active cells. The lower level of vapC expression in NC dormant cells is obviously due to their low metabolic level.

Figure 5.

Detection of VapC expression in Mycobacterium smegmatis. Cells of M. smegmatis were initially grown in LBT medium for 30 h and then were used as an inoculum for growth in potassium-limited HdB medium. Transition to the NC state was noted by CFU counts. qRT-PCR was performed using equal amounts of cells as described in Materials and Methods. Each point is the mean of at least five biological replicates (three technical replicates each). Error bars represent the standard deviations of the mean. Differences between vapC expression levels in transiting cells (hatched bars) and both the NC state (filled bars) and active cells (open bars) were statistically significant (Newman–Keuls multiple comparison test; *< 0.01)). Scale graduations correspond to 109 copy numbers per 108 cells.

Discussion

The molecular mechanisms underlying the transition of nonsporulating bacteria to the dormant state have not yet been characterized in detail (Young et al., 2005). In this regard, the possible role of TA loci in bacterial dormancy is now being widely discussed (Lioy et al., 2012; Tashiro et al., 2012). Proteomic studies of M. tuberculosis grown under nutrient-starved conditions have revealed a significant increase in a number of TA systems at the protein level (Albrethsen et al., 2013). It was also shown that successive deletion of all TA genes from E. coli chromosome leads to a gradual decrease in the number of persisting cells formed after antibiotic treatment (Maisonneuve et al., 2011).

In the present study, we demonstrated for the first time the involvement of a TA system in the development of mycobacterial dormancy. This conclusion is supported by the following observations.

Overexpression of the VapC toxin in M. smegmatis resulted in cell growth arrest and a significant decrease of transcription, as estimated by a decreased level of uracil incorporation in low metabolic activity judged by reduced respiration rate (Figs 2 and 3). These results partially confirm earlier published observations of growth arrest of M. smegmatis cells overexpressing VapC (Robson et al., 2009). However, in contrast to our results this was not accompanied by inhibition of transcription. Note that in our experiments, cells used for determination of transcription or respiration levels were kept statically for at least 2 weeks after VapC overexpression. This time was necessary to observe a majority of cells with ovoid shape. At the same time, Robson et al. (2009) used cells within 21 h after induction when earlier effects of VapC could be observed. We believe that toxin induction can result in translation inhibition at an earlier state of development of nonreplicating cells whilst a longer exposure time of VapC causes inhibition of transcription and metabolic activity.

VapC toxin overexpression resulted in formation of nonreplicating ovoid dormant cells (Fig. 1). It is important that these cells retained culturability, demonstrating reversibility of the toxin-inducible phenotype. We have previously shown that under certain conditions, transition to dormancy resulted in formation of non-replicating, but culturable ovoid forms of M. smegmatis (Anuchin et al., 2009, 2010) and of M. tuberculosis (Shleeva et al., 2002, 2011). Interestingly, formation of ‘lemon-shaped’ E. coli cells was documented after overexpression of YeeV (Tan et al., 2011) and CptA (YgfX; Masuda et al., 2012), toxins which target essential cytoskeleton proteins, FtsZ and MreB. Unfortunately, the viability or metabolic activity of such altered cells was not detected.

Overexpression of VapB antitoxin prevented transition to the dormant state under potassium limitation (Shleeva et al., 2004). In this model dormant cells temporarily became NC, although overexpression of VapB results in the complete loss of ability of M. smegmatisvapB to form NC cells (Fig. 4). Evidently, the effect of VapB is due to an imbalance between VapB and VapC as overexpression of the entire vapBC locus did not influence cell transition to the NC state (Fig. 4). This conclusion is in agreement with the reported growth-arresting effect of toxin expression in E. coli (Gotfredsen & Gerdes, 1998), M. smegmatis (Robson et al., 2009) and M. tuberculosis (Sharp et al., 2012), which was reversed by coexpression of the respective antitoxins. Thus, inactivation of the VapC toxin by abundant VapB antitoxin leads to the loss of the ability to transit to the NC state, highlighting that VapC is required for the transition of mycobacteria to dormancy. Moreover, transition to the NC state is accompanied by an increase of vapC transcription level (Fig. 5), which provides a mechanistic link between NC state formation and VapC activity.

We might consider that the M. smegmatis – ΔvapBC knockout strain would have the same phenotype as the overexpressing vapB antitoxin (M. smegmatisvapB) strain under potassium limitation. Unexpectedly, M. smegmatis – ΔvapBC strain lost (irreversibly) viability despite the absence of the toxin gene in the vapBC locus (Fig. S6).

These results may be explained by the action of toxins other than VapC that are also under the control of VapB. Indeed, Frampton et al. (2012) observed that overexpression of MazF led to M. smegmatis cell death. Moreover, it was shown that VapB antitoxin can bind MazF toxin (Zhu et al., 2010). In line with this hypothesis, we found that expression of VapB antitoxin prevented M. smegmatis – ΔvapBC cells from losing viability and developing the NC state (Fig. S5). However, additional studies are needed to elucidate a possible interplay between different TA systems in one cell.

In conclusion, the results presented here reveal the participation of the vapBC locus in the induction of mycobacteria transitioning to dormancy. Evidently, the TA system could be a part of the specific mechanism involving different enzymes and regulators, activation of which eventually results in formation of the dormant state and NC phenotype. This mechanism is not yet clear, and details of the participation of the TA system in this mechanism need to be elucidated.

Acknowledgements

The work was supported by ‘Molecular and cellular Biology’ Programme of the Russian Academy of Sciences: RFBR Grants 12-04-01604-a, 11-04-00713-a, 11-04-01440-a and 12-04-01760/12.

Ancillary