Multi-Stress Induction of the Mycobacterium tuberculosis MbcTA Bactericidal Toxin-Antitoxin System

MbcTA is a type II toxin/antitoxin (TA) system of Mycobacterium tuberculosis. The MbcT toxin triggers mycobacterial cell death in vitro and in vivo through the phosphorolysis of the essential metabolite NAD+ and its bactericidal activity is neutralized by physical interaction with its cognate antitoxin MbcA. Therefore, the MbcTA system appears as a promising target for the development of novel therapies against tuberculosis, through the identification of compounds able to antagonize or destabilize the MbcA antitoxin. Here, the expression of the mbcAT operon and its regulation were investigated. A dual fluorescent reporter system was developed, based on an integrative mycobacterial plasmid that encodes a constitutively expressed reporter, serving as an internal standard for monitoring mycobacterial gene expression, and an additional reporter, dependent on the promoter under investigation. This system was used both in M. tuberculosis and in the fast growing model species Mycobacterium smegmatis to: (i) assess the autoregulation of mbcAT; (ii) perform a genetic dissection of the mbcA promoter/operator region; and (iii) explore the regulation of mbcAT transcription from the mbcA promoter (PmbcA) in a variety of stress conditions, including in vivo in mice and in macrophages.


Introduction
Mycobacterium tuberculosis (Mtb), the etiological agent of tuberculosis (TB), is a strictly human-adapted obligate pathogen of major public health importance (WHO, Global Tuberculosis report, 2019). Indeed, TB was responsible for 1.5 million deaths and about ten million new TB cases were reported in 2018. Of particular concern, an estimated 484,000 new infections were due to Mtb strains resistant to one or more frontline antibiotics in 2018, and 80% of these Mtb strains were multi-drug resistant (MDR). Therefore, the development of novel therapies against TB is urgently needed. In line with this, we recently started to explore the potential utilization of intrinsic

MbcT Toxin Depletes NAD + and Is Bactericidal in Mycobacterium smegmatis
In order to express the MbcT toxin in M. smegmatis, we used an expression vector, pGMCS-TetR-P1-mbcT, that can stably integrate into the mycobacterial chromosome, within the glyV tRNA gene, through site-specific recombination catalyzed by the L5 mycobacteriophage integrase [12]. In this plasmid, mbcT is expressed under the tetracycline-inducible promoter P myc1 tetO [13], hereafter abbreviated P1. Upon transformation into a Mtb strain deleted for the endogenous mbcAT operon, pGMCS-TetR-P1-mbcT conferred anhydro-tetracycline (Atc)-dependent depletion of the essential metabolite NAD + , resulting in bacterial cell death [1]. In contrast, growth of a derivative of wild-type M. smegmatis strain mc 2 155 carrying this plasmid was not affected by addition of Atc to its growth medium (Figure 1a), presumably because the amount of MbcT produced in such conditions was not sufficient to deplete NAD + from the M. smegmatis cells (Figure 1d). To verify this hypothesis, we constructed plasmids expressing mbcT under the control of stronger Atc-inducible promoters, namely P606Pld and P606 (obtained from D. Schnappinger, Weill Cornell Medical College, NY, USA). Induction of P606Pld-mbcT had no effect on mc 2 155 growth (Figure 1b), although it reduced the NAD + content to 8% of the amount measured in the uninduced strain carrying pGMCS-TetR-P1-mbcT. In contrast, mbcT expression from the Atc-induced P606 promoter abolished mycobacterial growth ( Figure 1c) and, accordingly, reduced the amount of NAD + to less than 2%, compared to that in control cells (Figure 1d). Furthermore, it resulted in an approximately 99% loss of colony forming units (CFU) within few hours following induction (Figure 1e), and assessment of cell viability using the LIVE/DEAD BacLight stains followed by flow cytometry analysis demonstrated that the proportion of propidium iodide-permeable cells, which was very low in uninduced exponentially growing cells (Figure 1e), increased markedly by 8 or 24 h after the addition of Atc (Figure 1f-i). This confirmed the bactericidal effect of the NAD + depletion following the induction of P606-mbcT. In conclusion, providing that it is driven by a relatively strong promoter, production of MbcT in M. smegmatis can recapitulate the bactericidal effect of the toxin previously observed in Mtb.

Construction of a Dual Fluorescent Reporter System to Assess Gene Expression in Mycobacteria
In order to monitor the expression of mycobacterial genes of interest in in vitro cultures or during infection, we developed a dual fluorescent reporter system in which an integrative plasmid carries both (i) a constitutive reporter, serving as an internal standard or probe for the detection of the bacterial cells and (ii) an additional reporter, expressed from the promoter/operator region of the gene under investigation. This reporter system was constructed by multisite gateway recombination [14], using plasmid pDE43-MCS as backbone, and entry vectors carrying either a 5 , a middle or a 3 module, as depicted in Figure 2a. The 5 module consists of a fluorescent reporter, which can be either green fluorescent protein (GFP), mCherry, mTurquoise, or mVenus, expressed from the constitutive P1 promoter. The middle entry clone introduces a second fluorescent reporter gene, from the same list but different from the first one, fused to the promoter/operator region of the mbcAT operon. Finally, the 3 entry clone introduces either nothing, using an empty entry clone (pEN23A-MluI), or any gene(s) of interest, which effect on P mbcA is to be tested. Figure 2b depicts an example of such construct, pGMCS-P1-mCherry-P mbcA -GFP-mbcA, in which the mbcA gene is inserted to assess the effect of MbcA on its own transcription.
Toxins 2020, 12, x FOR PEER REVIEW 4 of 14 list but different from the first one, fused to the promoter/operator region of the mbcAT operon. Finally, the 3′ entry clone introduces either nothing, using an empty entry clone (pEN23A-MluI), or any gene(s) of interest, which effect on PmbcA is to be tested. Figure 2b depicts an example of such construct, pGMCS-P1-mCherry-PmbcA-GFP-mbcA, in which the mbcA gene is inserted to assess the effect of MbcA on its own transcription.

Auto-Repression of mbcAT
A set of three plasmids, namely pGMCS-P1-mCherry-PmbcA-GFP, pGMCS-P1-mCherry-PmbcA-GFP-mbcA, and pGMCS-P1-mCherry-PmbcA-GFP-mbcAT, was constructed and transformed into M. smegmatis mc 2 155. The resulting strains were grown until exponential phase in complete 7H9 medium and the fluorescence of mCherry and GFP were measured by flow cytometry. The untransformed mc 2 155 strain exhibited a very low background level of fluorescence ( Figure 3a). In contrast, the three transformed strains exhibited both red and green fluorescence (Figure 3b-d).
Comparison of the mean fluorescence intensities showed a similar level of fluorescence for mCherry in the three constructs, whereas the fluorescence for GFP was approximately 5-fold lower when the MbcA antitoxin was expressed, regardless of MbcT toxin co-expression ( Figure 3h). Therefore, we conclude that the transcription from PmbcA is auto-repressed by MbcA, independently of the presence of the toxin. We confirmed these observations in a qualitative assay, by observation of the bacterial cells in fluorescence microscopy ( Figure S1). In addition, we also introduced the same set of plasmids into M. tuberculosis H37Rv and observed the cells in fluorescence microscopy after fixation with pformaldehyde. As expected, in the wild-type H37Rv, expression of GFP remained very low in all our constructs because of the MbcA repressor produced from the chromosomal copy of the mbcAT operon. In contrast, in H37Rv ∆mbcAT::Kan R , the pattern of mCherry and GFP expression was comparable to that observed in M. smegmatis, demonstrating that autoregulation occurs similarly in the two mycobacterial species.

Auto-Repression of mbcAT
A set of three plasmids, namely pGMCS-P1-mCherry-P mbcA -GFP, pGMCS-P1-mCherry-P mbcA -GFP-mbcA, and pGMCS-P1-mCherry-P mbcA -GFP-mbcAT, was constructed and transformed into M. smegmatis mc 2 155. The resulting strains were grown until exponential phase in complete 7H9 medium and the fluorescence of mCherry and GFP were measured by flow cytometry. The untransformed mc 2 155 strain exhibited a very low background level of fluorescence ( Figure 3a). In contrast, the three transformed strains exhibited both red and green fluorescence (Figure 3b-d).
Comparison of the mean fluorescence intensities showed a similar level of fluorescence for mCherry in the three constructs, whereas the fluorescence for GFP was approximately 5-fold lower when the MbcA antitoxin was expressed, regardless of MbcT toxin co-expression ( Figure 3h). Therefore, we conclude that the transcription from P mbcA is auto-repressed by MbcA, independently of the presence of the toxin. We confirmed these observations in a qualitative assay, by observation of the bacterial cells in fluorescence microscopy ( Figure S1). In addition, we also introduced the same set of plasmids into M. tuberculosis H37Rv and observed the cells in fluorescence microscopy after fixation with p-formaldehyde. As expected, in the wild-type H37Rv, expression of GFP remained very low in all our constructs because of the MbcA repressor produced from the chromosomal copy of the mbcAT operon. In contrast, in H37Rv ∆mbcAT::Kan R , the pattern of mCherry and GFP expression was comparable to that observed in M. smegmatis, demonstrating that autoregulation occurs similarly in the two mycobacterial species.
Analysis of the MbcA sequence predicts a helix-turn-helix motif between amino-acids 20 and 40, forming a putative DNA binding domain. In order to test whether this motif is important for repression by MbcA, the Arg 33 and Arg 37 residues were substituted with glutamic acid by site-directed mutagenesis of pGMCS-P1-mCherry-P mbcA -GFP-mbcA, and the resulting plasmids were transformed into M. smegmatis mc 2 155. Both modifications abolished the repression by the MbcA variants, leading to an expression level comparable to that observed in the absence of MbcA (Figure 3h). When the R 33 E substitution was introduced in the plasmid pGMCS-P1-mCherry-P mbcA -GFP-mbcAT, the co-expression of the toxin MbcT with the R 33

Genetic Dissection of the mbcAT Promoter/Operator Region
In the plasmids described above (Figure 3), the mbcAT promoter is located on a 261-bp fragment (P mbcA(261) -GFP, labeled in red letters in Figure 4a) amplified with the oligonucleotides Fw1/Rv1 (Table S1). A shorter 103-bp fragment, extending between oligonucleotide Fw2 and Rv1, also contains the entire promoter and operator, because it produced the same pattern of expression (P mbcA(103) -GFP, labeled in green letters in Figure 4b). A putative -10 element of σ A -dependent promoters is present upstream from the mbcA transcription start site identified by RNA-seq experiments [15]. We confirmed this promoter location by generating two 239-bp DNA fragments amplified with the oligonucleotides Fw1/Rv2, with the wild-type sequence and Fw1/Rv3, with a mutation changing a consensus T into a non-consensus G in the -10 element. Whereas the first fragment still allowed the expression of GFP (P mbcA(239) -GFP, labeled in blue letters in Figure 4b), the T=>G substitution almost completely abolished the production of the downstream encoded GFP (P mbcA(239-G) -GFP, labeled in brown letters in Figure 4b). In addition, we observed that the Fw1/Rv2 fragment was much less sensitive to repression by MbcA. To further analyze the mbcA operator, we constructed mutations of 4 (Mut1; changing CAAA into TCGT) or 3 nucleotides (Mut2; changing ACA into GTG), and we observed that Mut1 resulted in a high expression in the presence of MbcA, whereas Mut 2 did not abolish repression by MbcA. In the plasmids described above (Figure 3), the mbcAT promoter is located on a 261-bp fragment (PmbcA(261)-GFP, labeled in red letters in Figure 4a) amplified with the oligonucleotides Fw1/Rv1 (Table  S1). A shorter 103-bp fragment, extending between oligonucleotide Fw2 and Rv1, also contains the entire promoter and operator, because it produced the same pattern of expression (PmbcA(103)-GFP, labeled in green letters in Figure 4b). A putative -10 element of σ A -dependent promoters is present upstream from the mbcA transcription start site identified by RNA-seq experiments [15]. We confirmed this promoter location by generating two 239-bp DNA fragments amplified with the oligonucleotides Fw1/Rv2, with the wild-type sequence and Fw1/Rv3, with a mutation changing a consensus T into a non-consensus G in the -10 element. Whereas the first fragment still allowed the expression of GFP (PmbcA(239)-GFP, labeled in blue letters in Figure 4b), the T=>G substitution almost completely abolished the production of the downstream encoded GFP (PmbcA(239-G)-GFP, labeled in brown letters in Figure 4b). In addition, we observed that the Fw1/Rv2 fragment was much less sensitive to repression by MbcA. To further analyze the mbcA operator, we constructed mutations of 4 (Mut1; changing CAAA into TCGT) or 3 nucleotides (Mut2; changing ACA into GTG), and we observed that Mut1 resulted in a high expression in the presence of MbcA, whereas Mut 2 did not abolish repression by MbcA.

Multistress Induction of mbcAT Expression
To assess the regulation of mbcAT expression, the plasmid pGMCS-P1-mTurquoise-P mbcA -mVenus-mbcA was introduced into wild-type Mtb H37Rv, and the resulting strain was grown to mid-exponential phase and subjected to various stress conditions. As shown in Figure 5, we could observe qualitatively that the expression from P mbcA was increased upon starvation following incubation of the cells in phosphate saline buffer (PBS), or upon treatment with H 2 O 2 or with the NO-generating reagent diethylenetriamine/nitric oxide adduct (DETA/NO).
Finally, Mtb strain H37Rv/pGMCS-P1-mTurquoise-P mbcA -mVenus-mbcA was used to infect mouse or human macrophages, and fluorescence microscopy of fixed infected macrophages demonstrated that transcription from P mbcA is stimulated upon Mtb phagocytosis by macrophages ( Figure 6).

Multistress Induction of mbcAT Expression
To assess the regulation of mbcAT expression, the plasmid pGMCS-P1-mTurquoise-PmbcA-mVenus-mbcA was introduced into wild-type Mtb H37Rv, and the resulting strain was grown to midexponential phase and subjected to various stress conditions. As shown in Figure 5, we could observe qualitatively that the expression from PmbcA was increased upon starvation following incubation of the cells in phosphate saline buffer (PBS), or upon treatment with H2O2 or with the NO-generating reagent diethylenetriamine/nitric oxide adduct (DETA/NO).
Finally, Mtb strain H37Rv/pGMCS-P1-mTurquoise-PmbcA-mVenus-mbcA was used to infect mouse or human macrophages, and fluorescence microscopy of fixed infected macrophages demonstrated that transcription from PmbcA is stimulated upon Mtb phagocytosis by macrophages ( Figure 6).

Multistress Induction of mbcAT Expression
To assess the regulation of mbcAT expression, the plasmid pGMCS-P1-mTurquoise-PmbcA-mVenus-mbcA was introduced into wild-type Mtb H37Rv, and the resulting strain was grown to midexponential phase and subjected to various stress conditions. As shown in Figure 5, we could observe qualitatively that the expression from PmbcA was increased upon starvation following incubation of the cells in phosphate saline buffer (PBS), or upon treatment with H2O2 or with the NO-generating reagent diethylenetriamine/nitric oxide adduct (DETA/NO).
Finally, Mtb strain H37Rv/pGMCS-P1-mTurquoise-PmbcA-mVenus-mbcA was used to infect mouse or human macrophages, and fluorescence microscopy of fixed infected macrophages demonstrated that transcription from PmbcA is stimulated upon Mtb phagocytosis by macrophages ( Figure 6).

Discussion
The dual fluorescent reporter system presented here, constructed using the Gateway cloning technology [12], allows for quantification and/or visualization of bacterial cells, and in parallel for following expression of a gene of interest ( Figure 2). As shown here, this system is stable and robust, since it is based on an integrative vector and exhibited only minor variations of the expression of the internal standard reporter (mCherry in Figures 3 and 4). Allowing both quantitative (Figures 3 and 4) and qualitative measurements (Figures 5 and 6), this system can be used to follow gene expression not only in bacterial cultures, or inside infected macrophages, as shown in this work, but also inside host tissues during infection of animal models [16]. Finally, it is also very versatile, since it offers a straightforward way to construct bacterial variants carrying a set of different reporters, as exemplified here with a variety of fluorescent proteins (Figure 2). We believe that it can be of broad interest to investigate gene expression and regulation in mycobacteria.
RNA-seq experiments indicated a potential transcription start site (TSS) of mbcA in M. tuberculosis (indicated as +1 in Figure 4) and located 34 nucleotides upstream from its translation initiation site (TIS) [15]. Based on bioinformatics analysis, it has been proposed that transcription of the mbcAT operon occurs from two overlapping promoters recognized by RNA polymerase holoenzyme using alternative sigma factors [17]. An upstream promoter, with a σ L -dependent -10 element (GGTTC) located at position -54 to -50 from the TIS, and a second promoter, using a σ H -dependent -10 element (CGTGTC), located at positions -50 to -45. The σ L and σ H alternative sigma factors are both present in M. smegmatis [18] and might therefore be involved in mbcAT expression in both mycobacterial species used in this study. Our data do not rule out the possibility of mbcAT expression from σ L -or σ H -dependent promoters, but they demonstrate that yet another -10 element is essential for the expression of mbcAT (Figure 4), which sequence TGTCATAAT fits (nucleotides in bold letters) the consensus of σ A -dependent promoters with an extended -10 motif (TGNTANNNT [15]). The involvement of this -10 element in mbcAT expression has been confirmed by our genetic analysis since its activity was abolished by a mutation changing a consensus base of the -10 element into a non-consensus one, at position -40 with respect to the TIS (Figure 4).
Many type II TA systems are regulated by auto-repression, with the antitoxin binding to an operator site in the vicinity of the operon promoter [2,4]. It is also the case for the mbcAT operon (Figures 3 and 4). In solution, MbcA and MbcT form a heterododecameric complex (3 × [MbcT-MbcA] 2 ) arranged around a 3-fold symmetry axis in which MbcA folds into a single structured domain consisting of seven α helices [1]. Helices α2 and α3 form a putative helix-turn-helix motif, and Arg 33 and Arg 37 , which are pointing out of helix 3, could be important for the recognition of the operator site. Consistently, substitutions of any of these two arginine residues totally abolished the transcription repression by MbcA (Figure 3). The mbcA promoter region harbors a direct repeat of two GACAAA motifs, separated by 12 bp (underlined with a magenta line in Figure 4a). Deletion of the downstream motif on promoter fragment Fw1-Rv2 abolishes the repression by MbcA and mutation Mut1, which disrupts this downstream motif, increases transcription in the presence of MbcA to the level observed in absence of the repressor. In contrast, modification of nucleotides downstream from this motif (mutation Mut2) did not relieve the repression by MbcA (Figure 4b). Although additional biochemical experiments are needed to determine the precise interaction of MbcA with its operator site, these observations suggest that MbcA could bind the GACAAA direct repeats and that both repeats are indispensable to constitute a functional operator.
Several autoregulated type II TA systems are subject to the so-called conditional cooperativity, i.e., a repression mechanism in which the antitoxin-repressor can participate in different types of complexes, alone, or with its cognate toxin, exhibiting different repression efficacies [2,4,19]. Conditional cooperativity has been proposed to contribute to several aspects of the TA systems biology [2]. In such systems, the antitoxin alone binds its operator site with low affinity, resulting in poor repression. When toxin and antitoxin are present in stoichiometric amounts, binding of the toxin to the antitoxin increases its affinity for the operator, leading to a more potent repression of transcription than is observed for the antitoxin alone. Finally, high amounts of toxin can result in a different heteromeric complex exhibiting low affinity for the operator, thereby relieving repression in excess of toxin. As already shown for a number of TA systems [20], our data suggest that the MbcTA system is not subject to conditional cooperativity, since the level of transcription repression by the antitoxin appeared independent on the production of the toxin (Figure 3). We note that in the dodecameric MbcTA complex, the helix-turn-helix motif of MbcA, necessary for the repression (Figure 3), is masked by an interaction with the α6 helix of MbcT [1]. In consequence, the dodecameric complex is most likely unable to participate in repression. In the wild-type H37Rv M. tuberculosis strain, the ectopic expression of MbcT under the control of the P1 promoter revealed not to be toxic [1], indicating that expression of the wild-type mbcAT operon leads to higher amounts of MbcA antitoxin than of MbcT toxin. Therefore, only a fraction of MbcA must be is involved in MbcT-neutralizing complexes, and it is likely that the remaining excess of MbcA can function as auto-repressor.
Previous work has already reported increased expression of mbcAT in stress conditions. The MbcT toxin (Rv1989c) has been identified by proteomic analysis in culture filtrates of PBS-starved M. tuberculosis cells, whereas it is not detectable during exponential growth [21]. Transcriptomic studies demonstrated increased amounts of mbcA and/or mbcT mRNA in stationary compared to exponential phase [15], during the so-called enduring hypoxic response [22], or in M. tuberculosis persisters surviving a treatment with D-cycloserine [8]. Our data are in agreement with these observations, and further demonstrate that transcription of mbcAT increases upon exposure to oxidative stress caused by H 2 0 2 or NO-generating compounds ( Figure 5), conditions mimicking those encountered inside macrophages [23][24][25]. In agreement, we also observed increased transcription from the mbcAT promoter within infected macrophages ( Figure 6). To date, the physiological role of the MbcTA system has not been firmly established. However, the existence of a gene mimicking the antitoxin MbcA in the genome of several myco-bacteriophages [1] strongly suggests that this system could belong to the family of TA systems used to fight against phage infections [26,27]. The multi-stress induction of mbcAT, leading to accumulation of both the toxin and the antitoxin, should not be toxic for the M. tuberculosis cells. However, it is reminiscent of the so-called general stress response observed in many bacterial species, which results in the expression of a variety of adaptive responses to stresses not yet encountered by bacterial cells but that contribute to reinforce their resistance if they eventually happen to deal with one of these stresses [28].
The M. tuberculosis genome is particularly rich in TA systems [7][8][9][10][11], and, as already proposed for other bacterial pathogens [29,30], it is tempting to speculate that at least some of these systems could be used as novel tools in adjunct therapies with classical antibiotics. However, a number of studies have proposed that activation of TA systems may trigger a non-growing physiological state, or so-called persistence, in which the bacterial cells become tolerant to antibiotics [31][32][33][34]. Although the role of TA systems in the development of tolerance is still a matter of debate [35], this raises a serious drawback to their potential therapeutic use. The MbcTA system attracted our attention because it is one of the only two TA systems of M. tuberculosis that harbor an essential antitoxin gene, i.e., which cannot be genetically disrupted [36]. We reasoned that this could be the consequence of an atypical and particularly toxic activity of MbcT. Indeed, we have demonstrated that MbcT activity, when expressed in recombinant conditions generating physiological expression levels of the toxin, is bactericidal for M. tuberculosis through depletion of NAD + [1], a metabolite known to be essential for mycobacteria survival, even in a dormant state [37][38][39]. In addition, we have shown that production of the MbcT toxin during infection could act in synergy with isoniazid treatment to reduce the bacterial load in mice infected by M tuberculosis, supporting the idea that MbcTA could be a novel therapeutic target [1]. The observations reported here strengthen this proposal, given that the induction of mbcAT during infection ensures the presence of the toxin inside the infecting M. tuberculosis cells, where targeting its activation could help fighting against TB.

Oligonucleotides and Plasmids Used in This Work
Oligonucleotides and plasmids used in this work are listed in Tables S1 and S2, respectively. Plasmids pEN-and pGMCS-were constructed by multisite gateway recombination [14] following the manufacturer's instructions (Life Technologies, Carlsbad, CA, USA). DNA fragments were amplified by PCR, using Phusion DNA polymerase (Thermo-Fisher Scientific, Waltham, MA, USA) with the templates and oligonucleotide pairs indicated in Table S1. Plasmids pEN41-and pEN12-were generated by "BP" cloning, using pDO41A-or pDO221A-as destination vectors, and DNA fragments flanked by attB4/attB1 or attB1/attB2, respectively. Plasmids pGMCS-TetR-P1-mbcT, pGMCS-TetR-P606Pld-mbcT and pGMCS-TetR-P606-mbcT were constructed by "LR" cloning reactions, using plasmid pDE43-MCS as destination vector, plasmid pEN41-TetR as 5 entry clone, plasmids pEN12-P1, pEN12-P606Pld, or pEN12-P606 as middle entry clone, and plasmid pEN23-mbcT as 3 entry clone. Plasmids pGMCS-P1-reporter1-P mbcA -Reporter2 were constructed by "LR" cloning reactions, using plasmid pDE43-MCS as destination vector, a plasmid pEN41-P1-Reporter as 5 entry clone, a plasmids pEN12-P mbcA -reporter as middle entry clone, and plasmids pEN23-MluI, pEN23-mbcA, or pEN23-mbcAT as 3 entry clone. Site directed mutagenesis of plasmids was performed by amplification by PCR of the plasmid to be mutated with a pair of overlapping oligonucleotides carrying the mutation to be introduced (Table S1). The linear plasmid DNA was purified on agarose gels and circularized using HD In-Fusion cloning reactions (Takara, Kusatsu, Japan). All plasmid constructs were verified by DNA sequencing.

M. Smegmatis Viability Assays and NAD + Measurement
Exponentially growing cultures (OD 600 between 0.05 and 0.2) of M. smegmatis mc 2 155 transformed with the desired plasmids were divided in two: half was left in the same medium (uninduced cultures) and the other half was treated with 200 ng mL −1 of Atc to induce expression from the tetracycline-inducible promoters. After various time post-induction, samples were harvested and centrifuged to eliminate residual Atc. Cells were resuspended in PBS buffer and dilutions were plated on L Agar supplemented with streptomycin, colonies were counted after 3 days at 37 • C. For labeling with LIVE/DEAD ® BacLight dyes (Molecular Probes, Eugene, OR, USA), cells were harvested 8 or 24 h post-Atc induction. Cells were centrifuged, resuspended in PBS buffer and stained as recommended by the manufacturer. Labeled cells were analyzed by fluorescence activated cell sorting using a BD FACS LSRFortessa X20 flow cytometer. Flow cytometry data analysis were performed using FlowJo software (Version 10; Becton, Dickinson and Company, Ashland, OR, USA).
Samples of the above cultures were harvested 8 h post-Atc induction and centrifuged. Cells were resuspended in PBS, adjusting the OD 600 to 0.2. 0.1 µm-diameter glass beads were added to tubes containing 500 µL of each cell suspensions, and cells were lysed by four 60-s pulses at full speed in a bead-beater device. The samples were centrifuged for 1 min at 20,200× g and 50 µL of the lysates were mixed with an equal volume of NAD/NADH-Glo™ Detection Reagent (Promega). Luciferin Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6651/12/5/329/s1, Figure S1: Auto-repression of the mbcAT operon expression in M. smegmatis and M. tuberculosis, Table S1: Primers used for plasmid constructions, Table S2: Plasmids used in this work.