Escherichia coli TOP10 and DH5α were used as strains for cloning and E. coli BL21 (DE3) was used as a host for protein expression. M. smegmatis mc²155 and M. tuberculosis H37Rv were used throughout this work. The E. coli strains were grown in Luria-Bertani (LB) broth, while the mycobacterial strains were cultured in Middlebrook 7H9 broth (Difco) containing 10% oleic acid-albumin-dextrose-catalase supplement (Becton Dickinson) and 0.05% Tween 80. All strains were grown at 37°C with shaking. Hygromycin (200 μg ml^-1 for E. coli and 100 μg ml^-1 for M. smegmatis), kanamycin (50 μg ml^-1 for E. coli and 25 μg ml^-1 for M. smegmatis), 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (Xgal 50 μg ml^-1) and sucrose (2% w/v) were used for selection or screening as appropriate. Acid-nitrosative multi-stress was induced in 7H9 buffered medium at pH 5.3 by the addition of NaNO₂ up to a final concentration of 5 mM for 5 h. Plasmids used throughout this work are listed in Supplementary Table S1. All primer sequences used in this study are listed in Supplementary Table S2.

The bioinformatic analysis of the MSMEG_3765 locus was conducted using BLAST and Clustal Omega. The analysis to identify putative TetR binding sites was conducted using MEME¹ (Bailey et al., 2009). MEME was set to find palindromic motifs with a minimum width of 6 bp and a maximum of 50 bp. MEME was set to return a maximum of three motifs.

The M. smegmatis ΔMSMEG_3765 mutant strain was isolated using a two-step homologous recombination strategy (Parish et al., 1999). A 804 bp fragment (up), containing the upstream flanking regions of MSMEG_3765 (from 3829208 to 3829941), was PCR-amplified from M. smegmatis mc²155 genomic DNA using the forward upMS3765f and reverse upMS3765r primers, with HindIII-BamHI sites respectively, to clone the up fragment into p2NIL, yielding the pFP2 plasmid. A 906 bp fragment (dw), containing the downstream flanking regions of MSMEG_3765 (from 3830395 to 3831300), was PCR-amplified from M. smegmatis mc²155 genomic DNA using the forward dwMS3765f and reverse dwMS3765r primers, with BamHI-PacI sites respectively, to clone the dw fragment into pFP2, yielding the pFP3 plasmid. To obtain the suicide delivery vector (pFP4), the PacI cassette from pGOAL19 was cloned into pFP3. pFP4 was electroporated into M. smegmatis mc²155 and single crossovers were selected using kanamycin, hygromycin and Xgal. A single blue kanamycin and hygromycin-resistant colony was streaked onto fresh media without selection, and incubated at 37°C for 3–5 days to allow for second recombination events, before selection on plates containing sucrose and Xgal. The white sucrose-resistant colonies were screened for kanamycin and hygromycin sensitivity, then analyzed by PCR to confirm the deletion in MSMEG_3765. The deletion event in M. smegmatis ΔMSMEG_3765 was verified by sequencing.

To complement M. smegmatis ΔMSMEG_3765, a DNA fragment containing the 636 bp coding sequence of MSMEG_3765 (including start and stop codons) was amplified from M. smegmatis mc²155 with forward cMS3765f and reverse cMS3765r primers. The forward primer included an optimized Shine–Dalgarno sequence (Andreu et al., 2010). The 636 bp fragment was cloned into the EcoRI-XhoI sites of pMV306hsp (Addgene#26155) to obtain pFP6, which was electroporated into M. smegmatis ΔMSMEG_3765. Recombinant clones were selected using kanamycin, and the insertion confirmed by PCR.

RNA for RT-PCR and RTq-PCR analyses were extracted from wild-type mc²155, knockout mutant and complemented strains from log-phase cultures under standard growth conditions. The bacteria were harvested, resuspended in 4 ml of RLT buffer (RNeasy kit – Qiagen), and vortexed using glass beads. Cell lysates were recovered by centrifugation and RNAs was extracted using RNeasy kit (Qiagen), according to the manufacturer’s instructions. RNA samples were treated with RQ1 DNase (Promega) for 30 min at 37°C, followed by heat inactivation. Finally, the quality and quantity of RNA was assessed using NanoDrop spectrophotometer analysis (Nanodrop, Thermo Scientific) and gel electrophoresis. Reverse transcription was performed for 15 min at 42°C in a total volume of 20 μl containing 1 μg total RNA (QuantiTect Reverse Transcription kit-Qiagen). As negative controls, samples without the reverse transcription step were used as template.

RT-PCR was performed to evaluate the transcriptional organization of the MSMEG_3765 locus. Oligonucleotides were designed to span the intergenic regions from MSMEG_3760 to MSMEG_3765 (Supplementary Figure S1). As positive controls, each region was amplified using M. smegmatis mc²155 chromosomal DNA as template. Analysis by Real time PCR was performed using SYBR green technology on a 7500 RT-qPCR system instrument (Applied Biosystems). The sigA gene was used as an internal standard for expression analysis. The PCR conditions included an initial denaturation at 95°C for 10 min, followed by 40 cycles of amplification of 15 s at 95°C, 1 min at 60°C, and 30 s at 72°C. RT-qPCR analysis was performed in triplicate, and each assay included standard curves for both internal control and target genes, obtained by amplifying serial dilutions (ratio 1:10) of the samples. Relative expression levels were normalized using sigA and calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001).

Total RNA was extracted from mid-log phase M. smegmatis wt, ΔMSMEG_3765 and complemented strains grown in 7H9/ADC at 37°C with shaking using the GTC/Trizol method (Waddell and Butcher, 2010). Samples were DNase-treated and purified using RNeasy columns (Qiagen), and the quality assessed by Agilent 2100 Bioanalyzer. Mycobacterial RNA (1 μg) was directly labeled with Cy3 fluorophore using the Universal Linkage System (ULS, Kreatech Diagnostics). Microarray hybridizations from three biological replicates (two for the complemented strain) were conducted as previously described (Tailleux et al., 2008) using an M. smegmatis microarray (Agilent Technologies) designed by the Bacterial Microarray Group at St. George’s (ArrayExpress accession number A-BUGS-39). To define the ΔMSMEG_3765 regulon, significantly differentially expressed genes were identified comparing ΔMSMEG_3765 to both wt and complemented strains using a moderated t-test (p-value < 0.05 with Benjamini and Hochberg multiple testing correction) and a >2.5-fold change. Fully annotated microarray data have been deposited in ArrayExpress (accession number E-MTAB-5869).

To identify the region containing putative MSMEG_3765-responding operators, we cloned: (1) a 277 bp DNA fragment containing 43 bp coding sequence and 234 bp upstream from the ATG codon of MSMEG_3760, using MS13f and MS13r primers; (2) a 218 bp DNA fragment containing 106 bp of MSMEG_3761 coding sequence, 36 bp of MSMEG_3762 coding sequence and 76 bp of intergenic non-coding region, using MS14f and MS14r primers. The PCR products, obtained with iProof high-fidelity Taq (Bio-Rad), were cloned into the BamHI and ApaI sites of the E. coli–mycobacteria shuttle vector pFPV27 (Valdivia et al., 1996), yielding pFP13 and pFP14 plasmids, respectively. To analyze the promoter region of the M. tuberculosis H37Rv Rv1687c gene, we cloned a 200 bp DNA fragment, containing 40 bp coding sequence and 160 bp upstream from the ATG codon of Rv1687c into the pFPV27 plasmid, using Rv10f and Rv10r primers, yielding the pFP10 plasmid.

The resulting recombinant plasmids, together with the pFPV27 negative control plasmid (empty vector) and the pFPV27hsp positive control plasmid (carrying the mycobacterial hsp60 promoter) (Dhandayuthapani et al., 1995), were electroporated separately into M. smegmatis wt and M. smegmatis ΔMSMEG_3765 cells. GFP fluorescence was measured as described previously (Carroll and James, 2009), with excitation at 490 nm and emission at 510 nm. To evaluate promoter activity of MSMEG_3762 under acid-nitrosative multi-stress, M. smegmatis (pFP14) log-phase cells were incubated for 5 h in 7H9 buffered medium at pH 5.3 and supplemented with 5 mM NaNO₂ before measuring fluorescence.

The MSMEG_3765 coding region (630 bp) was PCR-amplified from M. smegmatis mc²155 genomic DNA using eMS3765f and eMS3765r primers, and the product was cloned into the NdeI-XhoI sites of the pET-22b(+) expression vector. The recombinant plasmid, pFP5, was sequence verified and used to express and purify the C-terminally His-tagged TetR3765 protein. For expression, E. coli BL21 (DE3) cultures containing pFP5 were grown at 37°C to mid-exponential phase. Cultures were induced with 1 mM IPTG for 2 h at 37°C and harvested by centrifugation. The recombinant protein was purified using Profinity IMAC resins (Bio-Rad) following manufacturer’s instructions and concentrated using Centricon ultra-filtration spin columns (cut-off 10 kDa; Millipore) in 20 mM Tris HCl pH 8, 500 mM NaCl. The purified protein was correctly folded, as evaluated by circular dichroism spectroscopy (data not shown).

Binding assays were performed incubating increasing concentrations of purified recombinant MSMEG_3765 (from 1.5 to 10 pmol) in HEPES 40 mM (pH 8.0), NaCl 150 mM, MgCl₂ 20 mM, 10% glycerol with 0.5 pmol of MSMEG_3762 (218 bp) or Rv1687c (200 bp) upstream fragments (used in the GFP assays). The binding reaction was performed at 37°C for 20 min. The resulting DNA/protein complexes were resolved on native PAGE gels (8% acrylamide:bisacrylamide 30:1) and stained with SYBR Safe DNA Gel Stain (Invitrogen). To probe the specificity of MSMEG_3765 binding activity to MSMEG_3762 and Rv1687c upstream regions, reactions were incubated in the presence of 0.5 pmol MSMEG_3760 upstream fragment (127 bp, primers mot3760f1/MS13r) as non-specific competitor DNA. The MSMEG_3762 upstream fragment of 218 bp was refined into four shorter fragments for further EMSA analysis: (a) 133 bp containing 21 bp of the MSMEG_3761 coding sequence, 36 bp of MSMEG_3762 coding sequence and 76 bp of intergenic non-coding region (primers mot3762f1/MS14r); (b) 182 bp containing 106 bp of the MSMEG_3761 coding sequence and 76 bp of intergenic non-coding region (primers MS14f/mot3762r1); (c) 97 bp containing 21 bp of the MSMEG_3761 coding sequence and 76 bp of intergenic non-coding region (primers mot3762f1/mot3762r1); (d) 36 bp putative TetR binding motif.

Mycobacterium tuberculosis H37Rv Rv1685c and M. smegmatis MSMEG_3765 TetR transcriptional regulators share 71% amino acid identity. In order to identify their distribution among other mycobacteria, we searched for orthologs in other species. The regulator is conserved in mycobacteria with amino acid identities ranging from 62 to 73% (Figure 1A). In all species, the regulator is preceded by two genes in the same transcriptional orientation, annotated as an ABC transporter ATP-binding protein and an ABC transporter, that also share high percentage of identity between species (68–79%).

In order to define the transcriptional unit(s) in the MSMEG_3765 locus, an RT-PCR analysis was performed on total RNA extracted from log phase M. smegmatis mc²155 using oligonucleotide pairs designed to detect transcripts of individual or co-transcribed genes. MSMEG_3762, MSMEG_3763 and MSMEG_3765 were found to be co-transcribed defining the MSMEG_3762/63/65 operon (Figure 1B and Supplementary Figure S1). The analysis indicated that MSMEG_3760 and MSMEG_3761, were not part of the 3 gene operon but were transcribed as monocistronic units (Supplementary Figure S1).

Many TetR-like proteins have been shown to regulate adjacent genes (Balhana et al., 2015). RTq-PCR analysis was used to determine the effect of deletion of MSMEG_3765 on the expression of the surrounding genes. A M. smegmatis mutant strain carrying a 495 bp deletion in MSMEG_3765 was generated, ΔMSMEG_3765. The deletion was also complemented using the integrative recombinant pFP6 plasmid, harboring the MSMEG_3765 coding sequence under the control of the hsp60 promoter. Deletion or complementation of MSMEG_3765 did not affect growth in standard conditions, with a doubling time of ∼3 h for wt, mutant and complemented strains. Transcription of the MSMEG_3762/63/65 operon (as assessed by measuring the first gene in the operon MSMEG_3762) was upregulated by 55-fold in the mutant strain versus the wt (Figure 2A), whereas the expression of MSMEG_3760 and MSMEG_3761 was unchanged. The expression of the MSMEG_3762/63/65 operon was restored to wt levels in the complemented strain (Figure 2B). These data strongly suggest that MSMEG_3765 acts as a repressor of its own operon but does not control the expression of MSMEG_3760 or MSMEG_3761.

Genome-wide transcriptional profiling was applied as an unbiased approach to further describe the MSMEG_3765 regulon. The MSMEG_3765 regulon was defined as genes significantly differentially expressed in M. smegmatis ΔMSMEG_3765 compared to both wt and complemented strains to control for the possibility of polar effects of gene manipulation. Microarray analysis confirmed that MSMEG_3765 acts as a repressor of MSMEG_3762 and MSMEG_3763 with both genes induced in the ΔMSMEG_3765 genetic background (Table 1).

TetR proteins often bind palindromic regulatory sequences, therefore we searched for palindromic motifs upstream of the MSMEG_3762/63/65 operon. A conserved 36 bp region, containing a 34 bp palindrome, at the 5′ intergenic region extending into the coding sequence of M. tuberculosis Rv1687c, M. smegmatis MSMEG_3762, M. marinum MMAR_2486, M. avium subsp. paratuberculosis MAP_1393c, M. avium MAV_3085 genes was found (Figure 3A). This motif overlaps with a putative TATA box found 10 bp upstream of the translation start site for each gene.

TetR proteins often control divergently oriented genes, therefore we searched for additional motifs in the intergenic region upstream of mpg/amt1 in M. avium paratuberculosis and M. avium and mpg/MSMEG_3760 in M. smegmatis. Two copies of a 16 bp motif containing a 6 bp palindrome were found in this region in the M. smegmatis genome but not in the other genomes (Figure 3B). The RTq-PCR data suggested that this motif was not recognized by MSMEG_3765, however, reporter assays were performed to determine the functionality of both motifs.

Several reporter strains were made by cloning the promoter regions (including the motif regions) from MSMEG_3760, MSMEG_3762/63/65, and Rv1687c/85c/85c upstream of GFP to assay promoter activity. Promoterless GFP (empty vector) and GFP under the control of the hsp60 promoter were used as negative and positive controls, respectively. Expression of GFP from the MSMEG_3760 promoter resulted in a small but significant increase in fluorescence compared to empty vector (p < 0.01), but no significant difference was observed between wt and ΔMSMEG_3765 backgrounds (Figure 4A). Conversely, expression of GFP from the MSMEG_3762/63/65 promoter showed a significant increase (3.2-fold, p < 0.0001) in the ΔMSMEG_3765 compared to wt (Figure 4B). These results suggest that the promoter region upstream of MSMEG_3760 is active but not controlled by MSMEG_3765, while the MSMEG_3762/63/65 promoter is active and negatively regulated by MSMEG_3765. Since the motif identified in this region was also present in the M. tuberculosis Rv1687c/86c/85c gene cluster, a 200 bp fragment, containing the Rv1687c upstream sequence was assayed in M. smegmatis wt and ΔMSMEG_3765 strains. Promoter activity was detected in both genetic backgrounds with a fivefold increase (p < 0.0001) in expression in the mutant strain (Figure 4C), suggesting that the regulation of this gene cluster is conserved in M. tuberculosis.

Previous reports indicated that the Rv1687c/86c/85c gene cluster is induced by acid-nitrosative stress (Cossu et al., 2013). Therefore, we measured the promoter activity of the MSMEG_3762/63/65 upstream region under the same conditions. A small but significant increase in promoter activity was detected after exposure to 5 mM NaNO₂ at pH 5.3 (Figure 5), further evidence that this TetR-regulator is involved in the response to acid-nitrosative stress in M. smegmatis and M. tuberculosis.

The promoter studies suggested that MSMEG_3765 binds to the motif upstream of the MSMEG_3762/63/65 cluster but not to the motif identified upstream of MSMEG_3760. To verify the binding activity of MSMEG_3765, we conducted an electrophoretic mobility shift assay (EMSA) with the purified recombinant protein. A preliminary experiment using the three DNA fragments upstream of the MSMEG_3760, MSMEG_3762/63/65, and Rv1687c86c/85c genes (from the GFP assays) was performed. The upstream MSMEG_3760 fragment did not show a shift of electrophoretic mobility in the presence of the MSMEG_3765 recombinant protein, while the other two DNA fragments were shifted in the presence of the protein (Supplementary Figure S2).

Electrophoretic mobility shift assay experiments were then performed using the 127 bp upstream MSMEG_3760 fragment as a negative control in co-migration experiments either with MSMEG_3762 or Rv1687c upstream sequences. Increasing input of recombinant MSMEG_3765 protein resulted in enhanced shift of the MSMEG_3762 upstream fragment (218 bp) up to saturation (Figure 6A, lanes 2–4), while no shift was detected for the negative control fragment. In addition, a MSMEG_3765 concentration-dependent shift in electrophoretic mobility was observed for the Rv1687c upstream fragment (200 bp), also up to saturation with the highest concentration of MSMEG_3765 tested (Figure 6B, lanes 2–4). The EMSA analysis shows specific binding activity of MSMEG_3765 at the upstream regulatory regions of the M. smegmatis MSMEG_3762/63/65 and homologous M. tuberculosis Rv1687c/86c/85c operons.

The 218 bp MSMEG_3762 upstream fragment was truncated for further EMSA. Three DNA fragments were analyzed: (a) a 133 bp DNA fragment depleted of 85 bp MSMEG_3761 coding sequence (truncated at the 5′ end); (b) a 182 bp DNA fragment depleted of 36 bp MSMEG_3762 coding sequence (truncated at the 3′ end); (c) a 97 bp DNA fragment depleted of 85 bp MSMEG_3761 coding sequence and of 36 bp MSMEG_3762 coding sequence (truncated at the 5′ and 3′ ends) (Figure 7A). Only the 133 bp fragment, extending up to the 12th codon into the MSMEG_3762 coding region, retained binding activity for the MSMEG_3765 protein (Figure 7B), showing that truncation of the motif at the 3′ end destroyed binding.

To further characterize the DNA motif for MSMEG_3765 binding, the 36 bp MSMEG_3762 upstream fragment (Figure 3A) and a 40 bp MSMEG_3760 upstream fragment, containing the palindromic motif shown in Figure 3B, were tested by EMSA. The former fragment showed a MSMEG_3765 concentration-dependent shift in electrophoretic mobility, while the latter fragment was not affected by the presence of the protein (Figure 7C). These data correlate with the MEME analysis (Figure 3A) and demonstrate that MSMEG_3765, acting as transcriptional repressor, binds to a 36 bp TetR-like motif extending into the coding sequence of MSMEG_3762 and M. tuberculosis ortholog Rv1687c operons, controlling expression.