|Home | About | Journals | Submit | Contact Us | Français|
Tightly regulated gene expression systems represent invaluable tools for studying gene function and for the validation of drug targets in bacteria. While several regulated bacterial promoters have been characterized, few of them have been successfully used in mycobacteria. In this article we describe the development of a novel repressible promoter system effective in both fast- and slow-growing mycobacteria based on two chromosomally encoded repressors, dependent on tetracycline (TetR) and pristinamycin (Pip), respectively. This uniqueness results in high versatility and stringency. Using this method we were able to obtain an ftsZ conditional mutant in Mycobacterium smegmatis and a fadD32 conditional mutant in Mycobacterium tuberculosis, confirming their essentiality for bacterial growth in vitro. This repressible promoter system could also be exploited to regulate gene expression during M. tuberculosis intracellular growth.
Mycobacterium tuberculosis is still one of the most deadly microorganisms as it is responsible for ~2-million human deaths every year (1). The lack of an efficient vaccine makes antimicrobial treatment the principal strategy to fight tuberculosis, but the emergence of multidrug-resistant (MDR) (2), and, more recently, extensively drug-resistant (XDR) M. tuberculosis strains (3), is endangering this strategy and makes the development of new antitubercular drugs a global urgency.
Tightly regulated gene expression systems are powerful tools for studying gene function and for validating drug targets in bacteria. Moreover, they are required to characterize the function of essential genes, which cannot be deleted without obtaining a lethal phenotype. While many regulated promoters have been characterized in both Gram positive and Gram negative bacteria, only a few of them have been successfully used in mycobacteria. The first regulated gene expression system to be widely exploited in mycobacteria is based on the acetamide-inducible promoter of Mycobacterium smegmatis. It is very complex, based on two positive and one negative regulators, and can be induced by the addition of acetamide to the growth medium (4). It has been successfully used to overexpress proteins and to characterize conditional mutants in both M. smegmatis and M. tuberculosis, but it presents some major weaknesses such as the large size of its operator (~1.4 kb), sensitivity to catabolite repression, low stringency, and unfeasibility in experimental models of infection (5,6). Alternative systems for modulating gene expression in M. tuberculosis are those based on the TetR repressor (7–10), the Streptomyces coelicolor pristinamycin I repressor Pip (11), and the Rhodococcus rhodochrous nitrilase repressor NitR (12).
TetR-based systems are the most widely used in mycobacteria: the gene under investigation is placed under transcriptional control of a promoter containing one or two TetR operators (tetO). In the absence of tetracycline (Tc), TetR binds to tetO preventing the expression of the gene under investigation. On addition of Tc, TetR is inactivated and dissociates from tetO, leading to expression of the downstream gene. Since Tc and its less toxic derivative anhydrotetracycline (ATc) can cross biological membranes, while their analog doxycycline can be easily administered to animals, TetR-based systems can be successfully employed to control gene expression during intracellular growth and experimental infections. In 2005, three different groups independently reported the development of a Tc-inducible system for mycobacteria, and showed that it could be used to construct conditional mutants of M. tuberculosis as well as to characterize them during experimental infection (7–9). Even though these tools proved to be quite useful, they have the disadvantage that Tc functions as an inducer of gene expression and has therefore to be removed to silence the gene under investigation. For this reason Guo et al. (13) developed a modified system using a mutated TetR able to bind tetO only in the presence of Tc (revTetR). In this case repression could be achieved by addition, rather than removal, of Tc from the culture medium. This approach, even if less stringent than the classical TetR system, was successfully applied to study a secA1 conditional mutant of M. smegmatis (13) and a conditional mutant of M. tuberculosis affected in its ability to express the α and β subunits of the proteasome (14). This approach was recently improved by adapting the revTetR coding sequence to mycobacterial codon usage, but still requires the repressor gene to be provided on a multicopy plasmid to obtain efficient gene repression in slow growing mycobacteria (15).
Here, we report the development of an alternative promoter repressible system based on two different repressors (TetR and Pip) encoded at the chromosomal level. In this approach Tc-based induction of the Pip-encoding gene allows tight repression of the gene under investigation. We also show that this system is suitable for the construction of conditional mutants in essential genes of both M. smegmatis and M. tuberculosis.
The following bacterial strains were used: Escherichia coli Top10 (Invitrogen), E. coli DH5α and HB101 (laboratory stock), M. smegmatis mc2155 and M. tuberculosis H37Rv (laboratory stock).
Escherichia coli strains were grown at 37°C in Luria–Bertani (LB) broth or on LB agar plates. Mycobacterial strains were grown at 37°C in Middlebrook 7H9 broth (Difco) in 150 ml roller bottles with slow rotation (3 rmp) or 7H10 agar plates (Difco), supplemented with 0.2% glycerol and 0.05% Tween-80. For growth of M. tuberculosis, the medium was supplemented with 10% ADN (Albumin, Dextrose, NaCl) (16).
When needed, antibiotics were added to the media at the following concentrations: streptomycin (Sm): 20 µg ml−1; kanamycin (Km): 50 µg ml−1; hygromycin (Hyg) 150 µg ml−1 (E. coli) or 50 µg ml−1 (M. smegmatis and M. tuberculosis). ATc (Sigma) was added when required at final concentrations from 1 to 200 ng/ml. Pristinamycin I, a gift of Florence Bordon-Pallier (Sanofi-Aventis), was added when required at concentrations ranging from 20 to 2000 ng/ml.
Preparation of electrocompetent cells, electroporation and preparation of mycobacterial genomic DNA were performed as previously described (17).
pMY718 is an integrative plasmid containing the PfurA102 promoter (11) upstream of the gene encoding the Pip repressor, and the Pip-dependent Pptr promoter upstream of a promotorless lacZ gene. The region containing PfurA102 and pip was subjected to oligonucleotide-directed PCR mutagenesis as previously described (7) to introduce two TetR operators using the following primers (underlined sequences represent the operator sequences; HindIII restriction sites in RP558 and RP557 are shown in bold, while the PacI site of RP558 is shown in italics) (Supplementary Figure S1):
The mutagenized PCR product was cloned into pCR-Blunt II-TOPO (Invitrogen) and subcloned into pMY718 (an integrative vector able to integrate at the unique attB site of mycobacteriophage L5) (18) using HindIII to replace the PfurA102-pip region to obtain pFRA40 (Supplementary Figure S2, Supplementary Table S1).
The Tn10-derived tetR gene, obtained by NotI/EcoRI digestion of pMC1s (7) was cloned into the unique PacI site present in pFRA40, to obtain pFRA42A and pFRA42B, differing only for tetR orientation (Supplementary Figure S3, Table S1). The correct orientation of the two plasmids was verified by restriction analysis and DNA sequencing. Both plasmids were finally introduced in M. smegmatis and M. tuberculosis to obtain MS82, MS83, TB38.1 and TB38.2 respectively (Supplementary Table S2).
The 180-bp DNA fragment containing the ptr promoter region was amplified from pMY718 using primers RP707: 5′-AGATCTCCATCCTGACGGATGGCCTT-3′ and RP708: 5′-AGATCTATGCATTCAGGCTCCTTGTACGGTGTAC-3′ and cloned in pCR-Blunt II-TOPO. The resulting plasmid was digested with BglII and the fragment containing the promoter was inserted into a Hyg cassette-containing pBend-derived suicide vector (19) to obtain pFRA50. Subsequently, the first 704 bp of M. smegmatis ftsZ, and the first 754 bp of M. tuberculosis fadD32 were cloned downstream of the ptr promoter region in the suicide plasmid pFRA50 to obtain pFRA53, and pFRA51, respectively (Supplementary Table S1).
M. smegmatis mc2155 and M. tuberculosis H37Rv were electroporated with 2 µg of the two plasmids, in order to replace by insertional duplication the promoters of ftsZ or fadD32 with the Pip controlled promoter Pptr (Supplementary Figure S4). Recombinants were selected on 7H10 agar plates containing Hyg (50 µg ml−1). Integration of the plasmid via insertional duplication was confirmed by PCR. The resulting conditional mutant strains were finally transformed with the integrative plasmid pFRA42B (containing the TetR/Pip system) to obtain MS98 (M. smegmatis ftsZ conditional mutant) and TB47 (M. tuberculosis fadD32 conditional mutant) (Supplementary Table S2).
A single colony of MS98 was grown for ~20 generations on solid medium. Bacteria were resuspended, diluted and then plated with or without ATc to determine the viable counts. The proportion of mutants escaping ATc repression/generation was calculated using the following formula: percentage of mutants/(100 × number of generations) (20).
The level of repression after addition of ATc was determined in M. smegmatis MS82 and MS83. Their parental strain mc2155 was used as a control.
Similar experiments were performed with M. tuberculosis TB38.1 and TB38.2 using their parental strain H37Rv as a control. In this case the ATc concentration used ranged from 0 to 200 ng/ml.
β-galactosidase was assayed as previously described (21) and enzymatic activity expressed in Miller units (22) using the following formula: A= (377.77 × OD420)/(time of incubation at 37°C × sample volume × protein concentration in mg/ml).
Mycobacterium tuberculosis TB47 was grown to A600 0.1 and a sample (20 ml) was taken as reference for day 0. The culture was then split into two samples: 200 ng/ml ATc was added to one of them, whereas the other served as untreated control. Both were incubated at 37°C with gentle shaking (45 rpm) and samples (20 ml) were taken at 24 and 48 h after addition of ATc. Each cell pellet was resuspended in 0.3 ml Trizol (Invitrogen) and transferred to screw-cap tubes containing 150 µl zirconia–silica beads (Biospec Products) for disruption by bead-beating. The cell suspension was then transferred to a new tube, where chloroform–isoamylalcohol (24:1) extraction was performed. RNA was precipitated by adding 1/10 volume of sodium acetate 2 M pH 5.2 and 0.7 volume of isopropanol, washed with 70% ethanol, air-dried and resuspended in DEPC-treated water. DNase treatment was carried out using RQ1 RNase-free DNase (Promega), following the manufacturer’s recommendations and the reactions were subsequently stopped by phenol–chloroform extraction and ethanol precipitation. RNA was stored at −80°C in DEPC-treated water. The amount and purity of RNA were determined spectrophotometrically.
Mycobacterium tuberculosis RNA (1 µg) was incubated with specific reverse primers for sigA and fadD32 (1 µM each) at 70°C for 5 min. After cooling on ice, 1 mM dNTPs, 40 U RNase inhibitor and 25 U Transcriptor Reverse Transcriptase (Roche) were added in a final volume of 20 µl. The reaction was incubated at 55°C for 1 h and at 85°C for 5 min to inactivate reverse transcriptase. Samples without reverse transcriptase were also set up as negative controls.
All PCR primers and probes (molecular beacons) were designed by using Primer3 software (http://frodo.wi.mit.edu/primer3/), according to previously published guidelines (23,24), and their sequences are reported in Table 1. The 20 µl PCR reaction consisted of 1× GeneAmp PCR Gold buffer (Applied Biosystems), 2.5 mM MgCl2, 0.25 mM each dNTP, 0.5 µM each primer, 2.5 ng/µl molecular beacon, 1× ROX Reference Dye (Invitrogen), 1 U of AmpliTaq Gold DNA Polymerase (Applied Biosystems) and 2 µl of cDNA. Reactions were carried out in triplicate in an Applied Biosystems 7900HT Sequence Detection System with the following protocol: denaturation at 95°C for 10 min, 15 cycles of denaturation at 95°C for 30 s, annealing at 65°C for 30 s with 0.5°C step down each cycle, extension at 72°C for 30 s, 25 cycles of denaturation at 95°C for 30 s, annealing at 57°C for 30 s with data collection, extension at 72°C for 30 s.
Parallel reactions using different amounts of H37Rv chromosomal DNA were performed for each primer-probe set in order to obtain the standard curve correlating the threshold cycle with the number of template molecules. The resulting equation was used to quantify the target cDNA for fadD32 and sigA at each time point. Results were expressed as the ratio between the number of cDNA copies of the gene of interest and the number of sigA cDNA copies in the culture exposed to ATc divided by the same ratio in an RNA preparation obtained from cultures without ATc (fadD32 cDNA + ATc/sigA cDNA +ATc)/(fadD32 cDNA − ATc/sigA cDNA − ATc) (23,24).
THP-1 cells were grown at 37°C in a 5% CO2 atmosphere and were maintained in RPMI medium (Gibco) supplemented with 10% fetal bovine serum (Gibco). After expansion, THP-1 cells were differentiated into macrophages and infected with M. tuberculosis in 96-well plates with a multiplicity of infection of 1:20 cfu/macrophage as previously described (25). Bacteria were pre-grown for 48 h in ATc-containing medium before infection. After 90 min of incubation at 37°C, the medium was removed, and cells were washed twice with 100 µl of warm PBS to remove extracellular bacteria. Finally, 100 µl of warm RPMI with or without ATc (200 ng/ml), was added to each well and the plate was incubated at 37°C. RPMI with or without ATc was replaced every 48 h.
For 8 days, every 24 h, starting from 90 min after the initial washes, the medium was removed from three wells, and then intracellular bacteria were released by lysing the macrophages with 100 µl of 0.05% SDS. The suspensions obtained from the lysed macrophages were immediately diluted in 7H9 and plated to determine viable counts. About 95% of macrophages remained viable during the entire experiment, as determined by Trypan blue exclusion.
In order to develop a repressible mycobacterial promoter system, we constructed two integrative plasmids containing (i) the Tn10-derived gene encoding the tetracycline-sensitive transcriptional repressor TetR; (ii) the S. coelicolor gene encoding the pristinamycin I-sensitive transcriptional repressor Pip, and (iii) the reporter gene lacZ. In these constructs tetR is transcribed from a constitutive promoter, pip from a TetR-dependent promoter, and lacZ from a Pip-dependent promoter (Supplementary Figure S3). The tetR repressor was obtained from the TetR-dependent inducible system previously developed by Ehrt et al. (7), while pip and the Pip-dependent promoter were taken from the pristinamycin I-inducible system previously developed by Forti et al. (11). The TetR-dependent promoter was constructed by introducing two TetR operators (tetO) into the constitutive promoter PfurA102 (11,26) (Supplementary Figure S5). The two integrative plasmids (pFRA42A and pFRA42B), differing in the orientation of the genes encoding the two repressors, were introduced by electroporation into both M. smegmatis and M. tuberculosis to obtain the strains MS82, MS83, TB38.1 and TB38.2, respectively.
In the absence of Tc (or its less toxic derivative ATc), TetR can bind to its operators turning off pip transcription, and allowing lacZ expression. However, in the presence of Tc (or ATc), inactivation of TetR allows transcription of pip, resulting in repression of lacZ (Figure 1).
In order to evaluate the TetR/Pip OFF regulatory circuit, the M. smegmatis strains MS82 and MS83, containing the two different constructs, were plated on 7H10 plates containing X-gal and increasing concentrations of ATc. As shown in Supplementary Figure S6, both strains displayed clear β-galactosidase activity when grown in the absence of ATc, which decreased in the presence of 10 ng/ml of ATc and almost disappeared when the concentration was 50 ng/ml.
The system was subsequently tested in liquid medium. Bacteria were first grown for 12 h with or without 50 ng/ml ATc. Cultures grown with ATc were then diluted into fresh media containing the same ATc concentration, while cultures grown without ATc were diluted into fresh media with or without ATc. This allowed us to obtain cultures continuously exposed to ATc for 6 or 18 h, or not exposed at all. Bacteria were lysed and their β-galactosidase activity was determined. Results, shown in Figure 2A, clearly indicate that after 18 h of repression (~7.2 generations), β-galactosidase activity decreased ~65.6- (MS82) or 24.5-fold (MS83), with respect to the untreated sample, and was comparable to that of the negative control (M. smegmatis mc2155), while after 6 h of growth in the presence of ATc (~2.4 generations), it decreased ~5.1- (MS82) or 4.8-fold (MS83). These data, while suggesting that some time is needed to titrate out β-galactosidase, clearly indicate that tight lacZ repression is obtained since the beginning of exposure to ATc.
A similar experiment was performed in M. tuberculosis. In this case, bacteria pre-grown without ATc were diluted and grown with or without 200 ng/ml of ATc for 48 h (~2.6 generations). Here, we detected repression of β-galactosidase activity of ~7-fold for one construct and ~12-fold for the other (Figure 2B).
In both species, the activity of pFRA42A was higher than that of pFRA42B suggesting that the different orientation of pip and tetR interferes with pip expression.
In order to evaluate the ATc concentration-dependent repression of β-galactosidase, the M. smegmatis strain MS82 was grown overnight with 50 ng/ml of ATc, washed and then diluted into fresh media with different ATc concentrations. Every 24 h, an aliquot of the culture was collected for β-galactosidase activity measurement, while the culture was diluted into fresh media containing the same ATc concentration. Results, shown in Figure 3A, clearly indicate that 24 h after ATc removal some β-galactosidase activity was detectable only in cultures with ATc concentrations lower than 9 ng/ml. After 48 h this activity increased in all samples to remain stable for at least another 24 h. The level of activity was dependent on the ATc concentration. Activity in cultures grown with ATc concentrations higher than 3 ng/ml was comparable with that of the negative control (M. smegmatis mc2155) during the whole experiment showing that the TetR/Pip OFF system is extremely sensitive to very low doses of ATc and can be used to fine-tune gene expression.
A similar experiment was run with M. tuberculosis TB38.2. In this case bacteria were grown for 72 h with 200 ng/ml of ATc, washed and then diluted into fresh media containing different ATc concentrations. Every 48 h an aliquot of the culture was collected for β-galactosidase activity measurement, while the culture was diluted into fresh media containing the same ATc concentration. As shown in Figure 3B, in cultures exposed to 200 ng/ml of ATc, β-galactosidase activity remained low and, at least at the two later time points (96 and 144 h), when steady state was reached, it was comparable to that of the negative control (M. tuberculosis H37Rv). However, the culture exposed to 50 ng/ml of ATc showed higher β-galactosidase activity, which remained stable for the entire experiment. In cultures exposed to 10 ng/ml of ATc, β-galactosidase activity increased during the entire course of the experiment, suggesting that 10 ng/ml of ATc are not enough to obtain gene repression in M. tuberculosis. Finally in cultures grown in the absence of ATc β-galactosidase activity increased between 48 and 96 h to reach steady state. The β-galactosidase regulatory window observed at the two later time points of this experiment was ~60-fold.
In strains containing the TetR/Pip OFF system, the possibility of a leaky expression of Pip could result in partial repression of Pptr even in the absence of ATc. To investigate this possibility, we performed an experiment in which we compared the β-galactosidase activity of TB38.2 to that of TB29, a strain containing pMY696 (11) integrated into the chromosome. This plasmid is a progenitor of pFRA42B (present on TB38.2 chromosome) and carries lacZ under Pptr transcriptional control in the absence of both tetR and pip (Supplementary Table S1). Cultures pregrown in the absence of ATc were diluted in fresh media without ATc (TB38.2 and TB29) or containing 200 ng/ml of ATc (TB38.2). Every 48 h and for 6 days, one aliquot of each culture was diluted into fresh medium containing the same amount of ATc as before. β-galactosidase activity was measured 96 and 144 h from the beginning of the experiment. As shown in Figure 4, TB29 level of β-galactosidase activity was higher than that of TB38.2 grown without ATc (slightly more than 2-fold) suggesting that indeed some level of Pptr repression due to leaky expression of Pip may occur in this strain. It is worth noting that in TB38.2 cultures supplemented with ATc, the levels of β-galactosidase activity were comparable at 96 and 144 h, suggesting that at these time points β-galactosidase activity reaches steady state.
One of the most important features of a repressible promoter system is the possibility to construct conditional mutants allowing the study of essential genes. To demonstrate that the TetR/Pip OFF system was suitable for this purpose, we constructed an M. smegmatis mutant where ftsZ, a well-known essential gene involved in cellular division (7), was placed under transcriptional control of the Pip-dependent promoter Pptr (Supplementary Figure S4). The mutant was then transformed with the integrative plasmid pFRA42B carrying the TetR/Pip OFF system to obtain the conditional mutant MS98. When this strain was plated on solid medium containing 50 ng/ml of ATc, no growth was detected confirming that ftsZ repression was stringent enough to prevent growth (Figure 5A). Moreover, when this conditional mutant was grown in liquid media with or without ATc, after an initial phase during which optical density of the two cultures increased at the same rate (data not shown), bacteria growing in the presence of ATc started to clump and settled at the bottom of the tube (Figure 5B). When we looked at this culture under the microscope, the presence of very long and filamentous bacilli was observed, as expected for FtsZ-depleted bacteria (Supplementary Figure S7).
Mutants escaping ATc repression appeared at a frequency of ~2.3 × 10−5 or 1.2 × 10−8/generation depending on the absence/presence of streptomycin in the culture medium, respectively, suggesting a low level of instability of the integrative plasmid carrying pip and tetR in the absence of selection.
The same method used to obtain the ftsZ conditional mutant in M. smegmatis was used to obtain a conditional mutant in M. tuberculosis H37Rv where the fadD32 operon was placed under transcriptional control of the Pip-dependent promoter (TB47). The fadD32 operon, involved in mycolic acid biosynthesis, was previously demonstrated to be essential in mycobacteria (11,27). When plated on solid media containing 200 ng/ml of ATc, TB47 was unable to grow, confirming that the fadD32 operon is indeed essential in M. tuberculosis (data not shown). The mutant was then grown in liquid media containing different ATc concentrations ranging from 50 to 200 ng/ml: as shown in Figure 6A, under these conditions the fadD32 mutant stopped growing after 48 h of incubation. Growth was restored at different times, depending on ATc concentration in the media (Figure 6A). When these cultures were plated on solid media containing 200 ng/ml of ATc, we found no colonies, suggesting that growth resumption was due to ATc instability in these conditions and not to selection of mutants (data not shown). Interestingly, during the growth inhibition phase, bacteria lost their acid fastness (Supplementary Figure S8). Finally, we repeated this experiment by exposing the culture to 200 ng/ml of ATc for 10 days, followed by viable counts. While the presence of ATc was able to inhibit TB47 growth during the entire course of the experiment, a sharp decrease (~10-fold) in viable counts was detected between 48 and 72 h from the time of addition of ATc and it remained stable until the end of the experiment (Figure 6B).
In order to correlate cell growth and expression of the fadD32 operon, the relative amount of fadD32 mRNA was measured by quantitative RT-PCR during the first two days of exposure to ATc, before cells stopped growing and viable counts decreased. As shown in Figure 7, fadD32 was repressed ~12- and 37-fold 24 and 48 h after the beginning of exposure to ATc, respectively. These data indicate that the TetR/Pip OFF system works efficiently in M. tuberculosis by shutting down transcription from the Pptr promoter in the presence of ATc.
In order to determine if the TetR/Pip OFF repressible system could be exploited in controlling gene expression in intracellular bacteria, we infected THP-1-derived human macrophages with TB47. The strain was pre-grown for 48 h in 200 ng/ml ATc before the infection. When ATc was added to the cell culture medium, intracellular growth of the fadD32 mutant TB47 was strongly inhibited (Figure 8). These findings confirmed that ATc can penetrate macrophages and switch off transcription of the fadD32 operon, which is essential for intracellular growth of M. tuberculosis.
In the TetR/Pip OFF system the addition of ATc leads to induction of pip whose gene product represses the gene under Pptr control (fadD32, in case of TB47). Since Pip is released from its operator in the presence of pristinamycin I (28), the addition of this drug to TB47 ATc-treated cultures should be able to restore growth by alleviating Pip-mediated fadD32 repression.
To validate this hypothesis we grew TB47 in the presence of 200 ng/ml ATc, and after 48 h we added different concentrations of pristinamycin I (20, 200 and 2000 ng/ml). As shown in Figure 9, in cultures which received 20 or 200 ng/ml of pristinamycin I growth was rescued in a dose-dependent manner despite the presence of ATc in the media. However, no growth restoration was seen in the culture with the highest amount of pristinamycin I (2000 ng/ml) suggesting that at this concentration pristinamycin I is toxic (further experiments confirmed this hypothesis, unpublished observation). These findings clearly confirmed our hypothesis, underlining the extreme versatility of the TetR/Pip OFF system to fine-tune gene expression in slow-growing mycobacteria.
In this article we describe the development of a novel repressible promoter system effective in both fast- and slow-growing mycobacteria. This approach is based on two chromosomally encoded repressors dependent on tetracycline (TetR) or pristinamycin I (Pip). The gene of interest is placed under transcriptional control of Pptr, a Pip-dependent promoter from Streptomyces pristinaespiralis (28). The Pip structural gene is placed under transcriptional control of a TetR-dependent promoter. Addition of Tc (or ATc) to the culture induces pip leading to repression of transcription from Pptr (Figure 1).
In order to characterize this repressible system we constructed two M. smegmatis strains in which the reporter gene lacZ was placed under Pptr transcriptional control in slightly different genetic backgrounds: in one strain tetR and pip were integrated in the same orientation, while in the other one they were integrated in opposite orientation (head-to-head, Supplementary Figure S3). In both strains β-galactosidase activity was strongly repressed in the presence of ATc both in liquid and solid media (Figure 2A and Supplementary Figure S6). Similar data were obtained when the constructs carrying pip, tetR and the reporter gene were introduced in M. tuberculosis (Figure 2B). The strains carrying pip and tetR in the same orientation (MS83 and TB38.2), when grown in the absence of ATc, had a lower basal activity, demonstrating that the reciprocal position of the two genes has an effect on the system. These strains were chosen for further experiments and the plasmid used to generate them (pFRA42B) was used to construct the conditional mutants.
A dose–response experiment in which bacteria were exposed to different ATc concentrations showed that repression of lacZ was inducer concentration-dependent in both slow- and fast-growing mycobacteria. This is of great interest, since it demonstrates that the TetR/Pip OFF system can be used to fine-tune gene expression (Figure 3).
Two mutant strains were constructed, one in M. smegmatis (ftsZ) and the other in M. tuberculosis (fadD32). Addition of ATc affected the growth of the M. smegmatis ftsZ mutant even at very low concentrations (Figure 5), leading to the filamentous phenotype typical of FtsZ-depleted cells (Supplementary Figure S7), thus confirming the predicted essentiality of ftsZ, and showing that regulation of the TetR/Pip OFF system is tight enough to obtain conditional mutants. The same strain was also used to determine the rate at which mutants escaping ATc-mediated ftsZ repression could arise. Our data indicate that the number of mutants increased if cells were not subjected to antibiotic selection for the integrative plasmid carrying tetR and pip, suggesting that its excision was the main reason for their occurrence. Since the int gene present on L5-based integrative plasmids was previously shown to cause their intrinsic instability (29), the construction of an integrative plasmid with the int gene provided in trans by a replicative, curable vector should overcome this drawback. In this way, the TetR/Pip OFF circuit would be more suitable for long-term in vivo experiments in the absence of any kind of selection.
The system also performed well in M. tuberculosis as the fadD32 mutant stopped growing 2 days after ATc addition. In this mutant, the Pptr promoter is placed upstream of the operon encoding fadD32 and two additional genes, pks13 and accD4. Therefore, addition of ATc likely shuts off the whole cluster. The growth arrest may thus be due to essentiality of fadD32 or of any other gene in the operon. When ATc concentrations lower than 200 ng/ml were used in the experiment, growth of the fadD32 mutant resumed after 4 (50 ng/ml) or 5 days (100 ng/ml) from the initial exposure (Figure 6A). However, when the bacterial culture was exposed to 200 ng/ml of ATc, growth was inhibited for at least 10 days (Figure 6B), suggesting that this concentration should be regarded as the most effective in guaranteeing long-lasting performance. These data also imply that ATc is not stable in this condition and that if its concentration falls below a critical point, bacteria that survived fadD32 depletion could resume their growth. Forti et al. (11) recently reported the construction of a similar fadD32 conditional mutant in which the operon was positively regulated by Pip. The phenotype shown by this mutant was slightly different from that displayed by our mutant: in that strain fadD32 down-regulation caused clumping and a continuous decrease in optical density from day 3–13 after pristinamycin I removal. This difference could be due either to different kinetics of fadD32 silencing between the two strains or to a different stringency of the two systems. The first hypothesis is the most probable, since a preliminary experiment showed that the level of expression of the target gene in the repressed state is the same in the two systems (data not shown). Finally, since Forti et al. grew the mutant in medium with 2 µg/ml of pristinamycin I, a concentration that we showed to be slightly toxic for M. tuberculosis (Figure 9), the cells were already suffering when pristinamycin I was removed to repress fadD32, and this could have exacerbated their phenotype.
Our data confirmed that in M. smegmatis TetR-based regulated promoter systems are more sensitive to ATc than in M. tuberculosis (7). This could be due to several reasons: the cell wall of M. tuberculosis could be more permeable to ATc than that of M. smegmatis; moreover, M. tuberculosis encodes several efflux pumps, and at least two of them have been shown to increase the MIC to tetracycline when expressed in M. smegmatis (7): these pumps could reduce the ATc intracellular concentration. Finally, experiments with M. tuberculosis are usually longer than those performed with M. smegmatis, making them more susceptible to the decrease of ATc concentration due to its instability at 37°C.
When we analyzed cell viability during the growth inhibition phase, we found that viable counts dropped ~10-fold between the second and the third day of incubation in the presence of ATc, but then remained stable for seven additional days, when the experiment was terminated (Figure 6B). Interestingly, after 72 h of incubation in ATc, most bacteria lost their acid-fastness clearly supporting the data indicating involvement of the fadD32 operon in mycolic acid biosynthesis (Supplementary Figure S8) (27). Our findings suggest that fadD32 depletion is deleterious for most of the cells, but a small percentage of them can adapt and survive, even in the absence of active replication. However, additional data are required to confirm this hypothesis.
Growth of the fadD32 mutant stopped ~48 h after exposure to ATc, thus indicating that some time is required to reduce the concentration of the operon gene products below a critical point. This is most likely the result of several factors such as the rate of ATc penetration into the cells, fadD32 mRNA half-life and protein turnover. Indeed, quantitative RT–PCR showed that the level of the fadD32 transcript was reduced ~12- and 37-fold after 24 and 48 h from ATc addition, respectively, confirming that some time is required to completely ablate the mRNA, but that some repression occurs since the beginning of the experiment (Figure 7).
Finally, to verify that the TetR/Pip OFF system could be used to silence genes during macrophage infection, the fadD32 conditional mutant was used to infect THP-1-derived human macrophages. The growth of the fadD32 conditional mutant was strongly inhibited by the presence of ATc, but we could not detect the decrease in viable counts previously observed in vitro, suggesting that repression of fadD32 might be more detrimental for M. tuberculosis during growth in axenic culture than during growth in macrophages (Figure 8). Experiments to address this issue are in progress.
One of the appealing aspects of the TetR/Pip OFF system is that, being based on two repressors sensing different inducers, it is possible to turn it off and on by adding the two molecules in sequence. We verified this possibility by designing an experiment in which the fadD32 mutant was first exposed to ATc, and then pristinamycin I was added to the culture media. Results showed that indeed the effect of ATc on the fadD32 conditional mutant could be reversed by addition of pristinamycin I (without ATc removal), and that this effect is dose-dependent (Figure 9). This novelty could be extremely useful to fine-tune gene expression without perturbing the bacterial culture with centrifugation and washing steps.
In conclusion, we have developed a stringent, stable, chromosomally located, repressible promoter system, suitable for constructing conditional mutants in both fast- and slow-growing mycobacteria. The level of repression that was obtained in M. tuberculosis was ~60-fold. This is clearly lower than that reported by Forti et al. (11) for the Pip ON system. Since in the TetR/Pip OFF system the level of β-galactosidase activity after addition of ATc at steady state reaches the value of the negative control not encoding lacZ (Figure 3B), the smaller regulatory window showed by this system with respect to the Pip ON system is probably due to a lower level of expression in its ON state (in the absence of ATc) due to leaky expression of pip. This hypothesis is supported by the fact that an M. tuberculosis strain carrying lacZ under Pptr transcriptional control, but not encoding Pip, has higher β-galactosidase activity than TB38.2 (Figure 4). While this might complicate the construction of conditional mutants, if the target gene requires higher expression levels, this seems to be a minor problem since we have already used this system successfully to obtain 10 conditional mutants in M. tuberculosis (including fadD32) and two in M. smegmatis (including ftsZ) without difficulty (unpublished results).
The main advantage of the TetR/Pip OFF system with respect to the other repressible promoter systems that have been previously developed for mycobacteria (13,15) is that it does not have to be harbored by a replicative plasmid to obtain tight repression in slow-growing mycobacteria. This is probably due to amplification of the signal, thanks to the presence of two repressors working in series. This characteristic is likely to result in higher stability of the system, making it more suitable for long-term in vivo experiments. Moreover, its uniqueness, based on two different repressors, sensing different inducer molecules, results in a better versatility with respect to the other regulated promoter systems developed in the past for mycobacteria, likely facilitating the fine-tuning of gene expression.
Supplementary Data are available at NAR Online.
European Commission (LHSP-CT-2005-018923). Funding for open access charge: European Commission, VI Framework (contract No. LHSP-CT-2005-018923).
Conflict of interest statement. None declared.
The authors thank Sabine Ehrt for pMC1s, Daniela Ghisotti for pMY696 and pMY718, Florence Bordon-Pallier (Sanofi-Aventis) for pristinamycin I and Issar Smith for helpful discussion and reading the manuscript.