A recombinant Mycobacterium tuberculosis in vitro transcription system


  • Editor: Roger Buxton

Correspondence: Sébastien Rodrigue, Département de biologie, Centre de recherche sur les mécanismes du fonctionnement cellulaire, Université de Sherbrooke, Sherbrooke, QC, Canada J1K 2R1. Tel.: +819 821 8000 ext. 2798; fax: +819 821 8049; e-mail: sebastien.rodrigue@usherbrooke.ca


In vitro transcription constitutes an important tool in the study of the regulation of gene expression. Here, we present a fast and easy procedure to prepare Mycobacterium tuberculosis RNA polymerase for in vitro transcription assays. RNA polymerase is assembled from recombinant proteins expressed in Escherichia coli, thus eliminating the need for biosafety containment facilities, and is mixed with any of the 13 M. tuberculosisσ factors. We show that the recombinant RNA polymerase is free from contaminating σ factors, produces transcriptional start sites matching those characterized in vivo and allows the formal identification of σ factors involved in the expression of genes of interest.


In spite of long-standing efforts to prevent and treat tuberculosis, this disease remains a major public health concern. It is estimated that nearly a third of the human population is infected by its causal agent: Mycobacterium tuberculosis (Dye et al., 1999). Adaptation to the various conditions encountered by this pathogen in the establishment of a successful infection is thought to require a strict gene expression program. In prokaryotes, much of this control is obtained at the transcriptional level. Bacterial core RNA polymerase (RNAP) is composed of five subunits (α2ββ′ω) (Gruber & Gross, 2003). An additional subunit, the σ factor (σ), associates with RNAP and provides the promoter recognition function to the holoenzyme complex, thus allowing the expression of a particular group of genes (Gruber & Gross, 2003). Thirteen σ factors are encoded in the genome of M. tuberculosis (Cole et al., 1998). Although some have been characterized, much remains to be determined about their physiological roles and promoter recognition specificities (Manganelli et al., 2004). Furthermore, over 140 putative transcriptional regulators are presumably involved in gene expression modulation in this pathogen (Manganelli et al., 2004).

In vitro transcription assays have been extensively used in the study of transcription regulation in prokaryotes. They provide many mechanistic insights and faithfully mimic many aspects of in vivo transcription. Most importantly, in vitro transcription allows a direct investigation as to the role of RNAP, σ factor(s), promoter DNA or any regulator(s) that may contribute to the expression of a gene of interest. However, these studies may require large amounts of high-quality RNAP that can be difficult to obtain. A major difficulty in RNAP purification lies in complete elimination of contaminating σ factors. Biosafety concerns may also arise if large amounts of pathogenic microorganisms have to be grown in order to proceed with the purification. To circumvent these potential problems, some researchers have used commercially available Escherichia coli RNAP to study transcriptional regulation in other bacteria (Raman et al., 2001; Song et al., 2003; Sun et al., 2004). Such shortcomings could have important consequences as important protein–protein interactions may not be conserved between species (Steffen & Ullmann, 1998; Mencia et al., 1998; Lohrke et al., 1999). Moreover, RNAP from different bacterial species may have distinct properties (Artsimovitch et al., 2000).

Using recombinant proteins to reconstitute a functional enzyme is an interesting alternative to conventional RNAP purification from crude extracts (Tang et al., 1995). The procedure is fast, does not require biosafety containment facilities and yields a σ factor-free preparation. Moreover, hybrid RNAP can easily be generated, and mutations that could otherwise be lethal or could severely impair bacterial growth can be introduced in any of the proteins (Tang et al., 1995; Peck et al., 2002). This work describes the preparation and the use of a recombinant in vitro transcription system for the study of gene expression regulation in M. tuberculosis.

Materials and methods

Bacterial strains, plasmids and oligonucleotides

Oligonucleotides, plasmids and bacterial strains are described in Table 1. Escherichia coli strains were grown in Luria–Bertani (LB) medium supplemented with 100 μg mL−1 ampicillin, 50 μg mL−1 kanamycin and 34 μg mL−1 chloramphenicol when appropriate. Mycobacterium smegmatis MC2155 was grown in LB media supplemented with 0.05% Tween 80.

Table 1.   Plasmids and bacterial strains used in this study
 DescriptionSource or reference
133 (reverse RpoA)5′-CGGGATCCTAAAGCTGTTCGGTTTC-3′This study
19 (reverse N-term RpoB)5′-TTGGCGCGCAGCTCCATCCCGGTGCCCACCA-3′This study
18 (forward C-term RpoB)5′-TTGGCGCGCGGCGATCGACGCCGGCGACGT-3′This study
17 (reverse C-term RpoB)5′-CGGGATCCTTACGCAAGATCCTCGACACTTGC-3′This study
16 (forward N-term RpoC)5′-GGGAATTCCATATGGTGCTCGACGTCAACTTC-3′This study
13 (reverse C-term RpoC)5′-CGGGATCCCTAGCGGTAGTCGCTGTAGCCGTA-3′This study
30 (forward SigF)5′-CATGCCATGGACGGCGCGCGCTGCCGG-3′This study
6 (reverse SigF)5′-CGGGATCCTACTCCAACTGATCCCG-3′This study
70 (forward SigG)5′-CGGGATCCATGCGCACATCGCCGATG-3′This study
71 (reverse SigG)5′-GGAATTCTCACAGCGAATCGGGCAG-3′This study
61 (reverse SigH)5′-CCGGCTCGAGTCATGACGACACCCCCT-3′This study
72 (forward sigJ)5′-CCCGAGCTCATGGAGGTTTCCGAATTC-3′This study
73 (reverse SigJ)5′-CCCAAGCTTTCAATTCCGGTGATGCCT-3′This study
74 (forward SigK)5′-CGGGATCCATGACCGGACCGCCACGG-3′This study
75 (reverse SigK)5′-CCCAAGCTTTCATGACACGTCCAGGCA-3′This study
76 (forward SigL)5′-GGAATTCGTGGCTCGTGTGTCGGGC-3′This study
77 (reverse SigL)5′-CCCAAGCTTTCATCGAGTAACTCCCAG-3′This study
78 (forward SigM)5′-GGAATTCATGCCGCCACCGATTGGT-3′This study
80 (usfX extension)5′-TCTGGTCTTCGAGCTGGTCGGTCATGGTC-3′This study
97 (sin extension)5′-TTTCGATACCCCGGATGTCATCGCATCG-3′This study
pSR52pET16b derivative containing the rpoA gene encoding RpoAThis study
pJF09pET16b derivative containing the rpoB gene encoding RpoBThis study
pJF10pET30a derivative containing the rpoC gene encoding RpoCThis study
pSR01pET30a derivative containing the Rv2703 gene encoding SigAThis study
pSR02pET30a derivative containing the Rv2710 gene encoding SigBThis study
pSR03pET30a derivative containing the Rv2069 gene encoding SigCThis study
pSR28pET30a derivative containing the Rv3414c gene encoding SigDThis study
pSR04pET30a derivative containing the Rv1221 gene encoding SigEThis study
pSR05pET30a derivative containing the Rv3286c gene encoding SigFThis study
pSR29pET30a derivative containing the Rv0182c gene encoding SigGThis study
pSR06pET30a derivative containing the RV3223c gene encoding SigHThis study
pSR07pET30a derivative containing the Rv1189 gene encoding SigIThis study
pSR30pET30a derivative containing the Rv3328c gene encoding SigJThis study
pSR31pET30a derivative containing the Rv0445c gene encoding SigKThis study
pSR32pET30a derivative containing the Rv0735 gene encoding SigLThis study
pSR33pET30a derivative containing the Rv3911 gene encoding SigMThis study
pIS109pUC19 derivative containing a modified version of the Bacillus subtilissinIR locusPredich et al. (1995)
pYZ99pUC18 derivative containing the Mycobacterium tuberculosissigF locusDeMaio et al. (1996)
Bacterial strains
Escherichia coli BL21(DE3)pLysSB F- dcm ompT hsdS(rB-mB-) gal lambda(DE3) [pLysS Camr]Stratagene (La Jolla, CA)
Escherichia coli BL21(DE3) CodonsPlus-RPB F- ompT hsdS(rB-mB-) dcm+ Tetr gal lambda(DE3) endA Hte [argU proL Camr]Stratagene
Mycobacterium tuberculosis H37RvCommon lab strain, virulent and slow growingATCC 25618
Mycobacterium smegmatis MC2155Avirulent, fast-growing and saprophytic mycobacteriumSnapper et al. (1990)

Cloning of Mycobacterium tuberculosis RNA polymerase subunits

The M. tuberculosis rpoA gene was PCR amplified from H37Rv genomic DNA using primers 11 and 133. The 1052 bp fragment was digested in NdeI–BamHI and cloned into the corresponding sites of pET16b (Novagen, Madison, WI). The resulting plasmid, pSR52, contained an N-terminus hexahistidine-tagged version of the M. tuberculosisα subunit. The M. tuberculosis rpoB gene was cloned in pET16b using a trimolecular ligation procedure. The rpoB gene was divided into two parts that were PCR amplified from M. tuberculosis genomic DNA. The first amplicon was obtained using primers 19 and 20. The second part of the gene was amplified using primer 17 and 18. These PCR products were, respectively, cleaved with NcoI–BssHII and BssHII–BamHI and cloned into NcoI–BamHI-digested pET16b. Similarly, the M. tuberculosis rpoC gene was divided into two parts, using primers 15 and 16 or primers 13 and 14. These PCR fragments were, respectively, digested with NdeI–NotI and NotI–BamHI and cloned into pET30a (Novagen) using the NdeI and BamHI restriction sites, resulting, respectively, in plasmids pJF09 and pJF10.

Purification of recombinant Mycobacterium tuberculosis RNA polymerase subunits

Purification of RNAP α subunit

Escherichia coli BL21 (DE3) pLysS was transformed with pSR52, grown in 1 L of medium at 37°C until OD600 nm 0.6–0.8 was reached. Protein production was then induced by adding isopropyl β-d-thiogalactoside (IPTG) to a final concentration of 0.5 mM, and incubating for an additional 3 h. The culture was next harvested by centrifugation (5000 g, 5 min, 4°C), and the cell pellet was resuspended in 4 mL of buffer A (20 mM Tris-HCl, pH 7.9, 500 mM NaCl, 5 mM imidazole). Cells were disrupted by sonication, and the lysate was cleared by centrifugation (21 000 g, 15 min, 4°C). Hexahistidine-tagged α subunit was precipitated by gradually adding (NH4)2SO4 to a final concentration of 60%, collected by centrifugation (21 000 g, 20 min, 4°C) and redissolved in Buffer A plus 8 M Urea. The sample was adsorbed on 2 mL Ni2+-NTA agarose (Qiagen, Valencia, CA), washed twice with 10 mL of the same buffer containing 60 mM imidazole and eluted in 4 mL of the same buffer containing 500 mM imidazole.

Production of crude β and β′ subunits

Escherichia coli BL21 (DE3) CodonPlus-RP were transformed with pJF09 (β) or pJF10 (β′) and grown in 100 mL of LB broth at 37°C to an OD600 nm of 0.6–0.8, and protein expression was induced for 3 h by adding 0.5 mM of IPTG to the medium. Cells were then harvested by centrifugation (5000 g, 5 min, 4°C), resuspended in 1 mL buffer B (50 mM Tris-HCl, pH 7.9, 10 mM MgCl2, 10 μM ZnCl2, 1 mM ethylenediamine tetra-acetic acid (EDTA), 10 mM dithiothreitol, 10% glycerol) and disrupted by sonication. The lysates were next spun (21 000 g, 15 min, 4°C), and inclusion bodies were solubilized in 500 μL Buffer B plus 8 M Urea. Protein concentrations were estimated using the Bradford assay and gel electrophoresis.

Large-scale preparation of recombinant Mycobacterium tuberculosis RNA polymerase

The reconstitution protocol is based on a previously described method using E. coli RNA polymerase subunits (Tang et al., 1995). The reconstitution mixtures contained 2000 μg (∼50 nmol) of hexahistidine-tagged α, 7500 μg (∼50 nmol) of crude β and 15 000 μg (∼100 nmol) of crude β′. The reaction volume was completed to 50 mL with Buffer C (50 mM Tris-HCl, pH 7.9, 200 mM KCl, 10 μM ZnSO4, 1 mM EDTA, 5 mM 2-mercaptoethanol, 20% glycerol) plus 8 M urea. The reconstitution reaction was next step dialyzed against ∼750 mL of Buffer C using decreasing concentrations of urea (at least 2 h for eight to 10 successive twofold dilutions of urea starting with 4 M; the last two dialysis steps were performed in the absence of urea in Buffer C). The resulting solution was next incubated at 30°C for 45 min, and then cleared by centrifugation (16 000 g, 10 min, 4°C). Reconstituted RNAP and the free α subunit were adsorbed for 45 min with mild agitation onto 3 mL of pre-equilibrated Ni2+-NTA agarose resin (Qiagen). The resin was loaded into a column, washed three times with 16 mL of Buffer D (50 mM Tris-HCl, pH 7.9, 0.5 mM EDTA, 5% glycerol) plus 5 mM imidazole and eluted with 6 mL Buffer D plus 300 mM imidazole. Reconstituted RNAP was finally concentrated to a final volume of 1 mL, and cleared of free α subunit, using a Centricon-100 filter device (Millipore, Billerica, MA). The preparation was finally mixed with one volume of 100% glycerol, aliquoted and stored at −80°C.

Preparation of recombinant Mycobacterium tuberculosisσ factors

The 13 M. tuberculosisσ subunit genes were PCR amplified from M. tuberculosis genomic DNA with appropriate oligonucleotides (see Table 1), cleaved with appropriate restriction enzymes and cloned in corresponding sites of pET30a (Novagen). Escherichia coli BL21 (DE3) pLysS were transformed with the resulting plasmid and grown at 37°C in 1 L of LB broth to an OD600 nm of 0.6–0.8. IPTG was next added to a final concentration of 0.5 M and the cultures were incubated for an additional 3 h before being harvested (5000 g, 5 min, 4°C). Cells were disrupted by sonication and the lysate was cleared by centrifugation (21 000 g, 15 min, 4°C). All σ factors were resuspended from inclusion bodies using Buffer C containing 8 M urea. Insoluble material was collected by centrifugation (21 000 g, 20 min, 4°C), and the supernatant was adsorbed on 2 mL Ni2+-NTA agarose (Qiagen), washed twice with 15 mL of the same buffer containing 10 mM imidazole and eluted in 2 mL of the Buffer C plus 8 M urea containing 500 mM imidazole. σ factors were refolded by step dialysis against prechilled Buffer C, as described in previous section.

Mycobacterium smegmatis RNAP extraction and purification

Mycobacterium smegmatis RNAP extraction and purification were performed as described previously (Beaucher et al., 2002).

In vitro transcription assays

The promoter templates for in vitro transcription assays were a 798 bp BamHI–HindIII fragment from pIS109 (Predich et al., 1995; Beaucher et al., 2002) for the sinP3 promoter and a 269 bp NheI–HindIII fragment from pYZ99 (DeMaio et al., 1996; Beaucher et al., 2002) for the usfXP1 promoter. The in vitro transcription protocol used in this study was adapted from Beaucher et al. (2002) with the following modifications. Reaction were carried out in a final volume of 40 μL containing 45 mM Tris-HCl (pH 7.9), 5 mM MgCl2, 75 mM KCl, 1 mM dithiothreitol, 250 μg mL−1 bovine serine albumin (BSA) and 13% glycerol. Unless specified, each reaction contained 22.5 ρmol RNAP and 3 ρmol σ factor. Template-specific oligonucleotides were radiolabelled using γ-AT32P and T4 polynucleotide kinase and used for primer extension according to standard procedures.

Results and discussion

Production of recombinant Mycobacterium tuberculosis core RNAP and σ factors

High-quality reagents are essential to achieve robust and specific transcription in vitro. The first step in this attempt was to clone RNAP subunits and σ factors encoding genes in appropriate plasmids. To facilitate their purification, the RNAP α subunit and all σ factors were, respectively, N-terminally polyhistidine tagged to minimize the risk of affecting RNAP activity. Indeed, the amino moiety of most σ factors is thought to be less likely to play an important role in promoter recognition or in the interaction with transcription regulators, as compared with the C-terminus region (Paget & Helmann, 2003; Gruber & Gross, 2003).

Proteins were expressed in appropriate Escherichia coli BL21 (DE3) strains. The α subunit was purified from the soluble fraction of the lysate and mixed under denaturing conditions with β and β′-containing crude extracts. An approximate nonnative molecular ratio of 1 : 1: 2 (α : β : β′) was used for each RNAP reconstitution reaction. Other subunit stoichiometries were also tested without significantly changing the activity of RNAP (data not shown). Proteins were renatured by step dialysis using several buffer changes. Reconstituted RNAP were then purified by metal ion affinity chromatography. The preparation was dialyzed to remove chemicals that could possibly interfere with proper enzyme activity. RNAP was finally concentrated and free α subunits were cleared using a centricon-100 filter device (Fig. 1a). Each reconstitution yielded approximately 2 mg of RNAP (approximately 6 nmol), which is sufficient for at least 250 reactions. The stoichiometry of the recombinant enzyme was visually estimated from Coomassie-stained gels to 2 : 1 : 1 (α : β : β′) (Fig. 1a and data not shown). σ factors were next purified to near homogeneity using a Ni2+-NTA resin under denaturing conditions and then refolded by step dialysis (Fig. 1b).

Figure 1.

 Preparation of Mycobacterium tuberculosis recombinant RNA polymerase (RNAP) and σ factors. (a) Coomassie-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel showing individual RNAP subunits and some purification steps. MW, molecular weight marker; 1, purified His-α subunit; 2, overexpressed β subunit crude extract; 3, overexpressed β′ subunit crude extract; 4, flowthrough from Ni2+-NTA agarose purification of reconstituted RNAP; 5, flowthrough from α subunit removal using a Centricon-100 filter device; 6, reconstituted M. tuberculosis RNAP core enzyme. (b) Coomassie-stained SDS-PAGE gel showing purified recombinant M. tuberculosisσ factors. Letters from A to M designate the corresponding σ factors.

The activity of recombinant RNAP preparation was assessed using an in vitro transcription assay followed by a primer extension procedure. For this purpose, the Bacillus subtilis sinP3 promoter was chosen because it is efficiently recognized by housekeeping σ factors from many bacterial species, including M. tuberculosis (Predich et al., 1995; Beaucher et al., 2002). Increasing amounts of RNAP were added to the reaction and the radioactively labelled extension products were migrated on a gel. As shown in Fig. 2, the activity of recombinant RNAP is null in the absence of the appropriate σ factor (see lanes 1, 3, 5, 7 and 9). On the other hand, the activity of RNAP increases in a dose-dependent manner when a fixed amount of σA is present in the reaction (Fig. 2, lanes 2, 4, 6, 8 and 10). However, recombinant M. tuberculosis RNAP is roughly 10-fold less active than the best fractions of purified M. smegmatis RNAP, possibly because some reconstituted RNAP are not suitably refolded. In compensation for its weaker activity, the ratio between the specific signal and the no-σ signal is clearly to the advantage of recombinant RNAP. In fact, some signals could be observed in the negative control reaction (Fig. 2, lane 11) using M. smegmatis RNAP as it is difficult to eliminate contaminating σ factor completely from the preparations (Fig. 2 and unpublished results). Interestingly, although the ω subunit was proposed to be important for proper RNAP assembly (Mukherjee et al., 1999; Ghosh et al., 2001), it is obviously not essential for RNAP activity in vitro (Tang et al., 1995; Steffen & Ullmann, 1998; Mencia et al., 1998; Lohrke et al., 1999; Peck et al., 2002). Moreover, we observed no significant increase in either the yield or the activity of reconstituted M. tuberculosis RNAP when the ω subunit was added to reconstitution reactions, either as untagged protein from inclusion bodies, purified glutathione-S-transferase-fusion or purified hexahistidine-tagged fusion protein (data not shown). However, it is possible that the conditions were not appropriate for the activity of ω, thus precluding the potential benefits. Further work will be needed to test this possibility rigorously.

Figure 2.

 Assessment of RNA polymerase (RNAP) activity. Primer extensions using the Bacillus subtilis sinP3 promoter were performed by adding increasing amounts of Mycobacterium tuberculosis recombinant RNAP. Lanes 1–2, 3–4, 5–6, 7–8 and 9–10 contain, respectively, 1.5, 3.0, 7.5, 15.0 and 22.5 pmol of reconstituted M. tuberculosis (Mtb) RNAP. Primer extension using 8 pmol of purified Mycobacterium smegmatis (Msm) RNAP is shown in lanes 11–12 for comparison. Three picomoles of σA was added to even lanes only.

As the reaction conditions are crucial for optimal RNAP activity, the influence of all buffer components was tested according to variations reported in the literature. Table 2 details the tested possibilities, and reports optimal parameters.

Table 2.   Composition of in vitro transcription buffers tested
Tris-HCl14, 24, 4545 mM
MgCl25, 155 mM
KCl25, 75, 90, 16575 mM
Dithiothreitol0.1, 0.5, 1.01 mM
Glycerol9, 13, 1913%
Bovine serum albumin0, 250250 μg mL−1

Screening σ factors important for promoter activity

Relatively few mycobacterial promoters are reported in the literature (Jacques et al., 2005). Of these, few have been studied in great detail. In fact, some promoters were associated with particular σ factors or regulators, mostly by differential transcription profiling (Raman et al., 2001; Manganelli et al., 2001, 2002; Raman et al., 2004), and a few others were mutagenized to determine which nucleotides are important for the expression of a reporter gene (Kenney & Churchward, 1996; Roberts et al., 2004). Moreover, it is often tricky to establish sequence relationships between mycobacterial promoters, and the proposed σ factor consensus sequences somehow appear degenerated. However, using M. tuberculosis recombinant RNAP and σ factors, it has now become possible to better characterize these promoter requirements.

In order to demonstrate the capacity of the recombinant transcription system, we have determined which σ factor is involved in the recognition of some previously described promoters. We have individually tested all 13 M. tuberculosisσ factors at the B. subtilis sinP3 promoter template (Fig. 3a). No signal was observed in the absence of any σ factor. However, a clear band was visible when σA was added to the reaction (Fig. 3a). Interestingly, σB also allowed the detection of a transcript, most likely from the same transcriptional start site. This apparent overlap in σA and σB specificities is not so surprising considering that σB is categorized as a principal-like σ factor and that such a redundancy has also been described between σ70 and σ38 from E. coli (Gruber & Gross, 2003). We have next repeated this procedure using the usfXP1 promoter template. σF allowed the production of the only observed primer extension signal (Fig. 3b). As this promoter has been characterized previously in vivo, we performed a high-resolution mapping of the transcription start site to verify whether the in vitro reaction faithfully corresponds to what was formerly shown (Beaucher et al., 2002). The transcription initiation site derived from the recombinant RNAP perfectly matched the in vivo site thus suggesting that our in vitro transcription system can be considered as a reliable means of studying transcriptional regulation in M. tuberculosis. Moreover, as many σ factors were also shown to be functional in this type of assay (Fig. 3 and data not shown), it is reasonable to think that all recombinant M. tuberculosisσ factors are correctly refolded and should allow transcription initiation to occur, provided that an appropriate DNA template is used. Furthermore, the mycobacterial in vitro transcription system described here could become a valuable tool in the characterization of various transcription regulators, such as activators, repressors and anti-σ factors.

Figure 3.

 Recombinant Mycobacterium tuberculosis RNA polymerase (RNAP) allows screening for σ factors involved in transcription initiation of genes of interest. (a) and (b) primer extension reactions carried out with individual M. tuberculosisσ factors on the sin and the usfX promoter, respectively. Letters from A to M indicate the corresponding σ factors. (c) High-resolution mapping of the σF-dependent transcription start site from the usfX promoter template. No σ factor was added to the reaction shown in lane 1. Lane 2 contains 3 pmol of σF. The arrows indicate the transcription start site. The initiating nucleotide is shown in bold, and the proposed promoter boxes are underlined (Beaucher et al., 2002).


The authors would like to thank Pierre-Étienne Jacques and Annie Moisan for their valuable comments on the manuscript. This work was supported by funds from the National Science and Engineering Council of Canada and from the ‘Association pulmonaire du Québec’ awarded to R. B. and L. G. L. G. holds a Canada Research Chair on mechanisms of gene transcription. S. R. is the recipient of fellowships from the National Science and Engineering Research Council of Canada and from the ‘Fonds sur la Recherche en Santé du Québec’. J.-F.J. and S.R. contributed equally to this work.