Inhibition of Mycobacterium tuberculosis RNA Polymerase by Binding of a Gre Factor Homolog to the Secondary Channel

Abstract

Because of its essential nature, each step of transcription, viz., initiation, elongation, and termination, is subjected to elaborate regulation. A number of transcription factors modulate the rates of transcription at these different steps, and several inhibitors shut down the process. Many modulators, including small molecules and proteinaceous inhibitors, bind the RNA polymerase (RNAP) secondary channel to control transcription. We describe here the first small protein inhibitor of transcription in Mycobacterium tuberculosis. Rv3788 is a homolog of the Gre factors that binds near the secondary channel of RNAP to inhibit transcription. The factor also affected the action of guanosine pentaphosphate (pppGpp) on transcription and abrogated Gre action, indicating its function in the modulation of the catalytic center of RNAP. Although it has a Gre factor-like domain organization with the conserved acidic residues in the N terminus and retains interaction with RNAP, the factor did not show any transcript cleavage stimulatory activity. Unlike Rv3788, another Gre homolog from Mycobacterium smegmatis, MSMEG_6292 did not exhibit transcription-inhibitory activities, hinting at the importance of the former in influencing the lifestyle of M. tuberculosis.

INTRODUCTION

The transcription process of RNA polymerase (RNAP) is controlled by many regulators at different steps (6, 11, 28). These regulators include both general and operon-specific factors which determine the rate and extent of transcription (2, 3, 28). The functions of these regulators range from the activation of transcription to the repression of the process under different physiological conditions. Apart from transcription factors, different small molecules and antibiotics also target the RNAP to affect transcription. Being the integral component of the essential process, RNAP is a preferred target of a number of antibiotics. The mechanism of action and the binding site for these inhibitors in the multisubunit holoenzyme is distinct. The antibiotics that inhibit RNAP prevent the extension of the nascent RNA beyond the third nucleotide (rifampin and sorangicin) (7, 8, 31), prevent open complex formation (myxopyronin) (20), or perturb mobile elements in the active center (streptolydigin) (35). The antibacterial peptide microcin J25 inhibits transcription by binding within the RNAP secondary channel and inhibiting nucleoside triphosphate (NTP) uptake (21).

A number of structurally similar proteins have been identified from different bacteria which interact with RNAP through the secondary channel. The Gre proteins are the first members of this group of factors to assist RNAP in maintaining transcription accuracy by stimulating the cleavage of aberrant 3′ ends of the RNA to resume RNA synthesis (4, 5, 9, 17). Gfh1, DksA, and TraR are structural homologs of Gre factors but do not function like Gre; instead they inhibit the transcription process. Gfh1, which is present in Thermus sp. (13, 14, 16), inhibits both transcription initiation and elongation (16, 33), while DksA of Escherichia coli inhibits some of the rRNA promoters in conjunction with guanosine pentaphosphate (pppGpp) and activates transcription from several amino acid biosynthesis operons (19, 25, 26). TraR, found in conjugative plasmids, mimics the combined function of pppGpp and DksA in both the inhibition and activation of transcription (1). The secondary channel of RNAP is utilized by this group of structurally similar proteins for directly accessing the catalytic center of the enzyme (22, 27, 30, 33) and to influence the mobile elements, the bridge helix, and the trigger loop in the RNAP active center (23, 29, 36). However, not all of the proteins that interact with RNAP at the secondary channel are assigned a specific function. For example, Rnk from E. coli has a shorter N terminus, which is insufficient for it to reach the RNAP active center, hence it probably does not influence transcription (15). Other than the Gre factors and the proteins mentioned above, no other secondary channel binding proteins have been characterized to date in any of the bacterial genomes.

In the genus Mycobacterium, which includes several pathogenic species, open reading frames (ORFs) which showed similarity to the Gre factor were identified. The most similar in the genomes of M. tuberculosis and M. smegmatis are Rv3788 (9) and MSMEG_6292, respectively. Homology modeling of the ORFs indicate that their domain organizations are similar to that of the Gre factor and hence are likely to interact with RNAP in a similar fashion. However, these proteins do not have transcript cleavage stimulatory activity. Instead, Rv3788 of M. tuberculosis, but not MSMEG_6292 of M. smegmatis, inhibits transcription by binding near the RNAP secondary channel and possibly perturbing the catalytic center.

MATERIALS AND METHODS

Sequence alignment and homology modeling.

Multiple-sequence alignments were carried out by using ClustalW (http://www.ebi.ac.uk/Tools/msa/clustalw2). The figures were generated using the GenDoc multiple-sequence alignment editor (http://www.psc.edu/biomed/genedoc). Homology models of Rv3788 (using the Thermus thermophilus Gfh1 [TthGfh1] crystal structure as the template) and MSMEG_6292 (using the Escherichia coli GreA structure as the template) were generated using the comparative protein structure modeling program Modeler (version 9.3).

Expression and purification of the proteins.

The Rv3788 gene from M. tuberculosis was cloned in pET20b vector with a C-terminal His tag between the NdeI and HindIII sites (9). E. coli BL21 cells with pET20b-Rv3788 were grown to an optical density at 600 nm (OD₆₀₀) of 0.6 and induced with 0.3 mM isopropyl-β-d-thiogalactopyranoside (IPTG). The cells were lysed by sonication and centrifuged at 100,000 × g for 2 h. The supernatants were subjected to 45 to 60% ammonium sulfate precipitation and resuspended in 3 ml of TGE buffer (10 mM Tris-HCl, pH 8.0, 5% glycerol, 0.1 mM EDTA) and purified through DEAE-Sephacel chromatography by elution with a linear NaCl gradient of 50 to 400 mM. His-tagged Rv3788 was purified from E. coli BL21 cells transformed with pET20b-rv3788his using nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography. V36W, S126E, and D45A/D48A His-tagged mutants of the protein (see Table S1 in the supplemental material) were generated by the megaprimer inverse PCR method and purified by Ni-NTA chromatography. MSMEG_6292 was PCR amplified from M. smegmatis mc²155 genomic DNA using the primers listed in Table S1 in the supplemental material and cloned into pET20b (at NdeI and HindIII sites) with a C-terminal His tag and purified by following the same method as that for His-Rv3788. M. tuberculosis and M. smegmatis Gre (MtbGre and MsGre, respectively) proteins were purified by following the methods described earlier (9). M. tuberculosis and M. smegmatis RNAP (MtbRNAP and MsRNAP, respectively), which were used for in vitro transcription assays and protein-protein interaction studies, were purified by following the method described earlier (9, 10).

Ni-NTA pulldown assays.

MtbRNAP without any tag was used for interaction analysis with His-tagged Rv3788 or MSMEG_6292. Five-μg aliquots of both proteins were incubated at room temperature for 15 min in 50 μl of reaction buffer (50 mM Tris-HCl [pH 8.0], 100 mM potassium glutamate, 5% glycerol, and 20 mM imidazole). Twenty μl of Ni-NTA preequilibrated in reaction buffer then was added to the protein mixture and incubated for 30 min in a rotary mixer. The supernatant was separated and the pellet was washed thrice with 400 μl of the reaction buffer. The pellet was resuspended in 50 μl of the buffer followed by the addition of 6× SDS gel loading dye. The samples were incubated at 95°C for 5 min and centrifuged briefly, and the supernatant fractions were subjected to 11% SDS-PAGE. The gels were silver stained. The interaction of MsGre and MSMEG_6292 with the MsRNAP also was carried out by following a similar procedure.

In vitro transcription assays. (i) Stalled TEC preparation.

Ternary elongation complexes (TECs) were generated on a 5′ biotinylated T7A1 promoter DNA template using M. tuberculosis RNAP by following the method described earlier (9, 18).

(ii) Single-round transcription assays.

Fifteen nM T7A1 promoter containing template DNA and 50 nM RNAP were incubated in STB (50 mM Tris-HCl, pH 7.5, 5 mM magnesium acetate, 100 μM dithiothreitol [DTT], 5% glycerol, 50 μg ml⁻¹ bovine serum albumin [BSA], and 50 mM KCl) at 37°C for 15 min. Reactions were initiated by the addition of 100 μM NTP mix, 1 μCi of [α-³²P]UTP, 50 μg ml⁻¹ heparin and incubated further at 37°C for 15 min. In transcription inhibition reactions with Rv3788, the protein was preincubated with the template and MtbRNAP for 15 min at 37°C prior to the addition of the NTPs. Subsequently, the reactions were stopped with 2× gel loading buffer (0.025% [wt/vol] bromophenol blue, 0.025% [wt/vol] xylene cyanol FF, 0.08% amaranth, 5 mM EDTA, 0.025% SDS, and 8 M urea) and analyzed by 10% urea-PAGE. For carrying out transcription inhibition assays at different mycobacterial promoters, 50 nM different promoters containing DNA (gel purified) and 100 nM MtbRNAP were used. Transcription inhibition reactions were carried out as described above with 2 μM Rv3788.

Electrophoretic mobility shift assay (EMSA). (i) Closed complex.

M. tuberculosis P_rrnPCL1 promoter labeled at the 5′ end with [γ-³²P]ATP was incubated with MtbRNAP in the presence of increasing concentrations of Rv3788 on ice for 15 min in STB before being loaded onto a 4% native-PAGE run at 4°C.

(ii) Open complex.

To form competitor-resistant open complexes (RP_O), MtbRNAP, promoter, and increasing concentrations of Rv3788 were incubated for 15 min at 37°C, followed by the addition of 50 μg ml⁻¹ heparin. The DNA-protein complexes were resolved by native PAGE run at room temperature and quantified using Multi Gauge software (version 2.3; Fujifilm). The rate of RP_O decay was determined by initially incubating 100 nM RNAP and 1 nM promoter DNA in STB at 37°C for 30 min to form RP_O, which subsequently was challenged with 50 μg ml⁻¹ heparin and 2 μM Rv3788 protein. The aliquots were taken out at different time points between 1 and 32 min and subjected to 4% native PAGE run at room temperature.

(iii) Ternary elongation complex.

RP_O were formed with MtbRNAP and the P_rrnPCL1 promoter, followed by the addition of 100 μM NTP mix to form TECs. Two μM protein was preincubated with RP_O prior to the addition of NTPs in the reaction mixture containing Rv3788. For inhibition assays with pppGpp, RNAP (100 nM) was incubated with pppGpp (100 μM) in transcription buffer for 15 min, followed by the addition of 50 μg ml⁻¹ heparin, and resolved by 4% native PAGE. pppGpp was prepared by using M. tuberculosis Rel protein with GTP and ATP as substrates.

Western blotting.

Polyclonal antibodies raised in rabbits against Rv3788 were used for the detection of protein in M. tuberculosis H37Ra cells grown for 6, 8, 12, and 20 days. Antibody against M. tuberculosis σ^A was used as a control.

RESULTS AND DISCUSSION

Rv3788 and MSMEG_6292 have domain organizations similar to those of RNAP secondary channel binding proteins.

The homology model of Rv3788 was generated using T. thermophilus Gfh1 as the template (9). The model showed an overall sequence and structural similarity with the Gre factors (see Fig. S1A in the supplemental material), similar molecular weights, and the presence of the signature motifs of secondary channel binding proteins (see Fig. S1A and B). A coiled-coil domain with acidic residues in the loop between the two helices in the N terminus and a globular domain at the C terminus found in the protein are typical features of Gre factors (see Fig. S1A). The Rv3788 homologues were found only in slow-growing mycobacteria and were absent from fast-growing mycobacteria and other sequenced bacterial genomes (see Table S2 in the supplemental material). The protein which has been annotated as Rnk in the M. tuberculosis genome shows 21% identity and 33% similarity to MtbGre (Fig. 1A and B). Rnk is a secondary channel binding protein that was identified earlier in Gram-negative proteobacteria (15, 32). However, M. tuberculosis Rv3788 does not share much similarity with Rnk (only 12% identity and 25% similarity) (Fig. 1B), as demonstrated by the experimental results described below. The M. smegmatis Gre factor homolog MSMEG_6292 also has the structural features of a secondary channel binding protein (see Fig. S1A) and shares higher sequence similarity to the Gre factors than Rv3788 (21% identity to E. coli GreA, 39% identity to both M. smegmatis and M. tuberculosis Gre factors, and only 16% identity to Rv3788) (Fig. 1B). The homologs of MSMEG_6292 could be found in the genomes of all fast-growing isolates of mycobacteria sequenced to date except Mycobacterium abscessus. It is not found in the M. tuberculosis genome (see Table S2 in the supplemental material) and the other slow-growing species.

Fig 1.

Sequence comparison of Rv3788 and MSMEG_6292 to Gre factor and its homologs. (A) Multiple-sequence alignment of Rv3788 and MSMEG_6292 with E. coli GreA (Ec GreA), GreB (Ec GreB), Rnk (Ec Rnk); M. tuberculosis Gre (Mtb Gre); and T. thermophilus Gfh1 (Tth Gfh1) were carried out using Clustal W. (B) The table indicates the percentages of identity and similarity (shaded) of the proteins to each other.

The proteins interact with RNAP but lack Gre like activity.

Gre factors and Gfh1 bind RNAP at the entry to the secondary channel (24, 34, 37). The sequence similarity and resemblance in the overall domain organization of Rv3788 and MSMEG_6292 with these proteins suggest that they also interact with RNAP in a similar manner. The direct interaction of Rv3788 with MtbRNAP was tested by a pulldown assay as described in Materials and Methods (Fig. 2A). His-tagged Rv3788 bound MtbRNAP (lane 4 of Fig. 2A), while RNAP alone did not bind to the Ni-NTA matrix (lane 2 of Fig. 2A), confirming the direct physical interaction between the two proteins. MSMEG_6292 also interacted with the M. smegmatis RNAP in a similar fashion (Fig. 2B). From these data it is evident that the C-terminal domains (CTD) of both proteins are similar to Gre factors, and the differences (if any) in the properties lie in the N-terminal half of the protein.

Fig 2.

Interaction of RNAP with Rv3788 and MSMEG_6292. (A) Pulldown assays were carried out using His-tagged Rv3788 and MtbRNAP (without any tag). The scheme of the assay is described in the left panel. Proteins were incubated together at room temperature for 15 min, followed by the addition of Ni-NTA beads. Both pellet (pel) and supernatant (sup) fractions were loaded on an 11% SDS-PAGE. Lane 4 represents the pulldown of MtbRNAP in the pellet fraction along with Rv3788. (B) Pulldown assays also were carried out using His-tagged MSMEG_6292 and M. smegmatis RNAP without any tag. Lanes 4 and 6 represent the pellet fraction consisting of precipitated MsGre (lane 4) or MSMEG_6292 (lane 6) with MsRNAP. Gre homologs in mycobacteria lack transcript cleavage stimulatory activity. Rv3788 (C) and MSMEG_6292 (D) do not possess transcript cleavage stimulatory activity. Stalled elongation complexes of M. tuberculosis RNAP formed on the T7A1 promoter based template were incubated with the proteins, and the reaction products were resolved by 22% urea PAGE. Lane 1, only stalled elongation complex; lanes 2 and 3, stalled elongation complexes with 1 and 5 μM MtbGre (Fig. 2A) or MsGre (Fig. 2B), respectively; lanes 4 and 5, stalled complexes with 1 and 5 μM Rv3788 (A) or MSMEG_6292 (B), respectively.

The presence of conserved acidic residues in the tip of the N terminus coiled-coil domain and the similarity in overall domain organization of Gre factors led us to examine whether these homologs have Gre-like activity. Transcript cleavage assays were carried out with MtbRNAP using T7A1 20-nucleotide-long stalled elongation complexes. There was no detectable transcript cleavage inducing activity with Rv3788 (Fig. 2C) or MSMEG_6292 (Fig. 2D) in the standard assay conditions. The examination of multiple-sequence alignments of Rv3788 with E. coli and M. tuberculosis Gre factors showed that several residues important for Gre activity are absent from Rv3788 (Fig. 1A). These include Asn47, Tyr50 (of MtbGre), and several other basic residues in the N terminus (Fig. 1A). MSMEG_6292 also lacked several conserved residues in the N-terminal coiled-coil region flanking the acidic patch (see Fig. S2 in the supplemental material), resulting in a shorter N-terminal domain (NTD) compared to those of the Gre factors and Rv3788, which may be insufficient for it to reach the RNAP active center (see Fig. S1A).

Rv3788 is an inhibitor of transcription.

Although the Gre factor homologs did not exhibit any transcript cleavage stimulatory activity, their interaction with RNAP (Fig. 2A and B) is suggestive of their role in the regulation of transcription. In vitro transcription analysis showed that Rv3788 inhibited the runoff transcription from the T7A1 promoter (Fig. 3A). That the inhibition of transcription is an intrinsic property of Rv3788 and not due to any copurifying RNase was verified by incubating the transcription product with Rv3788 (see Fig. S3 in the supplemental material). The transcription inhibition also was not promoter specific, since Rv3788 efficiently inhibited transcription from different promoters from M. smegmatis (P_rrnPCL1, P_rrnB, and P_gyr) and M. tuberculosis (P_rrnPCL1, P_gyrP1, P_gyrR, and P_metU) with comparable efficiency (Fig. 3B). In contrast, the M. smegmatis protein MSMEG_6292 did not show any transcription inhibitory activity in similar experiments from T7A1 promoter-containing template (Fig. 3C), indicating the functional difference between the two Gre factor homologs from the two species of the same genus. The transcription inhibition activity of Rv3788 seems to be restricted to mycobacterial transcription machinery, as it did not inhibit E. coli RNAP (see Fig. S4A in the supplemental material). The lack of inhibitory activity was due possibly to its inability to interact with the E. coli RNAP (see Fig. S4B), indicating the requirement for species-specific protein-protein interactions for secondary channel-mediated regulation.

Fig 3.

Rv3788 inhibits transcription. (A) In vitro single-round transcription assays were carried out using T7A1 promoter-containing template in the presence of 0.5, 1, 2, and 5 μM Rv3788. The bar diagram represents the quantification of the runoff transcripts. Transcription in the absence of any factor was considered 100%. (B) In vitro transcription assays were carried out in the absence and the presence of 2 μM Rv3788 using templates having different promoters from M. smegmatis (Ms) and M. tuberculosis (Mtb). (C) Effect of MSMEG_6292 on in vitro transcription. Two different concentrations of the protein (2 and 5 μM) were incubated with MsRNAP, and a runoff transcription assay was carried out on T7A1 promoter-containing template. Transcription inhibition by 2 μM Rv3788 was used as a positive control (lane 2).

Rv3788 inhibits transcription during ternary complex formation.

We analyzed the mechanism of transcription inhibition by Rv3788 by carrying out EMSA on the M. tuberculosis P_rrnPCL1 promoter. MtbRNAP forms promoter-specific complexes on both P_rrnPCl1 and T7A1 promoters (see Fig. S5 in the supplemental material). Closed complex (Fig. 4A) or open complex (Fig. 4B) formation was not inhibited by Rv3788. Unlike the action of DksA, which affects the open complex stability of the E. coli RNAP (25, 27, 30), the stability of RP_O was not altered by the factor (Fig. 4C), indicating that in this case the inhibition of transcription is at a step beyond open complex formation. A reduction in the ternary elongation complex with NTPs in the presence of Rv3788 suggested that the protein inhibits that particular step (Fig. 5A). When NTPs were added to RP_O prior to the addition of the protein, inhibitory activity was not observed (Fig. 5A, lane 5). Further, Rv3788 also prevented the inhibitory activity of pppGpp on the M. tuberculosis P_rrnPCL1 promoter (Fig. 5B). These results together suggest that the protein prevents the access of the nucleotides and the effector molecules to the catalytic center directly or indirectly. Alternatively, Rv3788 could modulate the activity of the enzyme by inducing a conformational change in the active site.

Fig 4.

Mechanism of transcription inhibition. EMSAs were carried out using the end-labeled M. tuberculosis P_rrnPCL1 promoter and MtbRNAP. The effects of the increasing concentrations of Rv3788 on closed complex (RP_C) formation (A), open complex (RP_O) formation (B), and rate of decay of the open complexes (C) were determined by resolving the promoter-RNAP complexes in 4% native PAGE.

Fig 5.

Rv3788 inhibits NTP binding to RNAP, preventing formation of TEC. (A) EMSAs were carried using M. tuberculosis P_rrnPCL1 promoter-containing DNA to check the TEC formation between the RNAP promoter and the NTPs. Lane 1, RP_O; lane 2, RP_O in the presence of 2 μM Rv3788; lane 3, TEC formation in the presence of NTPs; lane 4, RP_O treated with 2 μM Rv3788 prior to the addition of NTPs; lane 5, NTPs were added to the RP_O before the addition of Rv3788. The bar diagrams in the right panel represent the quantification of the complexes. (B) Rv3788 prevents inhibition by pppGpp. The assays were carried out using M. tuberculosis P_rrnPCL1 to determine the action of pppGpp in the presence of Rv3788. Lane 1, RP_O; lane 2, RP_O incubated with pppGpp; lane 3, RP_O in the presence of 2 μM Rv3788; lane 4, RP_O in the presence of both pppGpp and Rv3788. The bar diagram in the right panel represents the quantification of % RP_O formation.

Transcription inhibition is mediated by the N-terminal domain.

Despite the conservation of the acidic residues at the tip of the predicted coiled-coil domain, instead of inducing cleavage, Rv3788 inhibited transcription. Thus, the property of Rv3788 is similar to that of TthGfh1 (16). The crystal structure of the Gfh1-RNAP complex revealed that the N-terminal coiled-coil domain enters the RNAP secondary channel and possibly modulates the bridge helix in the RNAP active center. Within the secondary channel, the Gfh1 NTD fits particularly well with the narrowest region of the secondary channel (34). Val 36 of Rv3788 probably is located in the narrowest region of the secondary channel, similarly to Leu33 of Gfh1 (Fig. 6A). A mutant of Rv3788 with Val36 replaced by Trp (V36W) significantly reduced transcription inhibitory activity (Fig. 6B). The bulky side chain of Trp probably prevents the Rv3788 NTD from penetrating into the secondary channel to reach the RNAP active center. However, it also is possible that the introduced bulky amino acid distorts the coiled-coil domain. However, this mutation did not abolish the interaction of the protein with RNAP (Fig. 6C), indicating that the interacting region in the C terminus is independent of the N terminus coiled-coil domain. Further, a mutant protein with changes in the acidic residues at the tip of the coiled-coil domain (D45A/D48A double mutant) showed significantly reduced inhibitory activity (Fig. 6B), indicating that the acidic residues are required for the transcription inhibitory activity.

Fig 6.

Rv3788 interacts with RNAP through its C-terminal domain and accesses the secondary channel of RNAP by using the N terminus to inhibit transcription. (A) The conserved residue V36 of Rv3788 was mutated to V36W to introduce a bulky group in the N-terminal coiled-coil domain. The S126E mutation was introduced in the C-terminal RNAP interaction domain. The acidic residues D45 and D48 in the coiled-coil domain were mutated to D45A/D48A (double mutant). (B) In vitro transcription inhibition assays with the WT and the mutant proteins using T7A1 promoter-containing template. (C) Pulldown assays were carried out using His-tagged WT, V36W, and S126E with MtbRNAP as described in Materials and Methods, and the samples were analyzed by 11% SDS-PAGE. Lanes 4, 6, and 8 represent the pellet fractions of the interaction assays with WT, V36W, and S126E proteins, respectively. The loss of RNAP interaction in the S126E mutant is indicated by the absence of RNAP from the pellet fraction (lane 8). S126E showed anomalous mobility in the gel, possibly caused by a structural change in the protein due to mutation.

The structure of Gfh1 in complex with RNAP revealed that the hydrophobic region in the CTD is involved in the interaction with the RNAP β′ coiled-coil domain at the edge of the secondary channel (34). The hydrophobic region of the Gfh1 CTD is conserved in Rv3788. In the TthRNAP, the interaction involves Phe982 and Leu983 in the β′ coiled-coil domain and Ile119, Met117, and Met125 in the Gfh1 CTD (34). The β′ coiled-coil domain loop region is highly conserved in MtbRNAP (9), suggesting the requirement of this conserved set of residues for its interaction with MtbGre and Rv3788. Ser123 and Ser126 of Rv3788 are two well-conserved residues found among the Gre factors and Gfh1. These conserved serine residues have been shown to be important for interactions with RNAP in the case of E. coli GreB (37), TthGfh1 (34), and M. tuberculosis Gre (9). The residues are located close to the conserved glutamate residue in the RNAP β′ subunit (34) and may hydrogen bond with it. Indeed, a mutation in the S126 residue of Rv3788 (S126E) (Fig. 6A) disrupted the interaction between the protein and MtbRNAP and also abolished its transcription inhibitory activity (Fig. 6B and C).

To test whether MtbGre and Rv3788 both compete for the same binding site in RNAP, interaction assays in the presence of increasing concentrations of MtbGre were carried out. MtbGre (no tag) could prevent Rv3788 (His tagged) from binding with RNAP, as indicated by the decrease in RNAP in the pellet fractions in the Ni-NTA pulldown assays (Fig. 7A). Similarly, when the transcription assays were carried out in the presence of Rv3788 and increasing concentrations of MtbGre, transcription inhibition was partially rescued by MtbGre (Fig. 7B). MSMEG_6292 also competed with MsGre for binding with RNAP. Increasing concentrations of MSMEG_6292 could compete with the Gre factor and inhibit the transcript cleavage activity, probably by preventing Gre factor accessibility to the secondary channel of the RNAP (Fig. 7C).

Fig 7.

Rv3788, MSMEG_6292, and the Gre factor share the same binding site on MtbRNAP. (A) Ni-NTA pulldown assays with MtbRNAP and Rv3788 were carried out in the presence of non-His-tagged MtbGre protein (5 and 10 μg) as described in Materials and Methods. (B) Inhibition of in vitro transcription from the T7A1 promoter by a fixed concentration of Rv3788 (1 μM) and in the presence of increasing concentrations (0.25 to 5 μM) of MtbGre (lanes 3 to 7). Lane 1, in vitro transcription from T7A1 promoter-containing template; lane 2, transcription in the presence of only Rv3788; lane 8, only MtbGre. (C) MSMEG_6292 competes with MsGre. The T7A1-TECs were preincubated with 1 to 5 μM MSMEG_6292 followed by the addition of 1 μM MsGre. Only MsGre was added in lane 2, and in lane 6 only 5 μM MSMEG_6292 was added.

Comparison of Rv3788 to Thermus species Gfh1.

Secondary channel binding proteins have emerged as one set of regulators of RNAP activity in different bacteria. Among them, Gfh1, DksA, and TraR are the only transcription factors that have been identified to date as inhibiting transcription (16, 30). Rv3788 shares the mechanism of inhibition with Gfh1, i.e., by preventing NTP binding. The inhibitory activity of Gfh1 was more robust at acidic pH, possibly due to the conformational toggle of the N-terminal domain at low pH (16). The conserved residue Gly86 in Gfh1 was predicted to be responsible for the conformational switch required for transcription inhibition (14), and residue Gly88 is also conserved in Rv3788. The inhibition assay carried out at different pHs showed that the transcription inhibitory activity of Rv3788 was greater at pH 6.0 than at pH 8.0 (Fig. 8A), indicating possible mechanistic similarity between Rv3788 and Gfh1. Similarly to Gfh1, Rv3788 also inhibited both runoff and abortive transcriptions (Fig. 8B). However, unlike Gfh1, it did not inhibit the intrinsic cleavage activity of RNAP (Fig. 8C). A few differences in the primary sequence of the N-terminal coiled-coil domains of the two proteins may account for these differences in activity. Gfh1 has four Asp residues in the loop region of the coiled-coil domain (Fig. 1), whereas the M. tuberculosis protein has only two acidic Asp residues at the same region. The Asp residues of Gfh1 were shown to interact with Mg(II) in the RNAP active center, rendering it unavailable to bind substrate NTPs and thus leading to transcription inhibition (16). A recent study, however, indicates that N-terminal acidic residues are not solely responsible for transcription inhibition by Gfh1 (34). The binding of the protein to the secondary channel and locking the RNAP in an alternative ratchet conformation is also considered to be important for transcription inhibition in conjunction with the acidic tip (34). Notably, in the case of Rv3788, mutations of the acidic residues in the N-terminal coiled-coil loop region significantly reduced transcription inhibitory activities but did not abolish the inhibition completely (Fig. 6B). Thus, it appears that both the Asp residues and the binding of the coiled-coil domain to the secondary channel, which possibly alters RNAP conformation, contribute collectively to the mechanism of inhibition. The major target of the secondary channel binding regulators is predicted to be the trigger loop of the RNAP β′ subunit, which undergoes substrate-induced refolding during the nucleotide addition cycle. The trigger loop has emerged as both a central regulatory element and as a key determinant for the processivity and fidelity of the transcription (29, 36). The high sensitivity of the structure of the loop to subtle alterations in neighboring domains and the modulation of its folding to trigger helix by many transcription factors appears to be a major mode of regulation of transcription. From the present studies, we suggest that it also could be the target of Rv3788 in the RNAP active center.

Fig 8.

Properties of Rv3788. (A) The inhibition of transcription from T7A1 promoter template at two different pHs, 6.0 and 8.0. The quantification of the transcription inhibition is represented in the graph. Transcription in the absence of Rv3788 is represented as 100%. (B) Inhibition of abortive transcription from M. smegmatis P_rrnPCL1 promoter by Rv3788. (C) Intrinsic cleavage activity of MtbRNAP in the absence (lanes 1 to 3) and in the presence (lanes 4 to 6) of 2 μM Rv3788 in a +20 T7A1-TEC.

Apart from Gfh1 and Rv3788, peptide inhibitor microcin J25 also was shown to inhibit transcription by binding to the RNAP secondary channel and occluding NTPs (21). Other inhibitors of bacterial RNAP described so far employ different mechanisms to perturb transcription (12). RNAP is evolutionarily conserved across different species, and a common mechanism of enzyme inhibition by proteins from diverse species (Gfh1 and Rv3788) indicates the importance of regulation through secondary channels. These results also suggest the coevolution of these proteins along with the RNAP to utilize the secondary channel. Despite similar mechanisms of action, these structurally similar proteins exhibit some differences in their properties, suggesting the evolutionary optimization of their function in different species to cope with the physiological needs to survive in a particular environment. A point in favor of this argument is that the Gre factor paralogs of the two species of mycobacteria are different. While one shows transcription inhibition, the other could not modulate the RNAP active center, possibly due to the reduced length of the N terminus. These results also are suggestive of specific requirements for transcriptional downregulation in M. tuberculosis under certain physiological conditions. The analysis of the M. tuberculosis cell lysates indicated that Rv3788 is present in all growth phases (see Fig. S6 in the supplemental material). However, the expression level was lower in the early exponential phase, increased in the mid-exponential phase, and remained constant thereafter.

In conclusion, Rv3788 inhibits transcription, possibly by interacting with RNAP to modulate activities at the active center. The presence of several secondary channel-binding proteins in the genome could result in a competition between them to access the active center through the channel. The inhibition of transcription by proteins from diverse species of bacteria by utilizing the RNAP secondary channel appears to be a general theme of transcription control. The identification of secondary channel-specific transcription inhibitors like Rv3788 in mycobacteria and the elucidation of the structural features of their interaction with RNAP could form the basis for the development of small-molecule inhibitors which block the channel and, as a result, inhibit RNAP.