The remarkable ability of Mycobacterium tuberculosis (Mtb) to survive inside human macrophages is attributed to the presence of a complex sensory and regulatory network. PrrA is a DNA-binding regulatory protein, belonging to an essential two-component system (TCS), PrrA/B, which is required for early phase intracellular replication of Mtb. Despite its importance, the mechanism of PrrA/B-mediated signaling is not well understood. In the present study, we demonstrate that the binding of PrrA on the promoter DNA and its consequent activation is cumulatively controlled via dual phosphorylation of the protein. We have further characterized the role of terminal phospho-acceptor domain in the physical interaction of PrrA with its cognate kinase PrrB. The genetic deletion of prrA/B in Mycobacterium smegmatis was possible only in the presence of ectopic copies of the genes, suggesting the essentiality of this TCS in fast-growing mycobacterial strains as well. The overexpression of phospho-mimetic mutant (T6D) altered the growth of M. smegmatis in an in vitro culture and affected the replication of Mycobacterium bovis BCG in mouse peritoneal macrophages. Interestingly, the Thr6 site was found to be conserved in Mtb complex, whereas it was altered in some fast-growing mycobacterial strains, indicating that this unique phosphorylation might be predominant in employing the regulatory circuit in M. bovis BCG and presumably also in Mtb complex.
Mycobacteriumtuberculosis (Mtb) is a successful intracellular pathogen and continues to be one of the leading causes of mortality worldwide [1,2]. Despite extensive tuberculosis management programs, it still accounts for 1.8 million deaths per year globally [1,2]. Growing evidence in recent years suggests that the incredible ability of this noxious pathogen to survive in the extremely unfavorable environment of the host is attributed to the presence of complex sensory and regulatory networks [3–5]. The Mtb genome encodes eleven functional Ser/Thr kinases (STPKs) from PknA to PknK (except PknC) and also the same number of two-component systems (TCSs) composed of histidine kinases (HKs) and their paired cognate response regulators (RRs) [3,6]. TCSs are the primary means to sense environmental stimulus in bacteria and provide adaptation to varying micro-environmental conditions [7–9]. Over the period of time, the TCSs are being considered crucial for the virulence and antibacterial drug resistance in several pathogenic bacteria [10,11]. Although some of the two-component systems such as MtrAB, MprAB, PhoPR, SenX3-RegX3 and dormancy-associated DosRS are extensively characterized towards understanding the remarkable ability of Mtb to survive in adverse environmental conditions [7,10,12–14], a large number of RRs are still poorly characterized. Basically, TCSs become activated through the auto-phosphorylation of sensor kinase at its histidine residue followed by phosphotransfer to the aspartate residue of cognate RR [8,15]. Phosphorylation of RRs at conserved aspartate residue induces conformational changes, creating an exposed hydrophobic surface to facilitate DNA–protein interaction . PrrA/B is one of the conserved TCSs in mycobacteria and is essential for the viability of Mtb when grown in an in vitro culture, and influences its early phase intracellular adaptation [17,18]. The reduced expression of PrrA and PrrB, in an Mtb transposon mutant Mt21D3, was correlated with the attenuated replication of Mtb in murine bone marrow-derived macrophages at early time points post-infection [17,18]. PrrA belongs to the OmpR family containing a characteristic C-terminal winged helix-turn-helix DNA-binding domain and N-terminal receiver domain (Supplementary Figure S1) . Structural analysis reveals that PrrA can exist either as a closed or as an open structure; the inter-convertible forms lead to the variable accessibility of DNA-binding recognition helix. Similar to the other members of the OmpR family, PrrA is also activated by accepting a phosphate group through its sensor kinase PrrB, which leads to an open conformation and presumably results in an increased DNA binding . Besides usual aspartate phosphotransfer, several RRs in pathogenic bacteria have been recently characterized that are regulated by additional non-aspartate phosphorylations. These additional non-aspartate phosphorylations have been found to be critical for cell growth, virulence and other metabolic processes [6,19,20]. Interestingly, the phosphorylation of Mtb PrrA at the N-terminal threonine residue has been evidenced recently in two independent studies of global phosphoproteome analysis of Mtb and M. bovis BCG (BCG) [9,21]. However, the significance of this phosphorylation event in survival and in the pathophysiology of mycobacteria has not been elucidated yet. In the present study, we have investigated the role of PrrA phosphorylation at aspartate and at non-aspartate residues, in the growth and in intracellular survival of mycobacteria. We have also demonstrated that such unusual post-translational modifications at the N-terminus play a crucial role in the regulation of the PrrA DNA-binding activity and transcriptional regulation.
Materials and methods
Cell line, bacterial strains and growth conditions
Peritoneal macrophages isolated from BALB/c mice, and the HeLa cells obtained from the American Type Culture Collection (ATCC), USA, were cultured in RPMI 1640 or in Dulbecco's modified Eagle's medium (2 mM L-glutamine, 1.5 g/l sodium bicarbonate, 1 mM sodium pyruvate, 10 mM HEPES and 4.5 g/l glucose), supplemented with 10% heat-inactivated fetal calf serum (Invitrogen) at 37°C and 5% CO2. Escherichia coli strains DH5α and BL21 (DE3) were procured from Invitrogen and were cultured in Luria–Bertani (LB) medium. Mycobacterial strains (Supplementary Table S2) were cultured in Middlebrook 7H9 medium (Difco) supplemented with 0.5% (v/v) glycerol, 10% (v/v) ADC, and 0.05% (v/v) Tween 80. The mycobacterial cultures used during the experiment for intracellular survival were plated on MB7H10 agar and incubated at 37°C to enumerate colony-forming units (CFU). The bacteria containing recombinant pMV261 or pMV361 plasmids, (Supplementary Table S3) were grown in the presence of 25 μg ml−1 kanamycin, and the strains transformed with the pML523 plasmid (for KO construction) were grown with 50 µg ml−1 hygromycin.
Construction of plasmids and protein expression
The ORFs of prrA and prrB genes were amplified by PCR using Mtb H37Rv chromosomal DNA as a template with the primers containing HindIII (in the forward) and XhoI (in the reverse) restriction sites (Supplementary Table S1). The gel-eluted PCR amplified products were digested with HindIII and XhoI and ligated into the pET28b . The phospho-ablative (PA) and phospho-mimetic (PM) mutants, in pET28b or pMV361, were generated using the PCR-based site-directed mutagenesis procedure as previously reported [22,23]. Briefly, the primers for the site-directed mutagenesis were designed by the Agilent QuikChange tool (Supplementary Table S1). The pET28b or pMV361 vectors harboring wild-type (WT) prrA were used as template to generate the desired site-directed mutants. The amplified products were treated with DpnI for removal of the template, subjected to heat inactivation at 60°C for 30 min, purified by gel elution, and introduced into E. coli DH5α. The recombinants were confirmed by sequencing. For the overexpression and purification of the engineered proteins, the cultures were grown in LB medium containing kanamycin (50 µg ml−1) and resulting cultures were incubated at 37°C with shaking until an optimum optical density of 0.5 was achieved. Isopropyl 1-thio-D-galactopyranoside (IPTG) (1 mM final) was then added to induce the expression, and the incubation was continued for an additional 3 h at 37°C. Purifications of the His-tagged proteins were performed as described previously . The proteins were eluted with an imidazole gradient of 0–500 mM. PrrA and its mutant proteins were obtained in the cytosolic fraction and were eluted in non-denaturing conditions, whereas PrrB was obtained in the membrane fraction and, therefore, eluted in 6 M urea buffer. The proteins were dialyzed against 50 mM phosphate buffer (pH 7.4) and 200 mM NaCl to remove urea and imidazole and then allowed to concentrate up to 10.0 mg ml−1.
Immunization of rabbits for polyclonal antibody production
Animal experiments were performed according to guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA) and CSIR–Central Drug Research Institute, Lucknow, India. The recombinant His-tagged PrrA and PrrB purified proteins (∼200 μg of each) were emulsified in Freund's incomplete adjuvant and used for immunization of female adult white New Zealand rabbits by subcutaneous injection . Rabbits were given booster injections of PrrA and PrrB antigens on day 0, 15, 30 and 45 respectively. On day 60, rabbits were bled, and the sera were separated by incubating at 37°C for 2–3 h . The antisera were confirmed for reactivity with PrrA and PrrB proteins using western blotting. The pSer/Thr (Qiagen) antibodies were procured commercially. The antibodies were used to look for the differential expression of the TCS proteins in mycobacteria.
Kinase assay was performed using a non-radioactive ADP-Glo™ Kinase Assay Kit (Promega) as reported previously . The relative extent of phosphorylation was evaluated indirectly by measuring the amount of ATP consumed in the reaction. Briefly, the enzyme (kinase) and substrate were incubated in the kinase buffer containing 40 mM Tris (pH 7.5), 2 mM MnCl2, 20 mM MgCl2, 2 mM DTT, 0.5 mg ml−1 BSA, and 10 µM ATP for 1 h at 26°C. ADP-Glo reagent was added and incubated for 30 min at room temperature. After the addition of KDR reagent (supplied with the kit), the luminescence was measured.
Bacterial cultures grown up to logarithmic phase were pelleted, and resuspended in buffer (50 mM Tris-Cl, pH 7.5, 300 mM NaCl, 5 mM EDTA), in the presence of the protease and phosphatase inhibitors. Cells were lysed by sonication and clear lysates were harvested by centrifugation at 13 000×g for 30 min at 4°C. The soluble extracts were incubated for 4 h with Sepharose beads pre-coupled with polyclonal antibodies against the PrrA at 4°C. Antigen/antibody complexes were eluted. The proteins were quantified and separated by SDS-PAGE, transferred on to the PVDF membrane and detected by immunoblotting with anti-PrrA or anti-pSer/Thr antibody.
Pull-down assay and circular dichroism spectroscopy
The full-length Mtb PrrA gene and its N-terminal and C-terminal truncated segments PrrA10ΔN and PrrA30ΔC were sub-cloned into pET28b at NcoI and XhoI sites to express the N-terminal His-tag free proteins. E. coli BL21 (DE3) strains were induced with 0.5 mM IPTG at 37°C for 4 h, and the clear cellular lysates were prepared in 1× binding buffer (20 mM Tris-Cl, pH 7.5, 50 mM KCl, 10 mM MgCl2 and 0.1% CHAPS and 10% glycerol) and 100 µg purified His-tagged PrrB was then added in each lysate containing 1.0 mg of total protein and incubated at 4°C for 1 h on a rocker. To extract the protein complexes, the lysates were applied to Ni-NTA resin and further incubated for 1 h at 4°C. The supernatant was removed and the resin was washed with 1× binding buffer containing 20 mM imidazole and eluted with 300 mM imidazole. Samples were resolved on 12% SDS-PAGE, transferred on to a PVDF membrane and immunodetected with anti-PrrB and anti-PrrA antibodies. For circular dichroism (CD) spectroscopy the His-tagged free PrrA (WT, N, and C-terminal truncated) proteins were purified by electroelution from a native polyacrylamide gel and dialyzed against 20 mM sodium phosphate buffer at pH 7.5. The CD spectra were acquired from 190 to 250 nm at 20°C temperature using a Jasco Scans 1500 instrument. The buffer contribution was subtracted from CD spectra expressed as the molar ellipticity [θ], in deg·cm2·dmol−1.
Mammalian two-hybrid system
A CheckMate™ Mammalian Two-Hybrid System (Promega) kit was used for protein–protein interaction studies. The genes were cloned in the vectors supplied with the kit as per instructions given in the user's manual (Promega). Briefly, the full-length genes of prrB, prrA and truncated forms of prrA were amplified from the genomic DNA of Mtb H37Rv using forward and reverse primers containing NotI and BamHI sites respectively (Supplementary Figure S6). The PCR products were digested and cloned into pBIND or pACT vector. HeLa cell line was transfected using Lipofectamine. The interaction of proteins was measured quantitatively using luciferase assays according to the manufacturer's instructions (Promega).
Electrophoretic mobility shift assays
Binding of PrrA WT and mutants to the promoter region of the prrAB operon was analyzed by electrophoretic mobility shift assays (EMSA). The prrAB promoter region between −250 to +20 was amplified from the genomic DNA of Mtb H37Rv strain using 5′-FAM-labeled PCR primers (Supplementary Table S1). EMSA reactions contained 1.0 ng of labeled purified PCR product in the buffer (20 mM Tris–HCl, pH 7.5, 50 mM NaCl, 4% glycerol, 3 mM MgCl2, 40 mg/ml BSA, 0.1 mM EDTA, 0.5 mM DTT) and increasing concentrations of proteins. For each phosphorylation reaction the enzyme and substrate were incubated at a ratio of 1 : 100 µg of total protein in the kinase buffer containing 40 mM Tris (pH 7.5), 2 mM MnCl2, 20 mM MgCl2, 2 mM DTT, 0.5 mg ml−1 BSA, and 0.1 mM ATP for 1 h at 26°C. The phosphotransfer was confirmed by kinase assay as described in a previous experiment. The working concentration of the phosphorylated proteins used in EMSA was estimated by subtracting the enzyme concentration. The mixtures of proteins and DNA were incubated for 30 min at room temperature [27,28]. Following the addition of 1.0 ml of loading buffer (0.025% bromophenol blue, 0.5× TBE and 10% glycerol), bound and free DNAs were resolved by fractionation on a pre-run (for 1 h at 10 mA) 8% native polyacrylamide gel in 0.5× TBE. DNA was visualized under fluorescence light using an LAS-500 instrument. The quantitative analysis of the bound DNA was carried out using ImageJ (1.47) software.
Analysis of promoter activity
The hsp60 promoter in the expression vector pMV261 was replaced with prrA promoter. The prrA promoter region was amplified from the genomic DNA of Mtb H37Rv using primers prrAP-F′ and prrAP-R′ (Supplementary Table S1). The luciferase reporter gene was sub-cloned into the pMV261 vector. The prrA promoter region was cloned upstream to the luciferase gene (luc) at the unique KpnI and EcoRI sites to generate pMV261_P::luc construct. The prrAB operon was amplified using primers prrA-F′ and prrB-R′ and cloned at the unique HindIII site downstream of the hsp60 promoter in the pMV361 vector. The desired mutations were created by site-directed mutagenesis and confirmed by sequencing. Constructs were electroporated into mc2 155. Fresh transformants were grown with aeration in 100 ml of Middle Brook 7H9 medium containing 25 µg ml−1 kanamycin, 0.05% Tween 80, and 10% ADC supplement (BD Diagnostic Systems), to an optical density of 0.6 at 600 nm. The cultures were centrifuged and lysed by sonication. Total protein was quantified by Bradford assay, and the luciferase assays were performed according to the manufacturer's instructions (Promega).
RNA isolation and qRT-PCR
RNA isolation from Mycobacterium smegmatis (MS) was performed using an RNeasy Mini Kit (Qiagen). Briefly, the log phase bacterial cultures were harvested and incubated with RNA protection solution for 1 h at 37°C. Cells were lysed using a bead beater. Further isolation steps were performed as described in the user's manual. qRT-PCR was performed by the SYBR Green method as described previously[18,22].
Gene deletion in M. smegmatis
For disruption of the MSMEG 5663-5662 operon, a two-step recombination strategy was employed . The first step involved the construction of the recombination delivery vector. The suicide vector pML523 was used to create prrAB knockout (KO) in M. smegmatis. A 1000-bp region upstream of the prrAB operon containing 23 bp 5′ prrA was amplified using the primers prrA_UpF′ and prrA_UpR′ and cloned into the pML523 vector at SwaI and SpeI restriction sites to create pML523.1 (Supplementary Table S1). Next, a ∼1000-bp fragment of downstream region containing the 70-bp prrB 3′ was amplified using primers prrB_DnF′ and prrB_DnR′ and cloned into pML523.1 to create pML523.2. In the second step, the vector was electroporated into mc2 155 and DCO recombinants were selected as described previously . Briefly, M. smegmatis mc2 155 was transformed with vector pML523.2 and grown at 37°C on MB7H9 medium containing OADC and Hyg (50 µg/ml). Transformants were grown in liquid cultures and plated for selection of DCOs. DCOs were screened for the presence of GFP, and XylE activity (DCO candidates were GFP positive, and XylE negative in 2% catechol). The deletion of genes was examined by PCR with multiple sets of primers.
Expression of proteins in mycobacteria
The bacterial cultures were harvested by centrifugation at 2500×g for 10 min. The bacterial cell pellets were washed twice with 1× PBS and were resuspended in extraction buffer (50 mM Tris, 100 mM NaCl, protease inhibitor cocktail) and sonicated. 4× Laemmli buffer was added to the cell lysates, boiled for 10 min and centrifuged at 9000×g for 5 min. Cell lysates were resolved by 12% SDS-PAGE and then transferred on to PVDF membrane (GE Healthcare Life Science, Germany). The membrane was blocked with 5% non-fat dry milk in TBST buffer (20 mM Tris-Cl, 50 mM NaCl and 0.5% Tween 20) for 2 h and incubated overnight at 4°C with primary antibody in 2.5% non-fat dry milk made in TBST. The membrane was then incubated for 1–2 h with the anti-rabbit or anti-mouse IgG-HRP conjugated secondary antibody (Santa Cruz Biotechnology), and developed with the ECL reagent (Amersham™, Prime, UK). The analysis was carried out using an Image Quant LAS 500 gel documentation system (GE Healthcare).
Isolation of peritoneal macrophages and infection with recombinant BCG
The isolation of peritoneal macrophages was performed as described previously . Briefly, the cells were isolated from 5–7 weeks old female BALB/c mice by washing the peritoneal cavity with 10 ml RPMI 1640 medium. A total of 1 × 105 cells per well were seeded in 12-well plates and infected at 1 : 10 m.o.i. After 2–3 h, the cells were washed twice with incomplete RPMI 1640 medium. The cells were then treated with 100 μg/ml amikacin to kill the non-phagocytosed bacilli. Samples for quantification of intracellular bacteria were taken at 0, 24 and 48 h post-infection by removing the supernatants. Cells were lysed in 0.5 ml sterile distilled H2O and incubated at 37°C until lysis was complete. The phagocytosis and intracellular persistence of mycobacteria were quantified by plating and determining the CFU.
The amino acid sequences of prrA and its orthologs in other mycobacterial species were obtained from the KEGG database (http://www.genome.jp) or TubercuList (tuberculist.epfl.ch). Sequences were aligned using the online T-Coffee server, applying the PAM matrix with default parameters. The ClustalW files were managed by the ESPript 3.0 server.
Dual phosphorylation of PrrA at Asp58 and Thr6 residues
Besides canonical aspartate phosphorylation by the cognate HK PrrB , the studies of comparative global phosphoproteomic analyses among Mtb, BCG and M. smegmatis have revealed that PrrA is phosphorylated in vivo at the Thr6 residue in Mtb and in BCG but not in M. smegmatis. The study by Prisic et al. has also demonstrated that the synthetic N-terminal peptide of Mtb PrrA is phosphorylated by the kinase domains of multiple STPKs at Thr6in vitro . However, for the progress of the present study , the identification of the exact location of aspartate phosphoacceptor residue and confirmation of phosphorylation of full-length PrrA protein by Mtb STPKs was obligatory. To predict the aspartate phosphoacceptor site, the amino acid sequences of PrrA orthologs from diverse mycobacterial and other related bacterial species were aligned and analyzed alongside the characterized phosphoacceptor sites in other orthologs. Therefore, we contemplate that Asp58 could be the potential site of phosphorylation owing to its location corresponding to the Asp63 residue in Rhodobacter sphaeroides  and high conservation among other orthologs (Figure 1A). Interestingly, the Thr6 residue was found to be conserved in all pathogenic species analyzed, including members of the Mtb complex, Mycobacterium leprae, Mycobacterium marinum, Mycobacterium africanum and in Mycobacterium ulcerans, but altered in some of the fast-growing strains such as M. smegmatis and in Mycobacterium abscessus (Figure 1A).
PrrA is phosphorylated at Asp58 residue by its cognate kinase PrrB.
To demonstrate the phosphorylation of PrrA at Asp58 and at Thr6 by HK and STPKs respectively, the Asp58 and Thr6 were replaced by alanine using site-directed mutagenesis (Figure 1B). The full-length protein of PrrA (WT), its PA isoforms (PrrA_D58A and PrrA_T6A), and the histidine kinase PrrB were purified from recombinant E. coli BL21 (DE3) (Figure 1C). The proteins were subjected to phosphorylation by PrrB. As expected the PrrA_WT and PrrA_T6A became phosphorylated by PrrB, whereas PrrA_D58A did not; this confirmed that Asp58 is the residue that is modified by phosphotransfer from its histidine kinase, PrrB (Figure 1D). Next, to confirm the non-aspartate phosphorylation of full-length PrrA by Mtb STPKs in vitro, the recombinant proteins of at least three STPKs (two cytosolic kinases, PknK and PknG, and a membrane-bound kinase PknJ) were also purified from E. coli BL21(DE3) as reported previously . The kinases were incubated with the protein in buffer containing ATP, and phosphorylation was detected by ADP-Glo kinase assay [22,26] and by immunoblotting with anti-pSer/Thr antibody as described previously . The WT PrrA was found to be phosphorylated by all of the three kinases used in this study, whereas the PrrA_T6A did not show this effect (Figure 2A). Following incubation of PrrA_T6A with PknG/PknK/PknJ, immunoblotting with anti-pSer/Thr antibodies revealed that a single point mutation at the alanine residue results in a total abrogation of phosphorylation (Figure 2B). Further, to investigate the status of the non-aspartate phosphorylation of PrrA in vivo, the protein was immunoprecipitated from the cell lysates of different mycobacterial strains grown up to the logarithmic phase using anti-PrrA antibodies, as described previously [34,35]. The reactivity of anti-pThr antibodies with the precipitated proteins confirmed that PrrA is phosphorylated in vivo at a threonine residue in Mtb H37Rv, Mtb H37Ra, and in M. bovis BCG. However, we did not observe phosphorylation in the protein harvested from M. smegmatis grown at similar physiological conditions (Figure 2C).
Phosphorylation of PrrA at Thr6in vitro and in vivo.
Deletion of N-terminal extension and effect on the physical interaction of PrrA with PrrB
Pull-down assays were performed to examine the physical interaction of PrrA with its cognate histidine kinase PrrB. The N- and C-terminal truncated PrrA proteins (His-tag free) were overexpressed in recombinant E. coli (BL21). The schematic of truncation and the generated fragments of the mutant proteins are shown (Figure 3A,B). PrrA_FL and ΔN interacted with PrrB, as the higher amount of protein was observed in the elution fractions than in the wash fractions (Figure 3C,D), whereas ΔC was found to be defective in enabling the physical interaction, evidenced by the presence of majority of the protein in wash fractions and not in the elution fractions (Figure 3E). The absence of PrrA in the elution fractions of the control experiments (in absence of PrrB) excluded the possibility of non-specific binding of the protein with Ni-NTA resin. A mammalian two-hybrid experiment also characterized the physical interaction between PrrA and its cognate partner PrrB. As shown in Figure 3F, an interaction of PrrA_FL and ΔN with PrrB was confirmed, as their co-transfection in HeLa cells resulted in enhanced luciferase activity over untransfected or empty-vector transfected controls, whereas C-terminal deletion abrogated the interaction, as no significant activity above background was observed. This suggested that PrrA might be interacting with PrrB through its C-terminus. Moreover, to investigate the effect of deletions on the overall conformation of the protein we performed far-UV CD spectroscopy with the FL, ΔN and ΔC proteins. The far-UV CD spectra of the N-terminal truncated forms exhibited the pattern of a typical α-helical protein similar to the WT, suggesting that the N-terminal overhang exerts minimal or no impact on overall folding of the proteins; contrary to this, the C-terminal deletion imparted significant conformational change (Figure 3G). Overall, our results revealed that the N-terminal unstructured region of PrrA affects neither the conformation of the protein nor its interaction with the PrrB, whereas the observed loss of interaction in case of PrrA_ΔC might be a consequence of improper folding of the protein upon C-terminal truncation.
Physical interaction of PrrA with its cognate histidine kinase, PrrB.
Effect of phosphorylations on DNA–protein interaction in vitro
The response regulator PrrA is considered to be activated by accepting a phosphate group at an aspartate residue through PrrB [16,32]. However, the impact of the non-aspartate phosphorylation on DNA binding required investigation. Considering the fact that, prrAB is an autoregulated operon , we used a prrAB promoter element for our DNA–protein interaction studies. The WT PrrA protein was able to shift the mobility of promoter sequence amplified (using specific 5′-FAM-labeled primers) from the genomic DNA of Mtb H37Rv (Supplementary Table S1). A competition assay with 10× or 50× non-specific and 10× specific DNA confirmed the specific binding of the protein with the target DNA (Figure 4A). The replacement of threonine with acidic residues such as aspartic acid (Asp) or glutamic acid (Glu) mimic the phosphorylation effect with regard to functional activity (PM mutation) [23,34–37]. The recombinant T6A and T6D mutations were generated by site-directed mutagenesis and the proteins were purified from recombinant E. coli BL21 (Figure 4B,C). The PrrA WT or mutant proteins T6A and T6D showed mobility shift of the target DNA probe to a comparable extent, suggesting that neither the residue (Thr6) nor its phosphorylation has a direct involvement in the interaction of PrrA with the target DNA. However, EMSA with the proteins, pre-incubated with PrrB in the buffer with or without ATP, revealed that the phosphotransfer by PrrB in the presence of ATP significantly enhances the DNA binding of PrrA WT and the mutants (Figure 4D). More interestingly, this enhancement was apparently higher in the case of T6D∼P as compared with that of WT∼P or T6A∼P (Figure 4D), suggesting that the two phosphorylations act synergistically to activate the protein for DNA binding. The percentage binding of proteins at each concentration of PrrA in both the unphosphorylated and phosphorylated conditions are shown (Figure 4E).
Effect of phospho-mimetic mutation at PrrA on DNA binding.
Phospho-mimetic mutation at Thr6 contributed to the promoter activation in vivo
To corroborate our results from in vitro DNA binding experiments with the in vivo system, a reporter gene-based promoter assay was performed in M. smegmatis mc2 155, using the luciferase gene as reported previously [38,39]. The M. smegmatis mc2 155 bacteria were transformed with the recombinant vector pMV261_PprrAB::luc, containing the luciferase gene, under the prrAB promoter. The resulting reporter strain P-luc was retransformed with pMV361 empty vector or with the recombinant plasmids expressing PrrA isoforms and PrrB (Figure 5A). The doubly transformed colonies were screened by colony PCR using Mtb prrA/prrB gene-specific primers and then confirmed by western blotting and the presence of luciferase activity. As expected, the recombinant strains were able to overexpress the proteins, whereas only basal level expressions of PrrA and PrrB were observed in the strain containing empty vector (Figure 5B). The bacterial cultures were grown up to mid-log phase in Middle Brook 7H9 medium supplemented with 10% ADC. Luciferase activity was measured as described in the Materials and methods section. As shown in Figure 5C, the P-luc strain harboring pMV361 vector alone exhibited basal promoter activity. However, as expected, the overproduction of PrrAB (WT), PrrAB (A:T6A and A:T6D) induced luciferase activity. Interestingly, the induction of the promoter activity was significantly higher in the strain producing PM PrrA protein. Conversely, the luciferase activity resulting from the overproduction of D58A remained comparable to the vector control. Next, the expression of an endogenous (chromosomally induced) copy of prrA/B operon in M. smegmatis was also monitored by qRT-PCR with the primers specific to the prrAMS, which do not target the transcription from ectopically expressed Mtb prrA. The normalized expression of prrAMS was slightly higher (∼1.5–2.0 fold) in the recombinant strains MS:PrrAB (WT) and in MS:PrrAB (A:T6D) as compared with the MS:pMV361 control (Figure 5D). Notably, unlike Mtb promoter, the chromosomal prrAMS promoter was induced by both Mtb PrrA WT and T6D to a comparable extent, suggesting the strain-specific regulation of the operon.
Phosphorylation-dependent promoter activation by PrrA in vivo.
MSMEG (5662-5663) operon is essential for the in vitro growth of M. smegmatis
The essentiality of PrrA/B (Rv0903c-Rv0902c) for the viability of Mtb H37Rv has been reported previously . Therefore, we tried to generate an MSMEG 5662-5663 (homologs of Mtb prrA/B in M. smegmatis) deletion mutant of M. smegmatis. For this, we used a homologous recombination strategy as reported in the previous studies [29,30]. An almost 1000 bp fragment spanning upstream and downstream of the MSMEG 5662-5663 operon was amplified by primers (Supplementary Table S1) and cloned into a suicide vector, pML523 (Supplementary Figure S2). The recombinant vector, pML523.2, containing homologous upstream and downstream regions of MSMEG 5663-5662, flanking the loxP-gfp2+m-hyg-loxP cassette, was transferred into mc2 155 by electroporation and colonies were selected on MB7H10/OADC/Hyg plates. The single isolated colony was picked up and inoculated into 10 ml MB7H10/ADC/Hyg medium. The liquid culture was allowed to grow at 37°C with shaking until the OD600 reached 0.6–0.8, diluted up to 1 × 10−6, plated on MB7H10/OADC/hyg plates and incubated at 40°C to select the single crossover (SCO) events. For the direct selection of double crossover (DCO) events, the diluted or undiluted cultures were plated on MB7H10/OADC/hyg supplemented with 2% sucrose and incubated at 40°C. SCO and DCO clones were screened for the presence of xylE and gfp. Also, the GFP and XylE positive clones (SCO intermediates) were passaged in MB7H9/ADC/Hyg medium containing 2% sucrose to allow for DCO events to take place. After each passage, the undiluted or 10−2-fold diluted cultures were plated on MB7H10/OADC/Hyg plates supplemented with 2% sucrose and incubated at 40°C to select for DCO clones. However, even after numerous independent attempts of both direct and indirect selections for DCOs, we were unable to recover DCO positive clones. Therefore, this led us to believe that PrrA/B TCS might be essential for in vitro growth of M. smegmatis as well. To further confirm this, we attempted the deletion of MSMEG 5663-5662 in the presence of an ectopic copy of Mtb prrAB. The recombinant pML523.2 was then transferred into mc2 155::prrAB merodiploid. The similar strategy of DCO selection was followed. The legitimate recombination and deletion of the desired fragment were confirmed by PCR (Figure 6A,B). In this attempt, ∼13% (4 out of 29) DCO (GFP+/XylE−/Suc+) candidates accumulated the desired mutation (Figure 6C), which confirmed that the deletion of MSMEG 5662-5663 is possible only in the presence of an ectopic copy of prrA/B. As the protein sequences in the two organisms (M. smegmatis and Mtb H37Rv) are more than 93% identical and the aspartate phosphoacceptor site and other functional domains are also conserved, the lethal phenotype of prrA/B gene deletion in MS is not surprising. Moreover, the upstream and downstream genes were amplified from the genomic DNA of mutants and sequenced to confirm the genomic integrity of both the upstream (MSMEG_5661) and downstream (MSMEG_5664) ORFs, which might have been affected due to the gene deletion.
MS prrAB is essential for its in vitro growth.
Overexpression of phospho-mimetic PrrA protein affects the growth of MS in culture and in intracellular survival of BCG
To assess the physiological implication of the PrrA phosphorylation at Thr6 and Asp58 positions, we examined the gene dosage response of different prrA alleles in fast- and in slow-growing mycobacterial species. Mtb PrrA has high similarity in amino acid sequence with its homolog in M. smegmatis except for a few N-terminal residues (Supplementary Figure S3). Therefore, by assuming the functional similarity of the protein in M. smegmatis we used it as a surrogate model to investigate the role of additional phosphorylation in Mtb. The M. smegmatis mc2 155 and M. bovis BCG were transformed with pMV361 containing the coding sequences of different prrA alleles (WT, T6A, T6D, and D58A) coupled with its cognate kinase gene, prrB under hsp60 promoter (Supplementary Table S2). The evaluation of CFU/ml of growing cultures at regular time intervals revealed that the overexpression of T6D but not T6A or D58A was accompanied by a slight decrease in the growth rate (Figure 7A) and change in colony morphology of M. smegmatis (Supplementary Figure S4). Unexpectedly, none of the recombinants showed any such growth defect or change in colony morphology of BCG (Figure 7B). The western blot experiments demonstrated the overexpression of the proteins in M. smegmatis (Figure 7C) and in BCG (Figure 7D). For intracellular survival studies, mouse peritoneal macrophages were infected with recombinant BCG strains (pMV361, WT, T6A, T6D, and D58A) at a multiplicity of infection 1 : 5. After 3 h of phagocytosis, the CFU recovered from the cells infected with all of the recombinants were comparable to the vector control. However, when intracellular replication of bacilli was monitored by estimating the CFU counts at 24 h and 48 h post-infection, a significant increase in CFU was noticed in the cases of prrAB(WT) and prrAB(A:T6D) merodiploids as compared with the vector control. The CFU of T6A of D58A mutants were comparable to the control (Figure 7E). This indicated that the dual phosphorylation of PrrA at both the aspartate and non-aspartate sites is important for the intracellular survival of mycobacteria.
Growth kinetics of M. smegmatis and BCG overexpressing PrrA variants.
Histidine–aspartate phosphotransfer is the universal mechanism of bacterial two-component systems, whereas a few examples of non-aspartate phosphorylation (NAP) of RRs are also reported in pathogenic bacteria. However, in Mtb, the only known example of RR bearing non-aspartate phosphorylation thus far is DosR. The NarL family RR DosR is phosphorylated by PknH at Thr65 prior to affecting the physiology of mycobacteria . The Mtb genome encodes a much smaller number of paired TCSs compared with other bacterial species such as E. coli and Bacillus subtilis . Through, the course of evolution, Mtb has reduced the number of TCSs [8,12] and therefore the integration of regulatory networks through the cross-talk between different arms of molecular signaling prevailed. The phosphorylation of RRs at aspartate residues by their histidine kinases is quite short lived and extinguishes quickly . Conversely, the non-aspartate phosphorylations such as serine, threonine and tyrosine are relatively more stable . Therefore, the convergence of the two distinctive phosphorylation pathways might help in the integration of cellular signals. Being a classical TCS response regulator, PrrA is also activated by the classical His-Asp phosphotransfer by PrrB, at an aspartate residue [16,17]. However, an investigation to ascertain the exact location of aspartate phosphorylation and the role of unusual N-terminus Thr6 phosphorylation in PrrA-mediated signaling has been reasonable. In the present study, the genetic and biochemical approaches have been used to prove that the phosphorylation of PrrA at Asp58 (an aspartate residue) and Thr6 (a non-aspartate residue) acts synergistically and exerts a differential effect on the growth of fast-growing and slow-growing mycobacterial species. To this end, site-directed mutagenesis was employed to generate point mutations to locate the exact position of phosphoacceptor sites. The conservation of Thr6 in the Mtb complex and variation in some fast-growing strains such as M. smegmatis suggested that this particular residue is somehow critical for PrrA activity in slow-growing MTBC strains. However, a large-scale genomic data analysis is needed to ascertain the role of Thr6 mutation in the evolution of MTB complex. Consistent with the previous report by Prisic et al., which showed that the phosphorylation of synthetic N-terminal peptide of PrrA can be achieved by multiple STPKs in vitro , PrrA was able to become phosphorylated by all the three STPKs used in the present study. However, because PknK is expressed only at 18 h post-infection, and regulates the early infection events , and PrrA is known for its crucial role in the early phase intracellular survival , we hypothesized that PknK could be the major kinase responsible for the phosphorylation of PrrA in vivo.
In several recent studies, the aspartate phosphorylation in bacterial RRs has been shown to be influenced by a distantly occurring NAP event [6,20]. Therefore, deciphering the interplay between the two distinctive phosphorylations, and determining how they ultimately influence bacterial growth and persistence could be critical to a complete understanding of mycobacterial pathogenesis. PrrA belongs to the OmpR/PhoB family RRs and acts by binding to the promoter sequence(s) of the target genes in a phosphorylation-dependent manner . Phosphorylations at N-terminal sensor or linker domains have been implicated in adjusting the RR′s function by inducing a conformational change in the DNA-binding domain [36,38,41]. For instance, the phosphorylation of the N-terminus Thr9 residue in Rv3875 abrogates its DNA-binding ability in vitro . In another study, the TetR family RR EthR is shown to be extensively phosphorylated at its unstructured N-terminus at Thr2, Thr3, Ser4 and Ser7. The DNA binding of EthR to its target promoter is found to be decreased upon mutation of phosphoacceptor residues . Although the identification of the PrrA-regulated genes remains warranted, its auto-regulation through the binding on its own promoter has been reported [8,16]. Not surprisingly, PrrA was able to bind to its own promoter DNA in both phosphorylated and unphosphorylated forms, which was consistent with the previous study . Although, the PA or PM forms of PrrA did not alter the DNA-binding affinity to the target DNA in the absence of aspartate activation, nevertheless, the differential activation of PrrA isoforms upon phosphotransfer by PrrB for in vitro DNA binding suggested the additive role of Thr6 phosphorylation in DNA–protein interaction. Moreover, the results from promoter assay experiments also supported that the two phosphorylations act synergistically in the activation process of the PrrA. The phosphorylations of TCS response regulators at non-aspartate residue(s) have been reported in several other pathogenic bacteria; for example, control of virulence regulator gene CovR in group A Streptococcus is phosphorylated at Thr65 . The Thr65 phosphorylation competes with the aspartate phosphorylation of CovR and subsequently dysregulates the positively auto-regulated CovR . Vancomycin-resistance-associated response regulator VraR of Staphylococcus aureus is phosphorylated at four non-aspartate residues present in its activator (Thr106 and Thr119) and DNA-binding domain (Thr175 and Thr178). The threonine phosphorylations in S. aureus VraR enhance the DNA binding and subsequent modulation of the target gene expression [43,44]. Interestingly, the Mtb RRs of both the OmpR and NarL family show conservation in their C-terminal effector domain whereas the N-terminal region is highly variable (Supplementary Figure S5). The role of an N-terminal variable arm of PhoP has been shown to be crucial for its activation . Although, deletion of the N-terminal overhang of PrrA did not affect the physical interaction of the protein with its cognate kinase PrrB, deletion of the C-terminal DNA-binding domain abrogated this interaction. The far-UV CD spectra of the N- or C-terminal truncated PrrA revealed that the N-terminal overhang exerts no obvious impact on the overall folding of the protein, whereas C-terminal truncation significantly altered the folding of protein compared with the WT. Noticeably, it seems that the misfolding of C-terminally truncated protein might have also contributed in the abrogation of this interaction. Another OmpR family RRs of Mtb, MtrA has also been shown to interact with MtrB and target DNA sequences through its C-terminal domain . Interestingly, MtrA shows very high structural similarity with that of PrrA . Hence, the study indicates a common mechanism of interaction in OmpR family RRs.
As prrA-prrB was previously reported to be an essential TCS for the growth of mycobacteria in an in vitro condition and in an intracellular milieu [17,18], we used M. smegmatis as a model to look into the physiological implication of the phosphorylations. The proteins PrrA and PrrB of Mtb share ∼93% identity with M. smegmatis, including the highly conserved output domain (DNA-binding domain), thus indicating that these proteins might be functionally conserved in both organisms. With an aim to characterize the physiological relevance of PrrA isoforms, we attempted to generate a deletion mutant in M. smegmatis. However, after several independent failed attempts of recovering the mutant we concluded that, like slow-growing Mtb H37Rv, this TCS is essential also in fast-growing M. smegmatis. Moreover, the PrrA/B two-component system is essential in Mtb H37Rv for the in vitro growth. The overexpression of PM Mtb PrrA (T6D) causes a growth defect in M. smegmatis but not in M. bovis BCG. However, this observation can be explained on the basis of genetic differences between the fast-growing and slow-growing mycobacteria strains. Despite the presence of homologs of almost all of the known STPKs and HKs in M. smegmatis, the global phosphoproteomes of slow-growing mycobacteria such as BCG and Mtb are larger than M. smegmatis, i.e. the proteomes of the slow-growing stains are more extensively phosphorylated than that of the fast-growing strains . Secondly, in M. smegmatis the endogenous copy is a mutant for the Thr6 phosphorylation, whereas BCG or Mtb have a phosphorylatable endogenous copy of PrrA. Therefore, the ratio of phosphorylated to unphosphorylated forms of the protein would be different in MS and BCG, which might be the reason for the obvious differential growth. Besides this, the overexpression of WT Mtb PrrA/B or a constitutively phosphorylated form of PrrA affected the intracellular replication of BCG at 24 and 48 h post-infection, suggesting a crucial regulatory role of Thr6 phosphorylation in the intracellular survival of mycobacterial species that are relevant to the present study.
In conclusion, the unique non-aspartate phosphorylation at N-terminal threonine residue is required for the optimal transcriptional activation of the target promoter, which eventually regulates the growth and intracellular survival of the mycobacteria. However, further investigation is underway to explore the PrrA regulon and its involvement in mycobacterial virulence. We confirmed the essentiality of this TCS in the non-pathogenic strain M. smegmatis, which suggests the involvement of the genes in basic physiological processes other than the pathogenesis. Our findings provide important novel information for understanding the unique regulatory mechanism used by pathogenic mycobacteria to survive in the unfavorable environment of macrophages during its infection cycle.
Mycobacterium bovis BCG
colony forming units
control of virulence regulator
electrophoretic mobility shift assays
vancomycin-resistance-associated response regulator
A.K.M. and K.K.S. conceived, analyzed and coordinated the study and wrote the manuscript. A.K.M. performed the cloning, KO, and molecular experiments. A.K.M. and S.M.Y. contributed to the preparation of figures. A.K.M., S.M.Y., R.K.D., and E.D. performed growth and survival experiments. All authors analyzed the results and approved the final version of the manuscript.
The work has been funded through CSIR network projects Splendid [BSC0104] and UNDO [BSC0103]. The CSIR-CDRI communication number allotted to this manuscript is 9585.
We thank Dr. Madhu Dikshit, Director of CSIR-CDRI for her encouragement and support. A.K.M. is a DBT India fellow and S.M.Y. is an AcSIR fellow. We thank Dr. Amit Singh (IISc Bangalore) for providing the pML523 vector. We also thank Dr. Pramod Singh (CDRI Lucknow) and Dr. Sameer Tiwari (CDRI Lucknow) for providing constructs of STPKs (pknG, pknK, and pknJ) used in this study.
The Authors declare that there are no competing interests associated with the manuscript.