The Mycobacterium tuberculosis Protein Kinase K Modulates Activation of Transcription from the Promoter of Mycobacterial Monooxygenase Operon through Phosphorylation of the Transcriptional Regulator VirS*

Mycobacterium tuberculosis encodes for 11 eukaryotic-like serine/threonine protein kinases. Genetic and biochemical studies show that these kinases regulate various cellular processes including cell shape and morphology, glucose and glutamine transport, phagosome-lysosome fusion and the expression, and/or activity of transcription factors. PknK is the largest predicted serine/threonine protein kinase in M. tuberculosis. Here, we have cloned, overexpressed, and purified protein kinase K (PknK) to near homogeneity and show that its ability to phosphorylate proteins is dependent on the invariant lysine (Lys55), and on two conserved threonine residues present in its activation loop. Despite being devoid of any apparent transmembrane domain, PknK is localized to the cell wall fraction, suggesting probable anchoring of the kinase to the cell membrane region. The pknK gene is located in the vicinity of the virS gene, which is known to regulate the expression of the mycobacterial monooxygenase (mymA) operon. We report here for the first time that VirS is in fact a substrate of PknK. In addition, four of the proteins encoded by mymA operon are also found to be substrates of PknK. Results show that VirS is a bona fide substrate of PknK in vivo, and PknK-mediated phosphorylation of VirS increases its affinity for mym promoter DNA. Reporter assays reveal that PknK modulates VirS-mediated stimulation of transcription from the mym promoter. These findings suggest that the expression of mymA operon genes is regulated through PknK-mediated phosphorylation of VirS.

Cell signaling is a process by which environmental signals are transmitted to cells, ultimately resulting in changes in gene expression and activity. One of the major mechanisms by which this takes place is by the reversible phosphorylation of cellular proteins. Protein phosphorylation in prokaryotes plays a regu-latory role in events as diverse as chemotaxis, bacteriophage infection, nutrient uptake, and gene transcription (1). The enzymes involved in these regulatory functions are either protein histidine kinases, phosphotransferases, or protein serine kinases. Analysis of the Mycobacterium tuberculosis genome sequence suggests the presence of 11 eukaryotic-like serine/ threonine protein kinases, two tyrosine protein phosphatases, and one serine/threonine protein phosphatase (2). Nine of these kinases contain a putative transmembrane domain, suggesting their localization to the cell membrane. To date, nine of these kinases have been biochemically characterized. They are PknA, PknB, PknD, PknE, PknF, PknG, PknH, PknI, and PknL (3)(4)(5)(6)(7)(8)(9)(10).
Although much work remains to be done toward understanding the role of the STPKs in M. tuberculosis biology, recent reports have established the role of these kinases in modulating cell shape and morphology, glucose transport, glutamine transport, glutamate metabolism, and regulation of the expression/activity of transcription factors (8,(11)(12)(13)(14)(15)(16)(17)(18)(19). Protein kinases A and B, encoded by pknA and pknB, respectively, are a part of the same operon that also contains the cistrons of protein phosphatase pstP, rodA (involved in cell shape control), and pbpA (involved in peptidoglycan synthesis). Overexpression of PknA in Mycobacterium bovis bacillus Calmette-Guerin (BCG) results in a deviation from normal cell morphology with the cells forming an elongated and branched structure, whereas overexpression of PKnB in M. bovis BCG results in the formation of widened and bulging cells (13). Recent antisense RNA-mediated PknF knockdown studies in M. tuberculosis and results from overexpression of PknF in Mycobacterium smegmatis show effects on cell division, growth rate, morphology, and glucose transport (14). Protein kinase G (PknG) has been shown to play a role in the survival of the pathogen in host macrophages by modulating phagosome-lysosome fusion (lysosomal transfer) after macrophages phagocytose mycobacterium (20).
The identification of the direct substrates of protein kinases is a key factor in determining their role in intracellular signal transduction. Several substrates of PknA and PknB have been identified. These include PbpA, a protein essential for cell division; conserved hypothetical proteins Rv1422 and Wag31; Forkhead-associated domain-containing proteins GarA and Rv0020c; mycolic acid synthesis pathway proteins KasA, KasB, and ␤-ketoacyl-ACP synthase III (mtFabH); the cell division protein FtsZ; MurD, a ligase involved in the process of peptidoglycan synthesis, N-acetylglucosamine-1-phosphate uridyltransferase (GlmU), a key enzyme that synthesizes UDP-Nacetylglucosamine; and sigma factor SigH and anti-sigma factor RshA (13,19,(21)(22)(23)(24)(25)(26)(27)(28). PknA-mediated phosphorylation events modulate the activities of the cell division protein FtsZ, and the proteins involved in mycolic acid synthesis, mtFabH, KasA, and KasB (24 -26). PknB-mediated phosphorylation of anti-sigma factor RshA negatively regulates its interaction with sigma factor SigH (19). We have recently shown that PknB-mediated phosphorylation of GlmU modulates its acetyltransferase activity (28). PknH, another transmembrane STPK, has been shown to control the embCAB operon through a regulatory protein EmbR (17). In addition, Rv0681 and DacB1 proteins have been identified as targets for PknH; however, the functional significance of these phosphorylations have not yet been worked out (29). PKnF has been implicated to play a role in glucose transport via regulation of an ABC transporter, Rv1747 (30,31). PknD-mediated phosphorylation of anti-anti-sigma factor Rv0516c inhibits its binding to antianti-sigma factor, Rv2638, suggesting a role for PknD in regulating transcription in M. tuberculosis (18).
In this study, we set out to characterize the serine/threonine protein kinase K and identify its substrates. PknK is a large protein (1100 amino acids). Although the N-terminal 290 amino acids are homologous to eukaryotic-like serine/threonine kinase domains, the C-terminal residues show homology with the regulatory region of Escherichia coli transcription regulator MalT. The PknK contains an ATP/GTP-binding site (P-loop) and a putative PDZ domain between amino acids residues 368 -375 and 465-533, respectively. Here, we show that the C-terminal region is important for the activity of PknK. Results show that the transcription factor VirS is a PknK substrate and further identify four proteins of the mymA (mycobacterial monooxygenase) operon to be novel substrates of PknK. We also demonstrate that VirS is a bona fide substrate of PknK, and PknK-mediated phosphorylation of VirS increases its affinity for mym promoter DNA by ϳ2.5-fold. Luciferase reporter assays show that PknK-mediated phosphorylation regulates VirS-mediated transcription from mym promoter region, suggesting a role for PknK in regulating the expression and possibly the activity of the mymA operon proteins.

EXPERIMENTAL PROCEDURES
Chemicals, Reagents, and Radioisotopes-The restriction and DNA-modifying enzymes pMAL-c2X vector and Amylose resin were purchased from New England Biolabs. Primers and analytical grade chemicals were purchased from Sigma-Aldrich. [␥- 32 Table S1) and Phu DNA polymerase (NEB), followed by cloning the amplicon so obtained into pENTR/ Directional-TOPO vector (Invitrogen). The kinase domain pknK-KD (ϳ1.2 kb) was similarly cloned using primers PknK-F1 and PknK-R2 (supplemental Table S1). The pknK and pknK-KD genes were subcloned into pMAL-c2X expression vector (NEB) using XbaI-HindIII to create the plasmids pMAL-PknK and pMAL-PknK-KD, respectively. pMAL-PknK-TATA and pMAL-PknK-K55M were generated by overlapping PCR mutagenesis using PknK-TATA F and R and PknK-K55M F and R primers (supplemental Table S1). The veracity of all clones was checked by sequencing. The pMAL expression plasmid constructs were transformed into E. coli BL21 (DE3)-CODON PLUS cells (Stratagene). Fresh transformants were grown in 500 -1000 ml of LB medium containing 100 g/ml ampicillin to a cell density of ϳA 600 of 0.6. The cells were induced with 1 mM IPTG, 5 and further grown for 12-16 h at 18°C. The maltose-binding protein (MBP)-tagged proteins so overexpressed were purified following the manufacturer recommendations (NEB).
Cloning, Expression, and Purification of VirS and Mym Operon Proteins-The genes encoding VirS (Rv3082), Mym (Rv3083), LipR (Rv3084), and FadD13 (Rv3089) as well as genes encoding for Rv3085, Rv3086, Rv3087, and Rv3088 were PCRamplified from M. tuberculosis H37Rv genomic DNA (kindly provided by the Tuberculosis Vaccine Testing and Research Materials Contract, Colorado State University), using the corresponding forward and reverse primers (supplemental Table  S1) and Phu DNA polymerase, followed by cloning of the amplicons into pENTR/Directional-TOPO vector (Invitrogen). The virS and mym genes were subcloned into pC6-2 expression vector (32) that has a caspase-6 cleavage site (VEMD) inserted into the BamHI site of pGEX-4T2 (Invitrogen), using NotI digestion. The genes lipR, Rv3085, Rv3086, Rv3088, and fadD13 were subcloned into pC6-2 using BamHI-XhoI digestion. The gene Rv3087 was subcloned into pMAL-C2X expression vector using XbaI-HindIII digestion. MBP-tagged Rv3087 was purified as described above. Caspase-6 site-containing GST-tagged proteins were overexpressed as described above. The harvested cells were then resuspended in (5 ml/1 g cells) STE buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 1 mM EDTA) containing protease inhibitors. Lysozyme was added to a final concentration of 100 g/ml, and lysate was incubated on ice for 15 min. DTT was added to a final concentration of 5 mM followed by the addition of N-laurylsarcosine to a final concentration of 1.5%. The contents were vortexed for 5 s followed by sonication to complete the lysis. Triton X-100 was added to a final concentration of 2%, and the samples were rocked at 4°C for 30 min. The supernatant was clarified by centrifugation at 10,000 ϫ g for 30 min. Binding to the GST beads and the elution of the GST fusion protein was essentially carried out according to the manufacturer's recommendations (GE Healthcare). Caspase-6 purification was carried out according to the protocol described (32). After purification, the proteins were dialyzed against storage buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 20% glycerol.
In Vitro Kinase Reaction and Phosphoamino Acid Analysis-In vitro kinase assays were performed by incubating 1-10 pmol of PknK, PknK-K55M, or PknK-KD, and 5 g of myelin basic protein (Mbp) in a 40-l reaction volume containing 25 mM HEPES (pH 7.4), 5 mM MnCl 2 , 15 mM MgCl 2 , 1 mM DTT, 1 M ATP, and 10 Ci of [␥-32 P]ATP for 15 min at 30°C. The reactions were stopped by the addition of 20 l of 4ϫ SDS sample buffer followed by boiling for 5 min. For the reactions with various putative substrates, kinase reactions were carried out with 5 pmol of PknK and 50 -75 pmol of substrate in 40 l of reaction volume containing 25 mM HEPES (pH 7.4), 20 mM magnesium acetate, 1 mM DTT, 10 M ATP, and 10 Ci of [␥-32 P]ATP. The reactions were resolved on 8 -15% gradient SDS-PAGE gels and transferred to nitrocellulose membranes followed by autoradiography. Bands corresponding to the labeled protein of interest were excised from the filter and digested with trypsin. Tryptic peptides were analyzed by two-dimensional resolution on thin layer cellulose plates as described (33). Aliquots of the tryptic peptide mixes were further processed for phosphoamino acid analysis as described (33).
Western Blot Analysis-The NdeI-SalI fragment of pknK was subcloned into pET22b, and thus the PknK-KD289 was overexpressed and partially purified. Polyclonal antibodies to PknK were raised in rabbit by injecting an emulsion containing Freund's incomplete adjuvant and gel-eluted PknK-KD289. For Western blot analysis, varying amounts of purified PknK-KD or M. tuberculosis whole cell extract or subcellular fractions were resolved by SDS-PAGE and transferred onto nitrocellulose membrane. After blocking the membrane with 5% nonfat dry milk in PBST (phosphate-buffered saline containing 0.1% Tween 20), the blot was incubated with rabbit anti-PknK polyclonal antibodies (1:3000 dilution) containing 2.5% bovine serum albumin and 2.5% nonfat dry milk in PBST, overnight at 4°C. After washing, the blots were incubated with donkey antirabbit horseradish polyclonal antibodies (1:5000 dilution) for 1 h at room temperature in PBST containing 5% nonfat milk. Following washing, the blots were developed using West Pico ECL kit (Pierce).
Purification of VirS and p-VirS-pMAL-C2X or pMAL-PknK constructs were digested with HindIII followed by end filling. These linear DNAs were digested with NdeI, and the fragment released (MBP or MBP-pknK genes) was cloned into NdeI-EcoRV sites of pDUET vector (Novagen). The gene encoding VirS was subcloned into the NotI site in the first multiple cloning site of pDUET vector. Expression constructs pDUET-MBP-VirS or pDUET-PknK-VirS were transformed into E. coli BL21 Codon Plus cells. Exponentially growing cultures (A 600 of ϳ0.6) were induced with 0.5 mM IPTG and further grown for 6 h at 18°C. The cells were harvested and lysed by sonication in buffer A (20 mM Tris-HCl, pH 7.4, containing 10% glycerol, 150 mM NaCl, and 5 mM ␤-mercaptoethanol and protease inhibitors). The cell lysates containing His 6 fusion proteins were mixed with equilibrated nickel-nitrilotriacetic acid-agarose affinity res-ins. After 2 h of incubation, resin was washed with 4 column volumes of buffer A containing 20 mM imidazole and 0.5 M NaCl. His-tagged VirS and p-VirS were eluted with buffer A containing 150 mM imidazole. The imidazole was dialyzed out using buffer A containing 20% glycerol. Purified proteins were estimated by Bradford assay, and the estimations were verified by resolving proteins on SDS-PAGE along with 0.3-1.5 g of bovine serum albumin.
Metabolic Labeling in E. coli-E. coli transformants harboring pDUET-MBP-VirS or pDUET-PknK-VirS were grown in 5 ml of LB medium containing 100 g/ml ampicillin to a cell density of ϳA 600 of 0.6. The cells were induced with 0.1 mM IPTG and further grown for 4 h at 18°C. The cultures were harvested, washed with 5 ml of low phosphate medium (0.2% bacto-tryptone, 0.4% casamino acids, 0.4% glucose, 100 mM Tris-HCl, pH 7.5, 20 mM KCl, 80 mM NaCl, 20 mM NH 4 Cl, 0.1 mM CaCl 2 , and 1 mM MgSO 4 ) and resuspended in 1 ml of low phosphate medium supplemented with 1 mCi of [ 32 P]orthophosphate and 0.1 mM IPTG and was further grown at 18°C for 4 h. The cells were harvested and lysed using a bead beater in buffer containing phosphate-buffered saline, 5% glycerol, 5 mM DTT, and protease inhibitors. The cell lysate was clarified, and the lysates containing His fusion protein were rocked with equilibrated nickel-nitrilotriacetic acid-agarose affinity beads for 4 -6 h at 4°C. The beads were then thoroughly washed with lysis buffer containing 20 mM imidazole and resuspended in SDS sample buffer. The digested samples were resolved on SDS-PAGE followed by transfer to nitrocellulose membrane and autoradiography.
Electrophoretic Mobility Shift Assay (EMSA)-The mym promoter region of M. tuberculosis was amplified from the BAC clone Rv48 using primers MymP-F and MymP-R (supplemental Table S1) and Phu DNA polymerase (NEB), followed by cloning the amplicon so obtained into pENTR/Directional-TOPO vector (Invitrogen). This was used as the template for amplifications to obtain substrate for use in EMSAs. The substrate so obtained was purified and radiolabeled in a 40-l reaction containing ϳ5 pmol of substrate, 20 Ci of [␥-32 P]ATP (specific activity 6000 Ci/mmol) and 20 units of polynucleotide kinase PNK in PNK buffer (Roche Applied Science), according to the manufacturer's recommendations. Radiolabeled PCR fragment was purified free of [␥-32 P]ATP and PNK using nucleotide removal kit (Qiagen). EMSAs were carried out in a 15-l reaction consisting of 7 ng (70 fmol; ϳ50000 cpm) of labeled DNA probe and various concentrations of VirS or VirS⌬ or pVirS in binding buffer (50 mM Tris-HCl, pH 6.8, 0.5 mM EDTA, 100 mM KCl, 0.5 mM DTT, 0.5 mM MgCl 2 , 100 ng/l poly(dI-dC), 250 ng/l bovine serum albumin, and 5% glycerol) for 10 min at 37°C. The reactions were electrophoresed at 4°C on 8% nondenaturing polyacrylamide gels (75:1-acrylamide:bis) in 1ϫ TBE buffer for 1 h at 200 V, followed by autoradiography.
Luciferase Reporter Assays-The Hsp60 promoter in the M. tuberculosis expression vector pVV16 (kindly provided by Tuberculosis Vaccine Testing and Research Materials Contract, Colorado State University) was replaced with UV15 promoter, which was amplified from pSE100 vector (a kind gift from Dr. Sabine Ehrt). The multiple cloning site of pVV16 was lengthened by cloning in an oligonucleotide con-taining desired restriction sites, and a second copy of the UV15 promoter was inserted after the unique NotI site (details of this new shuttle vector for expression in mycobacterium will be reported elsewhere). The mym promoter region was amplified from the BAC clone Rv48 using primers MymP-F2 and MymP-R1 (supplemental Table S1), and the amplicon obtained was cloned into pENTR/Directional-TOPO vector (Invitrogen). The luciferase reporter gene was amplified from pcDNA3-Luciferase reporter construct using primers LucF and LucR, and the amplicon was cloned into pENTR/Directional-TOPO vector (Invitrogen). The mym promoter region was cloned upstream of the luciferase gene (luc) to generate pENTR-ML. Luc or mym promoter-luc inserts were subcloned into the unique KpnI site in the modified pVV16 construct. PknK gene was amplified using primers PknK-F2 and PknK-R2 and cloned into the unique XbaI site downstream of the His 6 tag. The virS gene was amplified using primers VirS-F2 and VirS-R2 and cloned into the unique HindIII site upstream of C-terminal FLAG tag and downstream of a second copy of the UV15 promoter. The various constructs made are shown in Fig. 7A. These constructs were electroporated into M. smegmatis mc 2 155. Fresh transformants were grown in 100 ml of Middle Brook 7H9 medium (pH 6.8) containing 25 g/ml kanamycin, 0.05% Tween 80, and 1% ADC supplement (BD Diagnostic Systems) with aeration, to a cell density of ϳA 600 of 2.0. The cultures were centrifuged and lysed using a bead beater, and the luciferase assays were performed according to the manufacturer's recommendations (Promega). The reactions were analyzed using a Berthold luminometer.

Although Protein Kinase K Is Autophosphorylated Only on Threonine Residues, It Phosphorylates Mbp on Both Serine
and Threonine Residues-The gene encoding protein kinase K was cloned into pMAL-c2X vector, overexpressed, and purified as described above. Based on the sequence homology with other known kinases, Lys 55 of PknK was predicted to be the invariant lysine essential for catalytic activity (34). We mutated lysine 55 to methionine (K55M) and overexpressed and purified PknK-K55M. Both PknK and PknK-K55M were purified to ϳ90% purity (Fig. 1A). In vitro kinase reactions with the universal substrate myelin basic protein (Fig. 1B) demonstrated that PknK phosphorylated Mbp efficiently. A band corresponding to autophosphorylated PknK could also be detected (Fig. 1B). No phosphorylation of PknK or Mbp was detectable upon mutating the lysine 55 residue (Fig. 1B), indicating this lysine residue to be essential for kinase activity, as predicted. A slight anomaly in mobility of PknK-K55M was observed (Fig. 1A), which is most likely due to the differential phosphorylation status of the two proteins. Although the active PknK has the ability to autophosphorylate itself, the inactive PknK-K55M will be unable to do so. To determine which residue(s) were autophosphorylated in PknK, phosphoamino acid analysis was carried out. The results obtained clearly revealed that only threonine residues are so phosphorylated (Fig. 1C). Similar analysis carried out with phosphorylated Mbp showed phosphorylation on both serine and threonine residues (Fig. 1D).
Threonine Residues in the Activation Loop Are Required for PknK Activity-Protein kinases are broadly classified as RD kinases and non-RD kinases. RD kinases have a conserved arginine residue preceding a conserved aspartate in the catalytic loop. Many RD kinases are shown to be regulated by phosphorylation of their activation loop (35). The activation loop region in kinases is ϳ20 -35 amino acid residues in length and lies between two tripeptide motifs, DFG and APE (Fig. 2A). Other than PknG, all the STPKs in M. tuberculosis are classified as RD kinases (36). Using mass spectrometry analysis, the residues in the activation loop that are sites of phosphorylation have been identified in M. tuberculosis STPKs PknB, D, E, and F (shown by an asterisks in Fig. 2A) (37). Sequence alignment predicts three probable phosphorylation sites in the activation loop of PknK-two threonines residues (Thr 178 and Thr 180 ; Fig. 2A) and one serine residue (Ser 186 ; Fig. 2A). Because phosphoamino acid analysis data clearly demonstrated only threonine residues to be sites of autophosphorylation in PknK (Fig. 1C), we simultaneously mutated Thr 178 and Thr 180 in PknK to alanine residues by PCR mutagenesis. The PknK-T178A/ T180A (PknK-TATA) mutant thus created was overexpressed and purified to near homogeneity (Fig. 2B). PknK-TATA displayed no autophosphorylation (Fig. 2C). The results from kinase assays carried out with PknK and PknK-TATA using Mbp as the substrate (Fig. 2C) clearly indicated that these mutations led to a complete loss of the proteins activity (as did the K55M mutation). These results suggest that autophosphorylation events occurring at specific threonine residues in the activation loop are essential for PknK activity.

The PknK C-terminal Region Is
Phosphorylated-The C-terminal region (ϳ800 amino acids) of PknK is largely homologous to the transcription regulatory region of E. coli MalT. In the absence of any role played by this region in modulating the protein activity, the N-terminal PknK-kinase domain would phosphorylate Mbp with the same efficiency as the full-length protein. To address this question, maltosebinding protein-tagged PknK-KD403 (MBP-PknK-KD403) was overexpressed and purified. This protein retains an additional ϳ120 amino acids beyond the kinase domain to promote appropriate folding and stability. The full-length and PknK-KD (1-403 amino acids) proteins were overexpressed and purified under similar conditions ( Fig. 3A shows purified PknK-KD).
In vitro kinase assays carried out with Mbp as the substrate demonstrated that PknK-KD is active, although to a lesser extent than the full length (Fig. 3B). Upon examining their autophosphorylation patterns by tryptic peptide mapping using thin layer chromatography (Fig. 3C), two phosphorylated spots present in PknK were observed to be missing in PknK-KD (specified spots indicated in Fig. 3C by  dotted oval). These C-terminal specific phosphorylation events may play a role in regulating PknK activity.
PknK Is Localized to the Cell Membrane-The subcellular localization of a protein is usually related to its cellular function. Serine/threonine kinases containing putative transmembrane domains are believed to be involved in sensing extracellular environmental stimuli through their C-terminal domain, followed by transmitting the signals to the cytosolic kinase domain (11,38) to elicit the appropriate response. Among the 11 STPKs in M. tuberculosis, only PknG and PknK lack putative transmembrane domains. PknG has been shown to be localized to both the cytosol and cell membrane (7). To investigate the subcellular localization of PknK, anti-PknK polyclonal antibodies were raised in rabbit. The ␣-PknK antibodies thus obtained could effectively recognize PknK-KD at concentrations as low as 2 ng (Fig. 4A, lane 4). M. tuberculosis whole cell lysate, cytosol, cell wall, and the culture filtrate fractions (kindly provided by Tuberculosis Vaccine Testing and Research Materials Contract, Colorado State University) (39,40) were then analyzed for the presence of PknK by Western blotting. Detection of GlmU and PknB utilizing ␣-GlmU and ␣-PknB served as positive controls. GlmU is reported to be a cytosolic protein in E. coli (41). We detected GlmU in both cytosolic and cell wall fractions (Fig. 4B, middle panel). This is not very unusual because PknG, another cytosolic protein, has also been reported to be present in both cytosolic and cell wall fractions (7). As anticipated because of the presence of a transmembrane  domain in PknB, it was detected in the whole cell lysate and the cell wall fractions (Fig. 4B, bottom panel). However, contrary to our expectation that PknK would be cytosolic, we detected PknK in the whole cell lysate and the cell wall fractions only (Fig. 4B, top panel). PknK contains a putative PDZ domain between amino acids 465 and 533. PDZ domain-containing proteins are known to be involved in intramolecular or intermolecular interactions (42). Because PDZ domains have been shown to be involved in localizing various eukaryotic kinases to submembranous regions (43), it is possible that this domain plays a role in anchoring PknK to the cell membranes. Alternatively, PknK may be anchored to the cell membranes through interactions mediated by its C-terminal region.
PknK Phosphorylates VirS and MymA Operon Proteins-The gene encoding PknK is located ϳ1.5 kb from that of transcription factor VirS. VirS has been shown to regulate transcription of the mymA operon comprising of seven genes (44). The first gene in the mymA operon is mym, which encodes for the Mym protein that is a putative monooxygenase. Knocking out the mym or virS genes resulted in altered cell wall structure (45,46). Because transcription factors are known to be regulated by post-translational modifications such as phosphorylation (47), we decided to investigate the possibility of VirS being a PknK substrate. We also examined the possibility of the mymA operon proteins themselves being direct targets of PknK-mediated phosphorylation events. The genes encoding VirS, Mym, LipR, Rv3085, Rv3086, Rv3088, and FadD13 were cloned into the pC6-2 vector, a modified pGEX-4T2 vector that has a unique caspase-6 cleavage site inserted before the multiple cloning site (32). The gene encoding Rv3087 was cloned into pMAL-c2X. In vitro kinase assays were carried out using these proteins as putative substrates in reactions with PknK (Fig. 5A). To rule out the possibility of PknK phosphorylating the GST tag, reactions with GST protein as substrate were carried out (Fig. 5A, lanes 1 and 2). , and Rv3085 (lane 10) by PknK was 3-5-fold above background. Similar reactions carried out with MBP and MBPtagged Rv3087 revealed that Rv3087 is not phosphorylated by PknK (data not shown). These results definitively suggest that VirS, Mym, LipR, Rv3085, and Rv3088 are likely candidate substrates of PknK. To rule out the possibility of the identified candidates being spurious substrates caused by altered conformation upon GST tagging, the GST tag was cleaved off these five purified substrates by carrying out caspase-6 cleavage reactions. The caspase-6 reaction on various substrate GST-tagged proteins was carried out in the kinase buffer for 1 h (at 4°C) prior to performing in vitro kinase reactions. As is apparent from the silver-stained gel (Fig. 5B), the cleavage reaction was very efficient. Results of the kinase assays with these substrates (Fig. 5C) clearly demonstrated that VirS, Mym, LipR, Rv3085, and Rv3088 are indeed direct substrates of PknK.
PknK-mediated Phosphorylation of VirS Increases Its Affinity for mym Promoter-Because VirS is a transcription factor that had been reported to modulate the expression of the mymA operon genes, we investigated whether it is a bona fide substrate of PknK in vivo. Toward this, VirS was cloned into pDUET expression vector along with either MBP or MBP tagged PknK. E. coli BL21 codon plus cells transformed with pDUET-MBP-VirS or pDUET-PknK-VirS were metabolically labeled with [ 32 P]orthophosphoric acid. Western blot analysis of extracts made from these cells showed the expression of both PknK and VirS. However, we consistently observed decreased expression of VirS whenever PknK was co-expressed with it (Fig. 6A). A possible reason for this could be due to decreased availability of T7 RNA polymerase as well as translation machinery because MBP-PknK is a fairly large protein (molecular mass, 160 kDa). Nonetheless, VirS phosphorylation could only be detected when PknK was co-expressed with it, suggesting in vivo phosphorylation of VirS by PknK. These data validate in vitro results showing VirS to be a PknK substrate. VirS has been reported to reg- ulate the transcription of the mymA operon as well as autoregulate its own transcription (44). We investigated a possible role for PknK-mediated phosphorylation of VirS in the modulation of its DNA binding properties. Analysis of the primary sequence of VirS revealed the presence of a putative helix-turn-helix motif in the C-terminal 60 amino acids (Fig.  6B). We created a VirS mutant (VirS⌬) where the putative helix-turn-helix motif was deleted (Fig. 6B). A 156-bp DNA fragment spanning the intragenic region between virS and mym genes (Fig. 6B) was amplified, radiolabeled, and used as substrate in EMSA experiments using purified VirS, VirS⌬, and VirS purified from cells co-expressing PknK (p-VirS) (profiles of purified protein shown in Fig. 6C). Not surprisingly, we did not detect any protein-DNA complex using VirS⌬ (Fig. 6D). We detected a protein-DNA complex with both VirS and p-VirS (Fig. 6D). Interestingly, we consistently observed an increased affinity of VirS for the substrate DNA, when the protein was purified from cells co-expressing PknK (Fig. 6, D and E). Quantitation of EMSA experiments yielded apparent binding affinities (K d ) of VirS and p-VirS for the promoter DNA. It is evident from K d values that p-VirS has ϳ2.5-fold higher affinity for the promoter DNA compared with VirS (Fig. 6E). Taken together, these results suggest that PknK-mediated phosphorylation of VirS modulates its DNA binding properties, thus implicating a role for PknK in regulating the expression of both the mymA operon genes, as well as virS.
PknK Modulates VirS-mediated Transcriptional Activation from mym Promoter-To analyze the impact of increased interaction of VirS with mym promoter upon PknK-mediated phosphorylation, we used reporter assays to study transcriptional activation of the mym promoter. For this purpose we used reporter plasmids where the luciferase gene was cloned downstream of the mym promoter. Either virS or pknK or both were cloned downstream of constitutively active UV15 promoters in the same plasmid (Fig. 7A). Transcriptional activity of the promoter was measured by assaying for luciferase activity in the lysates of M. smegmatis transformants. Our results showed that mym promoter has significant basal activity, which is enhanced 2-3-fold because of expression of either PknK or VirS. However, when PknK was co-expressed along with VirS, we detected a further 1.5-2-fold increase in promoter activity. To determine whether the enhanced promoter activity is dependent on PknK kinase activity, we co-expressed VirS along with inactive PknK (PknK-K55M). Under these circumstances transcriptional activation levels dropped to that of VirS alone (Fig. 7C). Because the expression of PknK and VirS enhanced the transcription to a similar level, there exists a possibility that further increase in the transcription in M-L-K-V could be due to a combinatorial effect that is dependent on PknK activity. However, results in Figs. 5 and 6 showed that VirS is a substrate for PknK and PknK-mediated phosphorylation increases its affinity to mym promoter DNA. Taken together, the data suggest that increased transcription from mym promoter when PknK and VirS were co-expressed is likely due to PknK-mediated phosphorylation of VirS, thus implicating a role for PknK-mediated phosphorylation events in modulating VirS-mediated transcriptional activation from the mym promoter. . These reactions were resolved on 8 -15% gradient SDS-PAGE, and the gel was exposed for autoradiography. The bands corresponding to autophosphorylated PknK in the autoradiogram are indicated by arrows. The substrates showing 3-5-fold higher phosphorylation above the background GST phosphorylation are indicated by arrows. B and C, GST tag from the GST-VirS, GST-Mym, GST-LipR, and GST-Rv3087 proteins were cleaved by incubating these proteins with purified caspase-6 for 1 h at 4°C in kinase reaction buffer. Subsequent to the cleavage, in vitro kinase reactions were performed with or without PknK in 40 l of reaction volume for 15 min at 30°C. The reactions were resolved on 10% SDS-PAGE, silver-stained, dried, and exposed for autoradiography. B, silverstained gel. The bands corresponding to the cleaved proteins are indicated by arrows. C, corresponding autoradiogram. The bands corresponding to autophosphorylated PknK and phosphorylated substrates are indicated by arrows.

DISCUSSION
Eukaryotic STPK orthologs identified in bacteria have been shown to be involved in the regulation of development, stress response, and pathogenicity (11). Serine/threonine protein kinase orthologs in Yersinia pseudotuberculosis and Pseudomonas aeruginosa have been shown to be required for the virulence of these organisms (48,49). In M. tuberculosis, STPKs have been shown to modulate various cellular functions. Considering the important functions carried out by these characterized protein kinases, we decided to study the hitherto uninvestigated protein kinase PknK. Serine/threonine kinases are known to contain an invariant lysine in the Hanks' subdomain II (34), which is essential for efficient kinase activity, because it is required for orienting ATP (34). Accordingly, the conserved lysine residue at position 55 in PknK was mutated to methionine (K55M). This mutation, as predicted, abolished the kinase activity of the protein (Fig. 1B, fourth lane). Phosphorylation of specific serine and/or threonine residues in the activation loop of mycobacterial kinases has been clearly indicated to regulate the activity of these proteins. Conformational changes induced by the phosphorylation events allow appropriate positioning for substrate binding (35). Activation loop phosphorylations have been detected and the residues being phosphorylated have been identified in PknB, PknD, PknE, and PknF. Mutating one or more of the identified residues results in significant loss of activity (37). The activation loop region in serine/threonine kinases is one of the most variable regions. However, the sites of phosphorylation are relatively conserved (see sequence alignment in Fig. 2A). By mutating the threonine residues present at the pertinent positions, PknK activity was abolished (Fig. 2C, third and  fourth lanes), implying a role for activation loop phosphorylations in modulating PknK activity.
The C-terminal region of PknK shows homology with the transcription regulatory region of the transcription factor MalT, which activates transcription of the maltose regulon in E. coli (50). The MalT promoter interaction is positively regulated by ATP and maltotriose (51). The PknK C-terminal region shows homology to two domains of MalT- The position of putative helix turn helix DNA binding domain in VirS is indicated. DNA fragment used for EMSA studies is shown as a black bar. C, VirS⌬, VirS, and p-VirS (VirS purified from cells co-expressing PknK) purified as His 6 fusion proteins from E. coli were resolved on 10% SDS-PAGE and stained with Coomassie Blue. D, 7 ng of labeled DNA fragment was incubated with 125 nM VirS⌬ or different amounts of VirS or pVirS for 10 min at 37°C. DNAprotein complexes were resolved as described under "Experimental Procedures" and subjected to autoradiography. The bands corresponding to unbound DNA and protein-DNA complex are indicated by arrows. E, gel bands corresponding to bound and free DNA (unbound) were excised, and the counts were determined using a liquid scintillation counter. The percentage of DNA in complex was calculated using the formula: [counts in complex/(counts in the complex ϩ counts in free DNA)] ϫ 100. The values from three independent experiments were used to calculate the standard deviation. The data were fitted assuming a single binding mode, and the K d values were calculated using Graphpad Prism Software. domain 1 (PknK, amino acids 368 -580) and domain 2 (PknK, amino acids 581-775). Domain 1 in MalT is responsible for ATP binding and hydrolysis, whereas domain 2 forms part of the maltotriose-binding site (38). Deleting the C-terminal region of PknK abrogates its activity but does not abolish it (Fig.  3B, fourth lane). The possibility of the C-terminal region being a second site for ATP binding cannot be ruled out. Possible cooperative interactions between the two sites may be responsible for the almost 20-fold enhanced activity of the full-length protein that we detect, in comparison with kinase domain alone (data not shown). A comparative analysis of phosphopeptides of full-length PknK and PknK-KD (kinase domain of PknK) revealed specific sites of phosphorylation in the C-terminal region (Fig. 3C). Phosphorylation of the C-terminal region could be due to autophosphorylation or due to transphosphorylation of one kinase molecule by another. In eukaryotes, autophosphorylation of the C-terminal domain tyrosine residues in response to ligand binding in receptor tyrosine kinases regulates their interaction with various proteins required for downstream signaling (52). Phosphorylation of PknK in the C-terminal region may be important for its interaction with other cellular proteins.
Kinases mediate their effects on the cell by phosphorylating various cellular proteins. Thus, identifying the direct substrates of kinases is an area of rigorous investigations. Various methods have been used by different groups to identify kinase substrates. Substrates of PknA and PknB have been identified using a combination of two-dimensional gel electrophoresis and immunoblot analysis with a phosphor-(S/T)Q antibody (13). GarA, a substrate of PknB, was identified in an approach involving an in vitro kinase reaction with the kinase and mycobacterial whole cell lysate, followed by resolution of the reaction products on two-dimensional gels and identification of substrates by mass spectrometry (22). Recently, substrates of PknD have been identified by carrying out transcription profiling of wild type versus PknD-overexpressing cells with the help of microarrays (53). Novel substrates of PknA, PknB, PknH, and PknF have also been identified by carrying out in vitro kinase assays with the kinase and putative proteins encoded by neighboring genes (8,21,30). One of the major aims of this study was to identify substrates of PknK. The virS gene located near the pknK gene is not part of any operon. The virulence genes are involved in controlling pathogenicity. The mymA operon, located divergently from the virS gene, is named after the first gene in the operon, mym (mycobacterial monooxygenase), that is important for maintaining the integrity of the M. tuberculosis cell wall (46). The mymA operon encodes for seven cistrons that produce proteins proposed to be involved in the biosynthesis of mycolic acids (46). Transcription of the mymA operon was shown to be dependent on the presence of VirS (44). Disruption of virS or mym results in altered cell wall structure (46). When exposed to acidic pH, the mymA operon in the wild type M. tuberculosis shows 4-fold induction, suggesting a role for the proteins encoded by this operon in survival of the pathogen in the severely acidic conditions of activated macrophages (46).
The present study reveals VirS to be a substrate of PknK (Fig. 5, A and C) and further shows that PknK-mediated phosphorylation increases the affinity of this transcriptional activator for the promoter it activates. Because VirS is reported to induce transcription from the mym promoter under acidic conditions (44,45), luciferase activity was first assayed in the lysates of M. smegmatis transformants that were exposed to acidic conditions (data not shown). Our results showed that mym promoter has significant basal activity, which is enhanced 2-3-fold by VirS. However, co-expression of PknK did not further stimulate VirS-mediated transcriptional activation from FIGURE 7. PknK modulates VirS-mediated transcriptional activation from mym promoter at physiological pH. A, schematic representation of various luciferase reporter constructs used for this study. B and C, M. smegmatis mc 2 155 cells were electroporated with the reporter constructs, and the transformants were grown at physiological pH. B, loading controls showing the expression of PknK and VirS in various transformants. 25 g of lysates were resolved on 10% SDS-PAGE, transferred to nitrocellulose membrane, and probed with ␣PknK antibodies. FLAG tagged VirS was immunoprecipitated with FLAG-M2-agarose beads (Sigma) from 1 mg of lysate, and the immunoprecipitated samples were resolved on 10% SDS-PAGE, transferred to nitrocellulose membrane, and probed with ␣FLAG antibodies (Sigma). C, reporter assays carried out as described under "Experimental Procedures." The readings were normalized with respect to the total protein concentration and expressed as units/g of protein. The values from three independent experiments were used to calculate the standard deviation.
the mym promoter (data not shown). A possible explanation for this is an altered conformation of VirS in acidic conditions, resulting in higher levels of promoter activity that is independent of PknK. Interestingly, our results show that the co-expression of PknK and VirS stimulates transcription from the mym promoter under physiological conditions. This enhanced activation appears to be dependent on PknK kinase activity, implying that phosphorylation of VirS by PknK may be responsible for this effect (Fig. 7). In addition to VirS, several of the mymA operon proteins -Mym, LipR, Rv3085, and Rv3088are targets of direct phosphorylation by PknK (Fig. 5, A and C), suggesting the modulation of the activities of these target proteins by phosphorylation. The significance of PknK-mediated phosphorylations of all these identified substrates will be examined. The results of these investigations may help us gain useful insights into the role of PknK in M. tuberculosis biology.