Convergence of Ser/Thr and Two-component Signaling to Coordinate Expression of the Dormancy Regulon in Mycobacterium tuberculosis*

Signal transduction in Mycobacterium tuberculosis is mediated primarily by the Ser/Thr protein kinases and the two-component systems. The Ser/Thr kinase PknH has been shown to regulate growth of M. tuberculosis in a mouse model and in response to NO stress in vitro. Comparison of a pknH deletion mutant (ΔpknH) with its parental M. tuberculosis H37Rv strain using iTRAQ enabled us to quantify >700 mycobacterial proteins. Among these, members of the hypoxia- and NO-inducible dormancy (DosR) regulon were disregulated in the ΔpknH mutant. Using kinase assays, protein-protein interactions, and mass spectrometry analysis, we demonstrated that the two-component response regulator DosR is a substrate of PknH. PknH phosphorylation of DosR mapped to Thr198 and Thr205 on the key regulatory helix α10 involved in activation and dimerization of DosR. PknH Thr phosphorylation and DosS Asp phosphorylation of DosR cooperatively enhanced DosR binding to cognate DNA sequences. Transcriptional analysis comparing ΔpknH and parental M. tuberculosis revealed that induction of the DosR regulon was subdued in the ΔpknH mutant in response to NO. Together, these results indicate that PknH phosphorylation of DosR is required for full induction of the DosR regulon and demonstrate convergence of the two major signal transduction systems for the first time in M. tuberculosis.

oxide (NO) 5 (2). The survival and persistence of M. tuberculosis in this environment requires the ability to sense external signals and mount an effective adaptive response. M. tuberculosis possesses multiple families of signal transduction systems, including the Ser/Thr protein kinases (STPKs) and the two-component regulatory systems (TCSs) (3).
In a previous study, we found that the STPK PknH functions as an in vivo growth regulator (4). Hypervirulence was consistently detected in BALB/c mice infected with a pknH deletion mutant in M. tuberculosis after 3-4 weeks of infection (4), corresponding to the onset of adaptive immunity. Therefore, we hypothesized that M. tuberculosis uses the PknH kinase-mediated pathways to respond to host-induced signals to regulate its in vivo growth. Nitric oxide produced by the inducible nitricoxide synthase of the host macrophages plays a key role in controlling bacillary growth during the chronic phase of infection following activation of the host immune response (5). In vitro experiments revealed that the ⌬pknH mutant is more resistant to NO compared with WT (4), indicating that PknH may act as a sensor of NO to regulate M. tuberculosis growth in vivo.
Predictions from bioinformatics analysis and studies using in vitro kinase assays have identified three endogenous substrates of PknH kinase: EmbR (6), a transcriptional regulator of the embCAB genes involved in lipoarabinomannan and arabinogalactan synthesis; DacB1, a cell division-related protein; and Rv0681, a putative transcriptional regulator (7). However, the substrates and downstream effectors of PknH signaling in response to NO stimulus have yet to be discovered.
The DosR system, also known as DevR, is one of 11 pairs of TCSs present in M. tuberculosis (3). It is well established that DosR responds to hypoxia, NO, and CO via signaling through two cognate sensor kinases, DosS (DevS) and DosT (8,9) to activate transcription of a defined set of ϳ50 genes termed the "dormancy" or DosR regulon (10). Genes belonging to the DosR regulon, including dosR, are up-regulated in the Wayne model of dormancy (10,11), under low-oxygen tension (12)(13)(14), and in response to NO (10) and CO (15,16) and are believed to be involved in the adaptation of M. tuberculosis to a non-replicating persistent state in latent tuberculosis infection.
In this work, we demonstrate convergence of the two major signal transduction systems, the STPK and the TCS, for the first time in M. tuberculosis. Using a global proteomics approach, we identified members of the DosR regulon to be disregulated in a pknH deletion mutant in M. tuberculosis. We show that DosR is a substrate of PknH Thr phosphorylation and that cooperative DosS Asp phosphorylation and PknH Thr phosphorylation enhance DosR-DNA binding. Enhanced binding in vitro correlates with up-regulation of the DosR regulon in WT M. tuberculosis compared with ⌬pknH in response to NO. These results suggest that PknH and the Dos TCS coordinately regulate expression of a key physiological response of M. tuberculosis.

M. tuberculosis Growth/Stress Conditions-M. tuberculosis
H37Rv and a mutant strain lacking pknH, described previously (4), were grown in Middlebrook 7H9 broth supplemented with 10% albumin/dextrose/sodium chloride and 0.05% Tween 80. For iTRAQ analysis, strains were grown in rolling cultures to A 600 Ϸ 1.0, harvested, washed, and resuspended in acidified (pH 5.4) Middlebrook 7H9 broth/Tween 80/albumin/dextrose/ sodium chloride with 3.0 mM NaNO 2 and harvested after 48 h in standing cultures as described previously (4). For qRT-PCR analysis, cultures were grown in rolling cultures to A 600 Ϸ 0.3 and treated for 4 h with NaNO 2 or diethylenetriamine/NO as indicated. Cells were washed and resuspended when using acidified media.
iTRAQ and LC-MS/MS-The iTRAQ assay and phosphopeptide identification were performed by the University of Victoria Proteomics Centre (British Columbia, Canada; see supplemental "Methods").
RNA Extraction and qRT-PCR-Previously described procedures were followed for RNA extraction and qRT-PCR analysis (4). Primers for qRT-PCR are listed in supplemental Table S1. Results were analyzed using GraphPad Prism software. All values were normalized to cDNA expression levels of sigA.
Cloning, Expression, and Purification-Plasmids and primers used for cloning and site-directed mutagenesis are listed in supplemental Table S1. The genes pknH- , dosR, and dosS-(378 -578) were amplified from M. tuberculosis H37Rv genomic DNA using standard methods. The dosR gene was cloned into the pET22b vector; dosS-(378 -578) was cloned downstream of G-protein coding sequence into a modified pGEV2 vector (17), pJC8 (see supplemental "Methods"). Sitedirected mutagenesis was performed as described previously (7). For cell-based phosphorylation experiments, dosR was transferred into the pET30b kanamycin-resistant vector (producing an identical DosR recombinant protein), pknH was cloned into the pGEX-4T3 ampicillin-resistant vector, and both were cotransformed into Escherichia coli BL21. Expression of all proteins was carried out in E. coli BL21(DE3) as described (7), followed by purification on nickel-nitrilotriacetic acid columns (Qiagen) according to the supplied protocol.
Phosphoamino Acid Stability and Analysis-PknH-phosphorylated DosR was separated by SDS-PAGE and transferred onto 0.45-m PVDF membranes. Stability of the incorporated phosphate was tested by treating membranes with 1 N HCl, 3 N NaOH, or ddH 2 O overnight at room temperature and visualized by phosphorimaging. Phosphoamino acid analysis was performed as described (18) using cellulose plates and resolved in one dimension with isobutyric acid and 0.5 M NH 4 OH (5:3, v/v).
The Trp auxotrophic strain of M. smegmatis and plasmids pL240 and pL242, containing the N-and C-terminal fragments (N Trp and C Trp ) of N-(5Ј-phosphoribosyl)anthranilate isomerase, respectively, were generously provided by Helen O'Hare. The Split-Trp experiment was performed as described (20), with the following modifications. N Trp and C Trp were transferred from pL240 and pL242 into pALACE (21) and pPE207 (22) and designated pJC10 (hygromycin-resistant) and pJC11 (apramycin-resistant), respectively, to place the resulting fusion proteins under control of the inducible acetamidase promoter (see supplemental "Methods"). The indicated genes were PCR-amplified and cloned into pJC10 and pJC11. All inserts were sequenced. Cotransformed M. smegmatis Trp Ϫ was spotted (5 l) onto Middlebrook 7H10 broth, 1% glucose, 60 g/ml histidine, 50 g/ml hygromycin, and 30 g/ml apramycin plates; supplemented or not with either 0.02% acetamide or 120 g/ml Trp; and grown for 2-3 weeks at 30°C.
EMSA-Oligonucleotides corresponding to the combined central and proximal (CP) DosR boxes and the distal (D) box upstream of hspX were designed with 5Ј-guanine overhangs when annealed (supplemental Table S1). Radioactive [␣-32 P]dCTP was incorporated by Klenow (Fermentas) according to the supplied protocol. DosR (64 pmol) was phosphorylated by incubation with and without 0.2 g each of pre-autophosphorylated PknH, DosS, and both PknH and DosS, followed by incubation with 4 pmol of radiolabeled CP or D DosR boxes. Binding conditions were as described previously (23). Samples were resolved by 5% nondenaturing Tris borate/ EDTA PAGE. Gels were dried, and radiolabeled DNA bands were detected by phosphorimaging.

RESULTS
PknH-dependent Protein Expression-To draw a global picture of the regulation mediated by PknH kinase, we compared the protein expression profiles of WT and ⌬pknH M. tubercu-losis using the quantitative MS-based proteomics approach, iTRAQ (24). On the basis of our previous study showing that the ⌬pknH mutant survives better than WT M. tuberculosis and the complemented strain in standing cultures treated with acidified nitrite (NaNO 2 , an NO donor under acidic conditions) (4), we compared the global protein levels of the ⌬pknH mutant and its WT parental strain with and without a 48-h NaNO 2 treatment. We were able to identify and simultaneously compare the expression of 784 proteins using a cutoff of 95% probability in the identification of peptides (supplemental Table S2). Of these, 447 proteins were identified using at least two highconfidence peptides (Ͼ95%). Of the 331 proteins identified with a single high-confidence peptide, 262 proteins were identified with at least a second unique peptide of lower confidence (Ͻ95%), resulting in a cumulative unused protein score of Ͼ2.0. In total, 706 proteins were identified with high confidence based on the unused protein score of Ͼ2.0 (see supplemental "Methods" for data analysis). Fig. 1A shows the distribution of the ⌬pknH/WT protein level ratios based on their chromosomal location. The ratios of individual protein levels ranged from 0.67 to 2.36 for untreated samples and from 0.58 to 1.97 for NO-treated samples. To iden-tify proteins that differentially responded to NO through PknH signaling, we plotted the changes in protein expression due to NO treatment in ⌬pknH versus WT (⌬pknH/⌬pknH ϩ NO versus WT/WT ϩ NO) (Fig. 1B). Grouping the data into four clusters using the K-mean clustering algorithm, we identified the cluster with the highest mean attribute value to contain nine proteins having greater levels in the ⌬pknH mutant compared with WT and responding to NO treatment (Fig. 1B, circled). Strikingly, eight of the nine proteins are encoded within the DosR regulon. Further examination of the iTRAQ data revealed that of the 48 genes commonly regulated by DosR in response to NO, hypoxia, and the Wayne model of dormancy (10), 13 gene products were identified by iTRAQ, all of which had higher protein levels in the ⌬pknH mutant after 48-h standing conditions (Table 1). With the addition of NaNO 2 , these 13 DosR-regulated proteins were induced to similar levels in WT and ⌬pknH (Table 1).
PknH Phosphorylation of DosR on Thr-As each of the 13 proteins identified displayed the same expression pattern in ⌬pknH compared with WT, we reasoned that this pattern likely represents the entire DosR regulon and suggests that the mechanism of PknH regulation occurs at the level of the transcriptional regulator, DosR.
The transcriptional activity of DosR is dependent on phosphorylation of Asp 54 by its cognate histidine kinases, DosS and DosT, in response to hypoxia, NO, and CO (9,12). However, on the basis of our data, we hypothesized that PknH kinase regulates DosR activity by Ser/Thr phosphorylation. We therefore conducted in vitro kinase assays to test whether PknH phosphorylates DosR. As shown in Fig. 2A, DosR was phosphorylated when incubated with recombinant PknH, whereas DosR alone did not undergo autophosphorylation. Circled points represent proteins that clustered with the highest mean attribute value based on K-mean clustering of data points; eight of the nine points represent proteins encoded in the DosR regulon.

TABLE 1 iTRAQ comparison of DosR-dependent protein levels
Ratios (⌬pknH/WT) of DosR-inducible proteins identified by iTRAQ analysis are shown for cultures with and without treatment with 3.0 mM acidified nitrite. In untreated samples, DosR-regulated proteins that are commonly induced by NO (10), hypoxia (12), and standing conditions (32) all had elevated levels in the ⌬pknH mutant but approached ratios of 1.0 upon acidified nitrite treatment. This pattern of expression was not observed for proteins (Rv1177, Rv0231, and Rv3841) whose gene expression is DosR-dependent in response to hypoxia and standing conditions but not to NO. USPA, Universal Stress Protein A. Phosphoamino acid analysis identified that DosR was phosphorylated on Thr (Fig. 2B). Phosphorylation was acid-stable and alkali-labile, characteristic of Thr phosphorylation, but not of Asp phosphorylation (supplemental Fig. S2) (25).

MS/MS analysis identified PknH-monophosphorylated DosR at
Thr 198 (supplemental Fig. S1A) and Thr 205 (supplemental Fig. S1B) of the trypsin-digested 198 TQAAVFATELKR 209 peptide located in the C-terminal domain of DosR. Site-directed mutagenesis of DosR confirmed these findings: DosR(T198A) had reduced ability to be phosphorylated by PknH, and DosR(T205A) and DosR(T198A/T205A) were nearly abolished for PknH phosphorylation (Fig. 2C). Next, we used E. coli, which lacks any known STPKs, as a surrogate host to test PknH phosphorylation of DosR in a cellbased system. The active kinase domain of PknH and fulllength recombinant DosR were coexpressed in E. coli. MS/MS analysis of DosR purified from the PknH-expressing strain identified monophosphorylated (supplemental Fig. S1, C and D) and diphosphorylated ( Fig. 3 and supplemental Fig. S1E) DosR at the previously identified Thr 198 and Thr 205 residues, with diphosphorylation being the predominant species.
PknH Interaction with DosR in Mycobacteria-To determine whether PknH interacts with DosR in vivo, we performed two separate protein-protein interaction assays in M. smegmatis: the Split-Trp assay (20) and mycobacterial protein fragment complementation assay (19). In the former, protein-protein interaction leads to the reassembly of the N-and C-terminal fragments (N Trp and C Trp ) of N-(5Ј-phosphoribosyl)anthranilate isomerase, an enzyme required for Trp biosynthesis. In the  latter, reassembly of complementary fragments F1,2 and F3 of mDHFR confers resistance to trimethoprim.
As shown in Fig. 4A, using the mycobacterial protein fragment complementation system, coexpression of DosR-F1,2 and PknH-(1-401)-F3 reconstituted mDHFR expression as determined by trimethoprim resistance, indicating that PknH interacts with DosR in vivo. The interaction between PknH and DosR in M. smegmatis was slightly weaker than the positive control obtained with the interaction of the yeast Gcn4 dimerization domains but is consistent with the transient nature of kinase-substrate interactions. Using the less sensitive Split-Trp system, interaction between PknH and WT DosR was not observed; however, interaction between PknH and the phosphorylation-defective DosR(T198A/T205A) mutant restored growth of the Trp auxotrophic strain of M. smegmatis in the absence of exogenous Trp (Fig. 4B). This result suggests that the PknH-DosR interaction is dependent on the phosphorylation status of DosR. DosR(T198A/T205A) likely acted as a kinase-trapping mutant, where PknH was able to bind but not release DosR(T198A/T205A) due to its inability to be phosphorylated. This result is in agreement with Split-Trp studies related to another M. tuberculosis protein kinase, PknG, which interacts significantly better with its phosphorylation-defective substrate, GarA(T21A), compared with WT GarA (26). Taken together, these results provide further evidence that PknH interacts with and phosphorylates DosR in mycobacteria.
Enhanced DNA Binding of PknH-phosphorylated DosR-DosR is able to bind its cognate DNA sequence, the DosR box (12), and Asp phosphorylation enhances DosR-DNA binding (27). We therefore assessed the effect of PknH on DosR-DNA binding using EMSA. We compared the DNA-binding ability of PknH Thr-phosphorylated DosR with unphosphorylated and DosS Asp-phosphorylated DosR. We tested the binding of DosR to the D and CP DosR boxes in the promoter region of hspX, a DosR regulon member that we found to be disregulated in the ⌬pknH mutant. Thr phosphorylation of DosR by PknH enhanced binding of DosR to the D site of the hspX promoter in a manner comparable with Asp phosphorylation of DosR by DosS (Fig. 5A). The binding of DosR to the D site was further enhanced by the combined phosphorylation of DosR by both PknH and DosS (Fig. 5A). In the absence of DosR, PknH did not cause a shift to the DNA (data not shown). PknH phosphorylation of DosR also enhanced binding of DosR to the CP site, although DosS phosphorylation of DosR did not affect DosR binding to this site (Fig. 5B). This latter result may suggest that DosR has different affinities for different DosR boxes, although the absence of enhanced binding may also be a result of DosS dephosphorylation of DosR, as DosS catalyzes this reverse reaction very shortly after Asp phosphorylation (28).  We also tested the DNA-binding characteristics of DosR that had been phosphorylated by PknH in our cell-based system. Equal amounts of DosR (as determined by Bradford assay and Coomassie Blue staining of gel-separated purified protein) were purified from E. coli with or without coexpression of PknH and used in the EMSA assay. As shown in Fig. 5 (C and D), cell-based phosphorylation of DosR by PknH not only resulted in a significantly more intense band but also caused an additional shift to both D and CP DosR boxes. These results indicate that PknH phosphorylation of DosR enhances its binding to cognate DNA sequences.
The C-terminal domain of DosR binds to its DNA sequences as a tetramer of two DosR dimers (29). As PknH phosphorylation enhanced DosR-DNA binding, we wanted to see if phosphorylation enhances DosR dimerization in mycobacteria using the Split-Trp method. Constant Thr phosphorylation was mimicked by mutating the two Thr phosphoacceptors of DosR to Glu (DosR(EE)) (30). As shown in Fig. 4C, interaction between the two C-terminal domains of the phosphomimetic DosR(EE) mutant restored growth of M. smegmatis Trp Ϫ , whereas interaction between WT DosR proteins did not enable growth. Growth was not observed with either full-length WT or DosR(EE) proteins (data not shown) and was expected, as fulllength DosR exists in an inactive conformation (31).
Transcription Profiling of the DosR Regulon in M. tuberculosis-As we observed greater DosR-DNA binding upon phosphorylation of DosR by PknH, we hypothesized that this increase in DNA binding would correlate to increased DosR regulon transcription in M. tuberculosis. We therefore used qRT-PCR to measure DosR regulon expression in WT M. tu-berculosis compared with ⌬pknH. For a broad coverage of the DosR regulon spanning the M. tuberculosis genome, we looked at the expression of eight DosR regulon genes whose products were identified in our iTRAQ analysis and five additional DosR regulon genes not identified by iTRAQ. All values were normalized to the housekeeping sigA gene, whose expression is affected neither by NO stress (32) nor by acidic conditions (33).
A heat map of the ⌬pknH/WT ratios normalized to sigA expression for all 13 genes tested is shown in Fig. 6A (for graphical analysis, see supplemental Fig. S3). Basal level transcription of the DosR regulon in aerobic early log phase growth was unaffected by pknH deletion (first column). The addition of NaNO 2 resulted in an ϳ2-fold lower expression of the DosR regulon genes in ⌬pknH compared with WT M. tuberculosis (second column). Because NaNO 2 can also generate reactive oxygen intermediates, we verified these results using diethylenetriamine/NO, a specific NO donor, and found a similar 2-fold decrease in DosR regulon expression in ⌬pknH (third column). Expression of the regulon under standing conditions was also lower in the mutant, but to a lesser extent (fourth column). Finally, to mimic the combined low oxygen and presence of NO likely encountered in the host, standing conditions with the NO donors were tested and resulted in a similar decrease to DosR regulon expression in ⌬pknH (fifth and sixth columns).
The decreased DosR regulon expression in the pknH mutant was due to an impaired induction of each gene following NO treatment (Fig. 6B). Comparison of gene expression in cultures treated with NO relative to basal level transcription revealed strong induction of the DosR regulon in WT M. tuberculosis (mean of 22.5-fold, maximum of 83.0fold) but weaker induction in the ⌬pknH mutant (mean of 11.6-fold, maximum of 42.4-fold) (Fig. 6B). Although this ϳ2-fold difference is relatively moderate, it is comparable with the 40 -60% impaired induction observed in single knock-out mutants of DosS and DosT under hypoxic conditions (27,34). The modest expression may also be due to compensating function(s) of other STPKs present in ⌬pknH M. tuberculosis, as many of the STPKs appear to have substantial cross-talk activity (35). Nevertheless, these results indicate that PknH is required for full induction of the DosR regulon and agree with our EMSA analysis, suggesting that enhanced DNA binding of jointly Asp-and Thr-phosphorylated DosR leads to an increase in transcription of the regulon.

DISCUSSION
In this study, we have demonstrated for the first time that the dormancy regulon, an important and major regulatory response in the human pathogen M. tuberculosis, is controlled by two distinct signal transduction systems, the STPK and the TCS. We have shown that DosR is a substrate of PknH phosphorylation in vitro and in multiple cell-based systems. We also provide evidence that PknH phosphorylation of DosR enhances DosR dimerization and DNA binding, resulting in up-regulation of the DosR regulon in response to NO. A correlation between PknH and DosR has been suggested previously (36), and in this study, we provide the experimental basis to support this hypothesis.
Integration of these two types of signaling systems has been reported in other biological systems. In Streptococcus agalactiae, the STPK Stk1 phosphorylates the two-component response regulator CovR to repress CovR-dependent transcription of a secreted cytotoxin and to impede CovR transcriptional repression of a ␤-hemolysin/cytolysin gene (37) by inhibiting CovR-DNA binding (38). In Myxococcus xanthus, STPKs and a TCS coordinately regulate developmental changes in response to nutrient depletion. Expression of mrpC, encoding a transcription factor involved in fruiting body and myxospore formation, is transcribed by the TCS MrpAB but inhibited by Ser/ Thr phosphorylation by the Pkn8/Pkn14 STPK cascade (39,40). Convergence of STPKs and TCSs is also seen in eukaryotes where TCSs regulate activation of MAPK (Ser/Thr) signaling (41)(42)(43). Intriguingly, the HstK protein from the nitrogen-fixing Anabaena sp. PCC 7120 (44) and the NTHK2 ethylene receptor in tobacco plants (45) possess both Ser/Thr and histidine kinase activity. These examples demonstrate that STPKs and TCSs can be coupled to control a common signal transduction pathway.
Although further experiments are needed to elucidate the mechanism of action of PknH, the position of PknH phosphorylation suggests a potential means of post-translational regulation. Both phospho-Thr 198 and phospho-Thr 205 map to the critical regulatory helix ␣10 in the crystal structure of DosR (31). As suggested by Wisedchaisri et al. (31), DosR activation is dependent on the flexibility of this helix. In their model, helix ␣10 is in dynamic equilibrium in the closed-inactive conformation, bound to the N-terminal regulatory domain, burying the key Asp 54 residue, and in an open-inactive conformation, allowing Asp 54 to be solvent-exposed part of the time and thus available for Asp phosphorylation by DosS/T. Upon activation by Asp phosphorylation, helix ␣10 provides the DosR dimerization interface in an open-active conformation for DNA binding. Integration of PknH and DosS/T signal transduction systems controlling DosR activity would allow for tighter control of DosR-dependent activity. Activation of DosR by its cognate his-tidine kinases, DosS and DosT, results in a strong induction of the DosR regulon (27), and this induction is believed to be involved in metabolic changes that result in the pathogen entering a non-replicating persistent state. It is reasonable to expect mechanisms to be in place to prevent the pathogen from entering non-replicating persistence in the absence of an appropriate signal. Furthermore, nonspecific transcription and translation of the ϳ50 genes encoded in the DosR regulon would be considerably energy-costly. As a required second trigger (in addition to DosS/T) for full induction of the DosR regulon, PknH acts as a "molecular modulator" to repress nonspecific induction of the regulon and as an amplifier of the regulon in the presence of an appropriate signal.
The global proteomics approach proved to be a powerful tool for identifying key components in the PknH signal transduction pathway. However, our transcriptomic results seemingly contradict our proteomic data. The iTRAQ experiment was designed based on the enhanced survival of the pknH mutant in stationary phase growth exposed to lethal quantities of NaNO 2 (4) and was not designed to test specific DosR induction conditions. Furthermore, the 48-h time point tested in the iTRAQ experiment was well beyond the short-lived induction of the DosR regulon, whose gene expression largely returns to base line by 24 h (10,46). Therefore, due to the difference in conditions and time points tested, the iTRAQ and qRT-PCR data cannot be directly compared. The somewhat discrepant results may indicate, however, that PknH also plays a role in inhibiting or turning off DosR regulon expression beyond the 24-h induction period. Further experiments, including a time-dependent analysis of DosR regulon expression under controlled conditions, would be required to test this hypothesis.
Deletion of pknH results in hypervirulence after 3-4 weeks of infection (4), corresponding to the induction of the host adaptive immune response and production of NO (47). It is tempting to speculate that the hypervirulence observed in the ⌬pknH mutant may be mediated via signaling though DosR. An initial report indicated that deletion of dosR results in hypervirulence in mouse models (48). However, subsequent studies indicated that ⌬dosR displays either attenuation or no difference in pathogenicity in mice, guinea pigs, and rabbits compared with WT M. tuberculosis (46, 49 -51). Curiously, deletion of at least two members of the DosR regulon, hspX and Rv2623, each results in hypervirulence in mice (52,53). Up-regulation of the DosR regulon has also been associated with hypervirulence, as genes belonging to the DosR regulon are constitutively upregulated in the hypervirulent W-Beijing lineage of M. tuberculosis (54). It therefore remains a challenge to identify whether PknH signaling through DosR and/or the other known substrates contributes to the growth regulation and adaptation during the chronic or latent phase of infection.