Phosphorylation of Mycobacterial PcaA Inhibits Mycolic Acid Cyclopropanation

Background: Mycolic acid cyclopropanation plays an important role in mycobacterial virulence but molecular mechanisms controlling cyclopropanation remain unknown. Results: Phosphorylation of the cyclopropane synthase PcaA decreases mycolic acid cyclopropanation and intracellular survival and abrogates the phagosome maturation block (PMB). Conclusion: PMB modulation is dependent on mycobacterial protein phosphorylation. Significance: This study highlights Ser/Thr-dependent phosphorylation of a mycolic acid biosynthetic enzyme in mycobacterial virulence. Pathogenic mycobacteria survive within macrophages by residing in phagosomes, which they prevent from maturing and fusing with lysosomes. Although several bacterial components were seen to modulate phagosome processing, the molecular regulatory mechanisms taking part in this process remain elusive. We investigated whether the phagosome maturation block (PMB) could be modulated by signaling through Ser/Thr phosphorylation. Here, we demonstrated that mycolic acid cyclopropane synthase PcaA, but not MmaA2, was phosphorylated by mycobacterial Ser/Thr kinases at Thr-168 and Thr-183 both in vitro and in mycobacteria. Phosphorylation of PcaA was associated with a significant decrease in the methyltransferase activity, in agreement with the strategic structural localization of these two phosphoacceptors. Using a BCG ΔpcaA mutant, we showed that PcaA was required for intracellular survival and prevention of phagosome maturation in human monocyte-derived macrophages. The physiological relevance of PcaA phosphorylation was further assessed by generating PcaA phosphoablative (T168A/T183A) or phosphomimetic (T168D/T183D) mutants. In contrast to the wild-type and phosphoablative pcaA alleles, introduction of the phosphomimetic pcaA allele in the ΔpcaA mutant failed to restore the parental mycolic acid profile and cording morphotype. Importantly, the PcaA phosphomimetic strain, as the ΔpcaA mutant, exhibited reduced survival in human macrophages and was unable to prevent phagosome maturation. Our results add new insight into the importance of mycolic acid cyclopropane rings in the PMB and provide the first evidence of a Ser/Thr kinase-dependent mechanism for modulating mycolic acid composition and PMB.

Tuberculosis (TB) 3 caused by Mycobacterium tuberculosis (Mtb), remains a major threat to global health, claiming the life of two million individuals annually (1). Mtb is an obligate human pathogen predominantly growing within host phagocytes. In these cells, bacilli reside in phagosomes, which they prevent from undergoing maturation and fusion with lysosomes (2,3). In this manner, the bacilli remain in a weakly acidic and non-cytolytic environment. An early requirement for the phagosome maturation block (PMB) is the establishment of an all-around close apposition between the phagosome membrane and the mycobacterial surface that will occur only when phagosomes contain a single mycobacterium (loner phagosomes) (reviewed in Ref. 4). Whenever a phagosome contains more than one mycobacterium or mycobacterial clumps (social phagosome), the close apposition is not maintained in regions spanning adjacent bacteria. Such phagosomes systematically mature and fuse with lysosomes (reviewed in Ref. 4). Likewise, physical abrogation of the close apposition, by binding of par-ticles to the mycobacterial surface (3), leads to phagosome maturation and fusion with lysosomes. Mycobacteria are, however, not destroyed in phagolysosomes. Instead, they can rescue themselves from this cytolytic environment to again reside in non-maturing phagosomes that no longer fuse with lysosomes (Ref. 3,reviewed in Ref. 4). This, however, does not make the PMB redundant. The preferred site of residence for a mycobacterium in host macrophages remains a non-matured phagosome, regardless of whether this is achieved directly by the PMB or by re-establishing it after rescue from phagolysosomes (3). In addition, the PMB could be a strategy for sequestering pathogenic mycobacteria away from antigen-presenting compartments, thereby altering the host immune response (5).
The molecular basis for the PMB is the focus of intense research. Several mycobacterial determinants, including cell surface-associated lipids such as lipoarabinomannan (LAM) and cord factor, as well as proteins, were reported to take part in this process (6 -9). Among these, the mycobacterial Ser/Thr Protein kinase (STPK), PknG, secreted into macrophage phagosomes, is one of the candidates mediating both the PMB and intracellular survival of BCG and Mtb (10). It has, therefore, been proposed that modulation of phagosome-lysosome fusion by PknG could be mediated through phosphorylation of host proteins following secretion within the macrophage (10).
The family of STPK, expressed by both BCG and Mtb, is composed of 11 members (11). Whether membrane-associated STPKs play a role in regulating phagosome maturation through phosphorylation of mycobacterial proteins remains to be established. Recent studies have demonstrated the participation of STPKs in the regulation of many physiological pathways, such as those connected to cell shape/division and cell envelope biosynthesis (12). In particular, it was shown that all major type II fatty acid synthase (FAS-II) components are phosphorylated by STPKs and that, in general, phosphorylation inhibits the activity of the FAS-II enzymes (12)(13)(14)(15)(16). These studies culminated with the recent demonstration that InhA, the primary target of isoniazid, is controlled via phosphorylation and that phosphorylation of InhA severely impeded enoyl reductase activity, mycolic acid biosynthesis, and mycobacterial growth in a manner similar to isoniazid (17). Although these studies suggest that Mtb controls in a very subtle manner its FAS-II system by regulating each step of the elongation cycle, no information is available regarding whether STPK-dependent phosphorylation may also regulate the activity of enzymes involved in functional modification of mycolic acids, such as cyclopropane synthases. Previous studies identified the pcaA gene as required for mycolic acid cyclopropanation in both BCG and Mtb (18). The corresponding pcaA mutants failed to produce serpentine cords and, importantly, were unable to persist within and kill infected mice, indicating that PcaA is necessary for the establishment of a lethal and chronic infection (18). An attractive hypothesis is that regulation of PcaA activity by phosphorylation represents a strategy employed by pathogenic mycobacteria to regulate mycolic acid cyclopropanation and, as a consequence, mycobacterial pathogenesis.
Herein, we explored whether PcaA represents an endogenous substrate for mycobacterial STPKs and whether it participates in the PMB and intracellular survival of bacilli in human monocyte-derived macrophages. From our data, we propose that STPK-dependent phosphorylation of PcaA plays a critical role in regulating mycolic acid cyclopropanation, which in turn affects phagosome processing and intramacrophage survival.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Growth Conditions-Escherichia coli strains used for cloning and expression of recombinant proteins are listed in supplemental Table S1. Bacteria were grown in LB medium at 37°C. Media were supplemented with ampicillin (100 g/ml) or spectinomycin (100 g/ml) as required. BCG Pasteur 1173P2 and mc 2 2801 (BCG Pasteur pcaA::Tn5370) (18) strains were grown on Middlebrook 7H10 agar plates with OADC enrichment (Difco) or in Sauton's medium containing 0.05% tyloxapol (Sigma) with kanamycin (25 g/ml) or hygromycin (50 g/ml) when required.
Cloning, Expression, and Purification of PcaA, MmaA2, and Mutant Proteins-pcaA and mmaA2 were amplified by PCR using Mtb H37Rv chromosomal DNA as a template and primers listed in supplemental Table S2. Amplified products were cloned into the pETPhos vector (19), generating pETPhos_ PcaA and pETPhos_MmaA2. Site-directed mutagenesis was directly performed on pETPhos_PcaA using inverse PCR amplification with self-complementary primers (supplemental Table S2). The duet strategy (20) was used to generate phosphorylated PcaA in the pCDFDuet-1 vector harboring the PknF kinase domain. All constructs were verified by DNA sequencing. The different His-tagged recombinant proteins were overexpressed in E. coli C41(DE3) and purified using TALON Metal affinity resin (Clontech) as described previously (14). Cloning, expression, and purification of GST-tagged STPKs were described earlier (16).
Overexpression and Purification of PcaA from BCG-pcaA was cloned into the shuttle vector pVV16 (21) using the primers listed in supplemental Table S2. The resulting construct pVV16_PcaA_WT was used for directed mutagenesis to generate the pVV16_PcaA_T168A/T183A mutant. The constructs were electroporated into mc 2 2801. Transformants were grown and used for purification of the His-tagged PcaA protein as described above. The recombinant proteins were used for immunoblotting using anti-phosphothreonine antibody according to manufacturer's instructions (Santa Cruz Biotechnology).
In Vitro Kinase Assay-In vitro phosphorylation was performed as described (22)  Enzymatic Assays-Methyltransferase activity of PcaA and MmaA2 was measured by using the couple assay for mycolic acids as described (23) with slight modifications. Briefly, reactions were performed in a total volume of 600 l and followed spectrophotometrically at 412 nm to detect the conversion of SAH (methyltransferase product) to homocysteine by the Mtb SAH hydrolase. Assays were performed in presence of 100 mM (pH 7.5) phosphate buffer containing 1.5 M SahH, 250 M NAD, and 40 M of unsaturated fatty acids (cis, cis-11,14-eicosadienoic acid or linoelaidic acid, Sigma) pre-equilibrated for 5 min at 37°. After addition of 400 M 5,5Ј-dithio-bis(2-nitrobenzoic acid) (DTNB), the solution was blanked and added to pre-incubated 100 M SAM and 2 M of either PcaA or MmaA2 to start the reaction.
Complementation Studies and Immunoblotting-mc 2 2801 was transformed with pMV261, pMV261_PcaA_WT, pMV261_PcaA_T168A/T183A or pMV261_PcaA_T168D/ T183D (supplemental Table S1) constructed as follows. The pcaA gene (lacking the stop codon) and a 500-bp promoter sequence was cloned in fusion with an HA epitope in the C-terminal position by using the primers listed in supplemental Table S2. Site-directed mutagenesis was directly performed on pMV261_PcaA_WT using inverse PCR amplification with selfcomplementary primers (supplemental Table S2). Western blotting was performed on whole-cell extracts from cultures harvested at early stationary phase according to standard protocols with an anti HA-11 (12CA5) monoclonal antibody (Roche).
Analysis of the Mycolic Acid Profile-Mycobacterial cultures were metabolically radiolabeled with 1 Ci/ml [2-14 C]acetate (56 mCi/mmol, Amersham Biosciences) and added to mid logarithmic phase cultures for 5 h. Labeled mycolic acids were then extracted as described previously (24) and loaded onto thin layer chromatography (TLC) silica-coated plates. Mycolic acids were then separated using petroleum ether/acetone (95/5, v/v) and exposed overnight to a film.
Cording Assay-Approximately 50 colony-forming units (CFU) were plated on cord reading agar, incubated at 37°C for 3 weeks and observed under a light microscope MVX10 (Olympus).
Purification of Blood Monocytes, Differentiation into Macrophages (HMDM), and Infection-Human blood samples, purchased from the French National Blood provider of Montpel-lier, were collected from fully anonymized non-tuberculous control donors. This study was conducted according to the principles expressed in the Helsinki Declaration. Purified monocytes, isolated as described (25), were seeded onto 24-well plates at a density of 7 ϫ 10 5 /ml in complete culture medium (RPMI containing 10% FCS) and differentiated into macrophages (HMDM) with rh-M-CSF (10 ng/ml) for 7 days. HMDM were infected with exponentially growing BCG cultures (DO ϭ 0.8) at an MOI of 1:1 or 5:1. Single-cell bacterial suspensions were generated by 2 ϫ 10s pulses in a water-bath sonicator, followed by passage through a 26-gauge needle to disrupt remaining bacterial clumps. Before infection, residual bacterial aggregates were removed by low-speed centrifugation (120 ϫ g) for 2 min. Infected cells were washed with PBS and reincubated in bacteria-free medium. For CFU scoring, cells were lysed with 0.1% Triton X-100 in PBS at selected times post-infection. Serial dilutions of the lysate were plated onto Middlebrook 7H10 agar medium supplemented with OADC. Colonies were counted after incubation at 37°C for about 3 weeks.
Labeling of Endomembrane Compartments with Endocytic Tracers-Two methods were used as follows: (i) horseradish peroxidase (HRP) uptake. At 1 or 5 days post-infection, infected cells were incubated for 90 min at 37°C in complete medium containing 25 g/ml HRP (26). Endocytosis was stopped by fixing the cells overnight with 2.5% glutaraldehyde prepared in 0.1 M cacodylate buffer, pH 7.2, containing 0.1 M sucrose, 5 mM CaCl 2 and 5 mM MgCl 2 . Cells were then processed for HRP cytochemistry as described below (26). (ii) Latex bead uptake and transfer to lysosomes. Uninfected cells were washed twice with medium and exposed for 30 min at 37°C to hydrazidemodified latex beads of 0.1 m in diameter diluted in medium. After three washes in medium, cells were incubated for 2 h at 37°C in bead-free medium to chase the tracer to lysosomes (27). Cells were then infected with BCG at a MOI of 5:1 for 4 h,  washed, and re-incubated in BCG-free medium. One day later, cells were fixed with glutaraldehyde and processed for EM.
Horseradish Peroxidase Cytochemistry-Cells exposed to HRP and fixed with glutaraldehyde were washed overnight at 4°C with sucrose-containing buffer. After three washes with sucrose-free buffer, cells were incubated with 3,3Ј-diaminobenzidine tetrahydrochlorate (DAB)-H 2 O 2 as before (26) and then processed for EM.
Processing for Transmission Electron Microscopy-Cells fixed with glutaraldehyde were processed as described (28). Briefly, cells were washed with complete cacodylate buffer, and postfixed for 1 h at room temperature with 1% osmium tetroxide in the same buffer devoid of sucrose. They were washed with buffer, scraped off the dishes, concentrated in 2% agar in cacodylate buffer and treated for 1 h at room temperature with 1% uranyl acetate in Veronal buffer. Samples were dehydrated in a graded series of ethanol and embedded in Spurr resin. Thin sections (70 nm-thick) were stained with 1% uranyl acetate in distilled water and then with lead citrate and observed under the electron microscope (Zeiss 912).
Quantitation of Loner Phagosomes with Lysosomal Material-In all cases, 50 -100 different loner phagosomes taken at random from three different experiments were examined for the presence or absence of lysosomal material. Care was taken to avoid serial sections.

PcaA Is Phosphorylated in Vitro by Multiple Ser/Thr Kinases-
We first sought out to determine whether PcaA and MmaA2, that introduce cyclopropanes at the proximal and distal positions on the ␣-meromycolic acid, respectively (18,29) (Fig. 1A) were modified by phosphorylation. This changes the physicochemical properties of defined Ser or Thr residues by introducing negative charges that ultimately affect protein activity. The kinase domains of several Mtb STPKs (PknA to PknL), expressed as GST-tagged fusion proteins (16), were incubated with purified Mtb PcaA or MmaA2 and [␥-33 P]ATP. SDS-PAGE/autoradiography analysis indicated that PcaA, but not MmaA2, was phosphorylated by several kinases (Fig. 1B). Based on the intensity of the radioactive signal corresponding to phosphorylated PcaA, PknF appears to be the most efficient kinase to phosphorylate PcaA in vitro. Signals were weak with PknD, PknE, and PknH that all display various autokinase activities (16) but none were detected with PknA or PknB. Although MmaA2 is structurally highly related to PcaA (30), only PcaA is a specific substrate and interacts with PknF (Fig. 1B). This suggests that cyclopropanation of the ␣-mycolic acid proximal position may be regulated in mycobacteria by extracellular cues.
To address the relevance of PcaA phosphorylation in mycobacteria, recombinant BCG overexpressing PcaA_WT was analyzed by Western blotting using anti-phosphothreonine antibodies. Antibody specificity for the phosphorylated isoform was first assessed using PcaA purified from either E. coli (pET-Phos-pcaA) or E. coli co-expressing PknF and PcaA (pETDuet-pcaA). Phosphorylated PcaA from pETDuet-pcaA, but not unphosphorylated PcaA from pETPhos-pcaA, was specifically recognized by the antibodies (Fig. 3C). That PcaA from BCG carrying pVV16-pcaA was in a phosphorylated state was confirmed by a specific band recognized by the anti-phosphothreonine antibodies (Fig. 3C). Western blot analysis of recombinant PcaA purified from either exponential or stationary cultures of BCG carrying pVV16-pcaA showed similar levels of PcaA phosphorylation. This suggests that PcaA phosphorylation is growth-phase independent (Fig. 3C). No specific phosphorylation signals were detected in PcaA_T168A/T183A purified from BCG carrying pVV16-pcaA_T168A/T183A, thus excluding the existence of additional phosphorylation sites. Therefore, phosphorylation occurs at Thr-168 and Thr-183 both in vitro and in mycobacteria.
Phosphorylation Decreases PcaA Methyltransferase Activity-Multiple sequence alignments of PcaA orthologues from various mycobacterial species indicated that both Thr-168 and Thr-183 are conserved in Mtb, M. marinum, and M. leprae, whereas the second phosphorylation site (corresponding to Thr-183 in the Mtb sequence) is substituted by a proline in the M. smegmatis orthologue (Fig. 4A). Phosphorylation site mapping on the three-dimensional structure (30) emphasized the strategic roles of both threonine residues on PcaA activity (Fig.  4B). Thr-168 is located within the ␤-sheet of the core SAMmethyltransferase fold with the threonine side-chain facing a mobile protein segment that refolds upon cofactor-binding (33). Thr-183 was located in the so-called ␣2-␣3 specific motif of mycolic acid SAM-methyltransferases seemingly involved in the interaction with the mycobacterial acyl carrier protein AcpM (30,33). This motif, which delineates the entrance of the hydrophobic channel from the enzyme surface to the active site, displays the largest deviation when comparing the known structures of mycolic acid SAM-methyltransferases (33). The above data suggest that Thr-168 and Thr-183 are the unique phosphoacceptors in PcaA that directly influence its enzymatic

Regulation of Mycolic Acid Cyclopropanation
activity. This prompted us to compare the methyltransferase activity of phosphorylated and non-phosphorylated isoforms using a coupled assay originally developed for CmaA2 (23). This colorimetric assay is based on the detection of S-adenosylhomocysteine (SAH) conversion to homocysteine by SAH hydrolase. Two unsaturated fatty acids were used as substrates of Mtb PcaA and MmaA2; cis, cis 11,14-eicosadienoic acid and linoelaidic acid (Fig. 4C). Although these lipids are structurally distinct from the authentic unsaturated meromycolyl-AcpM substrates, both PcaA and MmaA2 catalyzed cyclopropanation of double bonds in presence of either lipid (Fig. 4, D and E). This reaction was inhibited by dioctylamine as reported earlier for CmaA2 (23) (data not shown). Importantly, the activity of phosphorylated PcaA, derived from E. coli carrying pETDuet-PcaA, was significantly reduced by 40 to 50% compared with nonphosphorylated PcaA from E. coli carrying pETPhos-PcaA, irrespective of the lipid substrate. Conversely, the MmaA2 methyltransferase activity remained unchanged whether or not it was produced from the pETPhos or pETDuet vectors, consistent with the fact that MmaA2 is not phosphorylated by STPKs (Fig. 1B). These results indicate that, in vitro, both PcaA and MmaA2 catalyze the introduction of cyclopropane rings on cis, cis 11,14-eicosadienoic acid and linoelaidic acid and support the critical role of STPK-dependent phosphorylation in regulating PcaA activity.
BCG PcaA Phosphomimetic Mutants Display Altered Phenotypes-Acidic residues such as aspartic acid (Asp) qualitatively mimic the phosphorylation effect with regard to functional activity (13-15, 17, 34). Therefore, phosphoablative (Thr to Ala replacements) and phosphomimetic (Thr to Asp replacements) pcaA alleles were generated and introduced into BCG mc 2 2801 (BCG Pasteur pcaA::Tn5370) (18). In these constructs, WT, phosphoablative (T168A/T183A) and phosphomimetic (T168D/T183D) PcaA isoforms were fused to a HA tag and their expression was monitored by Western blotting using anti-HA antibodies. As expected, BCG ⌬pcaA transformed with the empty vector (pMV261) failed to express PcaA whereas similar amounts of PcaA_WT, PcaA_T168A/T183A and PcaA_T168D/T183D were synthesized in the corresponding strains (Fig. 5A), allowing subsequent phenotype comparison. Whether phosphorylation of PcaA affects de novo mycolic acid biosynthesis was next investigated by labeling the various BCG strains with [1-14 C]acetate followed by mycolic acid extraction and thin layer chromatography (TLC)/autoradiography analysis. BCG mc 2 2801 exhibited an altered ␣-mycolic acid profile (18,35), that could be restored by complementation with either WT or phosphoablative PcaA (Fig. 5B). In contrast, introduction of PcaA_T168D/T183D failed to restore the parental mycolic acid profile ( Fig. 5B and supplemental Fig. S1) indicating that the phosphomimetic isoform is unable to complement the lack of PcaA activity in mc 2 2801, presumably as a consequence of its reduced methyltransferase activity (Fig. 4D). This effect was not due to a tertiary structure change, as suggested by modeling of the PcaA_T168D/T183D mutant structure. Indeed, introduction of Asp at position 168, unaccessible to solvent, and at position 183, that is solvent exposed, does not seem to induce steric or electrostatic conflicts that could alter the protein fold (supplemental Fig. S2). Thus, it can be inferred that PcaA phosphorylation in mycobacteria reduces/abolishes its activity, supporting the view that this post-translational modification plays a key role in regulating ␣-mycolic acid cyclopropanation.
The effect of constitutive PcaA phosphorylation on the colonial morphology of BCG was next examined. Although no differences were observed for the different BCG strains cultured in liquid medium (Fig. 6B), mc 2 2801 grew far more slowly than its parental strain on agar plates, as evidenced by colony size (supplemental Fig. S3). This growth defect could be fully restored by complementing the mutant strain with either PcaA_WT or PcaA_T168A/T183A, but not with the phosphomimetic mutant. Cording morphology, in which bacteria are intertwined into serpentine rope-like structures, is a distinctive feature of pathogenic mycobacteria. The characteristic mc 2 2801 cord formation defect (18) was restored following complementation with either WT or phosphoablative pcaA alleles. However, the phosphomimetic construct failed to restore cord formation (Fig. 5C). Thus, PcaA phosphorylation is strongly associated with a defective mycolic acid profile, subsequently affecting mycobacterial growth and serpentine cord formation.

Functional PcaA Is Required for Intracellular Survival and PMB-Although
PcaA cyclopropanation is required for Mtb persistence and pathology in late stages of infection in mice (18), its role during early infection remains unclear. This prompted us to investigate the ability of BCG ⌬pcaA to infect and replicate within human monocyte-derived macrophages (HMDM). Determination of intramacrophagic bacterial counts showed that survival of the mutant strain was severely compromised as compared with the WT strain (Fig. 6A). This was not due to an inherent growth defect as the mutant and parental strains replicated equally well in liquid medium (Fig. 6B).
Pathogenic mycobacteria survive in macrophages by residing in phagosomes which they prevent from maturing and fusing with lysosomes (2,36). We sought out to examine the putative role of PcaA in this process. Cells were exposed to electrondense endocytic tracers by two different approaches. In the first, small latex beads were chased to lysosomes prior to phagocytic uptake of the WT or the ⌬pcaA strains; in the second, the endocytic tracer horseradish peroxidase (HRP) was added at selected time points after infection (1 or 5 days post-infection) with the above strains and the cells were stained for HRP by EM cytochemical methods. Phagosome processing was analyzed in terms of fusion and intermingling of contents between phagosomes and the different endomembrane compartments, namely early endosomes (eEN) and late endosomes/lysosomes (indicated as Ly in text and figures) that can be easily distinguished from one another by the presence of latex beads (only in Ly) and the cytochemical staining pattern for HRP (37). To analyze phagosome processing, it is important to distinguish between loner and social phagosomes since, and as shown before, the PMB does not hold for social phagosomes that invariably mature and fuse with lysosomes (reviewed in Ref. 4). Therefore, when looking for the effects of PcaA on the PMB, observations and quantitations had to be limited strictly to loner phagosomes. One must also keep in mind that non-matured phagosomes fuse with eEN but are unable to fuse with Ly (4). With both strains, two types of loner phagosomes were encountered, i.e. (i) immature phagosomes either displaying a thin rim of HRP reaction product as is typical of eEN (not shown) or no marker (Fig. 6, C and F), and (ii) phagosomes that contain lysosomal material indicating that they have matured and fused with Ly (Fig. 6, D and G). As before, social phagosomes, containing two or more mycobacteria, all contained lysosomal material (Fig. 6, E and H). A quantitative analysis showed a 2-fold increase in the percentage of loner phagosomes displaying lysosomal material for BCG ⌬pcaA-containing phagosomes as compared with the WT strain (Fig. 6I). A more complete study done with HRP as endocytic tracer showed that the percentage of phagolysosomes remained stationary for at least 5 days post-infection (50% for the BCG ⌬pcaA mutant versus about 25% for the WT strain) (Fig. 6J). Overall, these results emphasize the major contribution of PcaA in the PMB.
Phosphorylation of PcaA Inhibits Intramacrophage Growth and Affects Phagosome Maturation-The above data indicate that phosphorylation of PcaA is linked to a defective cell wallassociated mycolic acid profile in BCG and affects its intracellular growth. We next determined the consequences of PcaA phosphorylation on both intramacrophage BCG growth (Fig.  7A) and the PMB (Fig. 7B). For both studies, HMDM were infected with the parental BCG strain, the ⌬pcaA mutant or the different BCG complemented strains. Over a 6-day period the CFU count increased 4 -5-fold for the parental strain and for both BCG ⌬pcaA strains complemented with the WT or phosphoablative isoforms. The phosphomimetic strain was unable to replicate within HMDM, similarly to the ⌬pcaA mutant (Fig.  7A). Concerning the PMB, only 30% of the loner phagosomes containing the ⌬pcaA mutant complemented with the WT or phosphoablative PcaA contained lysosomal material, similarly to the parental WT strain. In contrast, 60% of the loner phagosomes containing the phosphomimetic strain contained lysosomal material, as it was the case for the ⌬pcaA mutant (Fig.  7B). Altogether, these results indicate that phosphorylation of PcaA prevents growth and abrogates the PMB.

DISCUSSION
The present study extends previous work showing the regulatory role of phosphorylation on mycolic acid chain length and functionalization during its biosynthesis (12). More importantly, it provides the first evidence for a Ser/Thr kinase phosphorylation-dependent regulation of cyclopropane synthase. This mechanism is of special relevance as cyclopropanes are major contributors to the physiopathology of Mtb infection (18). Phosphorylation of Thr-168 and Thr-183 residues caused a strong decrease in methyltransferase activity in vitro. In the case of growing bacteria, a phosphomimetic mutant displayed a major alteration of the mycolic acid profile due to the lack of di-cyclopropanation of ␣-mycolic acids, similar to what is observed for the ⌬pcaA mutant. Likewise, impaired cyclopropanation was associated with altered colony morphology and inability to form serpentine cords (Fig. 8). Major phenotypical alterations have also been observed in in vivo situations. Inac- tivation of Mtb pcaA causes attenuation of virulence and a less severe granulomatous pathology in mice (18). In addition, PcaA-dependent cyclopropanation of TDM, a direct effector of Mtb pathogenesis, regulates its inflammatory activity (18,38). In particular, TDM purified from a PcaA mutant is hypoinflammatory for macrophages during early stages of infection (38).
We show here that phosphorylation of PcaA not only arrests intracellular mycobacterial replication but also abrogates the PMB. In addition to the requirement for a tight apposition of the phagosome membrane to the mycobacterial surface all around (3,4,39), which suggests an intimate cross-talk between the mycobacterium and its host cell, a large number of both cellular and mycobacterial molecules were seen to play a role in, or correlate with, the PMB and intracellular survival of mycobacteria (4,40,41). This leads to the question of whether any one of these is sufficient by itself and, if so, why is there such redundancy (3)? Present knowledge rests upon two types of experimental approaches: either eliminating the function of the molecular factor concerned and/or reconstituting the factor in an artificial system. With these approaches, i.e. deletion of pcaA and construction of phosphomimetic or phosphoablative mutants, we have studied phagosome maturation for a sufficiently long time (5 days versus a few hours only in many other studies) to conclude that mycolic acid cyclopropanation by PcaA maintains the PMB and that PcaA phosphorylation, by inhibiting cyclopropanation, abrogates this block indefinitely. It is noteworthy that part of the phagosomes, however, do not mature and fuse with lysosomes upon PcaA phosphorylation, suggesting that other mycobacterial factors such as TDM (9,42) or phospho-signaling proteins like PknG (10), the protein tyrosine phosphatase PtpA (43,44) and the secreted lipid phosphatase SapM (45) might also be important mycobacterial factors. Among these, only SapM and PtpA seem to interfere directly with host physiological processes leading to prevention of phagosome maturation, whereas PknG seems to affect phagosome maturation through phosphorylation of yet unknown host proteins following its secretion within macrophage phagosomes.
Previous genetic screens identified several Mtb mutants defective in PMB (8,46) and affected in various pathways, including lipid synthesis. Of notable interest is the Mtb pcaA::Tn mutant that failed to arrest phagosome maturation and trafficked to late phagosomal compartments in bone-marrow derived macrophages (46), thus supporting the view that the data obtained with BCG in the present study are relevant to Mtb.
With the PMB being the most conspicuous survival mechanism of pathogenic mycobacteria, why does Mtb phosphorylate PcaA? One must keep in mind that the fate of pathogenic mycobacteria does not rely exclusively on its own defense mechanisms. Host cells have developed several strategies for combating invading pathogens which, in turn, have developed several strategies for using the host environment to their advantage. In the present case, an attractive hypothesis would be that PcaA phosphorylation is driven by cell host factors located in the phagosome membrane and hence in direct contact with the mycobacterial surface due to the close apposition between the two structures. These factors would act as a signal for triggering autophosphorylation of mycobacterial STPKs, inducing in turn PcaA phosphorylation (Fig. 8). Mycolic acid structural alteration could affect the close apposition between the phagosome membrane and the mycobacterial surface. As a result, phagosomes would mature and fuse with lysosomes, as observed in the present study. However, whether PcaA phos-  phorylation is reduced/inhibited once the close apposition between the phagosome membrane and the mycobacterial surface is no longer maintained all around, cannot be determined with present technologies as it is not possible to separate the different types of phagosomes from one another.
For a successful infection, it must be assumed that pathogenic mycobacteria will not be destroyed in macrophages. In many situations, growth is arrested in phagolysosomes but mycobacteria survive in this cytolytic environment without significant loss of viability (2)(3)(4)39). Our study is in full agreement with the above situations as shown by the survival curves and the morphological appearance of bacilli within phagolysosomes.
In summary, this study provides conceptual advances in our understanding of the mycolic acid metabolic adaptation and regulatory events exploited by pathogenic mycobacteria to adapt their mycolic acid cell wall content. Although very challenging, future studies should now help to identify extracellular cues sensed by the different kinases and leading to PcaA phosphorylation. This work also strengthens the biological importance of PcaA in the physiology and virulence of the bacilli and provides evidence of a Ser/Thr kinase-dependent mechanism for modulating the composition of mycolic acids, a key component of the mycobacterial cell wall, and for maintaining mycobacteria in a non-matured phagosome, a hallmark for mycobacterial survival within host cells. Our results suggest that displacement of the unphosphorylated/phosphorylated PcaA balance in favor of the phosphorylated isoform rapidly leads to PMB inhibition and loss of intracellular survival. Thus, from an applied point of view and considering that PcaA has been proposed as a target for drug development against persistent bacilli (18,47), the selective inhibition of PcaA activity through constitutive phosphorylation may strongly impair Mtb survival, opening new and original perspectives for future anti-tuberculosis drug development. It is indeed noteworthy that small molecules, such as bryostatin (48) can activate STPK, which may be of great therapeutic value in inhibiting Mtb intracellular growth.