Characterization of Phosphopantetheinyl Hydrolase from Mycobacterium tuberculosis

ABSTRACT Phosphopantetheinyl hydrolase, PptH (Rv2795c), is a recently discovered enzyme from Mycobacterium tuberculosis that removes 4′-phosphopantetheine (Ppt) from holo-carrier proteins (CPs) and thereby opposes the action of phosphopantetheinyl transferases (PPTases). PptH is the first structurally characterized enzyme of the phosphopantetheinyl hydrolase family. However, conditions for optimal activity of PptH have not been defined, and only one substrate has been identified. Here, we provide biochemical characterization of PptH and demonstrate that the enzyme hydrolyzes Ppt in vitro from more than one M. tuberculosis holo-CP as well as holo-CPs from other organisms. PptH provided the only detectable activity in mycobacterial lysates that dephosphopantetheinylated acyl carrier protein M (AcpM), suggesting that PptH is the main Ppt hydrolase in M. tuberculosis. We could not detect a role for PptH in coenzyme A (CoA) salvage, and PptH was not required for virulence of M. tuberculosis during infection of mice. It remains to be determined why mycobacteria conserve a broadly acting phosphohydrolase that removes the Ppt prosthetic group from essential CPs. We speculate that the enzyme is critical for aspects of the life cycle of M. tuberculosis that are not routinely modeled. IMPORTANCE Tuberculosis (TB), caused by Mycobacterium tuberculosis, was the leading cause of death from an infectious disease before COVID, yet the in vivo essentiality and function of many of the protein-encoding genes expressed by M. tuberculosis are not known. We biochemically characterize M. tuberculosis’s phosphopantetheinyl hydrolase, PptH, a protein unique to mycobacteria that removes an essential posttranslational modification on proteins involved in synthesis of lipids important for the bacterium’s cell wall and virulence. We demonstrate that the enzyme has broad substrate specificity, but it does not appear to have a role in coenzyme A (CoA) salvage or virulence in a mouse model of TB.

pNPP), respectively. PptH failed to hydrolyze pNPP but hydrolyzed bis-pNPP to the chromogenic p-nitrophenol (pNP) product as detected by its absorbance at 405 nm. The reaction was carried out at pH 7.0 in the presence (Fig. 1B) and absence (see Fig. S1A in the supplemental material) of Mn 21 and Fe 21 . Phosphodiesterase activity on bis-pNPP was observed in the absence of exogenous metal ions ( Fig. 1C; Fig. S1B), probably because of copurification with metal ions (8), as supported by the decrease in activity after dialysis and addition of 2 mM EDTA to chelate residual metal ( Fig. 1C; Fig. S1B).
Activity was enhanced in the presence of Mn 21 or Fe 21 ( Fig. 1C; Fig. S1B), both of which were found in the active site of PptH crystals (8), as well as with Ni 21 ( Fig. 1C; Fig. S1B). As little as 1.0 mM Mn 21 or Fe 21 enhanced the specific activity of PptH (0.5 mM) by 2-fold or 3-fold, respectively, with maximum activity observed with 62.5 mM Mn 21 (8-fold increase) or Fe 21 (11-fold increase) (Fig. 1D). Mn 21 and Fe 21 each enhanced phosphodiesterase activity of PptH to a similar extent.
Effects of pH and ionic strength. Hydrolysis of bis-pNPP by PptH was tested over the pH range of 6 to 9, using MES (morpholineethanesulfonic acid) as the buffer at pH 6.0 and 6.5 and Tris over the range of 7.0 to 9.0. Phosphodiesterase activity was optimal at pH 7.0 to 7.5 (Fig. 1E) and was enhanced more than 4-fold compared to pH 6 or pH 9. Activity was highest (6 mmol/min/mg) in the absence of added NaCl and decreased with increasing ionic strength (Fig. S1C).
Kinetic parameters. Michaelis-Menten kinetic parameters of PptH with bis-pNPP were K m of 17.8 6 2.8 mM, V max of 8.6 6 0.8 mmol/min/mg, and k cat of 5.7 6 0.9 s 21 (Fig. 1F). The high K m highlights the low affinity of PptH for the artificial substrate.
Effect of potential modulators. Next, we tested the impact on the phosphodiesterase activity of PptH of cyclic GMP (cGMP), ATP, ADP, AMP, cAMP, guanosine tetraphosphate (ppGpp), coenzyme A (CoA) and polyphosphate (polyP). Only polyP and AMP reduced enzyme activity by .50% at 1 mM (97.5% and 66.8%, respectively) (Fig.  S2A). The 50% inhibitory concentrations (IC 50 s) of both polyP and AMP were high (109 mM and 456 mM, respectively) ( Fig. S2B and S2C), arguing against their potential to serve as physiologic modulators. We cannot rule out the possibility that some of these small PptH has phosphodiesterase activity when tested using the phosphomonoesterase and phosphodiesterase substrates para-nitrophenylphosphate (pNPP) and bis p-nitrophenyl phosphate (bis-pNPP), respectively. Reactions were carried out at pH 7.0 with Mn 21 and Fe 21 . (C) Metal dependence of PptH activity. Mn 21 , Fe 21 , and Ni 21 enhance PptH phosphodiesterase activity. (D) Dependence of PptH phosphodiesterase activity on the concentration of Mn 21 or Fe 21 . (E) Optimal phosphodiesterase activity of PptH at neutral pH. The y axis indicates milli-absorbance of the product, para-nitrophenol (p-NP), over time. (F) The kinetic parameters of PptH for substrate bis-pNPP; K m is 17.8 mM and V max is 8.6 mmol/ min/mg. Results show means 6 standard deviation (SD) from one experiment performed in at least triplicate. All experiments were performed at least two independent times. molecules serve as substrates for PptH, as is the case for AcpH and Rv0805, another annotated metallophosphoesterase from M. tuberculosis (4,16).
Probe-based assay for the identification of PptH substrates. After optimizing PptH phosphodiesterase activity on the nonphysiological substrate bis-pNPP, we investigated the activity of PptH on native substrates. The ability of PptH to catalyze the removal of the Ppt moiety from holo-AcpM was shown by SDS-PAGE gel shift of holo-to apo-AcpM and by mass spectrometric analysis of released Ppt (7). Unlike AcpM, most CPs do not show a gel shift upon conversion from holo-CP to apo-CP. To facilitate visualization of this reaction on additional substrates, we synthesized a modified CoA molecule (*CoA) with biotin attached to the sulfhydryl group of Ppt ( Fig. 2A). When added to apo-CP by a PPTase, this biotinylated Ppt (*Ppt) served as a probe for Ppt hydrolase activity using a gel-based assay with streptavidin detection ( Fig. 2A). P. aeruginosa AcpH was shown to remove coumarin-or rhodamine-modified Ppt from different CPs (17). Our biotinylated probe is reminiscent of maleimide (MAL)linked biotin-CoA conjugated probes (11,18), but we replaced the MAL linker with an acetamide (ACM) linker. The biotin-polyethylene glycol (PEG)-ACM-CoA probe (*CoA) avoided a problem encountered with the probe with a MAL linker, where reaction with CPs without addition of PPTase occurred upon denaturation of the samples for SDS-PAGE (not shown).
Upon incubation of the *CoA probe with apo-AcpM and Sfp, a broad-acting PPTase from Bacillus subtilis (9-12), we observed labeling of holo-AcpM ( Fig. 2B; *holo-AcpM) as indicated by the streptavidin-reactive band. After removal of unreacted probe and upon addition of M. tuberculosis PptH, we observed time-dependent, PptH-dependent release of the label from *holo-AcpM (Fig. 2C). Mass spectrometry corroborated disappearance of label from the substrate. A peak with a retention time of 4.8 min and mass of 773 Da, corresponding to *Ppt (biotin-PEG-ACM-Ppt), formed upon addition of PptH (Fig. S3). Another peak, with a retention time of 2.7 min and mass of 359 Da, corresponding to unlabeled Ppt, also formed. This likely arose from holo-AcpM in the substrate mixture that copurified with apo-AcpM.
Nonredundancy of PptH in M. tuberculosis. AcpH and PptH are structurally distinct Ppt hydrolases that appear to have evolved convergently. AcpH is expressed only in Gramnegative bacteria, while PptH is present in mycobacteria (3,7,8). Indirect evidence suggested that AcpH might be the only Ppt hydrolase in E. coli (3). To determine whether M. tuberculosis encodes multiple proteins with Ppt hydrolase activity, we added His-tagged holo-AcpM to whole-cell lysates of M. tuberculosis and asked whether Ppt is removed in lysates from wildtype bacteria and in lysates from M. tuberculosis lacking PptH activity, either because the gene encoding PptH was disrupted or because the cells expressed a mutant of PptH that conferred as much resistance to a PptT inhibitor as did the gene knockout (7). We monitored this reaction by visualizing the gel shift between apo-and holo-AcpM by immunoblot for the His-tag on AcpM. Holo-AcpM shifted to apo-AcpM in whole-cell lysates from wild-type bacteria but was stable in lysates lacking PptH or containing the inactive H246N mutant of PptH (7) (Fig. 4). These results suggest that PptH is the only Ppt hydrolase in M. tuberculosis that can remove Ppt from holo-AcpM under the conditions studied.
Lack of a role for PptH in CoA salvage. Ppt is an intermediate in the biosynthesis of CoA in M. tuberculosis; it is the product of the CoaC-mediated decarboxylation of 49-phosphopantothenoyl-L-cysteine (20). Ppt is also formed by the PanK-mediated phosphorylation  of pantetheine (21). We reasoned that the release of Ppt from CPs by PptH could serve as a salvage pathway to maintain CoA pools in M. tuberculosis. To test this hypothesis, we used a strain of M. tuberculosis in which addition of anhydrotetracycline (Atc) results in conditional knockdown (cKD) of panB to deplete the cell's CoA pool (22). PanB catalyzes the first step in the biosynthesis of CoA by converting 3-methyl-2-oxobutanoate to 2-dehydropantoate, and silencing of panB in M. tuberculosis is bacteriostatic (22). We identified pptH loss of function mutants in the panB cKD strain by selecting for resistant mutants to the reported PptT inhibitor, the amidinourea 8918 (MIC 90 , 3 mM) (7), and resequencing the locus encoding PptT and PptH. We identified an 8918-resistant strain with the point mutation C225R in PptH (Fig. 5A). In the crystal structure of PptH, C225R is located in the active site (8) within hydrogen bond distance to H248, which is involved in metal coordination (8). Mutation from cysteine to arginine is hypothesized to ablate PptH activity due to steric clash, leading to loss of metal binding, similar to what has been observed for other mutants resistant to 8918. We found that the level of PptH C225R is reduced compared to PptH wild type (WT), consistent with the proposed role of C225 in metal binding (Fig. S4). We next monitored the survival of strains upon silencing of panB in the background of PptH C225R or in the background of WT PptH, using both liquid and solid agar cultures. Atc-induced transcriptional silencing of panB resulted in growth attenuation in both culture formats ( Fig. 5B and C), as reported (22). There was no difference in growth between strains with or without functional PptH ( Fig. 5B and C). The results suggest that PptH does not play a role in recycling of Ppt in M. tuberculosis under the conditions tested.
Lack of a phenotype for PptH-deficient M. tuberculosis after aerosol infection of mice. PptH is conserved in mycobacteria, including Mycobacterium leprae with its greatly reduced genome, suggesting that PptH has a critical function that might be manifest in an infected host. To test this, we infected C57BL/6 mice with M. tuberculosis strains WT, DpptH, DpptH:pptH WT , and DpptH:pptH H246N . All strains achieved a comparable bacterial burden in lungs through 150 days of infection (Fig. 6A), as well as in spleen and liver ( Fig. S5A and S5B). Pulmonary histopathology was also comparable in mice infected with all four strains, as examined at day 150 postinfection (Fig. 6B).

DISCUSSION
Biologic and biochemical evidence indicated that PptH catalyzes the removal of Ppt from holo-CPs and converts at least one of them into apo-CPs (7). The crystal structure of PptH placed it in a new family of phosphohydrolases and revealed an active site occupied by a Mn-Fe binuclear center (8). In this study, we show that PptH has phosphodiesterase  . "%remaining" indicates the percentage of *holo-CP normalized to no PptH control at t = 0 (calculated using ImageJ). Experiments were performed three independent times; one representative is shown. activity dependent on Mn 21 and Fe 21 . E. coli AcpH, which is not homologous to PptH but catalyzes an analogous reaction, is also activated by divalent metal ions (2). The metal ions activate a water molecule to serve as the nucleophile in the hydrolysis reaction. In contrast to the similarity in metal dependence of these two characterized hydrolases, the pH profiles are distinct, with PptH active near neutral pH and AcpH in an alkaline pH range (2). The catalytic efficiency (k cat /K m ) of PptH is similar to that of PaAcpH (Ppt hydrolase from P. aeruginosa) on the artificial substrate bis-pNPP (4). A comparison of catalytic activity of PptH and Rv0805, another M. tuberculosis protein annotated as a metallophosphoesterase with a Mn-Fe binuclear center (8,16) underscores PptH's poor affinity for bis-pNPP (K m , 17.8 mM) and its low maximal velocity (V max , 8.6 mmol/min/mg) compared to those of Rv0805 (K m , 0.9 mM; V max , 74 mmol/min/mg). Although PptH phosphodiesterase activity is inhibited by polyP and AMP, this effect is only evident at concentrations above the physiologic range.  To test the activity of PptH on native substrates, we developed a biotin-linked probe (biotin-PEG-ACM-CoA) for an assay to visualize the release of Ppt from CPs. This probe is reminiscent of those developed previously (11,18) but afforded lower background in our assays. We used the broad-specificity Sfp PPTase to covalently attach the probe to the conserved serine of CPs (18). This assay revealed that M. tuberculosis PptH catalyzes the removal of Ppt from the M. tuberculosis holo-PKS13-ACP domain. PKS13 is essential for mycolic acid biosynthesis and is activated by PptT (14). M. tuberculosis PptH also hydrolyzed Ppt from nonnative holo-CPs, including those from FAS and NRPS, from E. coli, K. pneumoniae, and P. aeruginosa. AcpHs from other organisms are also reported to have broad specificity (3)(4)(5).
AcpH appears to be limited to Gram-negative bacteria, and PptH, to mycobacteria (3,7,8). M. tuberculosis does not encode a homolog of ATPHs, and since PptH has broad specificity, we reasoned that it may be the only Ppt hydrolase in M. tuberculosis. Whole-cell lysates from wild-type M. tuberculosis hydrolyzed holo-AcpM, whereas lysates from M. tuberculosis lacking PptH or lacking functional PptH (H246N) did not. These results suggest that PptH is likely the only protein responsible for removing Ppt from holo-AcpM in M. tuberculosis and may be the primary Ppt hydrolase in M. tuberculosis. It is possible that under other conditions, including conditions of stress, another hydrolase can compensate for the loss of PptH and perform an analogous function.
The physiological function of Ppt hydrolase activity remains to be determined (3,7). It is possible that the Ppt released by PptH is incorporated into the CoA biosynthetic pathway, but we saw no evidence that such a route, if it exists, makes a difference to the survival of M. tuberculosis whose growth is limited by reduced synthesis of CoA. Jackowski and Rock found that the rate of turnover of Ppt from holo-ACP is increased when CoA is low (23). However, they also showed that the released Ppt is excreted and not recoverable (24), as has been confirmed (3). It is possible that the Ppt produced by PptH in M. tuberculosis is also excreted and not recoverable. Neither AcpH nor PptH is essential in organisms in which they are encoded (3,7), and we show here that PptH does not contribute to the survival of M. tuberculosis in a mouse model of infection.
The lack of a phenotype for PptH-deficient M. tuberculosis in a standard mouse model of infection encouraged us to consider what important aspects of the life cycle of M. tuberculosis are not modeled by infection of experimental animals. The major part of the life cycle that goes unstudied in such models is aerosol transmission. Our current efforts to explore the biology of deactivation of holo-CPs by PptH are focused on modeling the changes in state associated with survival of M. tuberculosis in necrotic, hypoxic pulmonary lesions, partial reoxygenation in cavities, aerosolization in fully aerated, rapidly evaporating microdroplets of biologically relevant fluids, and rehydration in pulmonary alveolar lining fluid.

MATERIALS AND METHODS
Overexpression and purification of PptH (Rv2795c). An overnight culture of E. coli BL21-AI expressing truncated PptH (aa 1 to 310) (8) was diluted 1:100 in 2YT (yeast tryptone) broth. The culture was grown to an optical density at 600 nm (OD 600 ) of 0.8 followed by cold shock on ice for 20 to 30 min. Inducers (1 mM isopropyl-b-D-thiogalactopyranoside [IPTG] and 0.2% arabinose) were added to the culture, which was kept shaking overnight at 18°C. Bacteria were pelleted by centrifugation and lysed using a French press by resuspending the cell pellet in 50 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 5 mM MgCl 2 , 10% glycerol, and 5 mM imidazole. Cell debris was separated by centrifugation, and the supernatant was incubated with Ni-nitrilotriacetic acid (NTA) beads for 1 h at 4°C. Beads were loaded onto a gravity column and washed with buffers of 50 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 5 mM MgCl 2 , and 10% glycerol containing 10 mM, 20 mM, or 50 mM imidazole. Protein was eluted using 50 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 5 mM MgCl 2 , 10% glycerol, and 250 mM imidazole. The eluted protein was dialyzed in 50 mM Tris-HCl, pH 8.0, 20 mM NaCl, 5 mM MgCl 2 and 10% glycerol, applied to a HiTrap Q HP anion exchange column, and eluted with a linear gradient of 0 to 2500 mM NaCl. The protein yield was ;2.2 mg/liter of culture. The protein was flash-frozen in liquid N 2 and stored at 280°C until use.
Overexpression and purification of carrier proteins and Sfp. The genes encoding CPs from E. coli (ACP, EntB), K. pneumoniae (ACP), and P. aeruginosa (ACP) were cloned in a pET 28a vector and expressed in BL21-AI (primers are listed in Table S1). M. tuberculosis PKS13 ACP domain was cloned into pMCSG7 (primers are listed in Table S1) and expressed in BL21-AI. Overnight cultures of the transformants were inoculated into fresh 2YT medium and grown to log phase. The culture was put on ice for 20 to 30 min and induced with 0.3 to 1 mM IPTG and 0.2% arabinose for 3 h at 30°C with shaking. AcpM was expressed and purified as reported (25); however, induction was performed at 30°C for 3 h for the labeling experiments with the *CoA probe to increase the proportion of apo-AcpM (26).
Zhang (Proteomics and Metabolomics Core facility, WCM) for mass spectrometry and George Sukenick (NMR Analytical Core facility, MSKCC) for mass spectrometry and NMR.
Shilpika Pandey, conceived and performed experiments, wrote manuscript; Amrita Singh, performed experiments; Guangli Yang, probe synthesis and mass spectrometric analysis; Felipe B. d'Andrea, conceived and performed initial experiments; Xiuju Jiang, mouse infection; Travis E. Hartman, mass spectrometric analysis; John W. Mosior, PptH cloning and purification support; Ronnie Bourland, PptH cloning and purification support; Ben Gold, oversight and advice; Julia Roberts, experimental help; Annie Geiger, performed experiments with M. smegmatis whose results are not included here; Su Tang, PKS13 cloning; Kyu Rhee, oversight of mass spectrometric analysis; Ouathek Ouerfelli, oversight of probe synthesis and oversight of mass spectrometric analysis; James C. Sacchettini, oversight of project; Kristin Burns-Huang, oversight of project, wrote manuscript; Carl F. Nathan, oversight of project, wrote manuscript. This work was supported by the National Institutes of Health grant 1R21AI138939-01. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The Department of Microbiology & Immunology is supported by the William Randolph Hearst Trust. The Organic Synthesis Core Facility at Memorial Sloan Kettering Cancer Center is funded in part through an NCI P30 CA008748-53 core grant as well as an NCI R50 CA243895-01 grant to Ouathek Ouerfelli.
We declare no conflicts of interest.