The infectious cycle of Mycobacterium tuberculosis rests upon

Mechanism of phagolysosome biogenesis block by viable Mycobacterium tuberculosis Isabelle Vergne*†, Jennifer Chua*†, Hwang-Ho Lee*, Megan Lucas‡, John...
5 downloads 0 Views 689KB Size
Mechanism of phagolysosome biogenesis block by viable Mycobacterium tuberculosis Isabelle Vergne*†, Jennifer Chua*†, Hwang-Ho Lee*, Megan Lucas‡, John Belisle‡, and Vojo Deretic*§¶ Departments of *Molecular Genetics and Microbiology and §Cell Biology and Physiology, University of New Mexico School of Medicine, Albuquerque, NM 87131; and ‡Department of Microbiology, Colorado State University, Fort Collins, CO 80523 Edited by R. John Collier, Harvard Medical School, Boston, MA, and approved January 26, 2005 (received for review December 23, 2004)

macrophage 兩 phagosome 兩 tuberculosis 兩 lysosome 兩 phosphatidylinositol 3-phosphate

T

he infectious cycle of Mycobacterium tuberculosis rests upon the ability of this potent pathogen to parasitize host mononuclear phagocytic cells (1). In infected macrophages, M. tuberculosis resides within a phagosome that avoids the default maturation pathway leading to phagolysosome formation (2). The salient characteristics of the mycobacterial phagosome include (i) paucity of vacuolar H⫹ ATPase (3), (ii) attendant inefficient luminal acidification (3); and (iii) inadequate levels of mature lysosomal hydrolases (3, 4). These and additional (4–6) properties of the M. tuberculosis phagosome promote intracellular survival and growth of the tubercle bacilli and help avoid their immunological detection (1). The arrest of M. tuberculosis phagosome maturation has been studied at the membrane-trafficking level (2), with a focus on the small GTP-binding proteins, including Rab GTPases (7–9). Rabs direct intracellular trafficking by regulating activity and recruitment to organellar membranes of Rab-interacting partners and downstream effectors (10). The initial analyses of Rabs on mycobacterial phagosomes have indicated that the M. tuberculosis phagolysosome biogenesis arrest occurs between the stages controlled by the early endosomal GTPase Rab5 and its late endosomal counterpart Rab7 (7). A number of follow-up studies have indicated critical contributions of Rab5 effectors in mycobacterial phagosome maturation arrest, with a prominent role for the phosphatidylinositol 3-kinase (PI3K) hVPS34, its product phosphatidylinositol 3-phosphate (PI3P), and an array of PI3P-binding proteins (11–14). PI3P affects localization and function of proteins containing the PI3P-binding domains (FYVE, PH, and PX) (15). These proteins in turn execute various steps in membrane trafficking, endosomal protein sorting, and multisubunit enzyme assembly at the membrane, including phagosomal maturation (11, 16), early endosomal homotypic fusion (17), delivery of internalized plasma membrane receptors to late endosomes (18), formation of internal vesicles within late endosomal multivesicular bodies involved in termination of signaling events (19, 20), and phagocyte NADPH oxidase assembly at the membrane (21). PI3P is also important for the execution of the process of autophagy (22), which has recently been shown to restrict www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409716102

M. tuberculosis growth as a defense mechanism downstream of macrophage activation by IFN-␥ (23). In addition to hVPS34 interactions with Rab5, the recruitment of hVPS34 to endomembranes is controlled in macrophages by Ca2⫹, calmodulin (CaM), and Ca2⫹兾CaM kinase II (24). It has been shown that M. tuberculosis lipoarabinomannan inhibits cytosolic Ca2⫹ rise during phagocytosis (24, 25). A model of how M. tuberculosis blocks phagosome maturation has emerged, based on altered hVPS34 recruitment to mycobacterial phagosomes and altered PI3P patterns relative to the model, latex bead phagosomes (12, 14, 24). PI3P is essential for phagosome maturation into the phagolysosome, and inhibition of PI3P production arrests phagosome maturation (11, 16). However, the status of PI3P on phagosomes containing live vs. dead M. tuberculosis is not known. In this work, we show, using live cell imaging, four-dimensional (4D) confocal microscopy, and in vitro assays, that the principal difference between phagosomes harboring live or dead mycobacteria is the persistence of PI3P on phagosomes with killed microorganisms vs. the removal of PI3P from phagosomes harboring live bacilli. We show that, in addition to the known effects of the tubercle bacillus on suppressing Ca2⫹ fluxes (26), which in turn affect the recruitment of the PI3K responsible for generating PI3P on endomembranes (24), M. tuberculosis encodes a phosphatase that dephosphorylates PI3P and inhibits phagosome–late endosome fusion. These findings help explain how live M. tuberculosis maintains the phagosome maturation block and avoids lysosomal compartments. Materials and Methods Cell and Bacterial Cultures, Plasmid Constructs, Transfection, Microscopy, and Immunoblotting. RAW 264.7 cells were maintained in

DMEM, 4 mM L-glutamine, and 10% FBS. M. tuberculosis var. bovis bacillus Calmette–Gue´rin (BCG) was grown in 7H9 broth, and single-cell suspensions were prepared as described (11). For survival assays, mycobacteria were grown on 7H11 plates. Mycobacteria either expressed DsRed or were fluorescently labeled with 5 mg兾ml Texas red-X in PBS for 1 h. Bacteria were opsonized in DME supplemented with 10% FBS for 30 min. Mycobacteria were heat killed by incubation at 90°C for 10 min before labeling. The plasmid constructs and sources were as follows: P40PX-EGFP (M. Yaffe, Massachusetts Institute of Technology, Cambridge); 2xFYVE-EGFP (H. Stenmark, Norwegian Radium Hospital, Oslo); MTM1-EGFP (J. Dixon, University of Michigan, Ann Arbor), and MTMR3-EGFP (M. Clague, University of Liverpool, Liverpool, England). Macrophage transfection, immunofluorescence microscopy and 4D confocal microscopy were carried out as described (14). Live cell imaging is detailed in Supporting Text and Movies 1–5, which are published as supporting information on the PNAS web site. M. tuberculosis KatG antibody was from C. Barry This paper was submitted directly (Track II) to the PNAS office. Abbreviations: PI3K, phosphatidylinositol 3-kinase; PI3P, phosphatidylinositol 3-phosphate; 4D, four-dimensional; BCG, bacillus Calmette–Gue´rin; MtCFP, Mycobacterium tuberculosis H37Rv culture filtrate protein; RFU, relative fluorescence unit; MTM, myotubularin. †I.V. ¶To

and J.C. contributed equally to this work.

whom correspondence should be addressed. E-mail: [email protected].

© 2005 by The National Academy of Sciences of the USA

PNAS 兩 March 15, 2005 兩 vol. 102 兩 no. 11 兩 4033– 4038

CELL BIOLOGY

Live Mycobacterium tuberculosis persists in macrophage phagosomes by interfering with phagolysosome biogenesis. Here, using four-dimensional microscopy and in vitro assays, we report the principal difference between phagosomes containing live and dead mycobacteria. Phosphatidylinositol 3-phosphate (PI3P), a membrane trafficking regulatory lipid essential for phagosomal acquisition of lysosomal constituents, is retained on phagosomes harboring dead mycobacteria but is continuously eliminated from phagosomes with live bacilli. We show that the exclusion of PI3P from live mycobacterial phagosomes can be only transiently reversed by Ca2ⴙ fluxes, and that live M. tuberculosis secretes a lipid phosphatase, SapM, that hydrolyzes PI3P, inhibits phagosome–late endosome fusion in vitro, and contributes to inhibition of phagosomal maturation.

(National Institutes of Health, Bethesda). Affinity purified rabbit polyclonal antibody was raised against a peptide corresponding to the SapM residues 286–299 by using commercial services. PI3P Phosphatase Activity. Fluorescent phosphoinositide substrates

labeled with C6-NBD were from Echelon Research Laboratories (Salt Lake City). Phosphatase activity was monitored according to Taylor and Dixon (27) with fluorescent substrates in 50 mM ammonium acetate (pH 6.0) and 2 mM DTT for 30 min at 30°C. TLC analysis of reaction products was carried out as described (27). The release of phosphate was quantified by malachite green assay (Upstate Biotechnology, Lake Placid, NY) (28) in 50 mM Tris䡠HCl (pH 7.4) and 0.05% Triton X-100 at 37°C. Phagosome–Late Endosome Fusion Assay. Phagosomes and late

endosomes were purified, and fusion assay was carried out as described (5). Bacterial Culture Supernatants, Bacterial Extracts, and Purified Proteins. Culture supernatant was obtained by filtering through a

0.2-␮m filter. Mycobacterial pellets were homogenized by bead beating. When required, J774 were infected with BCG for 2 h and lysed, and a postnuclear supernatant was prepared according to Via et al. (7). The postnuclear supernatant was subjected to velocity sedimentation (1,000 ⫻ g, 45 min, 4°C) through 15% (wt兾wt) sucrose overlaid on a 50% (wt兾wt) sucrose cushion (7). The material at the 15–50% sucrose interface, containing mycobacterial phagosomes, was collected, and extracts were prepared by bead beating. M. tuberculosis H37Rv culture filtrate protein (MtCFP) and SapM were prepared as described (29, 30), respectively. Pure MptpA and MptpB (GST fusions) were from A. Koul (31). MtCFP was denatured by heating for 30 min at 95°C. Results and Discussion Persistence of PI3P on Phagosomes Containing Dead Mycobacteria.

Applying 4D confocal microscopy (14), we compared PI3P levels on phagosomes harboring either live or dead (heat-killed) M. tuberculosis var. bovis BCG. RAW 264.7 macrophages were transfected with constructs encoding GFP fusions to protein domains p40phox PX (P40PX-EGFP) or 2xFYVE from Hrs (2xFYVE-EGFP) that specifically bind PI3P (21, 32). Macrophages were infected with Texas red-labeled live or dead M. tuberculosis var. bovis BCG. The fluorescence intensity of the PI3P probe was imaged over time by rapid collection of z-stacks composed of dual color confocal optical sections. The images were analyzed, as described (14): (i) for each time point, the optical section corresponding to the maximum fluorescence intensity of the mycobacteria was identified; and (ii) this and additional confocal slices above and below were collapsed into a single x-y projection. Of the phagosomes containing dead mycobacteria monitored for 60 min, 100% (n ⫽ 19) remained positive for the PI3P probe at all times (Fig. 1A and Movie 1). Of the phagosomes in macrophages infected with live mycobacteria, 18% (n ⫽ 39) were scored as PI3P-positive (Fig. 1 A and Movie 2). Temporal analysis of the mean fluorescence PI3P probe intensity on phagosomes relative to the cytosol (RØ/C) is shown in Fig. 1B; relative fluorescence units (RFU) corrected for RFU of the cytosol are shown in Fig. 1C. The newly formed phagosomes displayed early and transient recruitment of the PI3P GFP probe, as reported (14), concomitant with the mycobacterial uptake by the macrophage. Past this initial period, dead and live mycobacteria showed dramatic differences, with the phagosomes containing dead mycobacteria being PI3Ppositive at all times, and with those containing live mycobacteria becoming and remaining fully PI3P-negative. The RØ/C value for dead mycobacterial phagosomes was maintained ⬎2, whereas live mycobacteria declined to an RØ/C of 1 (Fig. 1B). In keeping with the previously established role of PI3P in phagosomal maturation (11, 16), phagosomes harboring live mycobacteria (PI3P-negative) and 4034 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409716102

Fig. 1. PI3P persists on phagosomes containing dead but not live M. tuberculosis var. bovis BCG. (A) RAW 264.7 macrophages were transfected with P40PX-EGFP, allowed to phagocytose either live or dead (heat-inactivated) Texas red-labeled BCG, and analyzed by 4D confocal microscopy. Shown is quantification of PI3P positivity of phagosomes containing live or dead BCG (n ⫽ 45 live, n ⫽ 19 dead). (Insets) GFP fluorescence of the PI3P probe (grayscale) (Left) and merged images of GFP and red mycobacterial fluorescence (Right). **, P ⬍ 0.01. (B) Temporal quantification of phagosome fluorescence intensity relative to fluorescence of the cytosol. R␾/c, ratio between phagosome fluorescence intensity and cytosol fluorescence intensity. Shown are R␾/c obtained by 4D microscopy and live imaging of three different phagosomes harboring dead BCG (filled squares) and three different phagosomes in cells infected with live BCG (open triangles). (C) Quantification of fluorescence levels over time, expressed in relative fluorescence units (RFU; subtracted for RFU of the cytosol) of a phagosome with dead (filled squares) or live (open triangles) BCG. (D) Confocal immunofluorescence images of fixed specimens containing macrophages infected with live or dead BCG and immuno-stained for CD63. (E) Quantification of CD63 staining of phagosomes.

phagosomes with dead mycobacteria (PI3P-positive) matured differently. When the late endosomal兾lysosomal marker CD63 (33) was quantified, 32% of the phagosomes harboring live mycobacteria colocalized with CD63, whereas 86% of the phagosomes containing dead mycobacteria colocalized with CD63 (Fig. 1 D and E). Maintenance of PI3P on Phagosomes Containing Dead Mycobacteria Is PI3-Kinase-Dependent. Conversion of phosphatidylinositol to PI3P

is catalyzed by the type III PI3K hVPS34, a kinase that can be reversibly inhibited by LY294002 (34). RAW 264.7 cells transfected with the PI3P GFP probe were treated with 100 ␮M LY294002 and infected with M. tuberculosis var. bovis BCG. LY294002 did not inhibit mycobacterial uptake. After a chase period, LY294002 was washed out, and recruitment of the PI3P probe to phagosomes was monitored by live microscopy. LY294002 inhibited the generation of PI3P on phagosomes, resulting in a disappearance of PI3P (Fig. 2A). Upon LY294002 washout (Fig. 2B), PI3P was regenerated on phagosomes with dead mycobacteria (Fig. 2 A–C and Movie 3). The LY294002 addition兾washout cycle did not change the PI3Pnegative status of phagosomes with live mycobacteria (Fig. 2 D–F). Fig. 2G shows the kinetics of PI3P changes on phagosomes. Thus, Vergne et al.

Fig. 2. De novo generation and maintenance of PI3P on phagosomes harboring dead mycobacteria. RAW 264.7 cells were transfected with P40PX-EGFP, allowed to phagocytose either live or dead BCG (Texas red-labeled), and analyzed by 4D microscopy by using an UltraView microscope. Cells were treated with PI3K inhibitor (100 ␮M LY294002) as indicated (⫹LY). LY294002 washout (removal of the inhibitor) is indicated in the fluorescent image panels (⫺LY) and the graph (arrow). (Insets) Grayscale images of green (Left) and red (Right) fluorescence of areas with objects indicated by arrows. (A–C) Absence of PI3P probe on dead BCG phagosomes in cells treated with LY294002, and recruitment of PI3P probe upon LY294002 washout. (D–F) PI3P probe status (PI3P-negative) remains unchanged on live BCG phagosomes after a cycle of LY294002 treatment and washout. (G) Time course of phagosomal GFP fluorescence (expressed as percent of the maximum RFU) on dead BCG phagosomes with (open squares) or without (filled squares) LY294002 treatment. Arrow, LY294002 washout.

Increasing Intracellular Ca2ⴙ Only Transiently Restores PI3P Levels on Phagosomes with Live Mycobacteria. There is a correlation be-

tween mycobacterial entry into the macrophage and inhibition of cytosolic Ca2⫹ rise (26). The Ca2⫹ binding protein calmodulin (CaM) and its effector Ca2⫹兾CaM kinase II recruit the PI3K hVPS34 responsible for PI3P production to organellar membrane (24). We tested whether increasing intracellular Ca2⫹ can restore PI3P levels on phagosomes with live mycobacteria. RAW 264.7 cells transfected with the PI3P probe and phagocytosing live mycobacteria were treated with the Ca2⫹ ionophore A23187 as described (26), and analyzed by 4D confocal microscopy. Fig. 6, which is published as supporting information on the PNAS web site, shows a transient spike in PI3P increase on a phagosome containing live mycobacteria in cells treated with A23187. The PI3P spike occurred only in 10% of all phagosomes observed, and the majority of phagosomes did not show PI3P increase, despite the fact that Ca2⫹ rise was detected in ⬎95% of the cells treated with the ionophore (Fig. 6 K–M). Thus, Ca2⫹ alone cannot explain the difference in steady-state levels of PI3P on phagosomes harboring dead vs. live mycobacteria. Host PI3P Phosphatases and Absence of PI3P on Live Mycobacterial Phagosomes. Experiments with LY294002 addition兾washout (Fig.

2) indicate that PI3P levels on phagosomes represent a balance between PI3P synthesis and its turnover. We therefore tested the possibility that differential recruitment of a host PI3P-specific 3-phosphatase can explain differences between phagosomes harboring live vs. dead bacilli. Myotubularin 1 (MTM1) specifically dephosphorylates PI3P (35). RAW 264.7 macrophages, transfected Vergne et al.

with MTM1-GFP, were allowed to phagocytose live or dead mycobacteria, and the cells were analyzed by 4D confocal microscopy. MTM1-GFP was recruited to both types of phagosomes (Fig. 7, which is published as supporting information on the PNAS web site, and Movies 4 and 5). MTM1-GFP was present only early on and rapidly disappeared from either type of phagosomes. Similar phenomena were observed with the PI3P phosphatase MTMR3 (36). Both MTM1 and MTMR3 peaked and then desorbed from phagosomes at early time points. Neither MTM1 nor MTMR3 was connected to the later period of intense differences in PI3P levels between live and dead mycobacterial phagosomes. Thus, host cell PI3P phosphatases, MTM1 and MTMR3, cannot explain the observed differential removal of PI3P from phagosomes harboring live mycobacteria. Use of a bacteriostatic antibiotic, chloramphenicol, permitted a demonstration that a bacterial product was responsible for controlling the PI3P levels on mycobacterial phagosomes (Fig. 8, which is published as supporting information on the PNAS web site). Treatment of BCG with chloramphenicol increased PI3P positivity of mycobacterial phagosomes (Fig. 8 A, B, and D). A washout of chloramphenicol (which resulted in resumption of mycobacterial growth; Fig. 8E) reversed the PI3P status to the low levels seen with untreated live mycobacteria (Fig. 8 C and D). M. tuberculosis Secretes a PI3P 3-Phosphatase. In lieu of explanations based on Ca2⫹ or host cell phosphatases for differences in PI3P levels on phagosomes with live vs. dead mycobacteria, we tested whether a mycobacterial enzymatic activity was responsible for the lack of PI3P on phagosomes harboring live bacilli. Culture filtrate protein from virulent M. tuberculosis H37Rv, MtCFP, was incubated in an in vitro assay for PI3P phosphatase, and phosphoinositide products were resolved by TLC. Incubation with MtCFP resulted in a dose-dependent conversion of PI3P into its dephosPNAS 兩 March 15, 2005 兩 vol. 102 兩 no. 11 兩 4035

CELL BIOLOGY

PI3P generation on dead mycobacterial phagosomes is due to a PI3K activity, whereas live mycobacteria have the capacity to continually eliminate PI3P from the phagosomal membrane.

Fig. 3. M. tuberculosis secretes PI3P phosphatase SapM. (A) TLC (UV fluorescence) of reaction products when PI3P is incubated in the presence of MtCFP. Di-C6-NBD6-PI3P (1 ␮g) was incubated with different concentrations of MtCFP or boiled MtCFP. Lane 1, MtCFP incubated without Di-C6-NBD6-phosphoinositides; lane 2, Di-C6-NBD6-PI3P incubated without CFP; lane 3, Di-C6-NBD6-PI (product of PI3P hydrolysis) standard; lanes 4 – 6, Di-C6-NBD6-PI3P incubated with different concentrations of MtCFP; lane 7, Di-C6-NBD6-PI3P incubated with heat-inactivated MtCFP. (B) TLC (UV fluorescence) of reaction products after incubation of PI3P with SapM. Lane 1, Di-C6-NBD6-PI standard; lane 2, Di-C6-NBD6-PI3P incubated without SapM; lanes 3–7, Di-C6-NBD6-PI3P incubated with different concentrations of SapM. (C) PI3P phosphatase activity in MtCFP (40 ␮g兾ml) determined by using malachite green assay. Error bars represent SEM. (D) Comparison of SapM (1.5 ␮g兾ml) and MtCFP (40 ␮g兾ml) phosphatase specificity for phosphatidylinositol monophosphates (PI3P, PI4P, PI5P) determined by using malachite green assay. Error bars represent SEM. (E) Immunoblot comparison of SapM and KatG in MtCFP, bacterial cell extracts (BCG pellet), BCG culture supernatant (BCG Sup), and BCG phagosomes after 2-h infection of J774 macrophages.

phorylation product, phosphatidylinositol (PI) (Fig. 3A). Heat inactivation of MtCFP abrogated its ability to dephosphorylate PI3P (Fig. 3A). The breakdown of PI3P by MtCFP was quantified by using a colorimetric assay (28) (Fig. 3B). These results demonstrate that M. tuberculosis-secreted proteins contain PI3P phosphatase activity. We next considered candidate gene products encoded by the M. tuberculosis genome showing similarities to the active site in mammalian myotubularins. The characterized myotubularins are lipid phosphatases (with MTM1 having very poor tyrosine兾 dual specificity protein phosphatase activity), showing preference for monophosphorylated substrates such as PI3P, and contain the Cys-X5-Arg motif, including additional Asp residues within their active site (37). The closest match with this motif was found in the M. tuberculosis protein MptpB (encoded by Rv0153c), which was recently described as a secreted tyrosine phosphatase (31). When purified GST-MptpB was tested, no PI3P phosphatase activity could be detected (data not shown). A second secreted tyrosine phosphatase, MptpA (encoded by Rv 2234), also showed no PI3P activity, thus ruling out these proteins as possible candidates. Another secreted acid phosphatase activity has been described in M. tuberculosis culture filtrate (38) and was later purified and characterized as a nonspecific acid phosphatase (30). Although it was concluded in the initial report on SapM that it had no activity against phospholipids, based on the absence of activity with phosphatidylcholine, phosphatidylethanolamine, and phosphatidic acid (30), when purified M. tuberculosis SapM was tested with PI3P as a substrate, it showed strong activity comparable with that of MtCFP 4036 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409716102

(Fig. 3C). The preferred SapM substrate among phosphatidylinositol monophosphates was PI3P, matching the substrate specificity profile of MtCFP (Fig. 3D). When a panel of phosphatase inhibitors was tested, MtCFP and purified SapM showed identical sensitivity patterns with PI3P as a substrate (Fig. 9, which is published as supporting information on the PNAS web site). Thus, SapM is an M. tuberculosis PI3P phosphatase accounting for at least a part of the PI3P phosphatase activity in MtCFP. Antibodies raised against a SapM peptide showed that SapM was present in the CFP from M. tuberculosis and that it was enriched in BCG culture supernatant relative to a representative bacterial cytoplasmic protein, KatG, enriched in mycobacterial pellet (Fig. 3E). Furthermore, SapM contains a typical leader peptide proteolytically removed during its secretion兾maturation (30). SapM was present in vivo during macrophage infection. Macrophages were infected, and phagosomes werre prepared by collecting the material from the first sucrose gradient following the published phagosome purification procedure (7). Mycobacterial phagosomes from such preparations were positive for SapM (Fig. 3E). M. tuberculosis PI3P Phosphatase Is Required for Reduction of PI3P Levels on Phagosomes and Inhibition of Phagosome Maturation Containing Live Mycobacteria. We used the combined property of

molybdate to inhibit SapM (Fig. 9) and its inability to cross host cell membrane (39, 40) to inhibit specifically secreted bacterial products during mycobacterial uptake by macrophages. Infection of macrophages in the presence of molybdate resulted in increased PI3P positivity of mycobacterial phagosomes (Fig. 4 Vergne et al.

A–I). PI3P-positive phagosomes were more abundant in the samples with molybdate (Fig. 4J). A temporal analysis by 4D microscopy (Fig. 4K) confirmed that molybdate counteracted the action of a mycobacterial product affecting PI3P levels on phagosomes. Molybdate did not affect mycobacterial viability (Fig. 4L), excluding the possibility that its action was secondary to bacterial killing. The treatment with molybdate also translated into enhanced maturation of mycobacterial phagosomes (Fig. 10, which is published as supporting information on the PNAS web site). We conclude that a mycobacterial product sensitive to the phosphatase inhibitor molybdate, most likely SapM, has direct or indirect access in vivo to the PI3P substrate within the phagosomal membrane. M. tuberculosis PI3P Phosphatase SapM Inhibits PI3P-Dependent Phagosome–Late Endosome Fusion in Vitro. We next examined

whether SapM affects phagosome–late endosome fusion using Vergne et al.

purified organelles in an established in vitro system (5). SapM was added to the in vitro system with purified streptavidin bead phagosomes, and late endosomes were preloaded with the fluid phase marker biotin-horse radish peroxidase (HRP). HRP association with streptavidin beads occurs after membrane fusion between phagosomal and endosomal organelles and only upon mixing of the luminal content in the two compartments. Regulated organellar fusion between phagosomes and late endosomes is quantified as HRP activity bound to streptavidin beads at the end of reaction (5). Addition of SapM inhibited ATP-dependent regulated phagosome–late endosome fusion (Fig. 5A). In contrast, addition of purified GST-MptpA and GST-MptpB did not inhibit fusion relative to the control with GST alone (data not shown). The fusion depended on PI3P generation, because it was affected by the PI3K inhibitor wortmannin (Fig. 5 Inset). Thus, the PI3P phosphatase activity secreted by M. tuberculosis has an inhibitory effect on phagosome–lysosome fusion. PNAS 兩 March 15, 2005 兩 vol. 102 兩 no. 11 兩 4037

CELL BIOLOGY

Fig. 4. M. tuberculosis var. bovis BCG-secreted PI3P phosphatase is responsible for removal of PI3P from phagosomes harboring live mycobacteria. RAW 264.7 cells were transfected with P40PX-EGFP, allowed to phagocytose either live or dead mycobacteria (Texas red-labeled BCG), and analyzed by 4D microscopy. Infection with BCG was carried out in the presence or absence of sodium molybdate (1 mM) during phagocytosis and subsequent imaging. (A–C) Lack of PI3P probe recruitment to phagosomes harboring live BCG. (D–F) Recruitment of PI3P probe to phagosomes containing dead BCG. (G–I) Recruitment of PI3P probe on phagosomes containing live BCG treated with sodium molybdate. (J) Fraction (%) of live BCG phagosomes (n ⫽ 45) and live BCG phagosomes treated with sodium molybdate (n ⫽ 18) that recruited the PI3P probe. *, P ⫽ 0.02. (K) Temporal analysis and quantification of phagosome fluorescence intensity relative to cytosol fluorescence (R␾/c). R␾/c were obtained by processing 4D microscopy images of live cells with phagosomes: harboring dead BCG (filled squares), harboring live BCG-treated with sodium molybdate (filled triangles), or with live BCG (open triangle). (L) Colony-forming units (CFU) of live BCG vs. live BCG treated with sodium molybdate.

Fig. 5. M. tuberculosis-secreted PI3P phosphatase SapM inhibits phagosome–late endosome fusion in vitro. The standard phagosome–late endosome in vitro fusion reaction was carried as described (5), in the presence of 32 ␮g兾ml of purified SapM (Mt SapM). (Inset) Fusion reaction in the presence of 100 nM wortmannin (Wm). Ctrl, untreated control. Error bars represent SEM. *, P ⬍ 0.05, n ⫽ 3.

Conclusion In this study, we have uncovered the principal difference between live and dead intracellular mycobacteria. Live mycobacteria remove the PI3P from the phagosomal membrane, thus precluding organellar maturation. PI3P provides a marquee membrane tagging signal, earmarking phagosomes for maturation (11, 16). Dead mycobacteria have no ability to remove PI3P from their phagosomal membrane and mature down the phagolysosome biogenesis pathway. Live mycobacteria secrete a PI3P phosphatase activity that inhibits fusion between phagosomes and late endosomal兾lysosomal organelles. The secreted M. tuberculosis PI3P phosphatase is sensitive to heat inactivation, and correlates fully with the removal of PI3P from phagosomes harboring live but not dead mycobacteria. The M. tuberculosis PI3P activity was narrowed down to SapM, an acid phosphatase previously studied by Saleh and Belisle (30). Belisle and colleagues have shown that SapM is secreted by mycobacteria, a finding further extended in the present study. It remains to be determined how SapM, once secreted into the phagosomal lumen, gains access to PI3P within the cytofacial 1. Russell, D. G., Mwandumba, H. C. & Rhoades, E. E. (2002) J. Cell Biol. 158, 421–426. 2. Vergne, I., Chua, J., Singh, S. B. & Deretic, V. (2004) Annu. Rev. Cell Dev. Biol. 20, 367–394. 3. Sturgill-Koszycki, S., Schlesinger, P. H., Chakraborty, P., Haddix, P. L., Collins, H. L., Fok, A. K., Allen, R. D., Gluck, S. L., Heuser, J. & Russell, D. G. (1994) Science 263, 678–681. 4. Sturgill-Koszycki, S., Schaible, U. E. & Russell, D. G. (1996) EMBO J. 15, 6960–6968. 5. Vergne, I., Fratti, R. A., Hill, P. J., Chua, J., Belisle, J. & Deretic, V. (2004) Mol. Biol. Cell 15, 751–760. 6. Clemens, D. L. & Horwitz, M. A. (1996) J. Exp. Med. 184, 1349–1355. 7. Via, L. E., Deretic, D., Ulmer, R. J., Hibler, N. S., Huber, L. A. & Deretic, V. (1997) J. Biol. Chem. 272, 13326–13331. 8. Clemens, D. L., Lee, B. Y. & Horwitz, M. A. (2000) Infect. Immun. 68, 2671–2684. 9. MacMicking, J. D., Taylor, G. A. & McKinney, J. D. (2003) Science 302, 654–659. 10. Pereira-Leal, J. B. & Seabra, M. C. (2001) J. Mol. Biol. 313, 889–901. 11. Fratti, R. A., Backer, J. M., Gruenberg, J., Corvera, S. & Deretic, V. (2001) J. Cell Biol. 154, 631–644. 12. Fratti, R. A., Chua, J., Vergne, I. & Deretic, V. (2003) Proc. Natl. Acad. Sci. USA 100, 5437–5442. 13. Vieira, O. V., Harrison, R. E., Scott, C. C., Stenmark, H., Alexander, D., Liu, J., Gruenberg, J., Schreiber, A. D. & Grinstein, S. (2004) Mol. Cell. Biol. 24, 4593–4604. 14. Chua, J. & Deretic, V. (2004) J. Biol. Chem. 279, 36982–36992. 15. Lemmon, M. A. (2003) Traffic 4, 201–213. 16. Vieira, O. V., Botelho, R. J., Rameh, L., Brachmann, S. M., Matsuo, T., Davidson, H. W., Schreiber, A., Backer, J. M., Cantley, L. C. & Grinstein, S. (2001) J. Cell Biol. 155, 19–25. 17. Simonsen, A., Lippe, R., Christoforidis, S., Gaullier, J. M., Brech, A., Callaghan, J., Toh, B. H., Murphy, C., Zerial, M. & Stenmark, H. (1998) Nature 394, 494–498. 18. Siddhanta, U., McIlroy, J., Shah, A., Zhang, Y. & Backer, J. M. (1998) J. Cell Biol. 143, 1647–1659. 19. Gruenberg, J. & Stenmark, H. (2004) Nat. Rev. Mol. Cell Biol. 5, 317–323. 20. Katzmann, D. J., Odorizzi, G. & Emr, S. D. (2002) Nat. Rev. Mol. Cell Biol. 3, 893–905. 21. Kanai, F., Liu, H., Field, S. J., Akbary, H., Matsuo, T., Brown, G. E., Cantley, L. C. & Yaffe, M. B. (2001) Nat. Cell Biol. 3, 675–678.

4038 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409716102

membrane leaflet. It is possible that SapM is exported to the cytosolic side, or alternatively there may be a mechanism for PI3P presentation to SapM remaining on the luminal side. Either way, our data with molybdate, a membrane-impermeant (39, 40) inhibitor of SapM, indicate that, in vivo, SapM gains access to phagosomal PI3P. In the previous and present work, we and others have pinpointed the key membrane-trafficking processes targeted by mycobacteria (11, 14, 26). The mechanism of M. tuberculosis phagosome– lysosome fusion arrest seems to be at least a two-prong process converging upon PI3P. In earlier work (12, 24), we have identified that M. tuberculosis glycolipid, lipoarabinomannan, interferes with Ca2⫹ rise and recruitment兾activation of PI3K hVPS34 (24). The inhibition of Ca2⫹ rise (26) is important but not sufficient to maintain the maturation block, because M. tuberculosis has to preserve a PI3P-negative status during its long-term residence in infected macrophages. We have now uncovered the second component of this double-latch mechanism, in the form of a mycobacterial heat-sensitive enzymatic activity that removes PI3P from membranes. This finding explains the difference between heatkilled and live mycobacteria, because phagosomes containing heatinactivated bacilli, in sharp contrast to those with viable M. tuberculosis, remain strongly PI3P-positive and mature into the phagolysosome. The critical role of PI3P in elimination of intracellular mycobacteria was recently underscored in the context of the PI3P-dependent process of autophagy, which can be induced in infected macrophages by activation with the protective cytokine IFN-␥, overcoming the M. tuberculosis phagolysosome biogenesis block (23). Our identification of the M. tuberculosis enzymatic entity responsible for the PI3P removal from phagosomes as the previously characterized mycobacterial acid phosphatase SapM, provides a new target for drug and vaccine development. Such interventions, directed at preventing the establishment of intracellular M. tuberculosis, may disrupt the vicious cycle of tuberculosis persistence and propagation in human populations. We thank E. Roberts for suggestions regarding chloramphenicol treatment; M. Clague, J. Dixon, H. Stenmark, and M. Yaffe for plasmid constructs; and A. Koul for purified MPtpA and MPtpB. M. tuberculosis culture filtrate protein was prepared with the support of National Institute of Allergy and Infectious Diseases Contract NO1 AI-75320 titled ‘‘Tuberculosis Research Materials and Vaccine Testing.’’ This work was supported by National Institutes of Health Grant AI45148. 22. Petiot, A., Ogier-Denis, E., Blommaart, E. F., Meijer, A. J. & Codogno, P. (2000) J. Biol. Chem. 275, 992–998. 23. Gutierrez, M. G., Master, S. S., Singh, S. B., Taylor, G. A., Colombo, M. I. & Deretic, V. (2004) Cell 119, 753–766. 24. Vergne, I., Chua, J. & Deretic, V. (2003) J. Exp. Med. 198, 653–659. 25. Rojas, M., Garcia, L. F., Nigou, J., Puzo, G. & Olivier, M. (2000) J. Infect. Dis. 182, 240–251. 26. Malik, Z. A., Denning, G. M. & Kusner, D. J. (2000) J. Exp. Med. 191, 287–302. 27. Taylor, G. S. & Dixon, J. E. (2001) Anal. Biochem. 295, 122–126. 28. Maehama, T., Taylor, G. S., Slama, J. T. & Dixon, J. E. (2000) Anal. Biochem. 279, 248–250. 29. Sonnenberg, M. G. & Belisle, J. T. (1997) Infect. Immun. 65, 4515–4524. 30. Saleh, M. T. & Belisle, J. T. (2000) J. Bacteriol. 182, 6850–6853. 31. Koul, A., Choidas, A., Treder, M., Tyagi, A. K., Drlica, K., Singh, Y. & Ullrich, A. (2000) J. Bacteriol. 182, 5425–5432. 32. Gillooly, D. J., Morrow, I. C., Lindsay, M., Gould, R., Bryant, N. J., Gaullier, J. M., Parton, R. G. & Stenmark, H. (2000) EMBO J. 19, 4577–4588. 33. Metzelaar, M. J., Wijngaard, P. L., Peters, P. J., Sixma, J. J., Nieuwenhuis, H. K. & Clevers, H. C. (1991) J. Biol. Chem. 266, 3239–3245. 34. Brown, W. J., DeWald, D. B., Emr, S. D., Plutner, H. & Balch, W. E. (1995) J. Cell Biol. 130, 781–796. 35. Taylor, G. S., Maehama, T. & Dixon, J. E. (2000) Proc. Natl. Acad. Sci. USA 97, 8910–8915. 36. Walker, D. M., Urbe, S., Dove, S. K., Tenza, D., Raposo, G. & Clague, M. J. (2001) Curr. Biol. 11, 1600–1605. 37. Wishart, M. J., Taylor, G. S., Slama, J. T. & Dixon, J. E. (2001) Curr. Opin. Cell Biol. 13, 172–181. 38. Raynaud, C., Etienne, G., Peyron, P., Laneelle, M. A. & Daffe, M. (1998) Microbiology 144, 577–587. 39. Haque, S. J., Flati, V., Deb, A. & Williams, B. R. (1995) J. Biol. Chem. 270, 25709–25714. 40. Lerea, K. M., Tonks, N. K., Krebs, E. G., Fischer, E. H. & Glomset, J. A. (1989) Biochemistry 28, 9286–9292.

Vergne et al.

Suggest Documents