The myotubularin MTMR4 regulates phagosomal phosphatidylinositol 3-phosphate turnover and phagocytosis

Macrophage phagocytosis is required for effective clearance of invading bacteria and other microbes. Coordinated phosphoinositide signaling is critical both for phagocytic particle engulfment and subsequent phagosomal maturation to a degradative organelle. Phosphatidylinositol 3-phosphate (PtdIns(3)P) is a phosphoinositide that is rapidly synthesized and degraded on phagosomal membranes, where it recruits FYVE domain– and PX motif–containing proteins that promote phagosomal maturation. However, the molecular mechanisms that regulate PtdIns(3)P removal from the phagosome have remained unclear. We report here that a myotubularin PtdIns(3)P 3-phosphatase, myotubularin-related protein-4 (MTMR4), regulates macrophage phagocytosis. MTMR4 overexpression reduced and siRNA-mediated Mtmr4 silencing increased levels of cell-surface immunoglobulin receptors (i.e. Fcγ receptors (FcγRs)) on RAW 264.7 macrophages, associated with altered pseudopodal F-actin. Furthermore, MTMR4 negatively regulated the phagocytosis of IgG-opsonized particles, indicating that MTMR4 inhibits FcγR-mediated phagocytosis, and was dynamically recruited to phagosomes of macrophages during phagocytosis. MTMR4 overexpression decreased and Mtmr4-specific siRNA expression increased the duration of PtdIns(3)P on phagosomal membranes. Macrophages treated with Mtmr4-specific siRNA were more resistant to Mycobacterium marinum–induced phagosome arrest, associated with increased maturation of mycobacterial phagosomes, indicating that extended PtdIns(3)P signaling on phagosomes in the Mtmr4-knockdown cells permitted trafficking of phagosomes to acidic late endosomal and lysosomal compartments. In conclusion, our findings indicate that MTMR4 regulates PtdIns(3)P degradation in macrophages and thereby controls phagocytosis and phagosomal maturation.

Macrophages contribute to immune control of infection, tissue remodeling, and inflammation by undergoing a selective pathogen removal process, phagocytosis, which is critical for the clearance of apoptotic bodies and invading microorganisms (1,2). Phagocytosis includes the internalization of a particle followed by its subsequent processing within a membrane-bound organelle. The phagocytosis of IgG-opsonized particles is initiated when Fc␥ receptors (Fc␥Rs) 4 on the macrophage cell surface recognize the Fc portion of host IgG on microorganisms or cells. This induces Fc␥R clustering leading to localized actin polymerization that, in turn, propels the elaboration of pseudopods, which internalize the target into a membrane-bound vacuole called the phagosome (3). The phagosome is trafficked in a series of orchestrated interactions with other membrane-bound organelles and proteins, termed phagosomal maturation (3).
Phosphoinositides are membrane-bound signaling molecules involved in numerous cellular processes, including intracellular trafficking. In many eukaryotic cells, phosphatidylinositol 3-phosphate (PtdIns(3)P) predominantly localizes to the cytoplasmic face of endosomal and phagosomal membranes (4,5). PtdIns(3)P regulates endosomal and phagosomal maturation into degradative organelles that fuse with lysosomes. PtdIns(3)P recruits effector proteins, such as early endosome antigen 1 (EEA1) and hepatocyte growth factor-regulated tyrosine kinase substrate (Hrs), via their FYVE domains and phox (PX) motifs to endosomal and phagosomal membranes, where they regulate membrane fusion, sorting of cargo, and trafficking to subcellular compartments (5,6). PtdIns(3)P is dynamically turned over on phagosomes and rapidly appears and then disappears within a 10-min time frame (4). However, although the generation of PtdIns(3)P by the Class III phosphoinositide 3-kinase (PI3K) is well-described in mammalian cells (7), the molecular mechanisms underlying the removal of PtdIns(3)P from phagosomal membranes in macrophages are still unclear.
We have previously reported that MTMR4 regulates the endocytosis of transferrin and restricts the sorting of transferrin out of early endosomes (12). Intriguingly, MTMR4 also regulates the distribution of recycling endosomes and the degradation of several surface growth factor receptors, including epidermal growth factor receptor and transforming growth factor ␤ receptor (11,12,15,16). Murine macrophages express the cell-surface stimulatory receptors Fc␥RI, -III, and -IV and the inhibitory Fc␥RIIb (17,18). Following ligand binding, stimulatory Fc␥Rs undergo cross-linking with phosphorylation of associated immunoreceptor tyrosine-based activation motifs, thereby initiating signaling via tyrosine kinases and Class I PI3K to promote pseudopodal extension, which is required for phagocytosis via actin polymerization (19). Accordingly, initiation of phagocytosis depends on the presence of Fc␥Rs at the cell surface, which is highly regulated by endocytic recycling and degradation. However, whether MTMR4 regulates Fc␥R levels at the cell surface is unknown, and a potential role for MTMR4 in regulating phagosomal PtdIns(3)P and thereby phagosomal maturation has yet to be determined.
Here, we investigated whether MTMR4 regulates phagosomal PtdIns(3)P and plays a role in phagosomal internalization and maturation in mammalian macrophages. MTMR4 overexpression decreased the surface expression of Fc␥Rs and pseudopodal actin formation in RAW 264.7 macrophage cells, indicating that MTMR4 plays a role in phagocytosis initiation. Interestingly, MTMR4 overexpression was associated with a decreased efficiency of phagocytosis. We found that MTMR4 was dynamically recruited to the forming phagosome and negatively regulated the duration and amplitude of PtdIns(3)P signals on phagosomes. Finally, MTMR4 depletion accelerated the maturation of phagosomes containing pathogenic mycobacteria. These results identify MTMR4 as a novel regulator of macrophage phagocytosis that directs phagosomal PtdIns(3)P turnover.

MTMR4 negatively regulates the presence of Fc␥R on the macrophage cell surface
Endogenous MTMR4 expression in macrophages was assessed by immunoblotting of RAW 264.7 cell lysates using a polyclonal antibody against MTMR4 (12), which detected an immunoreactive polypeptide migrating at 165 kDa consistent with its predicted molecular mass (Fig. 1Ai). Similarly, RAW 264.7 cells transfected with recombinant HA-MTMR4 expressed a 165-kDa polypeptide detected by anti-HA antibodies (Fig. 1Aii). At steady state, Fc␥Rs undergo cycles of endocytosis and recycling back to the plasma membrane, whereas ligand-bound Fc␥Rs are diverted away from the recycling compartment to the degradative pathway (20). As MTMR4 is reported to regulate the trafficking of transferrin and epidermal growth factor receptor (12) , we next assessed whether  MTMR4 can regulate the cell-surface expression of Fc␥RI and  Fc␥RII/III at steady state. RAW 264.7 cells showed intense Fc␥RI  and Fc␥RII/III immunofluorescence staining at the cell membrane, whereas cells expressing HA-MTMR4 showed less intense  staining of Fc␥RI, but not Fc␥RII/III, when compared with non-transfected cells and the HA-vector control (Fig. 1B). Quantification of Fc␥RI immunofluorescence intensity showed 20% less surface Fc␥RI in cells expressing HA-MTMR4 compared with vector controls (Fig. 1C), whereas no significant difference in surface Fc␥RII/III was observed (Fig. S1A). In a similar experiment, siRNA targeting Mtmr4 was used to generate Mtmr4-knockdown RAW 264.7 cells. The efficiency of Mtmr4 knockdown was confirmed by quantitative RT-PCR (Fig. 1D), and decreased protein expression was confirmed by Western blotting (Fig. 1E). RAW 264.7 cells treated with Mtmr4 siRNA showed increased surface expression of extracellular Fc␥RI and Fc␥RII/III as assessed by flow cytometry (Fig. 1F). This was not a consequence of altered total cellular levels of Fc␥Rs, as this difference was not observed when intracellular receptors were included in immunostaining of permeabilized cells (Fig. S1B). This result was confirmed by immunofluorescence microscopy (Fig. 1G), which revealed a significant increase in immunofluorescent staining of Fc␥RI and Fc␥RII/III in cells treated with Mtmr4 siRNA (Fig. 1H). To confirm the specificity of this effect, RAW 264.7 cells were treated with Mtmr4 siRNA and subsequently transfected with HA-MTMR4 or HA-vector as a control before fixation and immunofluorescent assessment of Fc␥RI. Rescue of Mtmr4 knockdown by HA-MTMR4 overexpression, but no change in HA-vector control samples, verified the specific regulation of Fc␥R surface levels by MTMR4 (Fig. 1I).
One of the critical events initiated via clustered Fc␥R signaling is the local remodeling of the submembranous actin cytoskeleton (21), which contributes to the growth of pseudopods around the particle (22,23). We next assessed the presence of F-actin at phagocytic cups in cells overexpressing HA-MTMR4. RAW 264.7 cells transfected with either HA-vector or HA-MTMR4 were incubated with 6-m IgG-opsonized latex beads (bIgG-6m) to facilitate Fc␥R-mediated phagocytosis. Immunofluorescent assessment showed the distinct presence of phalloidin-stained F-actin at the phagocytic cups in control cells as well as HA-MTMR4expressing cells ( Fig. 2A). However, quantification revealed that cells expressing HA-MTMR4 showed a 20% reduction in the intensity of F-actin at phagocytic cups compared with vector controls (Fig. 2B). In contrast, cells treated with Mtrm4 siRNA showed a 57% increase in F-actin intensity at phagocytic cups (Fig. 2C). Therefore, MTMR4 overexpression decreases Fc␥RI at the cell surface and is associated with decreased phagocytic cup F-actin.

MTMR4 negatively regulates phagocytosis
One possible functional outcome of altered Fc␥R surface expression and actin polymerization is altered phagocytosis induction (1). Therefore, we next investigated whether MTMR4 regulates the efficiency of phagocytosis in macrophages. RAW 264.7 cells expressing HA-MTMR4 or HA-vector as a control were incubated with bIgG-6m, and the phagocytic index was determined as the number of fully internalized beads per 100 cells normalized to HA-vector control. The phagocytic index was MTMR4 regulates phagocytosis reduced in cells expressing HA-MTMR4 compared with vector controls (Fig. 3A). Because a similar effect was observed using 9-m (Fig. 3B) and 3-m (Fig. 3C) IgG beads, this effect was not dependent on bead size, suggesting that MTMR4 does not regulate phagocytosis by restriction of membrane delivery to the phagocytic cup (24). The catalytic activity of MTMR4 was required for A, lysates of nontransfected (i) and HA-vector-or HA-MTMR4 -transfected (ii) RAW 264.7 macrophages were subjected to SDS-PAGE and immunoblotted using a polyclonal antibody raised against recombinant MTMR4 or an anti-HA antibody for the detection of endogenous MTMR4 (arrowhead) and recombinant HA-MTMR4 (arrow), respectively. An anti-GAPDH antibody was used as loading control. B, cells were transfected with HA-vector or HA-MTMR4 as indicated and fixed. Unpermeabilized cells were stained with anti-Fc␥RI (red) and anti-Fc␥RII/III (green) antibodies and were thereafter permeabilized and stained again using an anti-HA antibody (blue) to detect transfected cells (marked with an asterisk). C, the Fc␥RI fluorescence signal in B was quantified in three independent experiments with 30 cells analyzed per condition. Within experiments, the fluorescence was normalized to that of the control condition, which was arbitrarily assigned a value of 100. D, RAW 264.7 cells were treated with negative control or Mtmr4 siRNA for 72 h (siRNA 1) or 48 h (siRNA 2 and siRNA 3), after which Mtmr4 mRNA levels were quantitated by RT-PCR analysis relative to Gapdh. mRNA levels were normalized to that of control siRNA cells, which was arbitrarily assigned a value of 1. E, siRNA-mediated knockdown of MTMR4 in lysates from control or Mtmr4 siRNA 1-treated cells was assessed by Western blotting using a polyclonal anti-MTMR4 antibody and anti-GAPDH antibody as loading control. F, cells were treated with control or Mtmr4 siRNA 3, and Fc␥RI and Fc␥RII/III signal fluorescence was quantified by flow cytometry in six independent experiments with Ͼ1000 cells analyzed. Fluorescence was normalized to that of control siRNA cells, which was arbitrarily assigned a value of 100. G, immunofluorescent micrographs of cells transfected with control siRNA or Mtmr4 siRNA 3, as indicated, and immunostained using anti-Fc␥RI and -Fc␥RII/III antibodies. H, the Fc␥RI signal fluorescence of cells transfected with control siRNA or Mtmr4 siRNA 2 or 3 was quantified in three independent experiments with 30 cells analyzed per condition. Within experiments, the fluorescence was normalized to that of the control condition, which was arbitrarily assigned a value of 100. I, the relative Fc␥RI fluorescence signal intensity of cells transfected with HA-vector or HA-MTMR4 and control or Mtmr4 siRNA 3 was quantified. Within experiments, the fluorescence was normalized to that of the control condition, which was arbitrarily assigned a value of 100. *, p Ͻ 0.05, two-tailed paired t test. Scale bars, 10 m. Error bars, S.E. Images are representative of at least three independent experiments.

MTMR4 regulates phagocytosis
phagocytosis regulation, as no difference in the phagocytic index was observed in cells transfected with catalytically inactive HA-MTMR4(C407A) (Fig. S2). Mtmr4 siRNA-treated cells showed a 16 -22% increase in the phagocytosis of bIgG-6m compared with control siRNA cells (Fig. 3D). To evaluate the extent to which MTMR4 negative regulation of phagocytosis depended on PI3K signaling, phagocytosis assays were carried out in cells pretreated with the PI3K inhibitor LY294002, which inhibits both phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P 3 ) and PtdIns(3)P generation by Class I, II, and III PI3Ks (with the exception of Class II␣) (7). LY294002 treatment resulted in significant inhibition of phagocytosis with no observed effect of Mtmr4 knockdown compared with control cells under these conditions (Fig. 3E). This result is consistent with the contention that MTMR4 negative regulation of phagocytosis occurs in part via a PI3K-dependent mechanism.

MTMR4 localizes to phagosomes in macrophages
To assess MTMR4 subcellular localization during phagocytosis, HA-MTMR4 -expressing RAW 264.7 cells were examined by immunofluorescence when phagocytosing bIgG-6m particles. HA-MTMR4 was found to be enriched on punctate and vesicular structures and on phagosomes with cytosolic signal also present (Fig. 4A), revealing that MTMR4 is recruited to phagosomes during Fc␥R-mediated phagocytosis.
We noted that phagosomal MTMR4 levels varied between phagosomes, consistent with the transient pattern of phagosomal PtdIns(3)P signal appearance and disappearance (4). To evaluate the spatiotemporal recruitment of MTMR4 to individual phagosomes, cells expressing YFP-MTMR4 were imaged live during phagocytosis of bIgG-6m ( Fig. 4B and Movie S1). As an experimental control, cells were cotransfected with CFP, a cytoplasmic marker, to ensure that YFP signal detected at the phagosome was the result of YFP-MTMR4 recruitment and not a consequence of morphometric changes due to pseudopodia and membrane ruffling (25,26). Under these conditions, mobilization of YFP-MTMR4-positive vesicles toward the base of the phagocytic cup was observed at the 1-min time point (Fig.   4B, arrows). Following engulfment, increased YFP-MTMR4 signal was detected around the phagosome (Fig. 4B). MTMR4 was also observed on motile structures with transient contacts made with each other and the phagosome, a pattern reminiscent of endosomal "kiss and run" behavior ( Fig. 4B (arrowheads) and Movie S1) (4). Similar changes were not observed in the cytoplasmic CFP signal, which remained evenly distributed over the entire time course (Fig. 4B). To quantify the recruitment of MTMR4 to the phagosome, the ratio of YFP-MTMR4 fluorescence to CFP fluorescence in the region of the phagosome (R p ) was calculated, and this was normalized to the corresponding YFP/CFP fluorescence of the entire cell (R c ) to control for variation in construct expression. The early recruitment of YFP-MTMR4 to the phagosome during engulfment was reflected by a steep increase in the R p /R c ratio, followed by a decline in YFP-MTMR4 fluorescence on the phagosome over a time scale of 5-10 min (Fig. 4C). Hereafter, the signal showed increased variability but remained above cellular background for Ͼ30 min in some cells. These results indicate a dynamic pattern of time-dependent recruitment of MTMR4 to the phagosome.

MTMR4 regulates the duration of PtdIns(3)P signaling on phagosomes
MTMR4 dephosphorylates PtdIns(3)P and PtdIns(3,5)P 2 in vitro and regulates endosomal PtdIns(3)P in vivo (12). However, whether MTMR4 plays a role in regulating the removal of phagosomal PtdIns(3)P is unknown. The temporal relationship between PtdIns(3)P and MTMR4 phagosomal recruitment was therefore assessed by cotransfecting cells with the PtdIns(3)P biosensor mCherry-2xFYVE (2xFYVE) and YFP-MTMR4, followed by live imaging during the phagocytosis of bIgG-6m. The 2xFYVE probe was enriched on the formed phagosome with peak signal and then a decline in signal within the first 10 min as reported ( Fig. 5Ai) (4). YFP-MTMR4 recruitment to the phagosome occurred after 2xFYVE enrichment and coincided with the decline in 2xFYVE signal (Fig. 5Aii). This was supported by quantification of 2xFYVE and YFP-MTMR4 sig- A, RAW 264.7 cells were transfected with HA-vector or HA-MTMR4, allowed to commence phagocytosis of bIgG-6m, fixed, and stained with phalloidin (green) and anti-HA antibodies (red) to detect F-actin and transfected cells, respectively. Immunofluorescent micrographs of transfected cells as well as merged and bright-field images are shown. Arrows, phagocytic cups. Scale bars, 5 m. B, phalloidin fluorescence was quantified on phagocytic cups in HA-vector-or HA-MTMR4 -transfected cells undergoing phagocytosis of bIgG-6m, after which they were fixed and stained as in A. 27-34 phagosomes were examined in three independent experiments, and fluorescence was normalized to the control condition, which was arbitrarily assigned a value of 100. C, quantification of pseudopodal phalloidin signal in cells treated with control or Mtrm4 siRNA 1 undergoing phagocytosis, fixed and stained as described in A. 176 -200 phagosomes were examined in four independent experiments, and fluorescence was normalized to the control condition, which was arbitrarily assigned a value of 100. Error bars, S.E. *, p Ͻ 0.05, two-tailed paired t test. Images are representative of at least three independent experiments.

MTMR4 regulates phagocytosis
nals on phagosomes, indicating a decrease in phagosomal PtdIns(3)P immediately after recruitment of YFP-MTMR4 to the phagosome (Fig. 5B). To characterize the role of MTMR4 in PtdIns(3)P phagosome kinetics, we assessed the accumulation of mCherry-2xFYVE during phagocytosis in cells coexpressing HA-vector control (Fig. 5C), HA-MTMR4 (Fig. 5D), or a catalytically inactive mutant HA-MTMR4(C407A) (Fig. 5E) that we have described previously (12). Again, the 2xFYVE signal around the phagosome was quantified in each experiment. In cells expressing HA-MTMR4, phagosomal PtdIns(3)P signals initially rose at a similar rate relative to HA-vector-transfected cells, but they peaked and declined more rapidly with MTMR4 expression (Fig. 5F and Movies S2 and S3). In contrast, cells expressing HA-MTMR4(C407A) showed increased phagosomal PtdIns(3)P levels and longer signal duration compared with controls ( Fig. 5F and Movie S4).
In a similar experiment investigating the effect of Mtmr4 depletion on PI(3)P levels during phagocytosis, phagosomal PtdIns(3)P signals were assessed live in cells cotransfected with mCherry-2xFYVE and control ( Fig. 6A and Movie S5) or Mtmr4 siRNA (Fig. 6B and Movie S6). The results were quantified and showed prolonged phagosomal 2xFYVE signal in Mtmr4 siRNA cells as compared with control (Fig. 6C). Therefore, overexpression of MTMR4 does not abolish PtdIns(3)P recruitment to the phagosome but attenuates its peak amplitude and accelerates its decline, whereas expression of a catalytically inactive MTMR4 variant delays the dephosphorylation of PtdIns(3)P on phagosomes, a result also observed using Mtmr4 siRNA. These results are consistent with the contention that MTMR4 is recruited to phagosomes and degrades PtdIns(3)P on phagosomal membranes.

MTMR4 regulates the maturation of phagosomes containing pathogenic mycobacteria
As MTMR4 appears to regulate key components of the early phagocytic pathway, including the surface expression of Fc␥Rs and phagosomal PtdIns(3)P levels, we also investigated whether MTMR4 is required for normal phagosomal maturation, which includes the acquisition of an acidic lumen and late endosome/lysosome markers (5,6). Pathogenic mycobacteria secrete 3-phosphatases that degrade PtdIns(3)P and thus arrest phagosomal maturation and the lysosomal degradation of the bacteria (27,28). Blockage of phagosomal maturation is observed with Mycobacterium tuberculosis or Mycobacterium marinum infection, but not Mycobacterium smegmatis, the latter a nonpathogenic mycobacteria (29,30). As MTMR4 suppresses phagosomal PtdIns(3)P association, we examined whether Mtmr4 knockdown, which leads to prolonged phagosomal PtdIns(3)P signaling, alters the maturation of phagosomes containing M. marinum. To this end, RAW 264.7 cells transfected with control or Mtmr4 siRNA were infected with live GFP-M. marinum, and LysoTracker Red was used to detect acidic organelles (Fig. 7A). Maturation of phagosomes was assessed by scoring the percentage of engulfed GFP-expressing mycobacteria that colocalized with LysoTracker-positive compartments, as reported (30). In control siRNA cells, 38% of GFP-M. marinum phagosomes colocalized with LysoTracker Red after 2 h of chase, indicating that the majority of GFP-M. marinum resided in a nonacidic, immature phagosome (not colocalized), suggesting mycobacterial phagosomal arrest (Fig. 7B). Conversely, Mtmr4 knockdown resulted in 53% colocalization (p Ͻ 0.05, Fisher's exact test), suggesting increased phagosomal maturation (Fig. 7B). These results were confirmed by quantifying the colocalization of GFP-M. marinum with an alternative marker for maturation that detects late endosomes and lysosomes. In this experiment, mature phagosomes were detected by preload and chase of transfected RAW 264.7 cells with Alexa Fluor 555-dextran, which is incorporated into late endosomes and lysosomes (Fig. 7C) (31). In control siRNA cells, only 21% of

Discussion
This is the first report describing the regulation of phagocytosis and phagocytic maturation in macrophages by the myotubularin PtdIns(3)P 3-phosphatase MTMR4. MTMR4 decreases macrophage surface levels of Fc␥Rs, which are critical for the

MTMR4 regulates phagocytosis
initiation of Fc␥R-mediated phagocytosis. Accordingly, MTMR4 depletion increases the efficiency of Fc␥R-mediated phagocytosis. Using live-cell imaging, we observed that phagosomal PtdIns(3)P formation precedes recruitment of MTMR4 to the phagosome, and MTMR4 temporally regulates the duration and amplitude of phagosomal PtdIns(3)P and, as a consequence, directs the maturation of the forming phagosome.
We were intrigued by previous findings that MTMR4 regulates the trafficking of transferrin in HeLa cells (12). Overexpression of MTMR4 blocks the sorting of transferrin out of

MTMR4 regulates phagocytosis
early endosomes, and MTMR4 regulates the distribution of recycling endosomes (12). In resting cells, ligand-free Fc␥Rs undergo constitutive cycles of endocytosis from the plasma membrane into early endosomes and sorting back to the plasma membrane via recycling endosomes (20). In this current study, surface levels of Fc␥RI were decreased in cells overexpressing HA-MTMR4. We did not see a significant change in the level of surface Fc␥RII/III in HA-MTMR4 -expressing cells, possibly because of experimental limitations making the microscopic observation of subtle changes in immunofluorescent staining challenging. Conversely, we found that surface levels of both Fc␥RI and Fc␥RII/III were increased in cells treated with Mtmr4 siRNA. There is an emerging function for PtdIns(3)P in regulating the transfer of cargo from the early endosomes into recycling compartments (8,32). One possible mechanism is via the sorting nexins, such as SNX4 and SNX17, and the interaction of their PX domains with early endosomal PtdIns(3)P to generate recycling tubules (8,33,34). Perturbed sorting of Fc␥Rs out of early endosomes would, in turn, reduce the recycling of Fc␥R back to the plasma membrane. The polymerization of F-actin is essential for pseudopod extension and particle engulfment. As further evidence of reduced receptor expression at the cell surface, we observed that MTMR4-overexpressing cells showed lower levels of F-actin at phagocytic cups, presumably as a consequence of decreased Fc␥R signaling. It is thus possible that MTMR4, which colocalizes with early and recycling endosomes (12), reduces phagocytosis initiation by decreasing surface Fc␥R levels and therefore signaling through regulation of endocytic recycling, thereby compromising the ability of macrophages to form phagocytic cups (see proposed model, Fig. 8A).
Although our results were generated using the widely used RAW 264.7 cell line as a representative for mammalian macrophages, validation using multiple cell lines would be of interest in future studies. Our finding that MTMR4 negatively regulates phagocytosis in RAW 264.7 macrophages is concordant with experiments in C. elegans demonstrating that the myotubularin, MTM-1, negatively regulates the internalization of apoptotic cells (35,36). The observed negative effect of MTMR4 on internalization may be mediated through regulation of PtdIns(3)P, and consistent with this, pretreatment with the PI3K inhibitor LY294002 abrogated the enhanced phagocytosis observed in Mtmr4 siRNA cells compared with controls. In the C. elegans model, Zou et al. (36) proposed that MTM-1 may regulate a pool of phagosomal PtdIns(3)P on the plasma membrane. However, although plasma membrane PtdIns(3)P has been noted in muscle cells and adipocytes during insulin signaling, we and other studies have found that the 2xFYVE probe does not localize to phagocytic pseudopods but appears at later time points on internalized phagosomal membranes (4,5,26,(37)(38)(39). Together, these observations indicate a PtdIns(3)P-dependent process determining phagocytic ability early in particle engulfment. Our presented results suggest that one such process might be the endocytic recycling of Fc␥Rs. Whether PtdIns(3)P also plays a direct role in the early stages of phagocytosis is unclear; however, its presence at the phagosome has been shown here and elsewhere (4).
Phagosomal PtdIns(3)P regulates phagosomal maturation and inhibition of the Class III PI3K disrupts phagosomal maturation (5,28). Mycobacteria subvert their own degradation as they reside in an immature phagosome and reduce phagosomal PtdIns(3)P via the secretion of bacterial 3-phosphatases, such as SapM and MptpB (28,40). As shown here, Mtmr4 siRNA cells exhibited increased maturation of M. marinum-containing phagosomes. This suggests that the delay in turnover of phagosomal PtdIns(3)P in Mtmr4 siRNA-treated cells is able to partially overcome the maturation block induced by mycobac-

MTMR4 regulates phagocytosis
teria and counteract the effect of the pathogen-secreted 3-phosphatase virulence factors. Inhibition of phagosomal maturation may thus represent a second mechanism by which MTMR4 regulates phagocytosis (Fig. 8B). Prolonged phagosomal PtdIns(3)P as observed in Mtmr4 siRNA cells may increase the recruitment of effectors such as EEA1 and Hrs to the phagosome cytosol interface, to promote membrane fusion and the development of multivesicular bodies, respectively (6,41). Additionally, there is evidence that some phagosomes have cyclic alterations in PtdIns(3)P (42,43). It is of interest to speculate that interruption of regulated turnover of PtdIns(3)P on phagosomes may alter the ordered recruitment of effectors or the conversion of early to mature Rab effectors (8).
During phagocytosis, there is a highly regulated temporal and spatial restriction of various phosphoinositides at the phagocytic cup and phagosome generated by distinct phosphoinositide kinases and phosphatases at the cytosolic/membrane interface (4,5,8,44,45). Ectopically expressed MTM1, which has a catalytic 3-phosphatase domain and a PH-GRAM domain, but lacks a FYVE domain, is recruited to the phagosome in RAW 264.7 macrophages and PLB-985 myeloid leukemia cells (28,46). Knockdown of MTM1 in PLB-985 cells led to
In summary, we have demonstrated the dynamic recruitment of a myotubularin 3-phosphatase, MTMR4, to the phagosome in RAW 264.7 macrophages and demonstrated MTMR4 regulation of Fc␥R levels and phagosomal PtdIns(3)P contributes to the regulation of phagosomal initiation and maturation. Therefore, our studies identify a novel mechanism for the coor-dinated termination of PtdIns(3)P signaling on macrophage phagosomes.

Materials
All reagents were from Sigma-Aldrich (Castle Hill, New South Wales, Australia) unless specified. DMEM and fluorescent conjugates were from Invitrogen (Mt. Waverley, Victoria, Australia). Anti-MTMR4 polyclonal antibody has been described previously (12). Anti-HA was from Covance (Princeton, NJ), anti-Fc␥R I and II/III were from BD (North Ryde, New South Wales, Australia), and 6 -9-m beads were from Bangs Laboratories (Carmel, IN).

Immunoblotting
To assess endogenous MTMR4 or recombinant HA-MTMR4 expression in macrophages, RAW 264.7 cells were lysed at 4°C in lysis buffer (50 mM Tris, 2 mM EDTA, 1% Triton X-100, 150 mM NaCl, protease inhibitor mixture (Roche Applied Science, catalog no. 11836153001)). After centrifugation, proteins in the Triton X-100 -soluble fraction were resuspended in SDS-PAGE sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 50 mM DTT, 0.01% bromphenol blue) and separated by 7.5% SDS-PAGE under reducing conditions, transferred onto a nitrocellulose membrane, blocked in Tris-buffered saline (TBS: 20 mM Tris, 150 mM NaCl, pH 7.4) supplemented with 5% skim milk, and then detected by incubation for 16 h at 4°C with a rabbit polyclonal antibody raised against a purified recombinant MTMR4 peptide (1:4) or a mouse monoclonal anti-HA antibody (0.1 g/ml), respectively. The blots were incubated for 1 h at room temperature with secondary antibodies (polyclonal anti-mouse IgG-horseradish peroxidase (Chemicon, Temecula, CA). All washing between the steps was carried out with 0.05% Tween 20 in TBS. Blots were developed using Enhanced Chemiluminescence reagent (Western Lightning Plus, PerkinElmer Life Sciences) and exposure to X-ray film for the appropriate times, and films were developed.

Immunofluorescence staining
RAW 264.7 cells were fixed in 3% paraformaldehyde, permeabilized with 0.1% Triton X-100 in PBS for 2 min and blocked with 5% goat serum or 3% BSA prior to incubation with primary and then secondary antibodies with PBS washes in between antibody incubation steps. Coverslips were mounted with Slowfade gold reagent (Invitrogen), and imaging was conducted by confocal or wide-field microscopy.
In experiments carried out to detect Fc␥Rs on transfected cells, unpermeabilized cells were first stained with mouse Alexa 647-conjugated anti-Fc␥RI (CD64) (catalog no. 558539, BD Pharmingen) and/or rat anti-Fc␥RII/III IgG 2b (clone 2.4G2; catalog no. 553141, BD Pharmingen). Cells were further stained with a fluorescent conjugated secondary chicken anti-mouse-647 and/or donkey anti-rat-488 antibody. In experiments where cells were transfected with HA-vector or HA-MTMR4, cells were permeabilized after Fc␥R staining and stained again with mouse monoclonal anti-HA (Covance, Princeton, NJ) and then antimouse IgG-Alexa Fluor 594 to detect transfected cells.
In experiments carried out to image and quantify phagocytosis using IgG-opsonized latex beads, fixed, unpermeabilized samples were first blocked then incubated with anti-human IgG-Alexa Fluor 594 secondary antibodies for 10 min to identify noninternalized beads. Samples were then washed, permeabilized, and reblocked before incubation with primary and secondary antibodies with fluorophores other than 594. The primary mouse anti-HA antibody was detected by secondary anti-mouse IgG Alexa Fluor 488, unless F-actin was assessed using Alexa 488-phalloidin, in which case the primary anti-HA antibody was detected by secondary anti-mouse IgG Alexa Fluor 594.

Quantitation of external Fc␥Rs in RAW 264.7 cells using immunofluorescent microscopy
RAW 264.7 cells stained with anti-Fc␥R antibodies were imaged with a Leica SP5 5 channel confocal microscope PL APO ϫ100 objective oil immersion NA 1.4, using LASAF software. Images were analyzed using ImageJ (National Institutes of Health), using the line tool. For overexpression studies, where cell numbers were limited by transfection, 50 lines were placed haphazardly transecting the plasma membrane, and the maximum fluorescence per line was recorded and averaged for each cell. For siRNA studies, where many more cells could be analyzed, four lines were placed per cell transecting the plasma membrane, and the maximum fluorescence was calculated and averaged for each cell.

RNA isolation and relative quantitative RT-PCR
For RT-PCR analysis, cells were lysed and RNA was purified using an RNeasy RNA extraction kit according to the manufacturer's instructions (Qiagen catalog no. 74104). RT-PCR was conducted on samples containing 10 ng of RNA in a two-step reaction using a Quantitect SYBR Green PCR kit (Qiagen) in 20-ml reactions run in triplicate in a Rotorgene 3000 real-time cycler (Corbett, Monash Biochemistry Imaging Facility). Primers were used to amplify mouse Mtmr4 and Gapdh cDNA (Quantitect Qiagen catalog nos. QT00148757 and QT01658692, respectively). The -fold difference in Mtmr4 mRNA in control versus Mtmr4 siRNA cells was calculated as described (55), using Gapdh as internal control.

Flow cytometry analysis of Fc␥RI and Fc␥RII/III in Mtmr4-knockdown cells
For cell-surface staining, Mtmr4 siRNA-and control siRNA-transfected cells were detached with 0.5 mM EDTA, 1% fetal bovine serum in PBS, pelleted by centrifugation, and then washed in PBS followed by one wash in stain buffer (2% BSA, 0.1% sodium azide in PBS). After centrifugation and aspiration of the supernatant, the cell pellet was gently resuspended in stain buffer at 1 ϫ 10 6 cells/100 l. 100-l aliquots of cells were then stained with Alexa Fluor 647-mouse anti-mouse Fc␥RI IgG 1 (clone X54 -5/7.1; 558539, BD Pharmingen), rat antimouse Fc␥RII/III IgG 2b (also called mouse BD Fc block; clone 2.4G2; 553141, BD Pharmingen), Alexa Fluor 647-mouse IgG 1 isotype control (clone MOPC-21; 557732, BD Pharmingen), or 647-rat IgG control (Thermo Fisher Scientific, catalog no. A-21247) on ice for 45 min away from light. Cells were then washed in stain buffer, centrifuged, and resuspended in 0.1% PFA, 2% glucose solution. Fc␥RII/III-stained cells were further stained with a fluorescent conjugated secondary anti-rat-647 antibody on ice for 45 min away from light. Cells were washed in stain buffer, centrifuged, and resuspended in 0.1% PFA, 2% glucose solution. Cells remained on ice away from light until analysis by flow cytometry. For intracellular staining, transfected cells were washed in PBS followed by fixation in 4% PFA for 10 -15 min at room temperature. After several washes in stain buffer, cells were resuspended in ice-cold 100% methanol

MTMR4 regulates phagocytosis
and permeabilized for 20 min on ice. Cells were then washed several times in stain buffer and resuspended in stain buffer at 1 ϫ 10 6 cells/100 l. Staining was then performed as described for cell-surface staining. 10,000 cells from controls and test samples were analyzed using a Cyan ADP analyzer and Summit version 4.3.1 software (Beckman Coulter, Lane Cove, New South Wales, Australia). Gates were set around unstained cells and cells stained with isotype controls based on their forward/ sideward light scatter pattern. Fc␥RI and Fc␥RIII/II cell-surface and intracellular expression levels (median fluorescence intensity) for Mtmr4 siRNA transfected cells were expressed relative to control cells. Data analysis was performed using FlowJo version 10.1 data analysis software (FlowJo LLC, Ashland, OR).

Phagocytosis of IgG-opsonized beads
3-, 6-, and 9-m latex beads were opsonized in human IgG by resuspension in thawed aliquots of 1 mg/ml human IgG with a volume ratio of 1:2.5 (45). RAW 264.7 cells were incubated with IgG-opsonized latex beads in 5% DMEM (for Accell siRNA experiments) or 10% FCS (for overexpression and Ambion siRNA experiments) on ice for 5 min and then at 37°C for 15 min. Cells were then fixed and stained with Alexa Fluor 594 anti-IgG to detect external beads. Internalized beads were identified by a combination of bright-field and fluorescence microscopy, and the phagocytic index was scored as the number of beads internalized per 100 cells. For the inhibition of PI3K signaling, cells were pretreated with 100 M LY294002 for 30 min. All results within independent experiments were normalized to that of the control condition (45). Cells were examined on an Olympus Provis wide-field microscope using UPLAN ϫ20 oil immersion objective NA 0.8 and UPLAN ϫ40 oil immersion objectives NA 1, acquired with an FView2 camera and Olympus analaSIS software, and images were analyzed using ImageJ.

Quantitation of actin in transfected cells
RAW 264.7 cells stained with Alexa Fluor 488-phalloidin were imaged using a Leica SP5 five-channel confocal microscope PL APO ϫ100 objective oil immersion NA 1.4, using LASAF software. For quantitation of phagosomal actin, only pseudopods that were extending parallel to the plane of the coverslip and had not yet fused circumferentially were measured, to ensure similar temporal stages between phagosomes. The ImageJ freehand tool was used to outline the pseudopods, and the average fluorescence in this region of interest was recorded.

Live-cell imaging of YFP-MTMR4 during phagocytosis
RAW 264.7 cells were grown on glass coverslips and transfected with 3 g of YFP-MTMR4 and 2 g of CFP empty vector (used as a cytoplasmic fluorescent control). 24 h later, cells were washed, and medium was replaced with DMEM HEPES containing 10% FCS and bIgG-6m added before live visualization. Imaging was performed using a Leica LX6000 wide-field microscope, 37°C humidified chamber, PL APO ϫ63 water immersion objective NA 1.2, using LASAF software. Quantitation of YFP-MTMR4 and CFP at the phagosome was performed using Metamorph software as described previously (26). The fluorescent images were each subjected to background subtraction. A binary mask was made by the addition of both fluorescent images and application of a manual threshold. The YFP image was multiplied by 1000 and divided by the CFP image to give a divided image. Binary and divided images were combined using the AND process, so that extracellular signal was excluded, and only signal in the divided channel present was within the mask. A square-shaped region of interest was placed around the phagosome and another slightly larger square around this for particle tracking in the bright-field channel using the cross-correlation centroid-tracking algorithm. Manual readjustment was performed where automatic tracking became erroneous. Regions tracked in the bright-field channel were transferred to the divided image, and measurements were taken of 1000 ϫ YFP/CFP signal in the phagosomal region (R p ). At each time point, the entire cell fluorescence in the divided image was also measured to give 1000 ϫ YFP/CFP (R c ). The phagosomal fluorescence at each time point was normalized to cellular fluorescence by calculating R p /R c .

Live-cell imaging of mCherry-2xFYVE during phagocytosis
For quantification of PtdIns(3)P during phagocytosis, cells on 25-mm glass coverslips were cotransfected with 5 g of HAvector, HA-MTMR4, or HA-MTMR4(C407A) together with 0.5 g of mCherry-2xFYVE. Cells were resuspended in DMEM-HEPES plus 10% FCS and imaged live following the addition of bIgG-6m. Time-lapse imaging was conducted with recording of mCherry fluorescence at 15-s intervals using a Leica LX6000 wide-field microscope, 37°C humidified chamber, PL APO ϫ63 water immersion objective NA 1.2, and LASAF software. Alternatively, cells seeded to poly-L-lysine-coated microchamber slides were transfected with Mtmr4 or control siRNA for 48 h in manufacturer-supplied serum-free media, followed by transfection with mCherry-2xFYVE for an additional 24 h. Cells were resuspended in 5% DMEM 1 h prior to experiments and then imaged live during phagocytosis suspended in DMEM-HEPES plus 5% FCS as above, using a LX6000 widefield microscope, 37°C humidified chamber, and FLUOTAR ϫ40 oil immersion objective NA 1.0, with images acquired at 30-s intervals.
Phagosomal mCherry-2xFYVE was quantified using ImageJ by measuring the average fluorescence in a circular region of interest manually placed around the phagosome at each time point. This was normalized for probe expression within the rest of the cell by dividing by the average cellular fluorescence less the average phagosomal fluorescence (region cell minus region phagosome (R c Ϫ p )), to give R p /R c Ϫ p at each time point.

Analysis of mycobacterial phagosomal maturation in Mtmr4knockdown cells
RAW 264.7 cells were incubated on ice for 20 min with 1 ϫ 10 7 mycobacteria, washed, and chased for 2 h at 37°C. 30 min prior to fixation, cells were incubated with LysoTracker Red at 1:5000. Alternatively, to label late endosomal/lysosomal compartments, cells were first incubated with 42 g/ml fluorescent dextran 10,000 M r for 1 h and chased for 2 h, prior to infection as above. The number of GFP-mycobacteria that colocalized with LysoTracker Red or, in alternative experiments, dextran was scored using a Leica SP5 five-channel confocal microscope, PL APO ϫ63 oil immersion objective NA 1.4, acquired with LASAF software.

Statistical analyses
Statistical tests were performed using GraphPad Prism version 5 (GraphPad software, La Jolla, CA) and are detailed in each figure legend. p Ͻ 0.05 was considered significant.

Image presentation
For image presentation, brightness/contrast adjustments were kept constant between different conditions and applied to the entire image, and no nonlinear adjustments were performed.