Phosphoinositide 5′-phosphatases have been implicated in the regulation of phagocytosis. However, their precise roles in the phagocytic process are poorly understood. We prepared RAW264.7 macrophages deficient in Inpp5e (shInpp5e) to clarify the role of this lipid phosphatase. In the shInpp5e cells, the uptake of solid particles was increased and the rate of phagosome acidification was accelerated. As expected, levels of PtdIns(3,4,5)P3 and PtdIns(3,4)P2 were increased and decreased respectively, on the forming phagocytic cups of these cells. Unexpectedly, the most prominent consequence of the Inpp5e deficiency was the decreased accumulation of PtdIns3P and Rab5 on the phagosome. The expression of a constitutively active form of Rab5b in the shInpp5e cells rescued the PtdIns3P accumulation. Rab20 has been reported to regulate the activity of Rabex5, a guanine nucleotide exchange factor for Rab5. The association of Rab20 with the phagosome was remarkably abrogated in the shInpp5e cells. Over-expression of Rab20 increased phagosomal PtdIns3P accumulation and delayed its elimination. These results suggest that Inpp5e, through functional interactions with Rab20 on the phagosome, activates Rab5, which, in turn, increases PtdIns3P and delays phagosome acidification.
Phagocytosis, an important process in innate and adaptive immune systems, refers to the uptake of large particles, including invading microorganisms and apoptotic self-cells, by specialized phagocytes [1–3]. The process of phagocytosis involves specific cell-surface receptors. One group of receptors, including mannose–fucose receptors, dectin1 and toll-like receptors, interact directly with the pathogens by recognizing conserved motifs on the microorganisms. Another set of receptors recognize opsonins, i.e. the serum components (e.g. antibodies and complement components) that interact with the surface molecules of foreign particles. The clustering of opsonin receptors on phagocytes induces the extension of the membrane around the particle, resulting in the formation of a phagocytic cup. This newly formed phagosome then undergoes maturation, a series of fusion and fission events that culminates in the formation of the phagolysosome, the next step for microbial elimination.
The relative abundances of certain phosphoinositide (PI) species fluctuate over the course of Fc-receptor-mediated phagocytosis [4–6]. For example, PtdIns(3,4,5)P3 transiently accumulates in the membrane during the formation of the phagocytic cup . This event is essential for the closure of the phagocytic cup when enveloping relatively large particles [8,9]. Shortly after the sealing of the phagocytic cup, PtdIns3P appears on the nascent phagosome [10–13] where it has an essential role in driving phagosome maturation [12,13]. The PtdIns3P on the phagosome is metabolized to PtdIns or phosphorylated to PtdIns(3,5)P2 during the subsequent maturation and acidification . Multiple PI-modifying enzymes have been implicated in the phagocytic process [15–17]. The sole pathway for producing PtdIns(3,4,5)P3 is the phosphorylation of PtdIns(4,5)P2 by class I PI 3-kinases. The formed PtdIns(3,4,5)P3 is degraded rapidly to either PtdIns(4,5)P2 or PtdIns(3,4)P2 by PI phosphatases. Cells lacking the PI 3′-phosphatase PTEN (phosphatase and tensin homologue deleted on chromosome 10) or the PI 5′-phosphatases, such as SHIP1 (Src homology 2 domain-containing inositol phosphatase 1), SHIP2 or Inpp5e, show an increased phagocytic activity [18–21]. Thus, decreased degradation of PtdIns(3,4,5)P3 on the forming phagocytic cup seems to activate phagocytosis. Interestingly, a defect in any one of the 5′-phosphatases is sufficient to increase phagocytosis [18–21], suggesting that each plays specific, non-redundant roles. However, little is known about their differential actions in the phagocytic process.
Members of the Rab family of GTPases play pivotal roles in phagosome formation and maturation through potential molecular links with PI species and PI-modifying enzymes [22,23]. More than 20 Rab proteins have been detected on phagosome membranes . Among them, Rab5 and Rab7 are the best characterized. Rab5 appears first on the phagocytic cup and promotes its transition to later stages . GTP-bound Rab5 associates with and activates class III PI 3-kinase (Vps34), leading to the accumulation of PtdIns3P . Rab5 also binds to a class I PI 3-kinase (p110β), two PI 5′-phosphatases [Inpp5b and oculocerebrorenal syndrome of Lowe (OCRL)] and a PI 4′-phosphatase, all of which may contribute to the regulation of phagosomal PI metabolism [25–28]. The early-to-late endosome transition is completed by the GAP-mediated inactivation of Rab5 and the displacement of Rabex5, a guanine nucleotide exchange factor for Rab5 [28,29]. After the dissociation of Rab5, Rab7 appears on the phagosome and facilitates its interaction with late endosomes and lysosomes . Recently, Rab20 has attracted attention as another regulator of phagosome maturation [30–32]. Expression of a dominant negative mutant of Rab20 was reported to impair phagosome maturation in macrophages [30,31]. However, another group reported that dominant negative Rab20 accelerated phagosome acidification .
Among the PI 5′-phosphatases, OCRL and Inpp5b bind to 16 distinct Rabs, whereas Inpp5e binds specifically to Rab20 . In the present study, we prepared RAW264.7 macrophages stably expressing shRNA against Inpp5e and detected a functional interaction between Inpp5e and Rab20. The cells lacking Inpp5e showed a decreased association of Rab5, Rab20 and PtdIns3P with the phagosomal membrane. The cells also showed increased rates of phagocytosis and accelerated acidification of the phagosome. These results suggest that Inpp5e regulates the association of Rab20 with the phagosome, keeping Rab5 active and ultimately delaying the early-to-late phagosomal transition.
MATERIALS AND METHODS
The sources of materials were as follows: zymosan, fluorescein isothiocyanate (FITC)-zymosan, LysoTracker Red and anti-zymosan IgG, RPMI 1640 medium (Life Technologies); sheep red blood cells (SRBCs) (Cosmo Bio); Alexa488-conjugated anti-rabbit IgG (Cell Signaling); protein assay kit (Bio-Rad); anti-SRBC (InterCell Technologies); Na251CrO4 (MP Biomedicals).
RAW264.7 cells (ATCC) lacking Inpp5e were prepared as follows. Oligonucleotides targeting Inpp5e (Supplementary Figure S1) were cloned into the pH1 vector downstream of the H1 RNA promoter, as previously described [18,34,35], to express siRNA hairpins. For each of the targeted sequences, a pair of oligonucleotides were synthesized (Hokkaido Life Sciences) with the following sequences: 5′-CCC(X)19TTCAAGAGA(Y)19TTTTTGGAAA-3′ and 5′-CTAGTTTCCAAAAA(Y)19TCTCTTGAA(X)19GGGTGCA-3′, where (X)19 is the coding sequence and (Y)19 is the complementary sequence. The oligonucleotide pair were annealed and ligated downstream of the H1 RNA promoter at the PstI and the XbaI sites of the pH1 vector. The vectors were transfected into RAW264.7 cells (5× 106 to 10×106 cells) at 250 V/950 μF (Gene Pulser II; Bio-Rad). At 24 h after transfection, puromycin (3 μg/ml) was added to the cells and the incubation was continued to select for resistant cells. To determine the efficiency of the gene silencing, total RNA was isolated with Sepasol (Nakarai Tesque) and the mRNA was quantified by reverse transcription PCR (RT-PCR) with the following primer pair: 3′-GGACGAGACAACATCCCATT-5′ and 3′-CCATCCTTTGCAGTGACCTT-5′. Clones showing >90% inhibition of mRNA expression were selected for the present study (Figure S1). Control cells were prepared, as above with a pH1 vector containing a 400-bp stuffer sequence instead of the target sequence. Cells deficient in SHIP1 were prepared as described previously . For the phagocytosis assay and the microscopic analysis, cells were seeded in 24-well plates (Becton Dickinson) or tissue culture-coated glass bottom dishes (Greiner Bio-One) in RPMI 1640 medium containing 4.5 g/l glucose and 10% FBS in a humidified 5% CO2 atmosphere at 37°C. Immediately before starting the assay, the medium was aspirated off and replenished with incubation buffer (complete RPMI 1640 medium without NaHCO3, fortified with 20 mM HEPES/NaOH, pH 7.4). Activities were then determined by incubating the cells at 37°C in a water bath at ambient conditions.
Preparation of IgG-coated RBCs and measurement of phagocytosis
RBCs were labelled with 51Cr as described previously . IgG-coated RBCs (IgG–RBCs) were prepared by incubating labelled cells with the rabbit anti-SRBC antibody at 37°C for 10 min in 5 mM veronal buffer (pH 7.5) supplemented with 0.1% gelatin, 75 mM NaCl, 0.15 mM CaCl2, 0.5 mM MgCl2 and 10 mM EDTA, followed by incubation on ice for 15 min. RBCs were washed thrice with 5 mM veronal buffer (pH 7.5) supplemented with 0.1% gelatin, 75 mM NaCl, 0.15 mM CaCl2, 0.5 mM MgCl2 (GVB) and finally suspended in incubation buffer (detailed above). Binding and phagocytosis of the IgG–RBCs were measured as reported previously . Monolayers of RAW264.7 cells (2×105 cells/well in a 24-well plate) were incubated with 51Cr-labelled RBCs (2×107 cells) at 37°C for the indicated periods of time. Monolayers were then washed thrice with PBS to remove unbound IgG–RBCs and then briefly exposed to 0.1 ml of hypotonic PBS (5-fold diluted). The radioactivity released into the supernatant during the hypotonic shock was measured as a proxy for the amount of IgG–RBCs bound to the surface of the phagocytes. Monolayers were washed an additional three times with PBS and finally solubilized in 0.5% Triton X-100. The radioactivity in the solution was measured to determine the amount of engulfed IgG–RBCs.
Phagocytosis of zymosan
FITC-labelled zymosan particles were mixed with an equal amount of unlabelled zymosan to expedite counting and sonicated for approximately 1 min. Particles were then opsonized or non-opsonized with normal mouse serum or anti-zymosan IgG at 37°C for 60 min before use. The buffer used for the serum opsonization was GVB, whereas EDTA–GVB was used for the IgG opsonization. The opsonized zymosan was washed thrice with GVB and finally suspended in incubation buffer. This prepared zymosan was sonicated for approximately 10 s immediately before use. RAW264.7 cells (105 cells/well in a glass bottom dish) were incubated at 37°C for the indicated durations with 5×106 zymosan particles. Phagocytosis was stopped by the addition of ice-cold PBS. Cells were then washed thrice with PBS, fixed with 4% paraformaldehyde for 15 min at room temperature and finally rinsed with PBS. Fluorescence images (excitation 488 nm, emission 525 nm) and phase-contrast images of at least 100 macrophages from three randomly selected fields were taken using confocal microscopy. The mean numbers of ingested zymosan particles are expressed as particles/100 RAW264.7 cell.
The tandem domain of myosin X, PH1(N)-PH2-PH1(C) [37,38], hereafter referred to as [PH (MyoX)], was isolated from mouse macrophages by RT-PCR with the following primer pair: 5′-AGATCTCCCTATTTCCACAGTTTTC-3′ and 5′-GAATTCCTACTTGGATCTCTGCAGCA-3′. This domain was then subcloned into pEGFP-C1 (CLONTECH) at the BglII and the EcoRI sites. Inpp5e was isolated from mouse embryos with the following primer pair: 5′-GCATCCTTGACAACAAGATTGC-3′ and 3′-AAGCCCATCCTTTGCAGTGACC-5′. The fragment was subcloned into pmCherry-C1 (CLONTECH). EGFP–[3×FYVE (EEA1)], EGFP-[2×PH (Tapp1)], dsRed-Rab5b and dsRed-Rab5b(Q79L) were kindly provided by Dr Sasaki (Akita University). mCherry-[3×FYVE (EEA1)] and EGFP-Rab5 were made by subcloning the [3×FYVE (EEA1)] and Rab5 fragments into pmCherry-C1 and pEGFP-C1 respectively. The EGFP-Rab20 construct was a kind gift from Dr. Fukuda (Tohoku University).
The plasmids were transfected with the Neon™ transfection system (Invitrogen) according to the manufacturer's protocol. Approximately 24 h after transfection, cells were subjected to microscopic analysis.
Monitoring PI dynamics during the course of phagocytosis
The cells were incubated with IgG–RBCs and placed on a BIOREVO BZ9000 microscope (Keyence) equipped with a CFI Plan Apo VC60xH oil immersion lens and phagocytosis was allowed to proceed at 37°C. Unless otherwise stated, the fluorescent images were collected every 1 min and the intensity of the phagosome-associated fluorescence was analysed using a BZ-II analysis system (Keyence).
Phagosome acidification analysis
In Figure 1, cells were loaded with 50 nM LysoTracker Red for 30 min, incubated with IgG–RBCs and centrifuged at 1500 g for 1 min, then incubated for an additional 10 or 30 min. When cells were incubated with IgG–RBCs for 2 h, the target RBCs were first added and fortified with LysoTracker for the last 30 min. The cells were then washed thrice with PBS, fixed with PBS containing 4% formaldehyde for 15 min at room temperature, permeabilized with PBS containing 0.3% Triton X-100 and 0.5% BSA for 60 min and incubated with Alexa488-anti-rabbit IgG. Upon imaging, Z-stacks were captured at 1-μm steps over a z-axis distance of 8 μm. Stacks were reconstructed, merged and analysed by BZ-H2C (Keyence application for BZ-9000), which allowed us to determine the fluorescent area. The numbers of engulfed IgG–RBCs were calculated from the area of Alexa488 fluorescence. Acidic phagosomes were counted in the merged images. We analysed five stacked images, each containing 100–120 cells. The data are shown as the means±S.D. For illustration, the images were contrast-enhanced, pseudo-coloured, merged, cropped and assembled.
Enhanced phagocytosis in shInpp5e cells
Inpp5e inhibits phagocytosis and phagosome maturation
It has been reported that macrophages deficient in Inpp5e exhibit enhanced phagocytosis . In the present study, we prepared RAW264.7 cells expressing shRNA against Inpp5e. Two lines of cells (shInpp5e-3 and shInpp5e-1795 cells in Figure S1) that produce shRNAs against different Inpp5e sequences, were prepared. As expected, the uptake of the IgG-opsonized RBCs (Figures 2A–2D) and the IgG-opsonized zymosan (Figure 1E) was increased in the shInpp5e cells. The binding of the IgG–RBCs to the macrophages was almost unaffected in the shInpp5e 3–6, but a little increased in the 1795c cells (Figure 1C). The increased Fcγ receptor expression in the 1795c cells may be due to the clonal variation. Phagocytosis of non-opsonized zymosan was also increased in the shInpp5e cells (Figure 1E and 1F). SHIP1 has been reported to negatively regulate both FcγR-mediated and CR3 (complement receptor)-mediated phagocytosis, whereas Inpp5e exclusively controls the FcγR-mediated pathway . In the present study, however, the shInpp5e cells showed a slight but significant increase in the CR3-mediated phagocytosis of serum-opsonized zymosan for unknown reasons (Figure 1E). We next analysed phagosome acidification, an important process for host defence and antigen presentation. The cells were loaded with LysoTracker Red for 30 min and then incubated with IgG–RBCs for 10 min. In the shInpp5e cells, most of the engulfed RBCs were found in the LysoTracker-stained acidic compartments (Figure 1A). In the vector-control cells, by contrast, only some of the target cells were observed in the acidic compartments. Longer incubation of the control cells increased the amount of RBCs in the acidic compartment but it was still lower than that of the shInpp5e cells, even after 30 min of incubation (Figure 1B). The phagosomes were almost 100% acidic in both the control and the shInpp5e cells, when determined 2 h after the addition of the target particles (Figures 1C and 1D).
Acceleration of phagosome acidification in shInpp5e cells
Inpp5e is involved in phagosomal PI metabolism
Based on kinetic analyses, Inpp5e is thought to be the most potent PtdIns(3,4,5)P3 5′-phosphatase [39–41]. Inpp5e can also hydrolyse the 5′-position phosphate of PtdIns(3,5)P2 and PtdIns(4,5)P2. We investigated the phagosomal accumulation of PI derivatives in the shInpp5e cells by live-cell imaging. For this purpose, EGFP-fused forms of [PH (MyoX)], [2×PH (Tapp1)] and [3×FYVE (EEA1)] were used as probes for PtdIns(3,4,5)P3, PtdIns(3,4)P2 and PtdIns3P respectively. In Figure 3(A), the control and the shInpp5e cells were transfected with EGFP-[PH (MyoX)]. After the addition of IgG–RBCs, the fluorescence around the engulfed RBCs was monitored. In control cells, PtdIns(3,4,5)P3 accumulated rapidly on the forming phagocytic cup but disappeared within 5 min after the onset of phagocytosis (Figures 3A and 3B), as reported above. In the shInpp5e cells, the level of PtdIns(3,4,5)P3 was slightly higher than that of the control cells. Because the difference was small, we next examined whether the level of PtdIns(3,4)P2, a metabolite of PtdIns(3,4,5)P3, was altered (Figures 3C and 3D). Levels of phagosomal PtdIns(3,4)P2, as monitored by EGFP-[2×PH (Tapp1)], were markedly decreased in ΔInpp5e cells, suggesting that Inpp5e metabolizes the PtdIns(3,4,5)P3 to PtdIns(3,4)P2 on the phagosome. PtdIns(3,4)P2 can be hydrolysed to PtdIns3P by PI 4′-phosphatases. Thus, we next monitored PtdIns3P using EGFP-[3×FYVE (EEA1)]. The accumulation of PtdIns3P on the phagosome was extremely reduced in the shInpp5e cells (Figures 3E and 3F).
Phagosomal levels of PtdIns(3,4,5)
P3, PtdIns(3,4) P2 and PtdIns3 P in shInpp5e cells
Inpp5e is required for the recruitment of Rab5
It has been reported that at most 30% of the phagosomal PtdIns3P is produced from PtdIns(3,4,5)P3 through successive conversion from PtdIns(3,4)P2, whereas the rest is produced from PtdIns [22,27]. From this point of view, it is questionable whether the marked attenuation of PtdIns3P accumulation in the shInpp5e cells (Figure 3F) can be fully explained by the observed decrease in PtdIns(3,4)P2 (Figure 3D). Because the activity of Vps34, which produces PtdIns3P by phosphorylating the 3′-position of PtdIns, is regulated by Rab5, we next examined whether the recruitment of Rab5 to the phagosome was altered in the shInpp5e cells. In Figure 4, the cells were transfected with EGFP–Rab5b and then incubated with IgG–RBCs. In the control cells, EGFP–Rab5b accumulated around the engulfed RBCs for approximately 10 min (Figures 4A and 4B). In the shInpp5e cells, this accumulation was remarkably decreased (Figure 4A and 4B). Cells lacking SHIP1, another PI 5′-phosphatase, showed no changes in Rab5 recruitment (Figures 4A and 4B), suggesting a specific function of Inpp5e in Rab5 recruitment. To confirm the role of Rab5 in the impaired accumulation of PtdIns3P in the shInpp5e cells (Figure 3F), the cells were transfected with a constitutively active form of Rab5b (Q79L) and examined for PtdIns3P accumulation. Active Rab5 induced a long-lasting accumulation of PtdIns3P on the phagosome in both the control and the shInpp5e cells (Figure 4C and Supplementary Figure S2). These findings suggest that the failure of Rab5b recruitment to the phagosome in the shInpp5e cells resulted in the impaired accumulation of PtdIns3P.
Decreased recruitment of Rab5b to the phagosome in shInpp5e cells
Inpp5e regulates the recruitment of Rab20
Rab20 has been shown to associate with Inpp5e by yeast two-hybrid screening . In addition, a recent report indicated that Rab20 associates with the phagosome and regulates its maturation . Thus, we considered that Inpp5e may regulate PI metabolism and phagosome maturation through its effect on Rab20. In Figure 5A, wild-type RAW264.7 cells were transfected with mCherry-Inpp5e and EGFP–Rab20. Accumulation of mCherry-Inpp5e on the phagosome was observed, as previously reported . EGFP–Rab20 localized to perinuclear compartments, as well as phagosomes, and the merged image thus shows the co-localization of Inpp5e and Rab20 on phagosomal membranes (Figure 5A). To test whether the localization of Rab20 at the phagosome depends on Inpp5e, EGFP–Rab20 was transfected into the control and the shInpp5e cells and its association with the phagosomal membrane was monitored (Figures 5B and 5C). In the control cells, EGFP–Rab20 resided on the phagosome for at least 20 min after its first appearance (Figure 5C). In the shInpp5e cells, by contrast, EGFP–Rab20 appeared on the phagosome as fast as in control cells, but it disappeared rapidly within 10 min (Figure 5C). Thus, we speculate that Inpp5e plays a role in tethering Rab20 to the phagosome membrane. We finally investigated the affect of Rab20 over-expression on PtdIns3P metabolism. In Figure 6, wild-type RAW264.7 cells were transfected with GFP or GFP–Rab20 together with mCherry-[3×FYVE (EEA1)]. In the GFP-transfected control cells, the PtdIns3P present around the engulfed RBCs, as detected by mCherry-[3×FYVE (EEA1)], disappeared within 10 min after its first appearance (Figure 6A, see also Figure 3E). Over-expression of Rab20 markedly increased the duration of its residence (Figure 6B). Thus, Inpp5e appears to increase the association of Rab20 with the phagosome, thereby increasing phagosomal levels of PtdIns3P.
Co-localization of Inpp5e and Rab20 at the phagosome
Persistence of phagosomal PtdIns3
P in Rab20-transfected cells
In the present study, we reveal the first observation of a functional relationship between Inpp5e and Rab20 in intact cells. We observed that Inpp5e and Rab20 co-localized on the phagosome (Figure 5A). Both the Rab20 association (Figure 5) and the PtdIns3P accumulation (Figures 3E and 3F) on the phagosome were markedly decreased in the shInpp5e cells.
Inpp5e is a PI 5′-phosphatase that can metabolize PtdIns(3,4,5)P3 to produce PtdIns(3,4)P2. Because PtdIns(3,4)P2 is metabolized to PtdIns3P by PI 4′-phosphatases, one possible cause of the decreased PtdIns3P in the shInpp5e cells is the decreased production of PtdIns(3,4)P2. In agreement with this prospect, the accumulation of PtdIns(3,4)P2 was decreased in the shInpp5e cells (Figures 3C and 3D). However, this decrease cannot fully explain the marked (more than 50%) decrease in the PtdIns3P accumulation because the successive conversion pathway has been reported to produce at most 30% of the phagosomal PtdIns3P. Vps34 is a class III PI 3′-kinase that phosphorylates PtdIns to produce PtdIns3P. A major role has been described for this enzyme in phagosome maturation. Thus, it is intriguing to consider that Inpp5e influences PtdIns3P accumulation by modulating Vps34 activity. We observed that the association of Rab5 and Rab20 with the phagosome was reduced in the shInpp5e cells (Figures 4 and 5). Rab5 is known to activate Vps34 in its GTP-bound form. Rab20 was recently shown to interact with Rabex5, a guanine-nucleotide-exchange factor (GEF) for Rab5 . Thus, Inpp5e may maintain Vps34 activity and enhance PtdIns3P production by tethering Rab5, Rab20 and Rabex5 to the phagosome.
The direct association of Rab20 with Inpp5e has been clearly demonstrated by the yeast two-hybrid system . However, it was not clear whether this direct binding could fully explain the Rab20 association with the phagosome in intact cells. We observed that the association of Inpp5e with the phagosome is transient. Inpp5e localized to the phagocytic cup and the early phagosome (Figure 5) but not the late phagosome. On the other hand, previous studies have shown that Rab20 persists on the phagosome with Rab5, LAMP1 and Rab7, indicating its long-term presence throughout the early to late phagosome stages . Inpp5e is thought to be necessary for Rab20 retention in early stages, but other proteins, such as Rab5 and Rabex5, are important at later stages.
Numerous PI 5′-phosphatases, including SHIP1, SHIP2, OCRL1, Inpp5b and Inpp5e, have been implicated in the regulation of phagocytosis [17–21,26]. Among them, SHIP1 and SHIP2 are structurally related. These enzymes contain an SH2 domain and associate with the ITIM (immunoreceptor tyrosine-based inhibition motif) of immune receptors . Interestingly, a deficiency in just one of these enzymes is sufficient to increase phagocytosis, suggesting that their functions are not fully redundant . OCRL1 and Inpp5b have a catalytically inactive RhoGAP domain and an ASH domain that binds to microtubules . These enzymes are different from SHIPs in that they play a role in the elimination of the PtdIns(4,5)P2 and have been implicated in clathrin-dependent endocytosis . They have also been reported to be recruited to the phagosome by Rab5 and to attenuate Akt signalling at the phagosome . However, it remains unclear whether deficiency in either OCRL1 or Inpp5b increases the uptake of solid particles. In the present study, we observed that the absence of another PI 5′-phosphatase, Inpp5e, increases the uptake of RBCs and zymosan particles (Figure 1). Our results suggest that Inpp5e has a specific function that is not shared by the other 5′-phosphatases.
We observed that the knockdown of the Inpp5e decreased Rab20 levels at the phagosome (Figure 5), which may be related to the accelerated acidification of the phagosome (Figure 2). This idea is in agreement with a previous study showing that acidification was inhibited by the wild-type Rab20 but increased by the dominant negative mutant or shRNA targeting Rab20 . In contrast, it has also been reported that expression of a dominant negative Rab20 mutant (T19N) arrests phagosome acidification in the macrophages [30,31]. Regarding this apparent conflict, the Pei et al.  stated that the Rab20 and Rab20T12N expression had no effect at later time-points, i.e. after 2 h of internalization. In agreement with this statement, the shInpp5e knockdown, which increased the phagosome acidification when determined 10–30 min after the addition of target particles (Figure 2B), had no effect after 2 h (Figure 2D).
A previous paper indicated that Rab20 plays a role in the early delay of phagosome maturation . The authors showed that interferon γ (IFNγ) stimulates the association of Rab20 with early phagosomes and delays phagosome maturation, which may be important for proper antigen cross-presentation in macrophages. Thus, it will be intriguing to examine the possible involvement of Inpp5e in the IFNγ-induced early delay of phagosome maturation.
We thank Dr Sasaki (Akita University) for EGFP-[3×FYVE (EEA1)], EGFP-[2×PH (Tapp1)], dsRed-Rab5b and dsRed-Rab5b(Q79L) and Dr. Fukuda (Tohoku University) for EGFP-Rab20.
Tomohiro Segawa and Kaoru Hazeki designed the experiments, performed most of the experiments and wrote the manuscript. Kiyomi Nigorikawa discussed the manuscript. Shin Morioka performed the experiments. Ying Guo, Shunsuke Takasuga and Ken Asanuma prepared the plasmids. Osamu Hazeki designed the experiments, critically analysed the data and wrote the manuscript.
This work was supported by Grant-in-Aid (Kakenhi) [grant numbers 23590078, 22114008 and 22590065] from the Japan Society for the Promotion of Science.