Phospholipase A2IVα Regulates Phagocytosis Independent of Its Enzymatic Activity*

Background: The mechanistic regulation of phagocytosis by phospholipase A2 (PLA2) remains to be defined. Results: A specific PLA2 isoform is activated during phagocytosis, translocates to the plasma membrane, and directly participates in phagosome formation, without involving its enzymatic activity. Conclusion: PLA2IVα regulates phagocytosis via a novel mechanism that requires membrane binding of its C2 domain. Significance: PLA2IVα mediates a novel mechanism of phagocytosis regulation. Group IVα phospholipase A2 (PLA2IVα) is a lipolytic enzyme that catalyzes the hydrolysis of membrane phospholipids to generate precursors of potent inflammatory lipid mediators. Here, the role of PLA2IVα in Fc receptor (FcR)-mediated phagocytosis was investigated, demonstrating that PLA2IVα is selectively activated upon FcR-mediated phagocytosis in macrophages and that it rapidly translocates to the site of the nascent phagosome. Moreover, pharmacological inhibition of PLA2IVα by pyrrophenone reduces particle internalization by up to 50%. In parallel, fibroblasts from PLA2IVα knock-out mice overexpressing FcγRIIA and able to internalize IgG-opsonized beads show 50% lower phagocytosis, compared with wild-type cells, and transfection of PLA2IVα fully recovers this impaired function. Interestingly, transfection of the catalytically inactive deleted PLA2IVα mutant (PLA2IVα(1–525)) and point mutant (PLA2IVα-S228C) also promotes recovery of this impaired function. Finally, transfection of the PLA2IVα C2 domain (which is directly involved in PLA2IVα membrane binding), but not of PLA2IVα-D43N (which cannot bind to membranes), rescues FcR-mediated phagocytosis. These data unveil a new mechanism of action for PLA2IVα, which demonstrates that the membrane binding, and not the enzymatic activity, is required for PLA2IVα modulation of FcR-mediated phagocytosis.

The phospholipases A 2 (PLA 2 s) 5 are lipolytic enzymes that catalyze hydrolysis of the sn-2 acyl bond of membrane phospholipids to generate free fatty acids and lysophospholipids (1,2). As the fatty acids released include arachidonic acid, PLA 2 has a critical role during immune responses through generation of the precursors of the eicosanoids, including the leukotrienes, prostaglandins, and thromboxanes, which are potent inflammatory mediators (3,4). At the same time, PLA 2 generates lysolipids as the ether-linked lysophospholipids, which are precursors of the inflammatory mediator platelet-activating factor.
The PLA 2 family includes several relatively unrelated proteins, among which are the following: the low molecular weight secreted enzymes (sPLA 2 s), which have been implicated in eicosanoid generation, inflammation, host defense, and atherosclerosis (5); the cytosolic Ca 2ϩ -independent group VI PLA 2 s, which have been reported to have roles in phospholipid homeostasis and membrane remodeling (6); and the group IV cytosolic PLA 2 s, which include six high molecular mass enzymes in Mus musculus (as PLA 2 IV␣ to PLA 2 IV), with three newly cloned isoforms (7). PLA 2 IV␣ has an essential role in initiation of the arachidonate pathway, and it is mainly involved in inflammation, intestinal ulceration, acute lung injury, anaphylaxis, and parturition, as shown through the PLA 2 IV␣ knockout mouse approach (8 -10). PLA 2 IV␣ activity requires up to micromolar concentrations of intracellular Ca 2ϩ to promote the binding of two calcium ions to its N-terminal C2 domain, thus inducing a conformational change and translocation of PLA 2 IV␣ from the cytosol to the cell membranes, where it can then access its substrates (11). The initial theory of PLA 2 IV␣ membrane translocation was based on hydrophobic interactions of its C2 domain with phosphatidylcholine-enriched membranes in response to an intracellular Ca 2ϩ increase (12). However, this has been revised through more recent studies that have shown that the cationic ␤-groove of the C2 domain can bind to ceramide 1-phosphate on internal membranes (13). Furthermore, the cationic cluster of the PLA 2 IV␣ catalytic domain (Lys-488, Lys-541, Lys-543, and Lys-544) can bind to anionic phospholipids, although the requirement for this interaction for membrane translocation or regulation of enzymatic activity is still under debate (14,15). The majority of these studies have documented translocation of PLA 2 IV␣ to internal cell membranes, such as the nuclear envelope, Golgi complex, and endoplasmic reticulum, whereby PLA 2 IV␣ contributes to the structure and function of these cell compartments (16 -18). Reports have also shown PLA 2 IV␣ translocation to the plasma membrane (19,20).
The role of the phosphorylation of PLA 2 IV␣ in induction of membrane translocation is not fully defined, although a requirement for PLA 2 IV␣ phosphorylation for its full enzymatic activation has been reported by different groups (15,21,22). Serine phosphorylation on PLA 2 IV␣ (Ser-505 or Ser-727) is mediated mainly by the mitogen-activated protein kinases (MAPKs) ERK1/2, and by the stress kinases p38 and JNK, and this has been shown to increase the intrinsic enzymatic activity of PLA 2 IV␣ (23,24).
In addition, although phosphatidylcholine is generally considered to be the PLA 2 IV␣ substrate (25), the activity of PLA 2 IV␣ does not show selectivity for the polar head group of any specific phospholipid substrate, with the only requirement being arachidonic acid in the sn-2 position (26). Indeed, PLA 2 IV␣ has been shown to be the specific enzyme that produces glycerophosphoinositols from the membrane phosphoinositides upon hormonal stimulation and oncogenic transformation (24,27,28).
The PLA 2 s have been reported to participate in immune cell functions also through the regulation of phagocytosis, as initiated by interactions of ligand-coated particles (particles opsonized with immunoglobulins or complement) with specific receptors expressed on the surface of phagocytic cells (Fc receptors (FcRs) or complement receptors) (29). Phagocytosis induces recruitment of a variety of signaling enzymes and adaptor proteins to the site of particle ingestion, thus resulting in a multifactorial process that triggers the reorganization of the actin cytoskeleton and internalization of the bound particles (30). Moreover, as well as PLA 2 involvement, the signaling pathway activated during phagocytosis varies according to cell type, the nature of the stimulus, the number and type of receptors involved, and the ability of certain microorganisms to subvert the phagocytic process (19,(31)(32)(33)(34)(35).
Here, we demonstrate that among the early events of macrophage activation triggered by FcR-mediated phagocytosis, there is the selective activation of PLA 2 IV␣, which is shown to translocate to the site of nascent phagosomes and to participate directly in the modulation of the phagocytic process, independent of its enzymatic activity.
Cell Transfection and Immunoblotting-Immortalized mouse lung fibroblasts (IMLFs) (39) were transfected with the Lipofectamine-PLUS reagent (Invitrogen), according to the manufacturer's instructions. To evaluate the overexpression of GFP-tagged PLA 2 mutants in IMLF cells, 100 g of total cell lysates were blotted with an anti-GFP antibody (Abcam, Cambridge, UK). The activation status of ERK1/2, p38, and JNK was determined as described previously (40). For PLA 2 IV␣ quantification, total cell lysates together with the recombinant PLA 2 IV␣ protein purified from SF9 cells (for methods, see Ref. 24) were blotted with an anti-PLA 2 IV␣ antibody.
Arachidonic Acid and GroPIns Analyses-These procedures were performed as reported in Refs. 40 -42. For stimulation by FcR-mediated phagocytosis, after overnight labeling, the cells were incubated for a further 2 h in serum-free DMEM plus 0. 25  Phagocytosis Assay-Three m latex beads were opsonized with 1 mg/ml rabbit IgG for 2 h at 37°C in phosphate-buffered saline (PBS) and then with 5 ng of Alexa546-or Alexa488-conjugated anti-rabbit IgGs for a further 1 h; excess IgGs were extensively washed out with PBS, and the beads were resuspended in DMEM.
The cells were serum-starved for 2 h and preincubated or not for 15 min with the inhibitors. The opsonized beads (100 beads/ cell) or Texas Red-zymosan A particles (100 particles/cell) were then added to the cells, incubated for 30 min at 4°C, washed with PBS, and incubated in DMEM at 37°C. Time course experiments of phagocytosis demonstrated that the plateau is reached after 60 min, and 30 min was used in all of the experiments shown. BAPTA-AM addition started 5 min before the incubation at 37°C to avoid cell detachment.
Phagocytosis was measured by two separate assays. For the morphological analysis, upon completion of phagocytosis the cells were fixed with 4% paraformaldehyde, washed twice in PBS, and permeabilized with 0.5% BSA, 50 mM NH 4 Cl, 0.05% saponin in PBS, pH 7.4, and incubated with 0.1% trypan blue. The trypan blue was allowed to freely diffuse into the cells and to counter-stain the internalized beads (visualized only by phase-contrast microscopy), so after trypan blue wash-out, the external beads were still fluorescent (43,44). Excess dye was removed (two washes with PBS), and phagocytosis was visualized using fluorescence microscopy. On each coverslip, 80 or 180 Raw264.7 cells or transfected IMLFs, respectively, were scored, and ϳ350 beads per sample were analyzed. Phagocytosis was quantified as the phagocytic index, the number of ingested beads/total number of cells analyzed.
Time-lapse images were obtained using a Zeiss LSM510 inverted confocal microscope system (Carl Zeiss, Gottingen, Germany). Images were acquired at 30-s intervals for a total time of 30 min.
In the second protocol, phagocytosis was assessed by fluorescence-activated cell sorter (FACS) analysis (45,46). Fluorescence intensity was recorded before (binding) and after 0.1% trypan blue addition (phagocytosis) to quench the fluorescence of the extracellular bound beads. Typically, 2 ϫ 10 5 events/ sample were counted, and the percentages of binding, as the numbers of the fluorescent-positive events (number of cells with beads), were expressed as percentages of total events (number of cells analyzed), and the percentages of phagocytic cells, as the numbers of fluorescent-positive events after trypan blue addition (number of cells with phagocytosed beads), were expressed as percentages of total events.
For the biochemical separation of phagosomes, the procedure was performed as in Ref. 47, with minor modifications. Briefly, phagocytosis was performed as described previously using tosyl-activated M-280 Dynabeads (2.8 m diameter, Invitrogen). At the indicated times, phagocytosing Raw264.7 cells were scraped into ice-cold lysis buffer (25 mM Tris-HCl, pH 7.4, 0.25 M sucrose, 2.5 mM DTT, 2.5 mM EDTA, 20 mM NaF, 1 mM sodium orthovanadate, 1 mM p-nitrophenyl phosphate, 5 mM imidazole, 50 mM o-phenanthroline, 2 mM pepstatin, and 1 mM phenylmethylsulfonyl fluoride) and sonicated (two times, 20-s pulses). The beads were isolated on ice using a magnet; the unbound material was removed, and the beads were washed with lysis buffer and then solubilized in Laemmli sample buffer (bead-associated membranes). The unbound material was centrifuged 45 min at 100,000 ϫ g at 4°C; the supernatant (cytosol) was recovered; the pellet was washed with lysis buffer and centrifuged again, and then the final pellet was resuspended in ice-cold lysis buffer and sonicated (one time, 20-s pulse) (nonbead associated membranes). The PLA 2 IV␣ levels in the different fractions were analyzed by Western blotting with the anti-PLA 2 IV␣ rabbit antibody and normalized by the presence of rabbit IgGs (used in the opsonization step) with an HRP-conjugated anti-rabbit antibody (for the bead-associated membranes), by the presence of GM130 with a mouse monoclonal anti-GM130 antibody (Transduction Laboratories, Lexington, KY) for the non-bead associated membranes, and by the presence of GAPDH with a mouse anti-lapine GAPDH antibody (AbD Serotec, Pucheim, Germany).
Phagocytosis of Listeria monocytogenes was quantified using a gentamycin-protection assay, as reported previously (48).
Statistical Analysis-All of the data are expressed as means Ϯ S.E. (error bars). Significance was calculated using paired, twotailed Student's t tests, with p values Ͻ0.05 considered significant.

FcR-mediated Phagocytosis Selectively Stimulates PLA 2 IV␣-
Macrophage responses to E. coli lipopolysaccharide (LPS) include the well documented release of arachidonic acid, which is considered to be the link between PLA 2 activity and the inflammatory response (3). Thus, the involvement of PLA 2 in FcR-mediated phagocytosis was investigated here in parallel with LPS stimulation in Raw264.7 cells. PLA 2 activity, as assessed by arachidonic acid release, was stimulated to similar extents by FcR-mediated phagocytosis of IgG-opsonized latex beads (3 m) and treatment with 20 g/ml LPS, with 1.1-and 0.7-fold increases over basal levels, respectively (Fig. 1A). To identify which specific PLA 2 is activated, a panel of inhibitors was used that have specificities for the different isoforms of the PLA 2 superfamily. Pretreatment of Raw264.7 cells with 0.5 M pyrrophenone, a selective and reversible inhibitor of the class IV PLA 2 s (38), nearly completely abolished the effects of both of these stimuli (Fig. 1A). In contrast, 50 M sPLA 2 IIA inhibitor-I, which inhibits class II and V secretory PLA 2 s (49), and 1 M bromoenol lactone, which inhibits group VI PLA 2 s (50), showed no inhibition of arachidonic acid release stimulated by FcR-mediated phagocytosis ( Table 1).
Previous studies from our laboratory characterized PLA 2 IV␣ as the specific PLA 2 isoform that mediates GroPIns production from the membrane phosphatidylinositol, through the transient formation of lysophosphatidylinositol (lyso-PtdIns) (24). Thus, the selective activation of PLA 2 IV␣ by FcR-mediated phagocytosis can be further monitored through the cellular levels of GroPIns, as quantified by HPLC analysis after equilibrium labeling with myo-[ 3 H]inositol (see under "Experimental Procedures"). At steady state, GroPIns levels are ϳ120 M in Raw264.7 cells, as assessed using mass spectrometry (for methods, see Refs. 42,51). Indeed, here the FcR-mediated phagocytosis and LPS stimulation specified above also increased the intracellular levels of GroPIns (0.5-and 1.2-fold over basal, respectively; Fig. 1B). This GroPIns production was PLA 2 IV␣dependent, as it was completely inhibited by treatment with 0.5 M pyrrophenone (Fig. 1B); in contrast, the inhibitors of secretory PLA 2 s and group VI PLA 2 s did not inhibit this GroPIns production ( Table 1).
As PLA 2 IV␣ phosphorylation has been associated with its activation (52), the levels of phosphorylated PLA 2 IV␣ were followed by Western blotting of Raw264.7 cell lysates. As shown in Fig. 1C, phosphorylated PLA 2 IV␣ can be distinguished from its nonphosphorylated form as it runs as a distinct upper band (24). Moreover, an anti-phospho-PLA 2 IV␣ antibody was used to specifically monitor phosphorylation of PLA 2 IV␣ on Ser-505. These approaches showed that under basal conditions, the level of phosphorylated PLA 2 IV␣ was higher than the nonphosphorylated form, and upon FcR-mediated phagocytosis and LPS stimulation, PLA 2 IV␣ was completely shifted to the higher molecular weight, and there was a corresponding ϳ70% increase in Ser-505 phosphorylation (Fig. 1C). Among the kinases known to regulate PLA 2 IV␣, both the MAPKs ERK1/2 and the stress kinases p38 and JNK were phosphorylated by FcR-mediated phagocytosis and LPS stimulation and consequently activated (Fig. 1C). In addition, inhibitors specific for these three kinases (U0126 for ERK1/2; SB203580 for p38; and SP600125 for JNK; all at 10 M) reduced the phagocytosis-induced PLA 2 IV␣ phosphorylation (Fig. 1D) and blocked the phagocytosis-induced arachidonic acid release (Ͼ90% inhibition) ( Table 1), which indicates that these MAPKs and stress kinases are upstream of PLA 2 IV␣ activity.
These analyses of arachidonic acid and GroPIns production demonstrate that during FcR-mediated phagocytosis in Raw264.7 cells PLA 2 IV␣ is selectively activated and regulated by multiple kinases. In addition, quantification of PLA 2 IV␣ protein levels demonstrated that PLA 2 IV␣ is an abundant pro-tein in these cells, with ϳ7 ng/g total cell lysate protein (Fig.  1E).
PLA 2 IV␣ Participates in FcR-mediated Phagocytosis-To determine the role of PLA 2 IV␣ in the uptake of IgG-opsonized particles, we first analyzed this in Raw264.7 professional phagocytic cells and then in fibroblasts modified for phagocytosis. This was carried out by immunofluorescence, using fluorescence microscopy for morphological analyses and FACS for automated quantification. These two analytical procedures produced similar results.
Pretreatment of the Raw264.7 cells with 0.5 M pyrrophenone induced a 50% reduction in opsonized particle uptake, which was seen as a reduced phagocytic index and percent of phagocytosis ( Fig. 2A and Table 2), with no interference in particle binding to the cells (Table 2). This inhibition by pyrrophenone was, however, not as large as that of 10 M cytochalasin D, a potent inhibitor of actin polymerization, and consequently of phagocytosis ( Fig. 2A and Table 2), but it was retained during the overall time course of the phagocytosis analysis (from 0 to 120 min). In contrast, treatments with bromoenol lactone and sPLA 2 IIA inhibitor-I were ineffective (Table 2). In parallel with this pharmacological approach, we also attempted to silence PLA 2 IV␣ expression using both small interfering siRNAs and vectors coding for short hairpin shRNAs; however, in both of these cases, the reduction in PLA 2 IV␣ protein levels reached was never sufficient to perform biochemical assays (data not shown).
First, we verified the endogenous expression levels of PLA 2 IV␣ in the IMLF ϩ/ϩ cells by Western blotting, along with the equal overexpression levels of the Fc␥R both without and

H͔arachidonic acid release and ͓ 3 H͔GroPIns production induced by FcR-mediated phagocytosis and LPS stimulation in Raw264.7 cells
Raw264.7 cells without or with 15-min preincubations with the different inhibitors, as indicated, were left untreated or were stimulated by FcR-mediated phagocytosis or with 20 g/ml LPS for a further 30 and 45 min, respectively, at 37°C (see "Experimental Procedures"). ͓ 3 H͔Arachidonic acid release as percentage of total ͓ 3 H͔arachidonic acid cell content, with basal release before phagocytosis and LPS stimulation of 0.93 Ϯ 0.06 and 1.12 Ϯ 0.07% of total cell content, respectively. Intracellular ͓ 3 H͔GroPIns production as a percentage of total water-soluble ͓ 3 H͔inositol-labeled metabolites, with the basal levels before phagocytosis and LPS stimulation of 9.98% Ϯ 0.60 and 8.12% Ϯ 0.62%, respectively. Data are means Ϯ S.E. of at least four independent experiments, each performed in duplicate. In parallel, under the conditions reported above, lysophosphatidylinositol levels were monitored by TLC (for the method see Ref. 42), with an increase of 0.3-and 0.8-fold over basal for stimulation by FcR-mediated phagocytosis and LPS, respectively, effects that were completely blocked by pyrrophenone treatment. ND means not determined. ͓ 3 H͔arachidonic acid release (% total cell content) with the PLA 2 IV␣ knock-out (Fig. 2B). Equal Fc␥R expression was also confirmed by the similar levels of binding of IgG-opsonized beads with these IMLF ϩ/ϩ and IMLF Ϫ/Ϫ cell lines ( Table 2). In the IMLF Ϫ/Ϫ cells, Fc␥R-mediated phagocytosis was 50% of that seen for the IMLF ϩ/ϩ cells and was insensitive to treatment with 0.5 M pyrrophenone (Table 2). Instead, in the IMLF ϩ/ϩ cells, 0.5 M pyrrophenone reduced the Fc␥Rmediated phagocytosis by 50%, thus corresponding to the baseline phagocytosis of the IMLF Ϫ/Ϫ cells (Table 2). This inhibition of phagocytosis was similar to that seen in the Raw264.7 cells ( Fig. 2A and Table 2). In addition, transfection of PLA 2 IV␣ in IMLF Ϫ/Ϫ cells promoted increased phagocytosis, which indeed reached levels comparable with the wild-type IMLF ϩ/ϩ cells (Table 2). With these IMLF Ϫ/Ϫ cells co-transfected with Fc␥R and GFP-PLA 2 IV␣, we also took advantage of the possibility to monitor the PLA 2 IV␣ cellular localization during Fc␥R-medi-ated phagocytosis using confocal microscopy. Here, time-lapse video microscopy revealed that GFP-PLA 2 IV␣ was initially diffuse in the cytosol. After the promotion of phagocytosis through binding of the opsonized particles to the Fc␥R in IMLF Ϫ/Ϫ cells, the GFP-PLA 2 IV␣ translocated to the nascent phagosomes at the level of the phagocytic cup; it then completely surrounded the particles as they were endocytosed. At the end of the process, GFP-PLA 2 IV␣ was again diffuse in the cytosol (Fig. 2C and supplemental Movies S1.A and S1.B). In the control IMLF Ϫ/Ϫ cells transfected for GFP alone, this GFP remained diffuse in the cytosol, without changing its localization, during the entire process of phagocytosis (supplemental Movie S2). Similar in vivo video microscopy studies in Raw264.7 cells could not be performed due to the abundant endogenous PLA 2 IV␣, which interferes with the correct localization of overexpressed GFP-PLA 2 IV␣. Instead, in these Raw264.7 cells the association of endogenous PLA 2 IV␣ with the phagosomal membranes was directly evaluated using a biochemical approach. During synchronized phagocytosis, PLA 2 IV␣ was enriched in the membrane fractions in a time-dependent manner (from 1 to 20 min), with a parallel decrease of its levels in the cytosolic fraction (not shown).
Because Ca 2ϩ is considered the main regulator of PLA 2 IV␣ membrane translocation (11), the effects on PLA 2 IV␣ localization of an intracellular Ca 2ϩ chelator (10 M BAPTA-AM) and also of pyrrophenone were monitored. By Western blotting, both PLA 2 IV␣ phagosome translocation (bead associated membranes in Fig. 2D) and total membrane translocation (nonbead associated membranes, Fig. 2D) were analyzed in phagocytosing Raw264.7 cells. Treatment with BAPTA-AM and pyrrophenone inhibited PLA 2 IV␣ membrane translocation by ϳ80 and 50%, respectively (Fig. 2D). A similar inhibition by BAPTA-AM and pyrrophenone on PLA 2 IV␣ membrane association was confirmed also by confocal microscopy in IMLF Ϫ/Ϫ cells expressing GFP-PLA 2 IV␣, and whenever a phagosome was present in cells treated with both agents, no enrichment of PLA 2 IV␣ was observed at the phagosomal level (Movies S3A-C). The effects of BAPTA-AM on PLA 2 IV␣ membrane trans-location was seen as 50% inhibition of FcR-induced phagocytosis without any significant effects on bead binding (Table 2).
Altogether, these data indicate that PLA 2 IV␣ is a component of the phagocytic machinery that is formed consequent to FcR activation, thus regulating opsonized particle uptake. However, pyrrophenone inhibition of both PLA 2 IV␣ enzymatic activity and membrane translocation questioned the possible mechanism of action of PLA 2 IV␣ in the regulation of the phagocytosis process.
PLA 2 IV␣ Regulates FcR-mediated Phagocytosis via a Nonenzymatic Mechanism-A number of studies have suggested an involvement of PLA 2 s in the process of phagocytosis, through the actions of PLA 2 metabolites on membrane curvature (19,32). Indeed, PLA 2 hydrolysis of the outer leaflet of the membrane bilayer can result in the conversion of cylindrical and cone-shaped phospholipids into inverted cone-shaped lysophospholipids, which can stimulate or initiate membrane bending (16).
The requirement for the enzymatic activity of PLA 2 IV␣ in FcR-mediated phagocytosis was directly assessed using the coexpression of Fc␥R and GFP-tagged PLA 2 IV␣ wild-type or mutated versions (Fig. 3A) in the IMLF Ϫ/Ϫ cells, with the GFP allowing the monitoring of the transfected cells. As verified by Western blotting, under these conditions the overexpression levels of GFP-PLA 2 IV␣ were always lower than the endogenous levels in Raw264.7 cells (Fig. 3B and its legend). Unexpectedly, overexpression of the GFP-PLA 2 IV␣(1-525) deletion mutant of PLA 2 IV␣ (without a complete catalytic domain, and therefore without enzymatic activity; Fig. 3, A and C, and Ref. 36) rescued the impaired phagocytosis of these IMLF Ϫ/Ϫ cells ( Fig.  3D and Table 2).
To investigate whether the membrane binding of PLA 2 IV␣ is sufficient to regulate Fc␥R-mediated phagocytosis even in the absence of PLA 2 IV␣ catalytic activity, another catalytically inactive PLA 2 IV␣ mutant was used: GFP-PLA 2 IV␣-S228C, with a single point mutation in the catalytic domain (55), and the GFP-PLA 2 IV␣-C2 domain, because membrane localization of PLA 2 IV␣ is mainly driven by the insertion of its N-terminal C2 domain into the lipid bilayer (Fig. 3A) (11). When this GFP-PLA 2 IV␣-S228C mutant or the PLA 2 IV␣-C2 domain was cotransfected with Fc␥R in the IMLF Ϫ/Ϫ cells, phagocytosis increased to a level comparable with transfection of the wildtype GFP-PLA 2 IV␣ (Fig. 3, C and D, and Table 2).
To further confirm the necessity for PLA 2 IV␣ membrane binding for rescue of this impaired Fc␥R-mediated phagocytosis, another PLA 2 IV␣ point mutant was used, i.e. GFP-PLA 2 IV␣-D43N, with a single point mutation in the C2 domain that prevents its membrane translocation (Fig. 3, A and C) (12). GFP-PLA 2 IV␣-D43N co-expression with Fc␥R in these IMLF Ϫ/Ϫ cells induced a small increase in the basal phagocytosis index that was not comparable with that of the wild-type PLA 2 IV␣ protein (Fig. 3D). This might be a consequence of the high overexpression levels of this GFP-PLA 2 IV␣-D43N mutant, which would lead to its partial membrane mislocalization. None of these constructs significantly affected the binding of the opsonized particles to the IMLF Ϫ/Ϫ cells (Table 2).
Under these conditions, we also monitored the effects of pyrrophenone treatment on the rescue of FcR-mediated phagocy-

Modulation of FcR-mediated phagocytosis
Cells without or with a 15-min preincubation with different inhibitors, as indicated, were incubated in the absence or presence of IgG-opsonized latex beads for 30 min at 37°C, and both particle binding (FcR binding, before trypan blue addition) and uptake (FcR phagocytosis, after trypan blue addition) were quantified by FACS analysis. For BAPTA-AM, the addition started in the last 5 min of opsonized bead binding. FcR-mediated phagocytosis of untreated Raw264.7 cells (time 0) was 0.65 Ϯ 0.4%. IMLF cells were from PLA 2 IV␣ wild-type (IMLF ϩ/ϩ ) and PLA 2 IV␣ knock-out (IMLF Ϫ/Ϫ ) mice, and they were transfected with Fc␥R and the different constructs, as indicated in the left-hand column. When GFP-tagged constructs were expressed, only transfected cells were analyzed, instead of the total population (see under "Experimental Procedures"). Data are expressed as percentages of total analyzed cells and represent means Ϯ S.E. of at least four independent experiments. a , p Ͻ 0.05; b , p Ͻ 0.02; c , p Ͻ 0.001 (Student's t-test), with respect to their relevant controls. tosis by the different PLA 2 IV␣ mutants in IMLF Ϫ/Ϫ cells. Here, pyrrophenone impaired the recovery of phagocytosis consequent to overexpression of GFP-PLA 2 IV␣-S228C (ϳ50% inhibition) ( Table 2). Instead, pyrrophenone was inactive when the PLA 2 IV␣(1-525) deletion mutant or the PLA 2 IV␣ C2 domain was expressed in the same system (Table 2). Of note, neither of these two proteins contain a complete catalytic domain and hence the PLA 2 IV␣-binding site for pyrrophenone (56).
As an alternative to pyrrophenone, a combination of the inhibitors of the two main upstream kinases of PLA 2 IV␣ activation was used in the Raw264.7 cells (U0126 and SB203580; both at 10 M). Here, both phagocytosis-induced release of arachidonic acid and FcR-mediated phagocytosis were monitored. The PLA 2 IV␣ enzymatic activity was completely inhibited by this treatment (Table 1), whereas FcR-mediated phago-  (18,36,55). The point mutant GFP-PLA 2 IV␣-D43N cannot translocate to membranes under cell stimulation (12). B, representative Western blotting of 2.5-30 and 5-100 g total lysates from Raw264.7 and IMLF Ϫ/Ϫ cells overexpressing GFP-PLA 2 IV␣, respectively. IMLF Ϫ/Ϫ transfection efficiency was 30%, as verified by FACS analysis, independent of transfection construct, and the estimated total protein was ϳ520 and ϳ350 pg/cell for Raw264.7 and IMLF Ϫ/Ϫ cells, respectively. The PLA 2 IV␣ protein content was comparable between 3 g of Raw264.7 cell lysate and 100 g of IMLF Ϫ/Ϫ cell lysate. Thus, assuming uniform overexpression of GFP-PLA 2 IV␣ and taking into account the efficiency of transfection, the protein levels reached in PLA 2 IV␣-transfected IMLF Ϫ/Ϫ cells was one-tenth that of the endogenous level in Raw264.7 cells. C, representative Western blotting of transfection levels of GFP and GFP-PLA 2 IV␣ (wild-type and mutants, as indicated) in IMLF Ϫ/Ϫ cells, followed using an anti-GFP antibody (see "Experimental Procedures"). Data are representative of at least four independent experiments. D, quantification by morphological analysis of particle uptake (phagocytic index) in IMLF Ϫ/Ϫ cells transfected with Fc␥R Ϯ GFP or with different GFP-PLA 2 constructs (as indicated). Data represent means Ϯ S.E. of at least four independent quantifications. E, internalization of L. monocytogenes in Raw264.7 cells. Cells were infected with bacteria (multiplicity of infection of 10) for 1 h at 37°C, and internalization was quantified using the gentamicin protection assay (see "Experimental Procedures") in the absence (Ϫ) and presence (ϩ) of 0.5 M pyrrophenone (30-min pretreatment), and expressed as colony forming units per well. In control proliferation assays, pyrrophenone did not affect L. monocytogenes growth (up to 10 h of pyrrophenone treatment). Data are means Ϯ S.D. of nine independent experiments, each performed in triplicate. *, p Ͻ 0.05 (Student's t test). cytosis was not significantly affected ( Table 2). This further demonstrates that PLA 2 IV␣ modulates the phagocytic process independent of its enzymatic activity also in an endogenous expression system and thus under physiologically relevant conditions. PLA 2 IV␣ Metabolites Do Not Regulate FcR-mediated Phagocytosis-To further exclude a role for the PLA 2 IV␣ catalytic activity in FcR-mediated phagocytosis, some of the PLA 2 IV␣ metabolites that were detected in Raw264.7 cells during phagocytosis were tested and excluded for their involvement in this process ( Table 3). As shown under "FcR-mediated Phagocytosis Selectively Stimulates PLA 2 IV␣," in Raw264.7 cells FcR-mediated phagocytosis can induce the production of GroPIns and its precursor lyso-PtdIns (Table 1 and its legend), which might be a mediator of PLA 2 IV␣ activity on membrane curvature. To define the role of these metabolites in the regulation of the phagocytic process, lyso-PtdIns (10 -30 M) and for completeness also GroPIns, GroPIns4P, and GroPIns4,5P 2 (300, 50, and 50 M, respectively) were added exogenously to Raw264.7 cells, and the uptake of IgG-opsonized beads was monitored. As shown in Table 3, this uptake was insensitive to these lyso-PtdIns and GroPIns treatments; in contrast, it was inhibited by 25 and 30% by GroPIns4P and GroPIns4,5P 2 , respectively.
In addition, the alternative downstream pathways activated by PLA 2 IV␣ that mainly involves arachidonic acid metabolism were analyzed. Blockers of the arachidonate pathway and arachidonic acid derivatives were added to phagocytosing Raw264.7 cells, and they did not significantly affect the uptake of opsonized-particles (Table 3). Altogether, these data demonstrate that it is not the enzymatic activity, but it is the membrane binding of PLA 2 IV␣ that is required for the regulation of FcR-mediated phagocytosis.
PLA 2 IV␣ Involvement in Phagocytosis of Zymosan and Bacteria-In addition to the data shown for the opsonized latex beads, this study of the role of PLA 2 IV␣ was extended to different types of phagocytosis, including the uptake of zymosan particles (derived from the S. cerevisiae cell wall) and of L. monocytogenes. The uptake of zymosan particles was measured by FACS analysis, and the results show that 0.5 M pyrrophenone had little inhibitory effect on the internalization of zymosan particles in Raw264.7 cells (ϳ15% reduction). This indicates that PLA 2 IV␣ has only a marginal role in the pathway triggered by zymosan particle binding to the ␤-glucan receptor and in its consequent internalization in Raw264.7 cells. L. monocytogenes internalization was measured by incubating Raw264.7 cells with the living bacteria, and treatment with 0.5 M pyrrophenone induced a reduction in L. monocytogenes phagocytosis by 40% (Fig. 3E). Thus, we can conclude here that PLA 2 IV␣ is specifically required during FcR-mediated phagocytosis and for internalization of bacteria, such as L. monocytogenes.

DISCUSSION
This study shows the selective phosphorylation, activation, and translocation of PLA 2 IV␣ to the membrane of nascent phagosomes during FcR-mediated phagocytosis. Moreover, PLA 2 IV␣ is shown not only to be part of the FcR-mediated phagocytosis signaling machinery, but also to actively modulate the uptake of IgG-opsonized particles. This was achieved using two alternative approaches, a potent and selective PLA 2 IV␣ inhibitor (pyrrophenone) and cells from PLA 2 IV␣ knock-out mice (IMLF Ϫ/Ϫ cells). Indeed, under both of these conditions, although not completely blocked, FcR-mediated phagocytosis was reduced by 50%. The residual phagocytic activity that is maintained in the absence of PLA 2 IV␣ appears to be due to alternative regulatory pathways, producing a redundancy of the phagocytic processes. As phagocytosis is a central process for the survival of any organism, it can be expected that such an essential process also has "fail-safe" mechanisms (30).
An intriguing part of this study is the demonstration that PLA 2 IV␣ regulates FcR-mediated phagocytosis independent of its enzymatic activity, which highlights a new mechanism of action for PLA 2 IV␣. Thus, the inhibition of FcR-mediated phagocytosis observed using pyrrophenone in Raw264.7 cells is not simply due to inhibition of PLA 2 IV␣ catalytic activity. Here, we demonstrate that the interaction between pyrrophenone and PLA 2 IV␣ generates an allosteric block, with the consequent inhibition of membrane translocation of PLA 2 IV␣ and the consequent impairment of FcR-mediated phagocytosis. Moreover, the main products of PLA 2 IV␣ enzymatic activity (arachidonic acid, arachidonate pathway derivatives, lyso-Pt-dIns, and the glycerophosphoinositols) do not contribute in any positive ways to the modulation of FcR-mediated phagocytosis.

metabolites and arachidonate pathway inhibitors
Raw264.7 cells without or with a 15-min preincubation with different compounds, as indicated, were incubated in the presence of IgG-opsonized latex beads for 30 min at 37°C, and both the particle binding (FcR binding; before trypan blue addition) and uptake (FcR phagocytosis; after trypan blue addition) were quantified by FACS analysis. Data are means Ϯ S.E. of at least four independent experiments. 5-LOX, 5-lipoxygenase; PGE, prostaglandin E; TxB 2 , thromboxane B 2 ; LT, leukotrienes. Through these data, we have thus built a model in which the binding of PLA 2 IV␣ to the plasma membrane lipid bilayer appears sufficient to regulate phagosome formation. In addition, in the IMLF Ϫ/Ϫ cells from the PLA 2 IV␣ knock-out mice, the PLA 2 IV␣ N-terminal C2 domain was identified as the portion of PLA 2 IV␣ that is sufficient to induce a rescue of impaired FcR-mediated phagocytosis. This is in agreement with previous studies, also from our laboratory, that have indicated the C2 domain as the minimal PLA 2 IV␣ portion that can still translocate to cell membranes following increases in intracellular Ca 2ϩ (18).

Condition
Based on the data presented here, we speculate on a new mechanism of action of PLA 2 IV␣, whereby membrane insertion of the PLA 2 IV␣ C2 domain induces perturbation of the membrane phospholipid packing, potentially also generating the membrane bending that is necessary for phagosome formation (57). This mechanism of membrane bending has been typically described for proteins that have lipid-binding domains, and a reference model can be seen with synaptotagmin I (58). Comparative analyses between the C2 domains of PLA 2 IV␣ and synaptotagmin I also show strong similarities, both in terms of their sequences and their tertiary structures, which would indeed be in favor of a similar mechanism of action for PLA 2 IV␣ (59).
The model proposed in this study, which sees a lack of involvement of PLA 2 IV␣ enzymatic activity in FcR-mediated phagocytosis, is supported by data obtained with heterologous expression systems that take advantage of the PLA 2 IV␣ knockout mouse model and also with the professional phagocytosing cell line Raw264.7 cells. The molecular mechanisms through which membrane translocation of the PLA 2 IV␣ C2 domain is sufficient to promote phagocytic cup formation remain to be defined and will be part of further studies.
Because PLA 2 IV␣ metabolite production was monitored during phagocytosis, we propose that PLA 2 IV␣ enzymatic activity is not relevant for the phagocytosis process per se but is required for bacterial killing (through NADPH oxidase activation and the consequent peroxide production) and for proinflammatory responses (through the production of lipid mediators, such as the eicosanoids and platelet-activating factor).