Group V Secretory Phospholipase A2 Translocates to the Phagosome after Zymosan Stimulation of Mouse Peritoneal Macrophages and Regulates Phagocytosis*

We have previously reported that group V secretory phospholipase A2 (sPLA2) amplifies the action of cytosolic phospholipase A2(cPLA2) α in regulating eicosanoid biosynthesis by mouse peritoneal macrophages stimulated with zymosan (Satake, Y., Diaz, B. L., Balestrieri, B., Lam, B. K., Kanaoka, Y., Grusby, M. J., and Arm, J. P. (2004) J. Biol. Chem. 279, 16488-16494). To further understand the role of group V sPLA2, we studied its localization in resting mouse peritoneal macrophages before and after stimulation with zymosan and the effect of deletion of the gene encoding group V sPLA2 on phagocytosis of zymosan. We report that group V sPLA2 is present in the Golgi apparatus and recycling endosome in the juxtanuclear region of resting peritoneal macrophages. Upon ingestion of zymosan by mouse peritoneal macrophages, group V sPLA2 is recruited to the phagosome. There it co-localizes with cPLA 2 α, 5-lipoxygenase, 5-lipoxygenase-activating protein, and leukotriene C4 synthase. Using immunostaining for the cysteinyl leukotrienes in carbodiimide-fixed cells, we show, for the first time, that the phagosome is a site of cysteinyl leukotriene formation. Furthermore, peritoneal macrophages from group V sPLA2-null mice demonstrated a >50% attenuation in phagocytosis of zymosan particles, which was restored by adenoviral expression of group V sPLA2 but IIA not group sPLA2. These data demonstrate that group V sPLA2 contributes to the innate immune response both through regulation of eicosanoid generation in response to a phagocytic stimulus and also as a component of the phagocytic machinery.

Phospholipase A 2 (PLA 2 ) 2 hydrolyzes the sn-2 position of cell membrane phospholipids to release free fatty acids and lysophospholipids. Where the fatty acid is arachidonic acid, it serves as the precursor for biosynthesis of leukotrienes and prostaglandins (1). Free fatty acids also may signal through peroxisome proliferator-activated receptors (2), and lysophospholipids activate cell surface G protein-coupled receptors (3). The family of PLA 2 enzymes includes the cytosolic PLA 2 (cPLA 2 ) iso-forms that utilize a catalytic serine (4,5). cPLA 2 ␣ is critical for the release of arachidonic acid for leukotriene and prostaglandin generation (6,7). The calcium-independent isoforms of PLA 2 , which also use a catalytic serine, have been implicated in cell membrane remodeling (8), spermatogenesis (9), and regulation of store-operated calcium channels (10). There are 10 mammalian secretory PLA 2 enzymes (11)(12)(13)(14) that use a catalytic histidine and that have been implicated in amplifying cPLA 2 ␣dependent eicosanoid generation (15)(16)(17) and in regulating secretory granule exocytosis (18). In certain circumstances, sPLA 2 enzymes may regulate eicosanoid generation independently from cPLA 2 ␣ (19). Certain sPLA 2 isoforms may act as ligands for cell surface receptors (20). However, the roles of the sPLA 2 enzymes have largely been inferred from their biochemical properties, the effects of their overexpression in transfected cells, or the use of pharmacologic inhibitors. To elucidate the biological functions of one of the sPLA 2 enzymes, group V sPLA 2 , we generated mice in which the gene encoding this sPLA 2 was deleted by homologous recombination (15).
Macrophages contribute to the innate immune response to invading microorganisms by engulfing them through the process of phagocytosis (21). Subsequent fusion of lysosomes with the phagosome to form phagolysosomes and generation of reactive oxygen species leads to killing of the ingested pathogen (22). The process of phagocytosis in mouse peritoneal macrophages results in the formation of leukotriene (LT) C 4 and prostaglandin E 2 that initiate and amplify the ensuing inflammatory response (23)(24)(25)(26). The synthesis of eicosanoids in response to phagocytosis of zymosan requires the provision of the substrate, arachidonic acid, by cPLA 2 ␣ (27). Mouse peritoneal macrophages derived from group V sPLA 2 -null mice have a 50% attenuation of eicosanoid generation in response to zymosan compared with wild-type macrophages, consistent with a role for group V sPLA 2 in amplifying the action of cPLA 2 ␣ (15). Further, the early phase of plasma extravasation in response to zymosan, which is dependent on cysteinyl leukotrienes acting at the type 1 CysLT receptor, is markedly attenuated in Group V sPLA 2 -null mice (15). Thus, group V sPLA 2 contributes to the innate immune response through the regulation of eicosanoid generation in response to a phagocytic stimulus, thereby regulating the ensuing inflammatory response. We now report that group V sPLA 2 translocates to the phagosome upon zymosan stimulation of mouse peritoneal macrophages where it co-localizes with the other enzymes of leukotriene biosynthesis. We further show that group V sPLA 2 regulates phagocytosis of zymosan independent of its effects on eicosanoid generation.
Cell Culture-The peritoneal cavities of mice were flushed with 5 ml of ice-cold PCG buffer (25 mM Pipes, 110 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 g/liter glucose, pH 7.4). After washing, 3.3 ϫ 10 5 cells were plated on sterile glass coverslips in 24-well tissue culture plates in 500 l of PCG buffer containing 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin and incubated overnight at 37°C with 5% CO 2 . Nonadherent cells were removed by washing three times with PCG buffer.
Synchronized Phagocytosis-After removing nonadherent cells, the macrophages were incubated for another hour at 37°C. The cells were then washed once with cold PCG buffer and cooled on ice for 5 min before the addition of zymosan (10 particles/cell) for 15 min on ice. The cells were then warmed by adding warm PCG buffer containing 10% fetal bovine serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin and incubated at 37°C for the indicated period of time (30). In all of the experiments, nonopsonized zymosan was used.
Immunofluorescence Analysis-Macrophages were fixed with 2% paraformaldehyde in phosphate-buffered saline for 15 min at room temperature, washed once in Hanks' balanced salt solution without Mg 2ϩ or Ca 2ϩ containing 0.1% bovine serum albumin (HBA), permeabilized with 0.025% saponin in phosphate-buffered saline for 10 min at room temperature, and washed twice with HBA. The cells were then blocked in HBA containing 5% normal donkey serum (blocking buffer) for 1 h at room temperature. The macrophages were incubated with 5 g/ml rabbit polyclonal antibody to group V sPLA 2 (28) and mouse mAb GS28 (2 g/ml), mouse mAb against Syntaxin 6 (5 g/ml), mouse mAb against the transferrin receptor (2.5 g/ml), or rat mAb against LAMP-1 (2.5 g/ml) in blocking buffer for 2 h at room temperature. Negative control cells were instead incubated for 2 h with rabbit IgG, mouse IgG, or rat IgG (Jackson Immunoresearch) in place of specific primary antibodies. After 2 h, the cells were washed extensively with HBA and incubated for 1 h at room temperature with FITC-conjugated donkey anti-rabbit IgG (1:400) and Cy3-conjugated donkey anti-mouse IgG (1:200) or Cy3-conjugated donkey anti-rat IgG (1:200). The cells were washed five times with HBA, mounted in Vectashield TM mounting medium (Vector Laboratories, Burlingame, CA), and imaged using a Nikon C1 confocal system combined with an Eclipse TE2000-U inverted microscope with 60ϫ oil PlanApo NA 1.4 objective. Conventional Nomarski differential contrast images were taken using a SPOT-RT digital camera. The data were analyzed by Photoshop image software. To determine the localization of the enzymes of leukotriene biosynthesis after phagocytosis, macrophages were fixed and permeabilized as described above before and after incubation with zymosan and incubated with antibodies directed to group V sPLA 2 , cPLA 2 ␣, 5-lipoxygenase, FLAP, or LTC 4 synthase. After washing, the cells were incubated with FITC-or Cy3-conjugated donkey anti-rabbit IgG (1:400) and imaged using a Nikon Eclipse TE2000-U microscope coupled with a Spot-RT Digital Camera.
Intracellular Localization of LTC 4 after Zymosan Phagocytosis-To detect CysLTs at the site of their generation, we used a modification of a previous protocol (31). Briefly, macrophages were activated with zymosan as described above. At the end of the incubation, the medium was removed, and the cells were quickly fixed and permeabilized with a solution of 0.25% EDAC, 2% paraformaldehyde for 15 min at room temperature. The cells were washed three times with HBA, incubated for 2 h at room temperature with a rat mAb anti-Cys-LTs (1:100) or a rat IgG1 (BD Pharmingen, San Diego, CA) in blocking buffer, washed in HBA five times, and then incubated with Cy3-conjugated donkey antirat IgG (1:200) in blocking buffer for 1 h and imaged by confocal microscopy. In selected experiments, before the addition of zymosan, the cells were incubated for 5 min with AA861 (10 M) or MK886 (10 M).
Measurement of Phagocytosis-Peritoneal macrophages were obtained from group V sPLA 2 -null mice and from their wild-type littermates. Macrophages were used from mice on a C57BL/6ϫ129 Sv background, from mice bred to a BALB/c background for 3-10 generations and from mice bred to a C57BL/6 background for 10 generations. Timedependent zymosan stimulation was followed by staining with Diff-Quick (Dade Behring Inc., Newark, DE) and visualization under the light microscope. The phagocytic index was calculated by dividing the number of phagosomes by the total number of cells in a field, which was multiplied by the percentage of phagocytosing cells (30).
Introduction of PLA 2 Enzymes into Peritoneal Macrophages-Cells obtained from peritoneal lavage were incubated on glass coverslips with adenoviral vectors for 48 h in PCG buffer containing 2% fetal bovine serum, 2 mM L-glutamine, and 15% L cell-conditioned medium as a source of macrophage colony-stimulating factor. The efficiency of adenovirus-mediated gene transfer was assessed by the expression of GFP using fluorescence microscopy. Wells in which 60 -70% of cells were GFP-positive were used for subsequent experiments in which cells were washed and stimulated with Zymosan (10 ppc), fixed with 2% paraformaldehyde, and imaged using fluorescence microscopy. The phagocytic index was calculated for green fluorescence protein-positive cells.
Statistical Analyses-We performed two-way analysis of variance using the Prism 4 software to compare phagocytosis between populations of peritoneal macrophages derived from gene-disrupted mice and their wild-type controls. Where statistical significance was apparent, we performed post-hoc analyses of individual time points using paired Student's t test.

RESULTS
Group V sPLA 2 Co-localizes with Markers of the Golgi Apparatus and the Recycling Endosome-We have previously demonstrated, using immunofluorescence microscopy with a specific antibody, that group V sPLA 2 is present in the cytoplasm and in the juxtanuclear region of Group V sPLA 2 Regulates Phagocytosis mouse peritoneal macrophages (15). The specificity of the staining was evident from the lack of any staining of peritoneal macrophages from group V sPLA 2 -null mice (15). To ascertain the nature of the juxtanuclear region where group V sPLA 2 resides, we double-labeled macrophages with a rabbit polyclonal antibody specific to group V sPLA 2 (15,28) and with monoclonal antibodies specific for particular subcellular compartments. We found that group V sPLA 2 partially colocalizes with a cis-Golgi marker, GS28 (Fig. 1A), and to a lesser extent with Syntaxin 6, a trans-Golgi network marker (Fig. 1B). In addition to the Golgi apparatus, the juxtanuclear region contains recycling endosomes, subcellular organelles that localize in close proximity to the nucleus whose function is to recycle protein to the cell surface. Dual label immunofluorescence of group V sPLA 2 with the transferrin receptor, an endosomal marker, showed strong co-localization in the juxtanuclear region but not elsewhere in the cell (Fig. 1C). Group V sPLA 2 did not co-localize with protein-disulfide isomerase, a marker of the endoplasmic reticulum (Fig. 1D). Negative control experiments revealed no cross-reactivity of donkey anti-rabbit IgG with mouse mAbs or of donkey anti-mouse IgG with rabbit IgG (data not shown).
Group V sPLA 2 Translocates to the Phagosome during Phagocytosis-We have previously reported a role for group V sPLA 2 in the generation of eicosanoids by mouse peritoneal macrophages after ingestion of zymosan (15). The correct temporal and spatial location of the enzymes of eicosanoid biosynthesis is critical for their function, yet little is known about the location of group V sPLA 2 during eicosanoid biosynthesis. We therefore examined the location of endogenous group V sPLA 2 in macrophages following ingestion of zymosan. Group V sPLA 2 translocated from its juxtanuclear resting position to the forming phagosomes within 5 min of zymosan phagocytosis, localizing in the forming phagocytic cup (50% of cells) or completely surrounding the phagocytosed zymosan particles (Fig. 2A). By 15 min group V sPLA 2 completely surrounded the phagocytosed particles, whereas juxtanuclear staining for group V sPLA 2 was still visible in about 10% of the cells. At 30 min to 1 h, there was complete loss of juxtanuclear staining for group V sPLA 2 , which was most prominent surrounding peripheral phagosomes with loss of immunostaining around phagosomes that had moved away from the periphery. To confirm the localization of group V sPLA 2 to phago-somes, we determined whether group V sPLA 2 co-localized with LAMP-1, a lysosomal protein that moves to the maturing phagosome to form the phagolysosome (32) (Fig. 2B). In resting cells LAMP-1 had mainly lysosomal staining, as expected. By 30 min after zymosan stimulation, LAMP-1 co-localized with group V sPLA 2 around ingested zymosan particles, consistent with the movement of group V sPLA 2 to the phagosome.
Because group V sPLA 2 amplifies cPLA 2 ␣-dependent eicosanoid biosynthesis in response to zymosan and because cPLA 2 ␣ has been reported to translocate to the phagosome upon ingestion of zymosan, we determined whether other enzymes of eicosanoid biosynthesis also translocate to the phagosome (Fig. 3). Preliminary kinetic analyses of leukotriene biosynthesis revealed detectable leukotriene generation within 15 min, which peaked at 3 h as previously reported (15) (Fig. 3A). Extrapolation from these data indicated that the maximal rate of LTC 4 biosynthesis occurs within the first hour of zymosan stimulation. Therefore, immunostaining for eicosanoid biosynthetic enzymes was performed at these early time points. Representative data are shown for 15 min and 1 h. 5-LO was present in both the cytoplasm and the nucleus of resting macrophages and was visible at the forming phagosome within 5 min. By 15 min the protein was also visible in a perinuclear location (Fig. 3B). At 60 min 5-LO was visualized mainly surrounding the most peripheral zymosan particles. FLAP localizes to the nuclear envelope of resting cells (33). 5 min after zymosan stimulation FLAP was prominently visible at the phagosome and was still evident in the perinuclear location where it remained visible at 15 and 30 min. Staining at the phagosome was most prominent in newly formed, peripheral phagosomes, whereas phagosomes further from the plasma membrane showed only faint and diffuse staining. Perinuclear staining for FLAP was indistinct by 60 m. LTC 4 S, an integral perinuclear membrane protein with homology to FLAP (34), translocated to the phagosome with kinetics that appeared slower compared with FLAP. Immunostaining  A, peritoneal macrophages isolated from BALB/c mice were stimulated with unopsonized zymosan (10 particles/cell) for 5, 15, 30, or 60 min, fixed in 2% paraformaldehyde, permeabilized with 0.025% saponin, and incubated with rabbit anti-group V sPLA 2 followed by FITC-conjugated donkey anti-rabbit IgG. Immunofluorescence images are paired with the corresponding Nomarski image. B, peritoneal macrophages isolated from BALB/c mice were left unstimulated or were stimulated with unopsonized zymosan (10 particles/cell) for 30 min, fixed in 2% paraformaldehyde, permeabilized with 0.025% saponin, and incubated with rabbit anti-group V sPLA 2 and mAb to LAMP-1. The cells were washed and incubated with FITC-conjugated donkey anti-rabbit IgG and Cy3-conjugated donkey anti-rat IgG. Confocal images were obtained as described under "Experimental Procedures. " Autofluorescence of zymosan particles is visible within each phagosome.
for LTC 4 was weakly visible around ϳ50% of the ingested particles at 5 min (data not shown), was clearly visible at 15 min, and increased to 100% of particles by 30 -60 m. In contrast to FLAP, strong perinuclear immunostaining for LTC 4 was visible at all time points studied (up to 60 min after zymosan stimulation).
Immunodetection of LTC 4 at the Phagosome-To definitively establish whether the phagosome is a site of LTC 4 biosynthesis, we modified a previously published protocol (31) to visualize intracellular LTC 4 at its site of formation (Fig. 3C). LTC 4 is not detectable in any location in unstimulated cells, consistent with its formation de novo in response to an inflammatory stimulus (data not shown). Five min after zymosan phagocytosis, 20% of the cells showed perinuclear and/or periphagosome staining, indicating the presence of LTC 4 in the perinuclear region and at the phagosome. Interestingly, at 15 min almost 100% of the phagocytosed particles demonstrated bright periphagosome staining, whereas perinuclear LTC 4 was visible in some cells (Fig. 3C). By 60 min LTC 4 was most prominent in the most peripheral, most recently formed phagosomes, and perinuclear staining was no longer present. The specificity of the staining for LTC 4 was demonstrated by the lack of staining in macrophages isolated from LTC 4 synthase-null mice and in wild-type macrophages pretreated for 5 min with AA861, a 5-LO inhibitor, or MK886, a FLAP inhibitor (data not shown).
Group V sPLA 2 -null Peritoneal Macrophages Are Defective in Phagocytosis-Because group V sPLA 2 translocated to the forming phagosome, we quantified phagocytosis in group V sPLA 2 -null macrophages. Interestingly, before the addition of zymosan, adherent peritoneal macrophages from group V sPLA 2 -null cells had a narrower and more elongated appearance compared with wild-type control cells (Fig.  4A). After stimulation with zymosan, group V sPLA 2 -null cells tended to retain this narrower configuration (Fig. 4A) and appeared to ingest fewer particles of zymosan than wild-type cells. Formal quantification of the phagocytosis of zymosan over 3 h confirmed that macrophages lacking group V sPLA 2 were markedly impaired in their ability to ingest zymosan ( Fig. 4B; n ϭ 3; p ϭ 0.016 by analysis of variance). The impair-  /cell). B, peritoneal macrophages isolated from BALB/c mice were stimulated with unopsonized zymosan (10 particles/cell) for 15 and 60 min, fixed in 2% paraformaldehyde, permeabilized with 0.025% saponin, and incubated with specific antibodies for each of the enzymes of LTC 4 biosynthesis followed by FITC-conjugated secondary antibodies. C, to demonstrate LTC 4 formation at the phagosome, the cells were fixed with EDAC and immunostained with rat mAb anti-Cys-LTs followed by Cy3-conjugated donkey antirat IgG. Immunofluorescence images are paired with the corresponding Nomarski or bright field image. ment of phagocytosis by group V sPLA 2 -null macrophages was not strain-specific, being observed when cells were derived from mice bred to a BALB/c background (for 3-10 generations), from mice bred to a C57BL/6 background (for 10 generations), or from the first generation of C57BL/6ϫ129 group V sPLA 2 -null mice (data not shown). By comparison, peritoneal macrophages derived from cPLA 2 ␣-null mice bred to a BALB/c background (Fig. 5A) and from LTC 4 synthase-null mice bred to a C57BL/6 background (Fig. 5B) had no significant defect in phagocytosis compared with cells from wild-type strain-matched littermates (n ϭ 3; p ϭ 0.86 and p ϭ 0.53, respectively, by analysis of variance).
Expression of Group V sPLA 2 , but Not Group IIA sPLA 2 , in Group V sPLA 2 -null Macrophages Restores Phagocytosis-The embryonic stem cells used to derive group V sPLA 2 -null mice were derived from 129 mice. Certain strains of mice, including 129 and C57BL/6, but not BALB/c, have a natural disruption of the gene encoding group IIA sPLA 2 because of insertion of a thymidine residue in the open reading frame (35). The genes for group IIA sPLA 2 and group V sPLA 2 are tightly linked on the distal portion of mouse chromosome 4 (36). A search of the map of mouse chromosome 4 at the National Center for Biotechnology and Information website indicates that these two genes are separated by less than 20 kilobases. Therefore the disrupted group IIA sPLA 2 gene is always carried with the disrupted group V sPLA 2 gene during backcrosses with minimal chance of their separation by a recombination event (Table 1). Indeed, we confirmed by PCR that group V sPLA 2 -null mice crossed to a BALB/c background for 10 generations carried the disrupted group IIA sPLA 2 gene (data not shown). To confirm that the reduced capacity for phagocytosis in group V sPLA 2 -null mice was due to a lack of group V sPLA 2 rather than group IIA sPLA 2 or the combined deletion of both proteins, we used adenoviral constructs to reconstitute expression of each sPLA 2 in peritoneal macrophages of group V sPLA 2 -null mice bred to a BALB/c background for 10 generations. As illustrated in Fig. 6, reconstitution of group V sPLA 2 by adenoviral transfer augmented the phagocytic index of group V sPLA 2 -null macrophages (N10 BALB/c) by 47.21 Ϯ 10.72% (mean Ϯ S.D. for five time points in two separate experiments; p ϭ 0.03 by analysis of variance) compared with cells infected with adenovirus expressing LacZ in tandem with GFP. Reconstitution of group IIA sPLA 2 failed to augment the capacity of these cells to phagocytose zymosan.

DISCUSSION
Macrophages are an integral part of the host innate immune response and one of the first lines of defense against certain pathogens. They rapidly ingest pathogens, internalizing them in phagosomes that later  The expression (ϩ) or disruption (Ϫ) of the genes encoding group IIA sPLA 2 and group V sPLA 2 is shown for group V sPLA 2 -null mice and their wild-type (WT) littermates bred to a C57BL/6 or BALB/c background. Expression is also indicated for group V sPLA 2 -null cells, derived from mice bred to a BALB/c background, reconstituted with adenovirally expressed group IIA sPLA 2 (Group IIA sPLA 2 -Ad) or group V sPLA 2 (Group V sPLA 2 -Ad).
Group IIA sPLA 2 Group V sPLA 2 recruit lysosomal proteins to form phagolysosomes in which the pathogens are killed by an array of proteases and the generation of reactive oxygen species (21). The process of phagocytosis also elicits the rapid generation of proinflammatory mediators, including prostaglandins and leukotrienes derived from arachidonic acid, that elicit an acute inflammatory response (23)(24)(25)(26). Previous work has firmly established that leukotrienes and prostaglandins are formed in a perinuclear location in response to intracellular calcium flux elicited by diverse cell membrane receptors and calcium ionophores (31,33,(37)(38)(39). We have previously reported that group V sPLA 2 is prominently expressed in a perinuclear location in mouse mast cells (28) and peritoneal macrophages (15), and we speculated that this would facilitate its interaction with translocated cPLA 2 ␣. Our present data confirm that group V sPLA 2 is expressed in the Golgi of adherent mouse peritoneal macrophages, co-localizing with GS28, a cis-Golgi marker, and Syntaxin 6, a marker of the trans Golgi network. This is consistent with our prior results in which immunoelectron microscopy revealed group V sPLA 2 in Golgi cicsternae of the mast cell (28). Interestingly, group V sPLA 2 also co-localizes in the perinuclear region of peritoneal macrophages with the transferrin receptor, which is a marker of the recycling endosomal pathway. It has been suggested that the location of group V sPLA 2 in the Golgi may simply be a reflection of its synthesis, especially when overexpressed by transfection. Similarly, its co-expression with the transferrin receptor may reflect its secretion from the cell and subsequent internalization. However, its location in the Golgi would facilitate its interaction with translocated cPLA 2 ␣, and as we show in this paper, its location in the endosomal compartment allows it to co-localize with the other enzymes of leukotriene biosynthesis in the forming phagosome.
Using a gene disruption approach, we previously reported a role for group V sPLA 2 in amplifying eicosanoid biosynthesis by mouse peritoneal macrophages stimulated with zymosan (15). We were therefore interested to investigate its subcellular trafficking during phagocytosis and found that it rapidly moved to the forming phagosomes (Fig. 2). Leslie and co-workers (30) recently reported that cPLA 2 ␣ translocates to the phagocytic cup of mouse peritoneal macrophages in response to ingestion of zymosan. cPLA 2 ␣ is essential for the release of arachidonic acid from mouse peritoneal macrophages in response to zymosan (27). The demonstration that group V sPLA 2 translocates to the forming phagosome (Fig. 2) along with cPLA 2 ␣ (30) (Fig. 3) prompted us to test the hypothesis that the other enzymes of leukotriene biosynthesis translocate to the phagosome and that this is a site of leukotriene biosynthesis. We found that 5-LO, FLAP, and LTC 4 synthase were each detected at the forming phagosome as early as 5 min after the ingestion of zymosan particles and were prominent in this location during the first hour after zymosan stimulation (Fig. 3B), the time of maximal LTC 4 biosynthesis (Fig. 3A). The kinetics of translocation and predominant location varied between enzymes. Thus, LTC 4 synthase retained a prominent perinuclear location in addition to its presence in the forming phagosome. Weaker perinuclear staining was retained for FLAP and was also transiently appreciated for 5-LO at 5 and 15 min, similar to the weak and transient perinuclear staining for cPLA 2 ␣ reported in mouse macrophages after ingestion of zymosan (30) and consistent with the expected movement of these enzymes to the perinuclear region in response to intracellular calcium flux. To definitively establish that LTC 4 was synthesized at the phagosome, we examined the location of LTC 4 itself. We adapted a protocol published by Weller and co-workers (31) in which EDAC is used to fix LTC 4 to cellular proteins at its site of formation within the cell before its export by specific carrier systems (40). Consistent with the results of immunostaining for the enzymes of its synthesis, we demonstrated the presence of LTC 4 at the phagosome within 5 min of the onset of zymosan ingestion. Interestingly, transient perinuclear LTC 4 could also be detected, consistent with the transient perinuclear detection of 5-LO (Fig. 3B) and cPLA 2 ␣ (30), although this was less prominent than that observed at the phagosome.
Previous work has demonstrated that the plasma membrane does not have sufficient surface area to provide membrane for efficient phagocytosis, which requires the recruitment of membrane from other subcellular compartments, mainly early endosomes (41). The contribution of the ER is controversial (41)(42)(43). It is intriguing to note that this not only provides the membrane that is physically required for phagocytosis but also results in the assembly of the complete machinery necessary for leukotriene biosynthesis and indeed results in substantial leukotriene biosynthesis that is most prominent at the phagosome. The result is that leukotriene biosynthesis proceeds not merely as the inevitable consequence of calcium flux causing synthesis in a perinuclear location, but rather the biosynthesis of leukotrienes is physically linked to the process of phagocytosis. We can only speculate on the teleology of these events. Leukotrienes are prominent proinflammatory mediators that are essential for efficient host defense against certain pathogens. Mice that lack 5-LO are more susceptible to death from pulmonary infection with Klebsiella pneumoniae (44) at least in part because of impaired phagocytosis in neutrophils unable to synthesize LTB 4 . Neutrophils from mice that lack cPLA 2 ␣ are less able to kill E. coli (45), an effect that was attributed to impaired platelet activating factor biosynthesis. Thus, leukotriene generation in response to phagocytosis may not only amplify the ensuing inflammatory response but also contribute to the killing of the ingested pathogens. Cysteinyl leukotrienes have also been implicated in the migration of dendritic cells (46) providing a link between leukotriene biosynthesis by professional phagocytes and the adaptive immune response. It remains to be determined whether our findings extend to dendritic cells.
Because group V sPLA 2 localizes in the wall of the phagosome within minutes of zymosan stimulation (Fig. 2), we assessed its contribution to phagocytosis. Indeed, macrophages from mice deficient in group V sPLA 2 had an attenuated capacity for phagocytosis of zymosan, suggesting that the enzyme is required for efficient phagocytosis (Fig. 5). It has been reported that leukotrienes stimulate actin polymerization (47)(48)(49)(50). We therefore speculated that group V sPLA 2 might be acting through provision of leukotrienes at the forming phagosomal membrane. However, macrophages deficient in leukotriene C 4 synthase or cPLA 2 ␣, which is required for zymosan-induced leukotriene generation (27), demonstrated no deficiency in phagocytosis (Fig. 5). Girotti   also reported that inhibition of cPLA 2 ␣ did not impair the phagocytosis of zymosan by mouse peritoneal macrophages. Collectively, these data indicate that group V sPLA 2 does not contribute to phagocytosis through the provision of leukotrienes or other eicosanoids. Our data contrast with those from Peters-Golden and co-workers (44), who reported that alveolar macrophages from 5-LO-null mice had impaired phagocytosis and killing of bacteria leading to increased lethality from K. pneumonia. This could be largely corrected by exogenous LTB 4 (44), LTC 4 (51), or 5-hydroxyeicosatetraenoic acid (51). The role of LTB 4 appeared to be restricted to phagocytosis of IgG-opsonized bacteria and was not observed for complement-coated or unopsonized bacteria (52). Neutrophils from 5-LO-null mice were also defective in phagocytosis of opsonized bacteria, but in contrast to alveolar macrophages this could be corrected only by exogenous LTB 4 and not other 5-lipoxygenase products (60). That we did not show a role for endogenous leukotrienes in ingestion of unopsonized zymosan by mouse peritoneal macrophages may relate to the nature of the cell (peritoneal versus alveolar macrophages) or the target (unopsonized zymosan versus opsonized bacteria). Ongoing work is focused on determining the extent of the defect in phagocytosis in group V sPLA 2 -null mice.
The allele carrying the disrupted group V sPLA 2 gene was derived from a 129 embryonic stem cell and therefore carried the closely linked gene encoding group IIA sPLA 2 , which is disrupted in both 129 and C57BL/6 mice (35). Thus, group V sPLA 2 -null mice bred to a BALB/c background are effectively double knockouts of both group V sPLA 2 and group IIA sPLA 2 compared with their wild-type littermates ( Table 1). The observation that phagocytosis is defective in peritoneal macrophages derived from group V sPLA 2 -null mice bred to a C57BL/6 background, in which both wild-type and knockout mice lack group IIA sPLA 2 , suggested that group V sPLA 2 was indeed required for efficient phagocytosis. To definitively prove that the defect in phagocytosis in group V sPLA 2 -null macrophages is due to the loss of group V sPLA 2 , we used adenoviral constructs to reconstitute group V sPLA 2 expression. The demonstration that expression of group V sPLA 2 , but not group IIA sPLA 2 , restored the phagocytic capacity of group V sPLA 2 -null macrophages (Fig. 6) confirms its role in phagocytosis. That group IIA sPLA 2 did not substitute for group V sPLA 2 indicates that these enzymes do not serve merely redundant functions. This observation also suggests that certain structural features of group V sPLA 2 , such as its capacity for interfacial binding to phosphatidylcholine (which it does not share with group IIA sPLA 2 ) as opposed to its capacity to bind to proteoglycans (which is shared with group IIA sPLA 2 ) are important in determining the translocation and/or function of the enzyme at the phagosome. Nevertheless, this speculation requires experimental confirmation.
Dennis and co-workers (53-55) have extensively characterized the contribution of group V sPLA 2 and cPLA 2 ␣ to the release of arachidonic acid and prostaglandin E 2 generation in the P388D 1 macrophage. When these cells are stimulated with lipopolysaccharide, there is increased steady state RNA for group V sPLA 2 that requires the action of cPLA 2 ␣ and that in turn induces COX-2 expression (56). Using immunofluoresence staining, these authors revealed that 18 h after stimulation with lipopolysaccharide, group V sPLA 2 was present in granules that coimmunostained for COX-2 and caveolin-2 (57). Studies by Cho and co-workers (58) indicated that exogenous group V sPLA 2 triggered LTB 4 biosynthesis by human neutrophils through activating cPLA 2 ␣. Subsequent internalization of group V sPLA 2 occurred after release of arachidonic acid and was proposed as a degradation pathway for the enzyme. In contrast, exogenous group V sPLA 2 triggered LTC 4 biosynthesis in human eosinophils independent of cPLA 2 ␣ (59). Using a fluorescent substrate, the authors demonstrated the action of group V sPLA 2 in hydrolyzing cell membrane phospholipids first at the surface of the eosinophil and then at the nuclear envelope to which 5-LO had translocated. Taken together these studies suggest that group V sPLA 2 may act in an autocrine or paracrine fashion at different subcellular locations in the cell, depending on the cell type and the nature of the activating stimulus. Where one PLA 2 induces the expression of the other, their co-localization is not required. Nevertheless, in circumstances of rapid eicosanoid generation that is regulated by the coordinate action of sPLA 2 and cPLA 2 ␣, as in our present studies, their colocalization seems logical. Our studies reveal for the first time the association of group V sPLA 2 at the phagosome where it couples to cPLA 2 ␣ to amplify leukotriene biosynthesis and where it facilitates phagocytosis through mechanisms that are independent of cPLA 2 ␣.
In summary, we provide novel data to indicate that group V sPLA 2 is present not only in Golgi compartments but also in the recycling endosome of mouse peritoneal macrophages. We provide the first demonstration that the phagosome is a site of leukotriene biosynthesis. We further show for the first time that group V sPLA 2 is required for efficient phagocytosis of zymosan particles and provide evidence that this requirement is likely independent of its role in eicosanoid generation.