Calcium-mediated translocation of cytosolic phospholipase A2 to the nuclear envelope and endoplasmic reticulum.

Cytosolic phospholipase A2 (cPLA2) is activated by a wide variety of stimuli to release arachidonic acid, the precursor of the potent inflammatory mediators prostaglandin and leukotriene. Specifically, cPLA2 releases arachidonic acid in response to agents that increase intracellular Ca2+. In vitro data have suggested that these agents induce a translocation of cPLA2 from the cytosol to the cell membrane, where its substrate is localized. Here, we use immunofluorescence to visualize the translocation of cPLA2 to distinct cellular membranes. In Chinese hamster ovary cells that stably overexpress cPLA2, this enzyme translocates to the nuclear envelope upon stimulation with the calcium ionophore A23187. The pattern of staining observed in the cytoplasm suggests that cPLA2 also translocates to the endoplasmic reticulum. We find no evidence for cPLA2 localization to the plasma membrane. Translocation of cPLA2 is dependent on the calcium-dependent phospholipid binding domain, as a calcium-dependent phospholipid binding deletion mutant of cPLA2 (delta CII) fails to translocate in response to Ca2+. In contrast, cPLA2 mutated at Ser-505, the site of mitogen-activated protein kinase phosphorylation, translocates normally. This observation, combined with the observed phosphorylation of delta CII, establishes that translocation and phosphorylation function independently to regulate cPLA2. The effect of these mutations on cPLA2 translocation was confirmed by subcellular fractionation. Each of these mutations abolished the ability of cPLA2 to release arachidonic acid, establishing that cPLA2-mediated arachidonic acid release is strongly dependent on both phosphorylation and translocation. These data help to clarify the mechanisms by which cPLA2 is regulated in intact cells and establish the nuclear envelope and endoplasmic reticulum as primary sites for the liberation of arachidonic acid in the cell.

The 85-kDa cytosolic phospholipase A 2 (cPLA 2 ), 1 which selectively releases arachidonic acid from the sn-2 position of membrane phospholipids, is crucial to the initiation of the inflammatory response. cPLA 2 activity is stimulated by a wide variety of agents, including the proinflammatory cytokines in-terleukin 1 (1,2) and tumor necrosis factor (3), macrophage colony-stimulating factor (4), thrombin (5,6), ATP (5), mitogens (7)(8)(9)(10), and endothelin (11). The release of arachidonic acid is the rate-limiting step in the generation of prostaglandins and leukotrienes, the proinflammatory eicosanoids. Cleavage of arachidonoyl-containing phospholipids also results in the release of lysophospholipid, the precursor of the inflammatory mediator platelet-activating factor (12). cPLA 2 is expressed in many cell types. Many of these are associated with the inflammatory response, such as monocytes (4), neutrophils (13), and synovial fibroblasts (14). However, cPLA 2 is also expressed in kidney, spleen, heart, lung, liver, testis, and hippocampus (15). This diverse pattern of expression is consistent with accumulating evidence that in addition to its role in inflammation, cPLA 2 participates in signaling in processes such as platelet activation (6,16), tumor necrosis factor-induced cytotoxicity (17), and cell proliferation (8,10,18). cPLA 2 activity is regulated both transcriptionally and posttranslationally. Post-translational activation is thought to occur by two mechanisms. One mechanism involves agonist-induced MAP kinase phosphorylation of cPLA 2 , resulting in stimulation of its intrinsic enzymatic activity (19,20). The second involves a Ca 2ϩ -dependent translocation of cPLA 2 from the soluble to the membrane fraction of cells (21)(22)(23), allowing cPLA 2 access to its arachidonoyl-containing phospholipid substrate. As discussed below, the results presented in this study establish that both mechanisms are important for the stimulation of cPLA 2 -induced arachidonic acid release.
In vitro studies have strongly suggested that the membrane binding function of cPLA 2 resides in its Ca 2ϩ -dependent phospholipid binding (CaLB) domain (22,24), a region similar to the CII domain of protein kinase C. This domain has been shown to be necessary and sufficient for Ca 2ϩ -dependent membrane binding (22,24). In this study, we characterize the translocation of cPLA 2 in intact cells and show that this enzyme translocates to the nuclear envelope and endoplasmic reticulum. Deleting the CaLB domain abolishes the ability of cPLA 2 to translocate to these membranes, whereas a mutation at the MAP kinase phosphorylation site has no effect on translocation. Each of these mutations is shown to prevent cPLA 2induced arachidonic acid release. Our present data, together with recent reports localizing several arachidonic acid-metabolizing enzymes to the nuclear envelope and endoplasmic reticulum (25)(26)(27)(28)(29), establish these membranes as major sites of arachidonic acid production and metabolism in the cell.

EXPERIMENTAL PROCEDURES
Cell Culture and Antibodies-CHO cells were maintained as described previously (5). Monoclonal antibody 1.1.1 (19) was raised against human cPLA 2 purified from E5-CHO cells and was used for staining of parental and cPLA 2 -overexpressing CHO cells. Polyclonal antibody 7905, generated against cPLA 2 produced in Eschericia coli, was used for immunoblotting (22). PGHS-1-specific polyclonal antibody was generated against the N terminus of ovine PGHS-1, residues Ala-25 to Cys-35 (30). E-selectin monoclonal antibody was generated against soluble human E-selectin (31).
Immunofluorescence Staining-Cells grown on coverslips in growth medium were rinsed with serum-free medium prior to a 2-min treatment with 2 M calcium ionophore A23187 (Sigma). Treated and untreated cells were washed twice briefly with Tris-buffered saline containing 0.01% Triton X-100. In the experiments presented in this paper, all washes and incubations of A23187-treated cells contained 1 mM CaCl 2 in an attempt to lessen the chance that translocated cPLA 2 would detach from membranes. Experiments performed in the absence of CaCl 2 , however, yielded identical staining patterns (not shown). Cells were fixed for 2 min at room temperature in a 50:50 mixture of acetone and methanol. After two brief washes, blocking was performed in Trisbuffered saline, 20% goat serum for 1 h at 37°C. Cells were incubated with primary antibody (150 g of cPLA 2 antibody/ml, 100 g of Eselectin antibody/ml, or a 1:20 dilution of PGHS-1 antibody) for 1 h at 37°C and washed once briefly followed by 2 ϫ 10-min washes. Incubations with second antibody were performed for 1 h at 37°C using a 1:100 dilution of fluorescein isothiocyanate-conjugated goat anti-mouse or anti-rabbit antibody (Zymed, San Francisco, CA) and washed as above. After a brief wash with water, coverslips were inverted onto slides with Slowfade bleaching retardant (Molecular Probes). The pattern of cPLA 2 localization was similar using a variety of staining procedures, including starving cells of serum overnight prior to staining, eliminating Triton from all wash buffers, or fixing the cells with 4% paraformaldehyde (followed by permeabilization with 0.2% Triton) (data not shown).
Fluorescence Confocal Microscopy-The subcellular location of cPLA 2 was visualized by fluorescence confocal microscopy. An Insight Bilateral scanning confocal microscope (Meridian Instruments, Okemos, MI) was used with an argon ion laser as the excitation source (32). The laser was used at a power ranging from 25-75 mW, depending on the level of fluorescence intensity. All images were photographed with a 100ϫ objective. Untreated and A23187-treated samples for a given cell line were analyzed using identical instrument settings.
Fractionation and Immunoblotting-Cells were starved overnight in serum-free media and treated with 2 M A23187 for 10 min. After washing with Tris-buffered saline, cells were collected by scraping and lysed in 1 ml lysis buffer (20 mM Hepes, pH 7.4, 1 mM EGTA, 0.34 M sucrose, 10 g of leupeptin/ml, 2 mM phenylmethylsulfonyl fluoride, 5 mM dithiothreitol) by Parr bombing, 600 p.s.i. for 5 min. Lysates were centrifuged for 1 h at 100,000 ϫ g, and protein (5 g) from the supernatant and pellet fractions was subjected to SDS-polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose, blotted with anti-cPLA 2 polyclonal antibody 7905 (22), and developed by chemiluminescence after incubation with horseradish peroxidase-conjugated protein A (Amersham).
Arachidonic Acid Release-Cells were plated onto 12-well cluster dishes at 1.5 ϫ 10 5 cells per well in growth medium. After a 20 h incubation, the medium was removed and replaced with 0.5 ml of ␣ medium (lacking serum) (Life Technologies, Inc., Grand Island, NY) containing 0.5 Ci of [ 3 H]arachidonic acid (100 Ci/mmol; New England Nuclear, Wilmington, DE) and incubated for another 20 h at 37°C. The cells were then washed three times with medium containing 0.1% bovine serum albumin and incubated with A23187 for 10 min. The medium was removed, and radioactivity was quantified by scintillation counting.

RESULTS
Ca 2ϩ Ionophore Induces cPLA 2 Translocation to the Nuclear Envelope and Endoplasmic Reticulum-cPLA 2 can be induced to associate with natural membranes in the presence of physiologically relevant Ca 2ϩ levels in vitro (21,22). This Ca 2ϩ -dependent translocation is thought to reflect a translocation that occurs in vivo when cells are stimulated with agents that increase intracellular Ca 2ϩ . As the translocation of cPLA 2 in response to Ca 2ϩ has not yet been characterized in intact cells, the identity of the membrane(s) to which cPLA 2 binds is not known. To address this issue, indirect immunofluorescence was performed using E5-4, a CHO cell line that overexpresses human cPLA 2 (5). In resting cells, cPLA 2 was found distributed throughout the cytoplasm. In contrast, after treatment with the Ca 2ϩ ionophore A23187, significant cPLA 2 staining appeared as a discrete ring surrounding the nucleus. A distinct ring was not apparent in untreated cells, although a low level of perinuclear staining was detectable in some cells. This suggests that an increase in cytosolic Ca 2ϩ induces a translocation of cPLA 2 from the cytoplasm to the nuclear envelope.
Interestingly, immunofluorescent staining of cPLA 2 in A23187-treated E5-4 cells was consistently more intense than that seen in untreated cells (quantitated in Fig. 1B). This is most likely due to a greater loss of soluble cPLA 2 than membrane-bound cPLA 2 throughout the staining procedure. A similar interpretation was offered for the results observed upon immunogold labeling of the arachidonic acid-metabolizing enzyme 5-lipoxygenase, which also translocates to the nuclear envelope in response to A23187 (27). In that study, 5-lipoxygenase was undetectable in unstimulated cells but was apparent at the nuclear envelope after ionophore treatment. Alternatively, binding of cPLA 2 to membranes may result in a better exposure of the epitope for the monoclonal antibody, resulting in more efficient antibody binding. In either case, the observed increase in cytoplasmic staining intensity suggests that cPLA 2 translocates not only to the nuclear envelope but also to a cytoplasmic membrane structure, most likely the endoplasmic reticulum. Indeed, in some experiments ionophore treatment resulted in an increase in reticular cytoplasmic fluorescence, consistent with endoplasmic reticulum staining.
The plasma membrane was consistently found to be devoid of cPLA 2 staining. To ensure that fixation had not disrupted the integrity of the plasma membrane, CHO cells overexpressing the plasma membrane protein E-selectin were stained. As shown in Fig. 2, E-selectin antibody was clearly able to label the plasma membrane. No labeling of the nuclear envelope was observed in these cells. As a control for the specificity of the anti-cPLA 2 monoclonal antibody, immunofluorescence was also FIG. 1. Differential localization of cPLA 2 in untreated and Ca 2ϩ ionophore-stimulated cells. CHO cells overexpressing wild type cPLA 2 were stimulated with 2 M Ca 2ϩ ionophore (ϩA23187) for 2 min or left untreated (ϪA23187). A, immunofluorescence was performed using a monoclonal antibody specific for cPLA 2 and a fluorescein-conjugated second antibody. Immunofluorescent staining was visualized by confocal microscopy (magnification, 330ϫ). B, quantitative fluorescence imaging of cytoplasm and nuclear envelope. The intensity of immunofluorescent staining was quantified by measuring fluorescence before and after ionophore treatment in equivalent areas of the cytoplasm and nuclear envelope (NE). ERϩCyto, endoplasmic reticulum ϩ cytoplasm. performed on the parental CHO cells used for these studies. Only faint staining was observed (Fig. 2), similar to that seen in controls in which the monoclonal antibody was eliminated from the staining protocol (data not shown). The lack of endogenous cPLA 2 staining is likely to be due both to the low level of cPLA 2 expressed in these cells and the failure of this antibody to recognize murine (and by inference, hamster) cPLA 2 efficiently. Fig. 2 also shows staining of CHO cells overexpressing PGHS-1, which metabolizes arachidonic acid to prostaglandin. Similar staining results were obtained with a CHO line expressing the isoform PGHS-2 (not shown). Both PGHS isoforms have been localized to the endoplasmic reticulum and nuclear envelope (25,29). As expected, neither E-selectin nor PGHS-1 staining patterns were affected by ionophore treatment.
CaLB, but Not Ser-505 Phosphorylation, Is Required for Translocation-In cell-free systems, the CaLB domain is required for the association of cPLA 2 with membranes (24). To confirm the importance of the CaLB domain for cPLA 2 translocation in intact cells, immunofluorescence was performed on a CHO line (⌬CII) that overexpresses a cPLA 2 deletion mutant lacking the CaLB domain (amino acids 1-134) (24). As shown in Fig. 3A, no increase in nuclear envelope staining was observed, indicating that this truncated cPLA 2 is unable to translocate in response to ionophore. Consistent with this result, ionophore did not induce an increase in cytoplasmic staining intensity in these cells (Fig. 3B). This result establishes the increase in staining intensity as a CaLB-dependent event (most likely, as discussed above, a result of membrane associ-ation). The effect of a mutation at the MAP kinase phosphorylation site, Ser-505, was also investigated. MAP kinase phosphorylation at this residue has previously been shown to increase the intrinsic enzymatic activity of cPLA 2 (19,20). To determine whether phosphorylation at this site was involved in regulating translocation, immunofluorescence staining of CHO cells expressing SA505-cPLA 2 , containing a serine-to-alanine mutation at this residue, was performed. The translocation of SA505-cPLA 2 in response to ionophore was indistinguishable from that observed with wild type cPLA 2 (Fig. 4). The increase in staining intensity seen with wild type cPLA 2 upon ionophore treatment was also observed in the SA505 line, further supporting the ability of this mutant to translocate to membranes.
To confirm that deletion of the CaLB domain, but not the mutation of Ser-505, abolished cPLA 2 translocation, subcellular fractionation was performed using cells expressing these mutants. After ionophore treatment, cell lysates were centrifuged at 100,000 ϫ g, and the soluble and particulate fractions were immunoblotted for cPLA 2 . As shown in Fig. 5, both cPLA 2 and the SA505-cPLA 2 mutant redistributed to the particulate fraction upon ionophore treatment, whereas the ⌬CII mutant did not. The insert shown below the ⌬CII portion of the figure represents a longer exposure of the blot, revealing the translocation of endogenous cPLA 2 in the ⌬CII line. These data confirm that phosphorylation at Ser-505 is not required for cPLA 2 translocation. Conversely, translocation is not required for Ser-505 phosphorylation, as evidenced by the presence of a cPLA 2 doublet in the ⌬CII but not the SA505 line. The upper band of the cPLA 2 doublet has previously been established to result from phosphorylation at Ser-505 (19). These data establish that translocation and phosphorylation occur independently.
ATP and thrombin have previously been shown to induce arachidonic acid release in CHO cells that overexpress cPLA 2 (5). However, in fractionation experiments using these agents, we were unable to detect the appearance of cPLA 2 in the pellet fraction (data not shown). This suggests that translocation  Fig. 1. B, quantification of fluorescence intensity. Staining intensity was measured as described in Fig. 1B. induced by ATP and thrombin is highly transient, consistent with the transient nature of the Ca 2ϩ increase induced by these agents (33).
Both Translocation and Ser-505 Phosphorylation Are Required for Stimulation of Arachidonic Acid Release-To investigate the relative importance of translocation and Ser-505 phosphorylation for cPLA 2 -mediated arachidonic acid release, the amount of arachidonic acid liberated upon A23187 treatment was measured in E5-4, ⌬CII, SA505, and parental CHO cells (Fig. 6). The level of arachidonic acid released from stimulated E5-4 cells was consistently at least two to three times greater than that released from parental CHO cells. As shown in Fig. 6A, cells expressing ⌬CII did not show enhanced arachidonic acid release in response to A23187 relative to CHO cells. Similarly, SA505-expressing cells showed no increase in arachidonic acid release over the parental line (Fig. 6B). The observation that each of these mutations, separately, can prevent arachidonic acid release establishes that both translocation and Ser-505 phosphorylation are indispensable to cPLA 2 function in intact cells. DISCUSSION The data presented here establish that in intact cells, cPLA 2 translocates to membranes in response to a rise in intracellular Ca 2ϩ . Specifically, cPLA 2 binds to the nuclear envelope and the endoplasmic reticulum. The translocation of cPLA 2 to these sites is consistent with previous work examining the subcellular distribution of cPLA 2 in Ca 2ϩ ionophore-stimulated macrophages (26). In this study, cPLA 2 from stimulated cells was detected both in the nuclear fraction and in the membrane fraction lacking nuclei, consistent with its presence in both the nuclear envelope and the endoplasmic reticulum. In another study (23), fractionation of neutrophils stimulated with granulocyte/macrophage colony-stimulating factor and the chemotactic peptide fMet-Leu-Phe revealed cPLA 2 to be in the membrane fraction, even after nuclei were removed by centrifugation. Although the translocation of cPLA 2 to the nuclear envelope could not be addressed in this study (the presence of cPLA 2 in the nuclear fraction was not examined), its presence in the membrane fraction supports the translocation of cPLA 2 to the endoplasmic reticulum. Taken together with our immunofluorescence results, including the lack of cPLA 2 staining at the plasma membrane, these data implicate both the nuclear envelope and endoplasmic reticulum as the primary sites for arachidonic acid production in the cell. Interestingly, PGHS-1 and -2 (25,29), the enzymes responsible for metabolizing arachidonic acid to prostaglandins, are also localized at these sites. In addition, 5-lipoxygenase, which metabolizes arachidonic acid to leukotrienes, as well as its activating protein, FLAP, are localized to the nuclear envelope (27). These data suggest that arachidonic acid is both produced and metabolized at the nuclear envelope and endoplasmic reticulum.
In the work presented in this paper, the localization of cPLA 2 was examined by immunofluorescent staining of cPLA 2 overexpressed in CHO cells. This approach was chosen to assess the FIG. 4. Translocation of SA505-cPLA 2 . Cells stably overexpressing SA505-cPLA 2 , which contains a serine-to-alanine mutation at the Ser-505 MAP kinase phosphorylation site, were probed with antibody to cPLA 2 . Indirect immunofluorescence was performed as described in the legend to Fig. 1.   FIG. 5. Fractionation of lysates from wild type (WT), ⌬CII, and SA505 cells. Cells overexpressing each cPLA 2 construct were stimulated with 2 M ionophore for 10 min (ϩ) or left untreated (Ϫ). Lysates were spun at 100,000 ϫ g, and 5 g of protein from the supernatant (S) and pellet (P) fractions were subjected to SDS-polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose and immunoblotted for cPLA 2 . Development was by chemiluminescence. The positions of full-length cPLA 2 and ⌬CII are indicated. The panel below the ⌬CII portion of the blot represents a longer exposure of these lanes, demonstrating that endogenous cPLA 2 translocates normally in the ⌬CII line. The band in the ⌬CII pellet fractions represents a small amount of insoluble ⌬CII. * indicates a cross-reactive species of unknown origin. effects of the ⌬CII and SA505 mutations on cPLA 2 translocation. Caution must be taken with this approach, however, as it is possible that overexpression of cPLA 2 may result in a wider pattern of expression than occurs physiologically. Interestingly, cPLA 2 translocation to the nuclear envelope was recently demonstrated in rat mast cells stimulated with A23187 or IgE/antigen (34). In this study, no obvious staining of the endoplasmic reticulum was observed. This observation raises the possibility that the localization of cPLA 2 to the endoplasmic reticulum is a consequence of overexpression. However, it is also possible that the relative proportion of cPLA 2 in the nuclear envelope and endoplasmic reticulum varies between cell types. Further studies will be necessary to clarify this issue. In any case, the observed lack of plasma membrane staining, both in mast cells and in cells overexpressing cPLA 2 , strongly suggests that cPLA 2 does not translocate to this membrane in response to increases in intracellular calcium.
Translocation of cPLA 2 is abolished upon deletion of the CaLB domain but not upon mutation of Ser-505. This suggests that translocation and phosphorylation regulate cPLA 2 independently, a conclusion supported by the observation that ⌬CII is phosphorylated in CHO cells (as evidenced by the doublet in Fig. 5). Both translocation and phosphorylation are critical for A23187-induced arachidonic acid release. The observation that these regulatory mechanisms function independently strongly supports a model in which phosphorylation at Ser-505 serves primarily to activate the enzymatic activity of cPLA 2 (19,20), and translocation allows access of the enzyme to its substrate.
This discussion focuses on the localization of cPLA 2 to the nuclear envelope and endoplasmic reticulum. How cPLA 2 is selectively localized to these sites is not known. The translocation of cPLA 2 to the nuclear envelope is consistent with previous data indicating that arachidonic acid is preferentially released from the nuclear envelope, as observed in pulse-chase experiments using [ 14 C]arachidonate-labeled HSDM 1 C 1 cells stimulated with bradykinin (35). Interestingly, electron microscopic studies using [ 3 H]arachidonic acid have shown that the nuclear membrane is the preferred site of initial arachidonic acid incorporation in these cells (36). Arachidonic acid is also incorporated rapidly into the endoplasmic reticulum; in contrast, transit to the plasma membrane is slow. This study shows that although arachidonic acid is preferentially incorporated at particular sites, it becomes evenly distributed throughout the cell. Thus, a correlation exists between the cellular localization of cPLA 2 and the sites at which arachidonic acid is initially incorporated into the cell. The significance of this correlation, however, is not yet known.
Several arachidonic acid-independent mechanisms could also explain the localization of cPLA 2 . Although Ca 2ϩ can induce purified cPLA 2 to bind membranes in the absence of additional protein (24), it remains possible that a "docking protein" localized to the nuclear envelope and endoplasmic reticulum allows more efficient binding of cPLA 2 at these sites. It is unlikely that a spatially localized release of Ca 2ϩ is responsible for the targeting of cPLA 2 , since our experiments revealed no translocation of cPLA 2 to the plasma membrane with A23187, whose effects are thought to include the triggering of a Ca 2ϩ influx across the plasma membrane. Another possibility is that some feature of the plasma membrane excludes cPLA 2 , and cPLA 2 simply translocates to all cellular membranes that it is capable of binding. Interestingly, the plasma membrane is known to be enriched in sphingolipids compared to the nuclear envelope and endoplasmic reticulum. It is conceivable that the increase in phospholipid packing density that results from a high sphingolipid content may prevent cPLA 2 binding. Consistent with this idea, Leslie and Channon (21) have shown that both the activity and calcium sensitivity of cPLA 2 are inhibited by the increased substrate packing density induced by sphingolipids. Clearly, a better understanding of the lipid and protein components of distinct cellular membranes will greatly facilitate our understanding of the mechanism(s) governing the localization of cPLA 2 .
As discussed above, the localization of cPLA 2 is similar to that reported for PGHS-1 and -2, which metabolize arachidonic acid to prostaglandins (25,29). PGHS-1 is expressed constitutively and is thought to perform certain physiological "housekeeping" functions. PGHS-2, whose expression is induced by cytokines, has been implicated in inflammation (37,38). cPLA 2 -mediated arachidonic acid release can couple to PGHS-2, since antisense oligonucleotide inhibition of cPLA 2 decreases the level of endotoxin-stimulated PGE 2 in monocytes (without affecting PGHS-2 activity) (39), and PGE 2 release in endotoxin-stimulated monocytes has been shown to be dependent on PGHS-2 activity even in the presence of PGHS-1 (40). However, in platelets, known from various studies to contain only PGHS-1, inhibition of cPLA 2 using arachidonoyl trifluoromethyl ketone blocks thromboxane B 2 formation, suggesting that cPLA 2 can also couple to PGHS-1 (41). Indeed, the association of cPLA 2 with the nuclear envelope and the endoplasmic reticulum correlates well with the localization of both PGHS-1 and PGHS-2. Although it is not yet clear which factors govern the coupling of cPLA 2 with each PGHS isoform, the colocalization of cPLA 2 with these enzymes and 5-lipoxygenase is certain to ensure the efficient utilization of arachidonic acid upon agonist stimulation, consistent with a central role for cPLA 2 in the agonist-induced biosynthesis of prostaglandins and leukotrienes.