Cellular Plasmalogen Content Does Not Influence Arachidonic Acid Levels or Distribution in Macrophages: A Role for Cytosolic Phospholipase A2γ in Phospholipid Remodeling

Availability of free arachidonic acid (AA) constitutes a rate limiting factor for cellular eicosanoid synthesis. AA distributes differentially across membrane phospholipids, which is largely due to the action of coenzyme A-independent transacylase (CoA-IT), an enzyme that moves the fatty acid primarily from diacyl phospholipid species to ether-containing species, particularly the ethanolamine plasmalogens. In this work, we examined the dependence of AA remodeling on plasmalogen content using the murine macrophage cell line RAW264.7 and its plasmalogen-deficient variants RAW.12 and RAW.108. All three strains remodeled AA between phospholipids with similar magnitude and kinetics, thus demonstrating that cellular plasmalogen content does not influence the process. Cell stimulation with yeast-derived zymosan also had no effect on AA remodeling, but incubating the cells in AA-rich media markedly slowed down the process. Further, knockdown of cytosolic-group IVC phospholipase A2γ (cPLA2γ) by RNA silencing significantly reduced AA remodeling, while inhibition of other major phospholipase A2 forms such as cytosolic phospholipase A2α, calcium-independent phospholipase A2β, or secreted phospholipase A2 had no effect. These results uncover new regulatory features of CoA-IT-mediated transacylation reactions in cellular AA homeostasis and suggest a hitherto unrecognized role for cPLA2γ in maintaining membrane phospholipid composition via regulation of AA remodeling.


Introduction
Arachidonic acid (cis-5,8,11,14-eicosatetraenoic acid; AA) is the precursor of the eicosanoids, a large family of compounds with key roles in the initiation and resolution of inflammation [1]. Since AA is not found in free fatty acid form in cells but esterified into the sn-2 position of membrane glycerophospholipids, the participation of phospholipase A 2 enzymes that liberate the fatty acid constitutes a limiting step for the synthesis of eicosanoids, a process which also depends on the expression levels and activity of the AA-metabolizing enzymes cyclooxygenases and lipoxygenases [2][3][4][5].
AA, which is usually the major polyunsaturated fatty acid in the membranes of innate immune cells, is not uniformly distributed among membrane glycerophospholipids. Rather, marked the plasmalogen-deficient RAW cell variants generated by Zoeller and coworkers [30][31][32]. Our previous studies with these cells uncovered essential roles for ethanolamine plasmalogens in regulating phagocytosis [33] and in the execution of lipopolysaccharide (LPS)-primed responses [29].
In the current work, we show that overall phospholipid AA remodeling is essentially independent of the amount of plasmalogen present in the cells, suggesting that compartmentalization of AA in innate immune cells may depend primarily on the relative distribution of the fatty acid between classes (PE versus PC versus PI) rather than on specific molecular species within classes (i.e., alkenyacyl versus alkylacyl versus diacyl species). Moreover, our results also implicate cPLA 2 γ in regulating phospholipid AA remodeling.

Cell Culture
RAW264.7 macrophage-like cells and their ether phospholipid-deficient variants RAW. 12 and RAW.108 (generously provided by Dr. R. A. Zoeller, Boston University) [30][31][32] and P388D 1 macrophage-like cells (MAB clone, generously provided by Dr. E. A. Dennis, University of California at San Diego) [38][39][40] were grown in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum, 100 U/mL penicillin, 100 µg/mL streptomycin, and 2 mM L-glutamine at 37 • C in a humidified atmosphere of 5% CO 2 , as previously described [41,42]. Mouse peritoneal macrophages from Swiss mice (University of Valladolid Animal House, 10-12 weeks old) were obtained by peritoneal lavage using 5 mL cold phosphate-buffered saline and cultured in RPMI 1640 medium with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin, as described elsewhere [43,44]. All procedures involving animals were undertaken under the supervision of the Institutional Committee of Animal Care and Usage of the University of Valladolid (No. 907046) and are in accordance with the guidelines established by the Spanish Ministry of Agriculture, Food, and Environment and by the European Union.
All experiments were conducted in serum-free media. When activated cells were used, the stimulus (150 µg/mL yeast-derived zymosan) was added 1 h after incubating the cells in serum-free media for the times indicated. Zymosan was prepared exactly as described [43,44]. Only zymosan batches that demonstrated no measurable endogenous phospholipase A 2 activity, as measured by in vitro assay under different conditions [45,46], were used in this study. Cell protein content was quantified according to Bradford [47], using a commercial kit (BioRad Protein Assay). To generate classically (M1) or alternatively (M2) polarized macrophages, the cells were treated with LPS (250 ng/mL) plus interferon-γ (500 U/mL) or with interleukin-4 (20 ng/mL) plus interleukin-13 (20 ng/mL) for 8 h, respectively [48]. When required, radiolabeling of the cells with [ 3 H]AA was achieved by including 0.25 µCi/mL [ 3 H]AA during the overnight adherence period (20 h). Labeled AA that had not been incorporated into cellular lipids was removed by washing the cells four times with serum-free Cells 2019, 8,799 4 of 20 medium containing 0.5 mg/mL albumin. The determination of [ 3 H]AA release from activated cells was carried out exactly as described [41,42,49].

Quantitative PCR
Total RNA was extracted from the cells with TRIzol reagent (Invitrogen) according to the manufacturer's instructions, and 2 µg RNA was reverse transcribed using random primers and oligo d(T) and Moloney murine leukemia virus reverse transcriptase (Ambion, Austin, TX, USA). Quantitative PCR was carried out with an ABI 7500 machine (Applied Biosystems, Carlsbad, CA, USA) using Brilliant III Ultra-Fast SYBR Green qPCR Master Mix (Agilent Technologies, Santa Clara, CA, USA). Cycling conditions were as follows: 1 cycle at 95 • C for 3 min and 40 cycles at 95 • C for 12 s, 60 • C for 15 s, and 72 • C for 28 s. The replicates were averaged, and fold induction was determined in ∆∆Ct-based fold-change calculations, with cyclophilin A as a control [50]. Primer sequences are available upon request.

Liquid Chromatography/Mass Spectrometry (LC/MS) Analyses of Eicosanoids
Analysis of eicosanoids by LC/MS was carried out exactly as described elsewhere [8,27], using an Agilent 1260 Infinity high-performance liquid chromatograph equipped with an Agilent G1311C quaternary pump and an Agilent G1329B Autosampler, coupled to an API2000 triple quadrupole mass spectrometer (Applied Biosystems, Carlsbad, CA, USA). Quantification was carried out by integrating the chromatographic peaks of each species and by comparing with an external calibration curve made with analytical standards [8,27].

Measurement of Phospholipid Arachidonate Remodeling
This was carried out exactly as described by Pérez et al. [60]. Briefly, the cells were pulse labeled with 1 nM [ 3 H]AA (0.25 µCi/mL) for 15 min at 37 • C. The cells were then washed with medium containing 0.5 mg/mL bovine serum albumin to remove the non-incorporated label. Afterward, the cells were placed in serum-free medium and incubated at 37 • C for the indicated periods of time. After lipid extraction, phospholipid classes were separated by thin-layer chromatography as indicated above. The spots corresponding to each phospholipid class were cut out and assayed for radioactivity by liquid scintillation counting. Scheme 1 provides a graphical description of this kind of experiment.

Liquid Chromatography/Mass Spectrometry (LC/MS) Analyses of Eicosanoids
Analysis of eicosanoids by LC/MS was carried out exactly as described elsewhere [8,27], using an Agilent 1260 Infinity high-performance liquid chromatograph equipped with an Agilent G1311C quaternary pump and an Agilent G1329B Autosampler, coupled to an API2000 triple quadrupole mass spectrometer (Applied Biosystems, Carlsbad, CA, USA). Quantification was carried out by integrating the chromatographic peaks of each species and by comparing with an external calibration curve made with analytical standards [8,27].

Measurement of Phospholipid Arachidonate Remodeling
This was carried out exactly as described by Pérez et al. [60]. Briefly, the cells were pulse labeled with 1 nM [ 3 H]AA (0.25 μCi/mL) for 15 min at 37 °C. The cells were then washed with medium containing 0.5 mg/mL bovine serum albumin to remove the non-incorporated label. Afterward, the cells were placed in serum-free medium and incubated at 37 °C for the indicated periods of time. After lipid extraction, phospholipid classes were separated by thin-layer chromatography as indicated above. The spots corresponding to each phospholipid class were cut out and assayed for radioactivity by liquid scintillation counting. Scheme 1 provides a graphical description of this kind of experiment.

Statistical Analysis
All experiments were carried out at least three times with incubations in duplicate or triplicate, and the data are expressed as means ± SEM. Statistical analysis was carried out by Student's t-test, with p < 0.05 taken as statistically significant.

Statistical Analysis
All experiments were carried out at least three times with incubations in duplicate or triplicate, and the data are expressed as means ± SEM. Statistical analysis was carried out by Student's t-test, with p < 0.05 taken as statistically significant. of carbons of a given fatty acid and x is an integer which, subtracted from n, gives the position of the last double bond of the molecule. The AA content was very similar in all three cell types tested. Also, no significant variations were detected in any other fatty acid, including the polyunsaturates of the n-3 series. When the various phospholipid classes were separated and analyzed for AA content, PE was found to constitute the major AA-containing class in all three types, representing almost half of total AA present in these cells. The fatty acid was found at comparable levels in PI and PC, and lower amounts were found in PS. No other phospholipid class contained significant amounts of AA ( Figure 1B). The relative distribution of AA among phospholipid classes of RAW264.7 cells and its plasmalogen-deficient variants, with PE predominating, is consistent with previous data for other phagocytic cells such as human monocytes [6] and murine peritoneal macrophages [7,8]. Collectively, these results show that cellular plasmalogen content influences neither cellular AA levels nor the relative distribution of the fatty acid among phospholipid classes. designated by their number of carbon atoms, and their number of double bonds are designated after a colon. To differentiate isomers, the n−x (n minus x) nomenclature is used, where n is the number of carbons of a given fatty acid and x is an integer which, subtracted from n, gives the position of the last double bond of the molecule. The AA content was very similar in all three cell types tested. Also, no significant variations were detected in any other fatty acid, including the polyunsaturates of the n-3 series. When the various phospholipid classes were separated and analyzed for AA content, PE was found to constitute the major AA-containing class in all three types, representing almost half of total AA present in these cells. The fatty acid was found at comparable levels in PI and PC, and lower amounts were found in PS. No other phospholipid class contained significant amounts of AA ( Figure  1B). The relative distribution of AA among phospholipid classes of RAW264.7 cells and its plasmalogen-deficient variants, with PE predominating, is consistent with previous data for other phagocytic cells such as human monocytes [6] and murine peritoneal macrophages [7,8]. Collectively, these results show that cellular plasmalogen content influences neither cellular AA levels nor the relative distribution of the fatty acid among phospholipid classes. The distribution of AA between phospholipid molecular species was measured by LC/MS ( Figure 2). Fatty chains within phospholipids are designated by their number of carbon atoms, and their number of double bonds are designated after a colon. A designation of O-before the first fatty chain indicates that the sn-1 position is ether-linked, whereas a P-designation indicates a plasmalogen form (sn-1 vinyl ether linkage) [61]. The results showed that PE plasmalogens constituted the major reservoir of this fatty acid in RAW 264.7 cells (Figure 2). High amounts of AA were also found in one particular PI species, namely PI(18:0/20:4), and in diacyl-PE species. Lower AA amounts were found in several PC and PS species ( Figure 2). Interestingly, despite the plasmalogen deficiency of RAW.12 and RAW.108 cells, the AA distribution by the phospholipid class in these cells was preserved due to a compensatory elevation of AA in diacyl species of PE and PC compared to wild type RAW 264.7 cells (Figure 2). The distribution of AA between phospholipid molecular species was measured by LC/MS ( Figure 2). Fatty chains within phospholipids are designated by their number of carbon atoms, and their number of double bonds are designated after a colon. A designation of O-before the first fatty chain indicates that the sn-1 position is ether-linked, whereas a P-designation indicates a plasmalogen form (sn-1 vinyl ether linkage) [61]. The results showed that PE plasmalogens constituted the major reservoir of this fatty acid in RAW 264.7 cells (Figure 2). High amounts of AA were also found in one particular PI species, namely PI(18:0/20:4), and in diacyl-PE species. Lower AA amounts were found in several PC and PS species ( Figure 2). Interestingly, despite the plasmalogen deficiency of RAW.12 and RAW.108 cells, the AA distribution by the phospholipid class in these cells was preserved due to a compensatory elevation of AA in diacyl species of PE and PC compared to wild type RAW 264.7 cells ( Figure 2).

Importance of Plasmalogen Content for Phosphospholipid AA Remodeling
In mammalian cells, plasmalogen enrichment with AA is thought to occur primarily via CoA-IT-mediated reactions, which transfer a fatty acyl moiety from a phospholipid donor, primarily AAcontaining PC species, to a lysophospholipid acceptor, very often an ethanolamine lysoplasmalogen, without using CoA or forming a free fatty acid intermediate [11][12][13][14]. This reaction also appears to be instrumental for AA mobilization responses, as inhibition of CoA-IT leads to marked inhibition of AA release [8,23,62]. To characterize this route, RAW 264.7 cells were labeled with [ 3 H]AA for 15 min and, after extensive washing to remove non-incorporated fatty acid, the movement of labels between phospholipid classes was analyzed. Immediately after the 15-min labeling period, PC was the major [ 3 H]AA-containing phospholipid, followed by PI and PE. [ 3 H]AA incorporation into PS was considerably lower ( Figure 3A). The amount of labeled AA in PC underwent a rapid decrease with time, which was paralleled by an increase of similar magnitude of AA in PE, reflecting the action of CoA-IT. Levels of labeled AA in PI and PS remained unchanged along the time course of the experiment. To make direct comparisons between various conditions and in accord with previous work [55] we have defined the time at which the amount of [ 3 H]AA in PC equals that in PE as the "remodeling time" and found it to be 21 ± 4 min (mean ± S.E.M., n = 6). Importantly, examination of the rate of AA remodeling from PC to PE in the plasmalogen-deficient variants RAW.12 and RAW.108 revealed essentially the same kinetics as in RAW 264.7 cells and, hence, nearly identical remodeling times ( Figure 3B). Thus, these results indicate that phospholipid AA remodeling from PC to PE is not influenced by the cellular plasmalogen content. For comparative purposes, remodeling experiments under identical conditions were also carried out using another murine macrophage-like

Importance of Plasmalogen Content for Phosphospholipid AA Remodeling
In mammalian cells, plasmalogen enrichment with AA is thought to occur primarily via CoA-IT-mediated reactions, which transfer a fatty acyl moiety from a phospholipid donor, primarily AA-containing PC species, to a lysophospholipid acceptor, very often an ethanolamine lysoplasmalogen, without using CoA or forming a free fatty acid intermediate [11][12][13][14]. This reaction also appears to be instrumental for AA mobilization responses, as inhibition of CoA-IT leads to marked inhibition of AA release [8,23,62]. To characterize this route, RAW 264.7 cells were labeled with [ 3 H]AA for 15 min and, after extensive washing to remove non-incorporated fatty acid, the movement of labels between phospholipid classes was analyzed. Immediately after the 15-min labeling period, PC was the major [ 3 H]AA-containing phospholipid, followed by PI and PE. [ 3 H]AA incorporation into PS was considerably lower ( Figure 3A). The amount of labeled AA in PC underwent a rapid decrease with time, which was paralleled by an increase of similar magnitude of AA in PE, reflecting the action of CoA-IT. Levels of labeled AA in PI and PS remained unchanged along the time course of the experiment. To make direct comparisons between various conditions and in accord with previous work [55] we have defined the time at which the amount of [ 3 H]AA in PC equals that in PE as the "remodeling time" and found it to be 21 ± 4 min (mean ± S.E.M., n = 6). Importantly, examination of the rate of AA remodeling from PC to PE in the plasmalogen-deficient variants RAW.12 and RAW.108 revealed essentially the same kinetics as in RAW 264.7 cells and, hence, nearly identical remodeling times ( Figure 3B). Thus, these results indicate that phospholipid AA remodeling from PC to PE is not influenced by the cellular plasmalogen content. For comparative purposes, remodeling experiments under identical conditions were also carried out using another murine macrophage-like cell line, P388D 1 , and using resident murine peritoneal macrophages. In keeping with previous estimates [63][64][65], the remodeling time of P388D 1 cells was found to be similar to that of RAW 264.7 cells and their variants and considerably lower than that of murine peritoneal macrophages ( Figure 3B). cell line, P388D1, and using resident murine peritoneal macrophages. In keeping with previous estimates [63][64][65], the remodeling time of P388D1 cells was found to be similar to that of RAW 264.7 cells and their variants and considerably lower than that of murine peritoneal macrophages ( Figure  3B).

Role of Plasmalogens in Functional Responses of Macrophages to Receptor Stimulation
In previous work from our laboratory, we took advantage of the RAW 264.7 cells and the plasmalogen-deficient variants RAW.12 and RAW.108 to establish the key roles of ethanolamine plasmalogens in regulating the phagocytic activity [33] and in the execution of LPS-primed responses [29] of macrophages. Here, we extended our studies on the role of plasmalogens in cellular physiology by studying other functional responses of macrophages. Figure 4 shows the quantitative analysis of eicosanoids produced by zymosan-stimulated cells. The profile of eicosanoids produced was similar, both qualitatively and quantitatively, to that previously reported by Buczynski et al. [66]. Only products of the cyclooxygenase pathway were found, with PGD2 constituting the major eicosanoid detected. Eicosanoid production was the same in the RAW 264.7 cells as in the plasmalogen-deficient variants (Figure 4), thus suggesting that plasmalogen status does not influence the eicosanoid biosynthetic response of the macrophages.

Role of Plasmalogens in Functional Responses of Macrophages to Receptor Stimulation
In previous work from our laboratory, we took advantage of the RAW 264.7 cells and the plasmalogen-deficient variants RAW.12 and RAW.108 to establish the key roles of ethanolamine plasmalogens in regulating the phagocytic activity [33] and in the execution of LPS-primed responses [29] of macrophages. Here, we extended our studies on the role of plasmalogens in cellular physiology by studying other functional responses of macrophages. Figure 4 shows the quantitative analysis of eicosanoids produced by zymosan-stimulated cells. The profile of eicosanoids produced was similar, both qualitatively and quantitatively, to that previously reported by Buczynski et al. [66]. Only products of the cyclooxygenase pathway were found, with PGD 2 constituting the major eicosanoid detected. Eicosanoid production was the same in the RAW 264.7 cells as in the plasmalogen-deficient variants (Figure 4), thus suggesting that plasmalogen status does not influence the eicosanoid biosynthetic response of the macrophages.  In the next series of experiments, we treated the cells with LPS/interferon-γ or interleukin-4/interleukin-13 to induce polarization/activation of the macrophages to pro-inflammatory (M1) or anti-inflammatory (M2) phenotypes [48], respectively, and changes in the gene expression levels of various markers associated to each phenotype were assessed by qPCR. Figure 5 shows that there were no differences between RAW 264.7 cells and the plasmalogen-deficient variants in the expression levels of any of the genes assayed.
These results, along with our previous data [29,33], underscore the differential involvement of plasmalogens in some, but not all, responses of macrophages, thus reflecting some sort of biological specificity of this kind of phospholipids. In the next series of experiments, we treated the cells with LPS/interferon-γ or interleukin-4/interleukin-13 to induce polarization/activation of the macrophages to pro-inflammatory (M1) or anti-inflammatory (M2) phenotypes [48], respectively, and changes in the gene expression levels of various markers associated to each phenotype were assessed by qPCR. Figure 5 shows that there were no differences between RAW 264.7 cells and the plasmalogen-deficient variants in the expression levels of any of the genes assayed.
These results, along with our previous data [29,33], underscore the differential involvement of plasmalogens in some, but not all, responses of macrophages, thus reflecting some sort of biological specificity of this kind of phospholipids.

Studies Utilizing AA-enriched Cells
As a consequence of continuous growth in culture, cell lines are known to exhibit diminished levels of polyunsaturated fatty acids, including AA, compared to normal cells [67,68]. For example, RAW 264.7 cells contain 2-3-fold less AA than murine resident peritoneal macrophages [58]. Since CoA-IT-driven fatty acid remodeling reactions determine the distribution of AA among membrane phospholipids, we reasoned that the much faster remodeling observed in RAW 264.7 cells compared to normal macrophages could be related to their relative deficiency in AA. To test this possibility, RAW 264.7 cells and the plasmalogen-deficient variants were cultured in media supplemented with 25 μM AA (complexed with bovine serum albumin at a 2:1 ratio) for 48 h. This procedure resulted in the cells increasing their AA content by about 2-3-fold, thus reaching levels comparable to those found in normal peritoneal macrophages [58]. Figure 6A compares the distribution of AA between phospholipid classes in the AA-enriched cells. Importantly, all of the AA-containing phospholipid classes in wild type RAW 264.7 cells and their variants incorporated exogenous AA to a similar extent. This resulted in the relative distribution profile of the fatty acid by class being preserved, i.e., PE comprising approximately 50% of cellular AA and PI constituting the second richest AA phospholipid, followed closely by PC. After the AA-enrichment period, the cells were spiked with [ 3 H]AA and the movement of AA from PC to PE was followed at different times. Notably, under these conditions, the remodeling process was markedly slowed in both wild type and plasmalogendeficient variants, with a remodeling time of approx. 2 h, i.e., comparable to that of resident peritoneal macrophages (cf. Figure 3B and Figure 6B). These data suggest that endogenous cellular levels of AA, not plasmalogen status, determines the rate of remodeling of AA among phospholipids.

Studies Utilizing AA-enriched Cells
As a consequence of continuous growth in culture, cell lines are known to exhibit diminished levels of polyunsaturated fatty acids, including AA, compared to normal cells [67,68]. For example, RAW 264.7 cells contain 2-3-fold less AA than murine resident peritoneal macrophages [58]. Since CoA-IT-driven fatty acid remodeling reactions determine the distribution of AA among membrane phospholipids, we reasoned that the much faster remodeling observed in RAW 264.7 cells compared to normal macrophages could be related to their relative deficiency in AA. To test this possibility, RAW 264.7 cells and the plasmalogen-deficient variants were cultured in media supplemented with 25 µM AA (complexed with bovine serum albumin at a 2:1 ratio) for 48 h. This procedure resulted in the cells increasing their AA content by about 2-3-fold, thus reaching levels comparable to those found in normal peritoneal macrophages [58]. Figure 6A compares the distribution of AA between phospholipid classes in the AA-enriched cells. Importantly, all of the AA-containing phospholipid classes in wild type RAW 264.7 cells and their variants incorporated exogenous AA to a similar extent. This resulted in the relative distribution profile of the fatty acid by class being preserved, i.e., PE comprising approximately 50% of cellular AA and PI constituting the second richest AA phospholipid, followed closely by PC. After the AA-enrichment period, the cells were spiked with [ 3 H]AA and the movement of AA from PC to PE was followed at different times. Notably, under these conditions, the remodeling process was markedly slowed in both wild type and plasmalogen-deficient variants, with a remodeling time of approx. 2 h, i.e., comparable to that of resident peritoneal macrophages (cf. Figures 3B and 6B). These data suggest that endogenous cellular levels of AA, not plasmalogen status, determines the rate of remodeling of AA among phospholipids.  A key step in phospholipid AA remodeling is the continuous generation of lysophospholipid moieties by phospholipase A2 enzymes, which act as fatty acid acceptors for the subsequent CoA-ITmediated transacylation reaction. Since macrophage activation by receptor-directed agonists, including zymosan, results in noticeable elevations of lysophospholipid levels compared to unstimulated cells [8,27,44,69,70], we investigated whether the activation state of the cells, with its attendant elevation of cellular lysophospholipid levels, exerted any influence on their phospholipid AA remodeling rate. To this end, the cells were labeled with [ 3 H]AA for 15 min and, after extensive washing, they were untreated (control incubations) or treated with zymosan (150 μg/mL) for different time periods and the transfer of label from PC to PE was measured. The experiments were carried out in otherwise untreated cells and in exogenous AA-treated cells, and both wild type and plasmalogen variants were used. No effect of cell activation could be ascertained on phospholipid AA remodeling under any of these conditions. The remodeling time of zymosan-activated cells was the same as that of resting cells, either using wild-type cells, the plasmalogen-deficient variants, or cells preloaded with exogenous AA. Thus, phospholipid AA remodeling rate is independent of the activation state of the cells.

Phospholipase A2 Inhibition Studies
Earlier studies attempting to identify the origin of the lysophospholipid acceptors used in the CoA-IT-mediated transacylation reaction took advantage of PLA2 inhibitors available at that time; however, given the uncertain specificity of some of the inhibitors used, these studies did not provide unambiguous responses [71,72]. Recently, much more selective phospholipase A2 inhibitors with improved properties have been developed [73] and we have used these to re-evaluate the involvement of phospholipase A2 isoforms. The inhibitors used were pyrrophenone for cPLA2α (at A key step in phospholipid AA remodeling is the continuous generation of lysophospholipid moieties by phospholipase A 2 enzymes, which act as fatty acid acceptors for the subsequent CoA-IT-mediated transacylation reaction. Since macrophage activation by receptor-directed agonists, including zymosan, results in noticeable elevations of lysophospholipid levels compared to unstimulated cells [8,27,44,69,70], we investigated whether the activation state of the cells, with its attendant elevation of cellular lysophospholipid levels, exerted any influence on their phospholipid AA remodeling rate. To this end, the cells were labeled with [ 3 H]AA for 15 min and, after extensive washing, they were untreated (control incubations) or treated with zymosan (150 µg/mL) for different time periods and the transfer of label from PC to PE was measured. The experiments were carried out in otherwise untreated cells and in exogenous AA-treated cells, and both wild type and plasmalogen variants were used. No effect of cell activation could be ascertained on phospholipid AA remodeling under any of these conditions. The remodeling time of zymosan-activated cells was the same as that of resting cells, either using wild-type cells, the plasmalogen-deficient variants, or cells preloaded with exogenous AA. Thus, phospholipid AA remodeling rate is independent of the activation state of the cells.

Phospholipase A 2 Inhibition Studies
Earlier studies attempting to identify the origin of the lysophospholipid acceptors used in the CoA-IT-mediated transacylation reaction took advantage of PLA 2 inhibitors available at that time; however, given the uncertain specificity of some of the inhibitors used, these studies did not provide unambiguous responses [71,72]. Recently, much more selective phospholipase A 2 inhibitors with improved properties have been developed [73] and we have used these to re-evaluate the involvement of phospholipase A 2 isoforms. The inhibitors used were pyrrophenone for cPLA 2 α (at least 3 orders of magnitude more potent for cPLA 2 α than for iPLA 2 β or sPLA 2 enzymes) [34]; FKGK18 for iPLA 2 β (at least 200 and 400 times more potent for iPLA 2 β than for cPLA 2 α and sPLA 2 s, respectively) [35]; GK436, also for iPLA 2 β (at least 1000-fold more potent for iPLA 2 β than for cPLA 2 α, and no appreciable effect on sPLA 2 enzymes) [36]; and GK241 for sPLA 2 (inhibits IIA and V forms, lacking appreciable inhibition against cPLA 2 α, iPLA 2 β, or any other sPLA 2 form) [37]. However, neither of these inhibitors had any measurable effect on the remodeling time of RAW 264.7 cells, thus providing further additional support for the lack of involvement of cPLA 2 α, iPLA 2 β, or sPLA 2 -IIA/V in phospholipid AA remodeling.
Recently, it has been speculated that a lesser known member of the group IV phospholipase A 2 family of enzymes, i.e., the group IVC form, also known as cytosolic phospholipase A 2 γ (cPLA 2 γ), could be involved in phospholipid AA remodeling [14]. This suggestion is made primarily on the basis of the biochemical properties of cPLA 2 γ (i.e., Ca 2+ -independent, manifests measurable CoA-independent transacylation activity in vitro, and permanently associated with membranes) [14]. To begin to address this question, conditions were established to achieve silencing of cPLA 2 γ by siRNA technology in RAW 264.7 cells. Since we have been unable to find reliable antibodies against murine cPLA 2 γ, the efficiency of siRNA knockdown was judged by qPCR. Using this technique, we were able to achieve as much as a 70-75% decrease in cPLA 2 γ mRNA under our conditions ( Figure 7A). Strikingly, even with this incomplete cPLA 2 γ silencing, cells still exhibited clear defects in AA remodeling from PC to PE, reflected by a statistically significant increase of their remodeling time ( Figure 7B). Thus, these data provide strong evidence for the involvement of cPLA 2 γ in phospholipid AA remodeling in macrophages. Conversely, upon stimulation with zymosan, cells deficient in cPLA 2 γ released similar amounts of AA as control cells ( Figure 8A), thus indicating that the enzyme is not a significant regulator of the AA release response. In keeping with previous data [8,41,42,[74][75][76][77], AA-release experiments conducted in parallel in the presence of the phospholipase A 2 inhibitors described above confirmed the role of cPLA 2 α but not of sPLA 2 or iPLA 2 β in AA release, as only inhibition of cPLA 2 α resulted in blockade of the response ( Figure 8B). In these experiments, we also used bromoenol lactone (BEL) for comparative purposes with data in bibliography [32,74]. BEL is a widely used irreversible inhibitor of calcium-independent phospholipase A 2 s, with little or no effect on calcium-dependent enzymes [78,79], although with potential off-target effects depending on cell type [80,81]. In our hands, BEL did not have any effect on zymosan-stimulated AA release ( Figure 8B). Collectively, these results provide evidence for separate roles for the group IV cytosolic family members cPLA 2 α and cPLA 2 γ in cellular AA homeostasis; while the former regulates AA release but not phospholipid AA remodeling, the latter does the opposite. least 3 orders of magnitude more potent for cPLA2α than for iPLA2β or sPLA2 enzymes) [34]; FKGK18 for iPLA2β (at least 200 and 400 times more potent for iPLA2β than for cPLA2α and sPLA2s, respectively) [35]; GK436, also for iPLA2β (at least 1000-fold more potent for iPLA2β than for cPLA2α, and no appreciable effect on sPLA2 enzymes) [36]; and GK241 for sPLA2 (inhibits IIA and V forms, lacking appreciable inhibition against cPLA2α, iPLA2β, or any other sPLA2 form) [37]. However, neither of these inhibitors had any measurable effect on the remodeling time of RAW 264.7 cells, thus providing further additional support for the lack of involvement of cPLA2α, iPLA2β, or sPLA2-IIA/V in phospholipid AA remodeling.
Recently, it has been speculated that a lesser known member of the group IV phospholipase A2 family of enzymes, i.e., the group IVC form, also known as cytosolic phospholipase A2γ (cPLA2γ), could be involved in phospholipid AA remodeling [14]. This suggestion is made primarily on the basis of the biochemical properties of cPLA2γ (i.e., Ca 2+ -independent, manifests measurable CoAindependent transacylation activity in vitro, and permanently associated with membranes) [14]. To begin to address this question, conditions were established to achieve silencing of cPLA2γ by siRNA technology in RAW 264.7 cells. Since we have been unable to find reliable antibodies against murine cPLA2γ, the efficiency of siRNA knockdown was judged by qPCR. Using this technique, we were able to achieve as much as a 70-75% decrease in cPLA2γ mRNA under our conditions ( Figure 7A). Strikingly, even with this incomplete cPLA2γ silencing, cells still exhibited clear defects in AA remodeling from PC to PE, reflected by a statistically significant increase of their remodeling time ( Figure 7B). Thus, these data provide strong evidence for the involvement of cPLA2γ in phospholipid AA remodeling in macrophages. Conversely, upon stimulation with zymosan, cells deficient in cPLA2γ released similar amounts of AA as control cells ( Figure 8A), thus indicating that the enzyme is not a significant regulator of the AA release response. In keeping with previous data [8,41,42,[74][75][76][77], AA-release experiments conducted in parallel in the presence of the phospholipase A2 inhibitors described above confirmed the role of cPLA2α but not of sPLA2 or iPLA2β in AA release, as only inhibition of cPLA2α resulted in blockade of the response ( Figure 8B). In these experiments, we also used bromoenol lactone (BEL) for comparative purposes with data in bibliography [32,74]. BEL is a widely used irreversible inhibitor of calcium-independent phospholipase A2s, with little or no effect on calcium-dependent enzymes [78,79], although with potential off-target effects depending on cell type [80,81]. In our hands, BEL did not have any effect on zymosan-stimulated AA release ( Figure  8B). Collectively, these results provide evidence for separate roles for the group IV cytosolic family members cPLA2α and cPLA2γ in cellular AA homeostasis; while the former regulates AA release but not phospholipid AA remodeling, the latter does the opposite.

Discussion
Phospholipid AA remodeling is necessary for cells of the innate immune system to distribute fatty acid within the appropriate cellular pools for its subsequent mobilization by phospholipase A2 enzymes. This is a key aspect in eicosanoid regulation because the nature and amount of eicosanoids produced under activation conditions may ultimately depend on the composition and subcellular localization of the phospholipid pool where the AA-hydrolyzing phospholipase A2 acts [1,2]. In innate immune cells, ether phospholipids, particularly the ethanolamine plasmalogens, are strikingly enriched in AA, which is thought to be due to the fact that these phospholipids are major acceptors for fatty acid transfer reactions between phospholipids that help shape the final distribution of the fatty acid in cells [11][12][13][14]. While the enrichment of ether phospholipids with AA suggests a key role for this kind of phospholipids in AA homeostasis, their role still remains relatively obscure. In fact, receptor stimulation of AA mobilization in plasmalogen-deficient cells is similar to that of normal cells [29,32], and none of the major AA-releasing phospholipase A2s described to date have been found to exhibit any particular preference for substrates containing an sn-1 ether bond [16]. Although these data suggested that plasmalogens may not be essential for the phospholipase A2-mediated AA mobilization process itself, the possibility remained that they are still instrumental for placing the fatty acid in the appropriate subcellular localizations via phospholipid transacylation reactions. Unexpectedly, the results presented in this study show that this is also not the case. We find no difference between plasmalogen-deficient and otherwise normal cells, thus suggesting that plasmalogen status has no influence on phospholipid AA remodeling. It should be noted in this regard that the relative distribution of AA among phospholipid classes (i.e., PE versus PC versus PI) is maintained in the plasmalogen-deficient cells because, in these cells, there is a compensatory

Discussion
Phospholipid AA remodeling is necessary for cells of the innate immune system to distribute fatty acid within the appropriate cellular pools for its subsequent mobilization by phospholipase A 2 enzymes. This is a key aspect in eicosanoid regulation because the nature and amount of eicosanoids produced under activation conditions may ultimately depend on the composition and subcellular localization of the phospholipid pool where the AA-hydrolyzing phospholipase A 2 acts [1,2]. In innate immune cells, ether phospholipids, particularly the ethanolamine plasmalogens, are strikingly enriched in AA, which is thought to be due to the fact that these phospholipids are major acceptors for fatty acid transfer reactions between phospholipids that help shape the final distribution of the fatty acid in cells [11][12][13][14]. While the enrichment of ether phospholipids with AA suggests a key role for this kind of phospholipids in AA homeostasis, their role still remains relatively obscure. In fact, receptor stimulation of AA mobilization in plasmalogen-deficient cells is similar to that of normal cells [29,32], and none of the major AA-releasing phospholipase A 2 s described to date have been found to exhibit any particular preference for substrates containing an sn-1 ether bond [16]. Although these data suggested that plasmalogens may not be essential for the phospholipase A 2 -mediated AA mobilization process itself, the possibility remained that they are still instrumental for placing the fatty acid in the appropriate subcellular localizations via phospholipid transacylation reactions. Unexpectedly, the results presented in this study show that this is also not the case. We find no difference between plasmalogen-deficient and otherwise normal cells, thus suggesting that plasmalogen status has no influence on phospholipid AA remodeling. It should be noted in this regard that the relative distribution of AA among phospholipid classes (i.e., PE versus PC versus PI) is maintained in the plasmalogen-deficient cells because, in these cells, there is a compensatory elevation of AA levels in diacyl species. This is an important concept because it suggests that, in terms of overall AA distribution, it is the substituent at the sn-3 position of the phospholipid (i.e., ethanolamine versus choline versus inositol versus serine) and not the chemical nature of the sn-1 bond (acyl versus alkyl versus alkenyl) that determines the incorporation of AA. Consistent with this notion, AA transacylation reactions involve only ethanolamine-and choline-containing phospholipids, thus indicating some sort of specificity at the level of the sn-3 substituent.
While the lack of sn-1 influence on the AA transacylation reaction in cells is unanticipated, it should be noted that CoA-IT, the enzyme catalyzing these transacylation reactions, may use 1-acyl-PE just as well as 1-alkenyl-PE [11][12][13][14] and this, in fact, may allow to explain satisfactorily why the plasmalogen-deficient variants are able to compensate their deficiency by directing the AA to diacyl phospholipids. On the other hand, the finding that plasmalogen-deficient cells and normal cells mobilize AA similarly in spite of their composition being so different at the molecular species level provides strong support to the idea that cells can sustain critical reactions with many different lipid compositions rather than a single composition [82]. Furthermore, our results also raise the intriguing possibility that the enrichment of plasmalogens with AA might not be necessarily related to regulatory aspects of AA homeostasis and eicosanoid metabolism but to biophysical effects and interaction of the phospholipid with other membrane components to sustain different biological responses. Our collective findings showing that plasmalogens participate in the execution of some responses [29,33] but not in others (this study) are fully consistent with this view. AA-containing ethanolamine plasmalogens are frequently found as components of specific membrane microdomains called lipid rafts [83][84][85]. The relative plasmalogen content within these domains may affect key properties such as fluidity, tendency to fusion, packing, thickness, and density, thereby influencing the biological behavior of membrane rafts in membrane transport and transmembrane signaling [83][84][85].
CoA-IT catalyzes the enzymatic step that is unique to the phospholipid AA remodeling pathway, i.e., the direct transfer of AA from a phospholipid donor to a lysophospholipid acceptor in the absence of CoA or ATP [11][12][13][14]. CoA-IT has been defined as a membrane-bound, calcium-independent enzyme. Based on biochemical and mechanistic commonalities, Yamashita and coworkers suggested that CoA-independent transacylation reactions in cells are catalyzed by (an) enzyme(s) of the phospholipase A 2 family and speculated that group IVC cytosolic phospholipase A 2 (also called cytosolic phospholipase A 2 γ, cPLA 2 γ) is a possible candidate [14,86,87]. The same authors noted that the cPLA 2 γ-catalyzed transacylase reaction works better when the AA donor is lysoPC instead of AA-containing diacyl-PC. On the other hand, Stewart et al. [88] noted that the enzyme has higher lysophospholipase activity than phospholipase A 2 activity. However, conditions of in vitro specificity assays are not necessarily translatable to the in vivo situation, where compartmentalization of substrates and products and the presence of competing enzymes may dramatically modify the specificities reported.
By using siRNA technology to specifically knock down cellular expression levels of cPLA 2 γ, we have tested experimentally the proposed involvement of cPLA 2 γ in regulating phospholipid AA remodeling. Our results clearly indicate that this is the case, as cells which were made deficient in cPLA 2 γ transfer AA from PC to PE significantly more slowly than control cells. These results constitute, to the best of our knowledge, the first report attributing cellular AA remodeling activity to a well-characterized molecular entity, i.e., the cPLA 2 γ enzyme. Thus, the finding is significant because it makes now possible to apply molecular biology approaches, such as overexpression or deletion, which should significantly expand our knowledge about the cellular and molecular regulation of phospholipid AA remodeling reactions. At this point, we cannot indicate whether cPLA 2 γ participates in the CoA-independent transacylation reaction by providing the lyso acceptors that initiate the reaction, by directly catalyzing the fatty acid transacylation, or by acting at both levels. Clearly, further work should be conducted to clarify these mechanistic issues and whether the enzyme is constitutively active or its activity increases after agonist stimulation [24]. Also, since our data clearly show that selective inhibition of cPLA 2 γ by siRNA slows but does not eliminate phospholipid AA remodeling, it seems likely that other enzymes in addition to cPLA 2 γ operate as CoA-IT in cells, and further work should be undertaken to identify them.
Another striking feature of this work is the finding that the rate of AA remodeling from PC to PE appears to be independent of the state of activation of the cells. We could not observe any change in the rate of AA remodeling from PC to PE in zymosan-stimulated cells compared to unstimulated cells. This finding was somehow unexpected because lysophospholipid availability initiates phospholipid AA remodeling in cells and zymosan-activated macrophages increase their intracellular lysophospholipid content as a consequence of cPLA 2 α activation [8,27,29]. Lack of an effect of cell stimulation on CoA-IT-mediated phospholipid remodeling is at variance with previous results in phorbol ester-stimulated platelets [89], tumor necrosis factor α-stimulated neutrophils [90], and antigen-stimulated mast cells [91] but is consistent with work in neutrophils on the regulation of platelet-activating factor synthesis via transacylation reactions, where it was found that CoA-IT activity did not increase as a consequence of cell activation but that regulation occurred at the level of increased substrate availability [92]. In this regard, we also report in this paper that increasing the cellular levels of AA results in decreased rates of CoA-mediated phospholipid AA remodeling. It is thus tempting to speculate with a scenario wherein an AA-containing phospholipid that is present in the AA-enriched cells but not in the otherwise normal cells may act to regulate CoA-IT-mediated transacylation reactions by directly impinging on the enzyme and/or by regulating substrate availability. We have previously shown that a short-lived AA-containing phospholipid produced by activated cells, namely 1,2-diarachidonoyl-glycerophosphoinositol, is able to regulate macrophage responses to innate immune stimuli [7].

Conclusions
We demonstrate in this work that, although plasmalogens constitute the major reservoirs of AA in normal cells, CoA-IT-mediated phospholipid AA remodeling reactions responsible for the asymmetric distribution of the fatty acid in the various phospholipid pools are not influenced by the plasmalogen content of cells. Compartmentalization of AA in cells appears to depend primarily on headgroup composition of membrane phospholipids rather than on specific molecular species. In addition, our work implicates cPLA 2 γ as a major enzyme involved in phospholipid AA remodeling. Taken together, our results provide new information to understand better the regulatory processes underlying cellular AA availability, which may be used to develop valid strategies to manipulate eicosanoid metabolism and signaling pathways in innate immunity and inflammation.