Lipid Droplet Biogenesis Induced by Stress Involves Triacylglycerol Synthesis That Depends on Group VIA Phospholipase A2*

This work investigates the metabolic origin of triacylglycerol (TAG) formed during lipid droplet (LD) biogenesis induced by stress. Cytotoxic inhibitors of fatty acid synthase induced TAG synthesis and LD biogenesis in CHO-K1 cells, in the absence of external sources of fatty acids. TAG synthesis was required for LD biogenesis and was sensitive to inhibition and down-regulation of the expression of group VIA phospholipase A2 (iPLA2-VIA). Induction of stress with acidic pH, C2-ceramide, tunicamycin, or deprivation of glucose also stimulated TAG synthesis and LD formation in a manner dependent on iPLA2-VIA. Overexpression of the enzyme enhanced TAG synthesis from endogenous fatty acids and LD occurrence. During stress, LD biogenesis but not TAG synthesis required phosphorylation and activation of group IVA PLA2 (cPLA2α). The results demonstrate that iPLA2-VIA provides fatty acids for TAG synthesis while cPLA2α allows LD biogenesis. LD biogenesis during stress may be a survival strategy, recycling structural phospholipids into energy-generating substrates.

Intracellular lipid droplets (LDs) 5 are cytosolic inclusions present in most eukaryotic cells, containing a core of triacylglycerol (TAG) and cholesteryl esters, surrounded by a phospho-lipid monolayer and by specific proteins, among which the best characterized belong to the perilipin family (1)(2)(3). The biology of LD has received increasing interest, due to the link between excess lipid storage in certain tissues and pathologies such as obesity, diabetes, or atherosclerosis (3,4). LDs have been shown to interfere with membrane translocation of the insulin-sensitive glucose transporter, an observation that might account for insulin resistance in type 2 diabetes (5). The recent identification of 132 genes controlling LD number, size, and distribution in Drosophila (6) illustrates the complexity of this organelle, whose dynamic nature is far from being understood fully.
LDs are formed in two very different environmental conditions and, presumably, the physiological significance in each case is different. First, cells accumulate LD in response to exogenous lipid availability (4), present in serum lipoproteins or as free fatty acids. There is general agreement that LD content arising from the medium has a storing purpose for energy generation and membrane building. Using this experimental paradigm of exogenous lipid loading, we have shown the implication of group IVA PLA 2 (cPLA 2 ␣) in LD biogenesis at a step beyond the synthesis of TAG (7). Second, many kinds of cellular stress, including inflammation, apoptosis induced by different insults or contact inhibition, also induce LD biogenesis. An account of this situation is the human T lymphoblastoid cell line (HuT 78) undergoing apoptosis triggered by Fas antibody (8). In these cases, the metabolic origins of LD-associated neutral lipids and the physiological function they subserve are not known.
Altering phospholipid metabolism has been shown to cause programmed cell death in some instances (9 -12), and, conversely, programmed cell death often alters phospholipid metabolism, resulting in the accumulation of lysophospholipids and fatty acids (13). Current evidence suggests that group VIA PLA 2 (iPLA 2 -VIA) is involved in the generation of lysophosphatidylcholine during programmed cell death, which may mediate attraction and recognition/engulfment signals for apoptotic cell clearance by phagocytes (13)(14)(15)(16)(17). Unlike cPLA 2 ␣, which shows a marked preference for arachidonic acid-containing phospholipids, iPLA 2 -VIA has no substrate specificity regarding the fatty acid residue at the sn-2 position, and its implication in eicosanoid synthesis may be minor. Instead, the enzyme has a housekeeping role mediating phospholipid remodeling through deacylation/reacylation reactions (18). Bearing this in mind, we hypothesized that during cellular stress part of the fatty acids released by phospholipid degradation could be incorporated into TAG and stored in LD. In essence, our results show that TAG synthesis associated with the formation of LD during stress depends on iPLA 2 -VIA. Further, we show that the implication of cPLA 2 ␣ in LD formation still holds in this new paradigm.
Cells-CHO-K1 cells were cultured as described (7). For the experiments, cells were seeded at a density of 30,000 cells/ml in 24-(0.5 ml) or 6-well (2 ml) plates and maintained in FBS-containing medium for 24 h. Before LD induction, cells were switched to serum-free medium for 24 h to set control conditions with minimal occurrence of LD (7). When indicated, cells (40 -70% confluence) were transfected with 1 g of plasmid/ml (pGFP-C3 from Clontech or iPLA 2 -VIA2-EYFP) using Lipofectamine Plus TM (Invitrogen), following the manufacturer's instructions. Nile Red Staining and Fluorescence Microscopy-Cells cultured on glass coverslips were washed with phosphate-buffered saline (PBS, Sigma-Aldrich), fixed with 3% paraformaldehyde for 10 min, and washed twice with PBS. Cells were overlaid with 0.5 ml of PBS, to which 2.5 l of a solution of Nile Red in acetone (0.2 mg/ml) was added, so that the final concentrations of Nile Red and acetone were 1 g/ml and 0.5%, respectively. Samples were kept in the dark until photographed in a Leica Qwin 500 microscope with a Leica DC200 camera, using the Leica DCviewer 3.2.0.0 software.
Flow Cytometry-Indirect quantification of LD by flow cytometry in Nile Red-stained cells was performed exactly as described (7). Briefly, paraformaldehyde-fixed cells were stained with 1 g/ml Nile Red during 45 min and analyzed with a Cytomics FC 500 (Beckman Coulter) equipped with an argon laser (488 nm), in the FL1 channel (505-545 nm). After gating out cellular debris, 30,000 events where acquired in all the assays, in linear scale. Fluorescence intensities in the different treatments were quantified as the median value of each distribution of events, and expressed as increase above the median of the control (serum-starved cells), which do not contain LD (7).
[ 3 (20). To separate the major lipid species, the chloroform phases were evaporated under vacuum, dissolved in 15 l of chloroform/methanol (3:1, v/v), and spotted onto silica gel G TLC plates (Merck), which were developed in hexane/diethyl ether/acetic acid (70:30:1, v/v), and stained with primuline spray. Identification of the major species was made by comigration with authentic standards. Quantification of radioactive TAG was done by scraping into vials the silica gel from regions corresponding to migration of the standards. Primuline-stained TAG was quantified by densitometric analysis after acquiring images under UV light.
Alkaline Hydrolysis of [ 3 H]TAG-Serum-starved cells in 6-well plates were labeled with [ 3 H]palmitate (5 Ci/ml or 0.1 M) and treated with cerulenin. Pooled lipid extracts from 6 wells were separated by TLC. Silica gel containing [ 3 H]TAG was scraped into glass tubes, and lipid was extracted with 1 ml of chloroform/methanol (1:1, v/v). 2 ml of chloroform and 0.5 ml of 1 M KOH in methanol/water (19:1, v/v) were added, and the tubes were left 1 h at room temperature. After neutralization with HCl, 2.5 ml of chloroform, 1.5 ml of methanol, and 1.5 ml of water were added to split organic and aqueous phases.
Calcium Imaging-Cells grown onto polylysine-coated coverslips were incubated with the calcium indicator Fura-2/AM at 4 M in Krebs buffer of the following composition (in mM): 119 NaCl, 4.75 KCl, 5 NaHCO 3 , 1.2 MgSO 4 , 1.18 KH 2 PO 4 , 1.3 CaCl 2 , 20 Hepes, and 5 glucose, pH 7.4). After 1 h, cells were washed and coverslips were mounted in a static chamber on an inverted Nikon TE2000U microscope of a conventional epifluorescence system. Cells were excited alternatively at 340 and 380 nm, and emission light was collected at 510 nm every 4 -10 s using a 12-bit charge-coupled device ERG ORCA Hamamatsu camera. A ratio image of cells was analyzed using the Metafluor software (Universal Imaging). 14 -20 cells were analyzed in each experiment.
Confocal Microscopy-Serum-starved cells were labeled for 24 h with 1 M C 1 -BODIPY 500/510-C 12 . After washing with medium, cells were treated with drugs for 6 h, fixed, and photographed in a Leica TCS SP2 AOBS confocal microscope.
Statistical Analysis-Data analysis was carried out with Prism software (GraphPad). Responses among different treatments were analyzed with one-way analysis of variance followed by the Bonferroni multiple comparison test.

Increase of the m edian above control
Ceru C75 U73122  synthesis of fatty acids is abolished by the treatments, and there is no external source of lipid, making it simpler to consider possible sources of LD content. In these conditions, 30 M cerulenin and 20 M C75 reduced cell viability 50% over 24 h, and inhibited acetate conversion to fatty acids 85-90% over a 6-h treatment (Fig. 1). The same 6-h treatments, over the duration of time at which cell viability was reduced 10% as assessed by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reduction assay (not shown), induced the biogenesis of LD and increased levels of the LDassociated protein adipophilin (ADRP) (Fig. 2). Indirect quantification by flow cytometry showed that the occurrence of LD was about half that found after treatment with FBS (Fig. 2E). Unlike the induction of LD biogenesis with FBS, which depends on the uptake of serum lipoproteins (7), LD induction by FAS inhibitors in FBS-free medium necessarily has to take place from endogenous sources. To find out whether LD induction by FAS inhibitors takes place with synthesis of TAG, we measured total TAG content and synthesis. As shown in Fig. 3 (A 3, C and D), demonstrating new TAG synthesis from pre-existing fatty acids. In contrast, LD induction with serum did not involve synthesis of TAG from endogenous precursors. Alkaline deacylation of radioactive TAG confirmed its origin from pre-existing fatty acids and that the radioactive label was not incorporated into glycerol (Fig. 3, F  and G).

Induction of LD Biogenesis and TAG Synthesis during Stress Does Not
Require Fatty Acid Synthesis-The observation that LD biogenesis and TAG synthesis occur in the absence of fatty acid synthesis made us consider whether this would hold in other models of cellular stress that do not involve the inhibition of FAS. To test this, we exposed the cells to various kinds of stress: pH 6, which has been shown to induce LD (22,23), C 2 -ceramide, which induces cell death after the early inhibition of phosphatidylcholine synthesis (9, 10), tunicamycin, which inhibits N-glycosylation and induces unfolded protein response (24,25) and autophagy (26), and deprivation of glucose, a condition that induces apoptosis and autophagy (27)(28)(29). As shown in Fig. 4, all the treatments induced LD and stimulated the synthesis of TAG, and there was a high correlation (r 2 ϭ 0.952) between TAG synthesis and LD promoted by all the stress inducers (Fig. 4N). Treatment with stress inducers in the presence of FAS inhibitors did not decrease LD formation or TAG synthesis but actually increased both effects. These results confirm that LD biogenesis and TAG synthesis during cellular stress do not require fatty acid synthesis.
Phospholipases D and C Are Not Involved in TAG Synthesis and LD Biogenesis during Stress-To address the origin of TAG we considered the role of phospholipases D and C, which generate phosphatidic acid and 1,2-diacylglycerol (DAG), respectively. Formation of these products could account for the incorporation of phospholipid-linked fatty acids into TAG, either directly or after the action of phosphatidate phosphohydrolase (PAP). To test this, we treated serum-starved cells with cerulenin or C75 and in the presence of the phospholipase C inhibitor U73122 (30) or 1-butanol (31), which competes with water in a phospholipase D-catalyzed transphosphatidylation reaction, generating phosphatidylbutanol instead of phosphatidic acid. These compounds did not affect LD biogenesis or TAG synthesis (Fig. 5, A and B), yet butanol inhibited phospholipase D-generated phosphatidic acid while increasing phosphatidylbutanol, and U73122 blocked inositol 1,4,5-trisphosphate-mediated calcium responses (Fig. 5, C and D).
Groups VIA and IVA Phospholipase A 2 (iPLA 2 -VIA and cPLA 2 ␣) Are Involved in TAG Synthesis and LD Biogenesis, Respectively-The release of fatty acids by a PLA 2 -mediated process might also account for TAG synthesis from pre-formed fatty acids. Fig. 6A shows that treatment with cerulenin or C75 stimulated the release of AA, in a manner that was sensitive to the cPLA 2 ␣-selective inhibitor py-2 at 1 M and to the iPLA 2 inhibitor BEL at 10 M, suggesting at first glance the activation of cPLA 2 ␣ and iPLA 2 . AA, however, was not incorporated into TAG (Fig. 6C). In contrast, palmitic acid was not released to the medium (Fig. 6B) but instead was used for TAG synthesis in a manner that was inhibited by BEL but not by py-2. This pharmacological evidence suggests the role of iPLA 2 but not cPLA 2 ␣ in providing fatty acids for TAG synthesis. On the other hand, py-2 inhibited LD biogenesis induced by cerulenin or C75 (Fig.  6K). This agrees with our previous finding that LD biogenesis from exogenous lipid requires cPLA 2 ␣ (7) and further validates it when LDs are induced by cellular stress. Interestingly, also  BEL inhibited LD formation induced by cerulenin or C75, monitored by microscopic examination (Fig. 6, E-J), flow cytometry (Fig.  6K), or by the expression of ADRP (Fig. 6L). Inhibition of iPLA 2 also precluded TAG synthesis and LD biogenesis induced by acidic pH, C 2 -ceramide, tunicamycin, or deprivation of glucose (not shown). The iPLA 2 family consists of two members in mammals, namely Group VIA PLA 2 (iPLA 2 -VIA) and Group VIB PLA 2 (iPLA 2 -VIB) (14). The S-enantiomer of BEL has been shown to be more potent on iPLA 2 -VIA than on iPLA 2 -VIB, whereas the R-isomer was more potent on iPLA 2 -VIB (32). In our system, 10 M S-BEL blocked incorporation of palmitate into TAG, and the generation of LD as stimulated by cerulenin or C75, but the same concentration of R-BEL was without effect (Fig. 6, M and N). Taken together, these data suggest the implication of iPLA 2 -VIA in the release of fatty acids required for TAG synthesis in our model. Cerulenin and C75 promoted the phosphorylation of cPLA 2 ␣ at Ser 505 , which is necessary for enzyme activation during LD biogenesis (7), and it was inhibited to some degree by BEL (Fig.  6L), suggesting that iPLA 2 -VIA acts upstream of cPLA 2 ␣. With this in mind, and taking into account that cPLA 2 ␣ is specific for AA whereas iPLA 2 shows no substrate specificity for the fatty acid at the sn-2 position (14), the inhibition of AA release by BEL (Fig. 6A) agrees with iPLA 2 -VIA playing a cPLA 2 ␣-activating role.
Down-regulation of the Expression of iPLA 2 -VIA Abrogates TAG Synthesis and LD Biogenesis-A major precursor of DAG for TAG synthesis is phosphatidic acid, which is dephosphorylated by PAP. Although BEL is very specific for the Group VI enzymes of the PLA 2 family, the drug also inhibits PAP (18,33,34). A more specific approach to inhibit iPLA 2 -VIA came from the use of three siRNA

Group VIA PLA 2 and LD Biogenesis in Stress
designed against the human sequence. iPLA 2 -VIA occurs in human tissues in at least five different splicing variants but only two have enzymatic activity, termed VIA-1 and VIA-2 (14). Although iPLA 2 -VIA-1 appears to be the only variant expressed in CHO cells (35), the siRNA sequences were designed against the common exons to ensure downregulation of both variants.
As shown in Fig. 7A, all three siRNA reduced iPLA 2 -VIA expression. Unlike cells transfected with control siRNA, which when challenged with cerulenin or C75 increased expression of ADRP (Fig.  7A), synthesis of TAG (Fig. 7B), and LD (Fig. 7C), knockdown of iPLA 2 -VIA rendered the cells unresponsive to cerulenin or C75 in terms of ADRP expression (Fig. 7A), TAG synthesis (Fig. 7B), and LD biogenesis (Fig. 7C), but it did not affect LD biogenesis induced by FBS (Fig. 7D) or increased expression of ADRP (not shown). Again, the results show that, unlike the paradigm of LD formation from exogenous lipid, iPLA 2 -VIA is required for TAG synthesis taking place during LD biogenesis under stress.
Overexpression of iPLA 2 -VIA Increases TAG Synthesis and LD Occurrence-Transient transfection of a construct encoding a fusion protein consisting of iPLA 2 -VIA2 followed by EYFP (iPLA 2 -VIA2-EYFP) (Fig. 8A) increased free fatty acid availability, as evidenced in the presence of the acyl-CoA synthetase inhibitor triacsin C (Fig. 8B). In contrast to serum-deprived GFP-transfected cells, which had little LD content (Fig. 8, C and F), overexpression of iPLA 2 -VIA-EYFP induced LD in serum-deprived conditions (Fig. 8, D and F) to levels higher than those attained in non-transfected cells under stress (compare with Fig. 4). As expected, this high LD content was abolished by 10 M BEL (Fig. 8,  E and F). Increased LD occurrence under basal conditions in iPLA 2 -VI-A-EYFP-transfected cells also mirrored the expression levels of ADRP (Fig. 8G) and TAG synthesis (Fig. 8I). Again, increased TAG synthesis was blocked by BEL but not by py-2 (Fig. 8I), FIGURE 8. Overexpression of iPLA 2 -VIA increases TAG synthesis and LD biogenesis. Cells were transfected with GFP or with a fluorescent fusion protein containing group VIA2 human iPLA 2 followed by EYFP (GVIA2-iPLA 2 -EYFP), as shown in panel A. GVIA2-iPLA 2 -EYFP-transfected cells that were labeled 24 h with [ 3 H]palmitate had increased free fatty acid content after a 6-h treatment with triacsin C (B). Unlike cells transfected with GFP, which had little LD content in the absence of stress inducers (C), LD occurrence in GVIA2-iPLA 2 -EYFP-transfected cells was well apparent (D) and sensitive to BEL inhibition (E). GVIA2-iPLA 2 -EYFP-transfected cells also had increased ADRP expression that was BEL-sensitive (G). LD indirect quantification by flow cytometry (F) shows that increased LD occurrence in GVIA2-iPLA 2 -EYFP-transfected cells was inhibited also by the cPLA 2 inhibitor py-2. Radioactive TAG also increased in [ 3 H]palmitate-labeled cells (I), but unlike LD occurrence, this effect was not inhibited by py-2. Increased LD in GVIA2-iPLA 2 -EYFP-transfected cells was mirrored by increased cPLA 2 activity (J) and phosphorylation at Ser 505 (H) that were inhibited by BEL. Overexpression of iPLA 2 -VIA also enhanced LD biogenesis (K) and TAG synthesis (L) during a 6-h treatment with cerulenin. H]palmitate, they were washed with medium and left untreated (control) or treated for 6 h with 20 M C75 alone or in combination with 1 M py-2. Cells were then disrupted by a 10-s sonication and centrifuged 60 min at 20,000 ϫ g to obtain pellet and supernatant fractions that were enriched in flotillin-1 and GAPDH, respectively (E). Radioactive TAG in the supernatant increased 4-fold due to C75 treatment, but this was reduced by half after the inhibition of LD biogenesis with py-2 (F). Inset in panel F shows that py-2 treatment did not affect total [ 3 H]TAG. *, significantly different (p Ͻ 0.05).
whereas increased LD occurrence was inhibited by BEL and py-2 (Fig. 8F). Importantly, overexpression of iPLA 2 -VIA increased the phosphorylation of cPLA 2 ␣ at Ser 505 (Fig. 8H) and the release of AA (Fig. 8J), in a BEL-sensitive fashion, the latter effect being also sensitive to py-2. Treatment of iPLA 2 -VIA-EYFP-transfected cells with cerulenin further increased LD occurrence and TAG synthesis (Fig. 8, K and L). Taken together, these results confirm the implication of iPLA 2 -VIA in TAG synthesis and LD biogenesis, and suggest again a role for the enzyme in the events leading to phosphorylation and activation of cPLA 2 ␣.
TAGs in LD Arise from iPLA 2 -VIA-The previous results demonstrate that LD biogenesis during stress takes place with TAG synthesis after fatty acids released by iPLA 2 -VIA and strongly suggest that the newly formed TAGs are present in LD. To further assess this, we labeled cells for 24 h with the fluores-cent fatty acid C 1 -BODIPY 500/ 510-C 12 and treated them with C75. As shown in Fig. 9 (A-D), C75 induced the formation of fluorescent LD, whereas iPLA 2 -VIA inhibition retained the fluorescent label in perinuclear membranes. Similarly, when LD biogenesis was induced by FBS, which is an iPLA 2 -VIA-independent process, fluorescence remained associated to membranes. A more direct evidence was obtained from a fractionation experiment in [ 3 H]palmitate-labeled cells: after a 24-h labeling period, LD were induced with C75, and cells were disrupted and centrifuged to obtain crude supernatant and membrane fractions. ADRP was induced by C75 treatment, and it was found mainly in the supernatant (Fig. 9E), indicating the enrichment of this fraction in LD. We have reported that inhibition of cPLA 2 ␣ during exogenous lipid loading results in the retention of TAG in membrane fractions (7). In agreement with this, the increase in [ 3 H]TAG in the LD-enriched supernatant induced by C75 was partially inhibited by py-2, whereas the converse was true regarding the membrane fraction (Fig. 9F). Taken together, the results show that TAG in LD formed during cellular stress arise, at least in part, from phospholipid hydrolysis by iPLA 2 -VIA.
LD Biogenesis during Glucose Deprivation May Be a Survival Strategy-To address a physiological significance of LD biogenesis during stress, we monitored cell viability after deprivation of glucose under conditions that prevent LD formation. As shown in the preceding experiments, this can be done either by inhibiting iPLA 2 -VIA, which prevents TAG synthesis and cPLA 2 ␣ activation, or by inhibiting cPLA 2 ␣, which does not affect TAG synthesis but blocks LD biogenesis. Other than iPLA 2 , BEL also inhibits PAP, which that can account for the toxicity of this compound (12,13). In fact, in CHO-K1 cells maintained in serum-free medium, 10 M BEL decreased cell viability some 70% in 8.5 h (Fig. 10A). We therefore chose cPLA 2 ␣ inhibition with 1 M py-2, which is not toxic over a 24-h period (Fig. 10A, left panel). When GFP-transfected cells were switched to glucose-free Krebs buffer, viability decreased 30% at 8.5 h and the presence of py-2 decreased it further 20% (Fig. 10A, middle panel). Glucose deprivation during 8. VIA-EYFP (Fig. 10A, right panel), but py-2 still was toxic (Fig.  10A, right panel). Also, glucose deprivation over 20 h induced cellular shrinkage in GFP-transfected cells (Fig. 10B), which underwent further shrinkage and nuclear condensation in the presence of the cPLA 2 ␣ inhibitor (Fig. 10C). During glucose starvation, py-2 also induced cellular shrinkage in iPLA 2 -VIA-EYFP-transfected cells (Fig. 10E), which were morphologically unaltered in the absence of the inhibitor (Fig. 10D). These results suggest that LD biogenesis during glucose deprivation has a survival effect.

DISCUSSION
Our results show that in the absence of de novo synthesis and with no external source of fatty acids, induction of cellular stress leads to the synthesis of TAG from pre-existing fatty acids in a process requiring iPLA 2 -VIA, and to LD biogenesis, which requires cPLA 2 ␣. In contrast, induction of LD by FBS is independent of iPLA 2 -VIA. Further, phosphorylation and activity of cPLA 2 ␣ decreases after iPLA 2 -VIA inhibition and increases after iPLA 2 -VIA overexpression, suggesting that enzyme control mechanisms for TAG synthesis and LD generation operate in a concerted manner.
LD biogenesis and synthesis of TAG is a hallmark of cellular stress, and it was suggested that alterations in phosphatidylcholine metabolism could be the source of TAG (8,36). TAG synthesis takes place from DAG and a CoA-activated fatty acid, by the action of DAG acyltransferases (37). In proliferating cells, acetyl-CoA carboxylase and fatty acid synthase control the supply of de novo DAG, which can be converted either to TAG or phospholipids. In fact, when phosphatidylcholine biosynthesis decreases, TAG accumulates (38). Iorio and colleagues (8) proposed that the mechanism for TAG synthesis during programmed cell death could involve a reduction in phospholipid biosynthetic pathways and/or the activation of specific phospholipases. Both mechanisms could increase DAG that would account for LD biogenesis at the expense of phospholipid pools. As for the second mechanism, we show that phospholipases C and D are not implicated. Regarding the first mechanism, a very early inhibition of the CDP-choline pathway for phosphatidylcholine synthesis has been shown to occur in some models of toxicity with the amphiphilic drugs ET-18-OCH 3 (39), hexadecylphosphocholine (40), or C 2 -ceramide (9, 10), or with cerulenin (38), and in fact we show that the two latter drugs induce TAG synthesis and LD formation. Further, CHO-MT58 cells, which express a thermosensitive CTP:phosphocholine cytidylyltransferase, divert newly synthesized DAG to the TAG pool when cultured at the restrictive temperature (38), and we have observed that under these conditions they develop LD. 6 Nevertheless, it is unlikely that all kinds of cellular stress inhibit phospholipid synthesis. Also, it is difficult to envisage how cerulenin or C75-mediated inhibition of fatty acid synthase could increase de novo formation of DAG unless fatty acids come from alternative sources. In this regard, our results show that (a) fatty acid availability increases after transfection of iPLA 2 -VIA, and (b) TAG synthesis takes place from pre-existing fatty acids and is dependent on iPLA 2 -VIA. We have been unable, however, to detect decreases in total phospholipids mirroring the increase in TAG synthesis (Fig. 3E), even with an additional 24-h incubation of the cells in fresh medium after labeling with [ 3 H]palmitate, to allow full incorporation of the tracer (not shown). This suggests that iPLA 2 -VIA provides fatty acids from a minor phospholipid pool.
Little is known about the control iPLA 2 -VIA (14). Various splice variants coexist in cells, and two of them (VIA-1 and VIA-2) have enzymatic activity. The enzyme contains eight (VIA-1) or seven (VIA-2) ankyrin repeats, which may allow oligomerization and full enzymatic activity. These two variants work similarly in our hands, because iPLA 2 -VIA-1 is the main form expressed in CHO cells, and transfection of iPLA 2 -VIA-2 has the same effect regarding TAG synthesis for the biogenesis of LD. Another two variants lack enzymatic activity but may act as dominant-negative inhibitors by precluding association of the active variants (41). It would be interesting to determine whether differential expression of active and inactive variants may account for LD generation during stress. In U937 leukemia cells undergoing apoptosis, iPLA 2 -VIA is cleaved by caspase-3, generating a fragment responsible for increased activity (42). This is not the mechanism taking place in our model, because we have observed that (a) treatment of CHO-K1 cells with 2 M actinomycin D, which induces strong caspase-3 proteolytic processing and activity, does not induce LD formation, and (b) LD induction by cerulenin is not inhibited by co-treatment with the pan-caspase inhibitor benzyloxycarbonyl-VAD-fluoromethyl ketone at 100 M (not shown).
Among the different groups of PLA 2 enzymes, iPLA 2 -VIA plays a significant role in the Lands cycle of basal phospholipid remodeling through deacylation/reacylation reactions (18,43). An account of the role of iPLA 2 -VIA in phospholipid homeostasis is the observation that overexpressing CTP:phosphocholine cytidylyltransferase increases phosphatidylcholine content to a much lesser extent than expected, due to accelerated degradation by iPLA 2 -VIA, which appears to respond to excess phospholipid synthesis (44,45). This housekeeping role of iPLA 2 -VIA, which opposes that of CTP:phosphocholine cytidylyltransferase in membrane phospholipid turnover, is important for cell cycle progression (46). Also, iPLA 2 -VIA-mediated phospholipid remodeling allows mitochondrial membrane repair during oxidative damage (47). Release of fatty acids in U937 cells undergoing apoptosis is mediated by iPLA 2 -VIA (42,48), and in this case the enzyme works promoting cell death. Also, overexpression of iPLA 2 -VIA in the same cells accelerates apoptosis induced by oxidative stress (49). In both cases, however, pharmacological inhibition or down-regulated expression did not rescue at longer times, showing that the enzyme is not determinant for apoptosis. Rather, it is responsible for the generation of lysophosphatidylcholine, which is recognized by macrophages for clearance of dying cells (13)(14)(15)(16)(17). Transgenic mice expressing human iPLA 2 ␥ (iPLA 2 -VIB) in a cardiac myocyte-selective manner accumulate great amounts of TAG during a brief caloric restriction, however this enzyme contributes to mitochondrial dysfunction (50). Our data show that overexpression of iPLA 2 -VIA delays cell death under glucose deprivation; further, preventing LD biogenesis after the inhibition of cPLA 2 ␣ opposes the protection elicited by iPLA 2 -VIA. Taken together, the results suggest a surviving role of LD biogenesis during glucose deprivation.
LD biogenesis has a protecting role against lipotoxicity derived from fatty acid accumulation (51,52). This condition requires the overload of high concentrations of exogenous fatty acids and does not apply to our experimental paradigm. Mammals survive starvation by activating proteolysis and lipolysis. Other than muscle, where protein is degraded mainly by the proteasome system and is the major source of precursors for hepatic gluconeogenesis, fasting activates autophagy in most tissues (53). Lipolysis, on the other hand, takes place primarily in adipocytes, which are the major source of circulating free fatty acids. This mechanism has received most attention over the past years (1), but all tissues and cells types must be able to release fatty acids from TAG packed in LD (54), conceivably to supply local energetic needs through ␤-oxidation. In this regard, a recent report (55) shows that neurons from female rats form LD and survive nutrient deprivation better than neurons from male animals, which undergo autophagic death. Our results agree essentially with this report, suggesting that LD biogenesis in cells under stress may represent a lipid counterpart of autophagy: fatty acids used for structural functions (membrane phospholipids) could be reused for TAG synthesis in a process mediated by iPLA 2 -VIA and packed in LD for energetic functions in a process requiring cPLA 2 ␣.