Cytosolic phospholipase A₂ regulates Golgi structure and modulates intracellular trafficking of membrane proteins

The Golgi complex and the trans-Golgi network are critical cellular organelles involved in the endocytic and biosynthetic pathways of protein trafficking. Lipids have been implicated in the regulation of membrane-protein trafficking, vesicular fusion, and targeting. We have explored the role of cytosolic group IV phospholipase A₂ (cPLA₂) in membrane-protein trafficking in kidney epithelial cells. Adenoviral expression of cPLA₂ in LLC-PK₁ kidney epithelial cells prevents constitutive trafficking to the plasma membrane of an aquaporin 2-green fluorescent protein chimera, with retention of the protein in the rough endoplasmic reticulum. Plasma membrane Na⁺-K⁺-ATPase α-subunit localization is markedly reduced in cells expressing cPLA₂, whereas the trafficking of a Cl^–/HCO₃^– anion exchanger to the plasma membrane is not altered in these cells. Expression of cPLA₂ results in dispersion of giantin and β-COP from their normal, condensed Golgi localization, and in marked disruption of the Golgi cisternae. cPLA₂ is present in Golgi fractions from noninfected LLC-PK₁ cells and rat kidney cortex. The distribution of tubulin and actin was not altered by cPLA₂, indicating that the microtubule and actin cytoskeleton remain intact. Total cellular protein synthesis is unaffected by the increase in cPLA₂ activity. Thus cPLA₂ plays an important role in determining Golgi architecture and selective control of constitutive membrane-protein trafficking in renal epithelial cells.

The polarity of epithelial cells derives from the selective distribution of their plasma-membrane proteins and lipids into distinct plasma-membrane domains (1, 2). Intracellular protein traffic from the endoplasmic reticulum (ER) to the cell surface through the Golgi apparatus is mediated by vesicles that transit between the organelles and different membrane domains. The newly synthesized membrane proteins may be directly addressed to their final destination from the ER (constitutive pathway) or stored in secretory vesicles until an extracellular stimulus induces their release to the membrane (regulated pathway) (1, 3). These pathways have crucial roles in the physiology of polarized cell function in the normal kidney and are perturbed in some pathological states. The asymmetry between basolateral and apical surfaces depends on mechanisms that control the movement of proteins to the correct target membrane in cells (1, 4). Despite intensive research, those processes remain poorly understood.

In addition to their classic roles in the production of second messengers in various aspects of signal transduction (5), lipids of intracellular compartments have been implicated in the regulation of membrane trafficking, vesicular fusion, and targeting (6). Phospholipases have also been implicated in this regulation. Phosphatidic acid, the main metabolite produced by hydrolysis of phospholipids by phospholipase D (PLD), has been proposed to have a direct action on target proteins involved in vesicle movement and to modify the lipid bilayer content and surface charge of membranes (7, 8).

The group IV 85-kDa cytosolic phospholipase A₂ (cPLA₂) acts at the 50-2 position of phospholipids to generate free arachidonic acid (AA). AA and its metabolites are important lipid second messengers (5, 9, 10). cPLA₂ is regulated by changes in intracellular free-calcium concentration within the physiological range. After release from phospholipids by PLA₂, AA can be converted by cyclooxygenases, lipoxygenases, and cytochrome P-450 monooxygenase to prostaglandins, leukotrienes, and hydroxyeicosanoic acids, products that are involved in the regulation of a number of cellular processes (5, 11). Lysophospholipids generated in vesicle membranes by PLA₂ may regulate the fusion of ER-transport vesicles with the Golgi apparatus (12). Furthermore, Tagaya et al. showed that PLA₂ regulates intra-Golgi protein transport in vitro (13).

We have shown recently that aquaporin 2 (AQP2)/green fluorescent protein (GFP) chimeras are useful constructs for studying intracellular sorting signals that direct this water-channel protein to specific cellular locations (14). When GFP is fused to the cytoplasmic amino-terminus of AQP2, GFP-AQP2(NT), the water channel enters a vasopressin-regulated (VP-regulated) pathway of plasma membrane insertion. In contrast, the COOH-terminal chimera, AQP2-GFP(CT), localizes constitutively on both apical and basolateral plasma membrane, independently of VP or forskolin stimulation, and traffics in a way that is indistinguishable from wild-type AQP1 (14).

Another protein that is constitutively targeted to the cell membrane in polarized epithelial cells is Na⁺-K⁺-ATPase, whose correct targeting is critical to normal cell function. Dopamine, which increases cPLA₂ activity, induces a decrease in Na⁺-K⁺-ATPase activity in proximal tubules with internalization of its α- and β-subunits into endosomes via a clathrin-dependent pathway (15).

The goals of this study were to evaluate the role of cPLA₂ in membrane-protein trafficking in a polarized renal epithelial cell. In cells expressing cPLA₂, cell volume is increased, and plasma membrane Na⁺-K⁺-ATPase α-subunit is markedly reduced. The amount of AQP2-GFP(CT) is decreased in the plasma membrane and increased in the rough ER. Golgi cisternae are markedly disrupted, and giantin and β-COP are dispersed from their normal, condensed Golgi localization. In contrast, there is normal VP-stimulated membrane localization of GFP-AQP2(NT) and normal constitutive membrane localization of a Cl^–/HCO₃^– anion exchanger. Furthermore, the distribution of tubulin and actin is not altered by cPLA₂, indicating that the microtubule and actin cytoskeleton remain intact. These results suggest that cPLA₂ plays an important role in the selective regulation of constitutive membrane-protein trafficking by its action on Golgi structure and function.

PLA₂ expression in cells infected with AdcPLA₂. There is a dose-dependent increase in cPLA₂ activity in AQP2-GFP(CT) LLC-PK₁ cells infected at a moi of 0 (uninfected controls), 10, 50, 100, 200, and 400 pfu/cell for 48 hours with AdcPLA₂. No increase is seen in uninfected cells or cells infected with AdLacZ. Normal MDCK and rat kidney mesangial cells have cPLA₂ activity within the range seen in LLC-PK₁ cells infected with AdcPLA₂ at a moi of 50 to 100 pfu/cell (Figure 1a). Because we wanted to establish conditions in which the expression of cPLA₂ did not exceed that normally measured in other renal epithelial and mesangial cells, we chose a concentration of 50 pfu/cell for our studies. PMA and A23187 treatment results in more AA release into the medium in AdcPLA₂-infected cells than in control cells or cells infected with AdLacZ (Table 1). In cells infected with AdcPLA₂, PLA₂ activity increases with time, with a peak of activity between 3 to 6 days after infection (Figure 1b). Immunoblot analysis confirmed that the progressive increase in cPLA₂ activity of cells exposed to increasing moi of AdcPLA₂ was a direct consequence of progressive increases in the expression of cPLA₂ (Figure 1c). When infected with 50 pfu/cell, the total amount of cPLA₂ protein was less in the infected LLC-PK₁ cells than it was in the MDCK or mesangial cells.

Figure 1

(a and b) Assay of cPLA₂ activity in LLC-PK₁ cells expressing LacZ and cPLA₂ and in uninfected cells. (a) AQP2-GFP(CT) LLC-PK₁ cells were incubated with either AdLacZ or AdcPLA₂ at different moi (10, 50, 100, 200, and 400 pfu/cell). Forty-eight hours later, the cells were lysed and centrifuged at 100,000 g for 1 hour at 4°C. Each assay was performed three times in duplicate. Cell lysates of MDCK and rat mesangial cells (MC) were processed at the same time and used to compare the level of activity in the different cell lines. cPLA₂ activity was significantly greater (P < 0.01) in cells infected with AdcPLA₂ at 10, 50, 100, 200, and 400 pfu/cell when compared with control noninfected cells and AdLacZ infected cells. (b) cPLA₂ activity was determined 1, 2, 3, 6, 10, 15, and 21 days after infection with AdLacZ or AdcPLA₂ at 50 pfu/cell or in control cells (n = 3). AdcPLA₂-treated cells had significantly greater (P < 0.01) cPLA₂ activity at each time point when compared with AdLacZ and control cells. (c) Immunoblot of cPLA₂ from control cells and cells infected with AdLacZ and AdcPLA₂. Cells were infected with AdLacZ or AdcPLA₂ at different moi. Forty-eight hours later, the cells were solubilized, and 25 μg of each lysate were separated by SDS-PAGE and a Western blot generated with 1:5000 diluted polyclonal anti-cPLA₂ Ab.

Table 1

AA release in LLC-PK₁ cells expressing cPLA₂ or β-galactosidase

Greater than 90% of the cells infected with AdLacZ or AdcPLA₂ expressed either β-galactosidase or cPLA₂, respectively, as assessed by immunocytochemistry (data not shown). There was no measurable toxicity over 48 hours after infection as assessed by the absence of an increase of lactate dehydrogenase (LDH) release (data not shown).

AQP2-GFP(CT), but not GFP-AQP2(NT), trafficking is disrupted by cPLA₂. Constitutive and hormone-regulated pathways of protein trafficking were examined using GFP/AQP2 chimeras as model systems. The AQP2-GFP(CT) chimera has been shown previously to traffic constitutively to the membrane. Constitutive AQP2-GFP(CT)–membrane localization is normal in AdLacZ-infected cells (Figure 2, a and b). By contrast, in cells expressing cPLA₂, there is a considerable reduction of AQP2-GFP(CT) on the plasma membrane and an increase in perinuclear, intracellular staining both in the absence (Figure 2c), or presence, of VP (Figure 2d).

Figure 2

Immunofluorescence localization of AQP2-GFP(CT) in transfected LLC-PK₁ cells. Two days after infection with either AdLacZ or AdcPLA₂, cells were fixed and examined for GFP expression using indirect immunofluorescence. In control cells (a) and cells expressing β-galactosidase (b), AQP2-GFP(CT) is localized on the basolateral and apical membrane. By contrast, in cells expressing cPLA₂ in the absence (c) or presence (d) of VP (10 nM, 30 minutes), there is a complete loss of AQP2-GFP(CT) basolateral-membrane staining. Bar, 5μm.

GFP-AQP2(NT) localization in intracellular vesicles is unchanged under unstimulated conditions in cells infected with AdcPLA₂ (Figure 3a) (14). A shift from intracellular vesicles to the basolateral membrane is observed in these cPLA₂-expressing cells upon stimulation with 10 nM VP for 30 minutes (Figure 3b) in a pattern identical to that seen in noninfected cells (14).

Figure 3

Normal vesicular trafficking of GFP-AQP2(NT) in cPLA₂-expressing LLC-PK₁ cells. Two days after infection with AdcPLA₂, cells were fixed and examined for GFP expression by immunofluorescence. Under nonstimulated conditions (a), GFP-AQP2(NT) was localized to intracellular vesicles. After stimulation with 10 nM VP for 30 minutes (b), staining shifted to the plasma membrane.

Immunoblot analysis of membrane fractions. To determine the intracellular localization of AQP2-GFP(CT) and GFP-AQP2(NT) in cells expressing LacZ or cPLA₂, total cellular membranes were separated by ultracentrifugation on a continuous iodixanol density gradient according to the technique described by van’t Hof et al. (20). Fractions were analyzed for enzyme activities identifying plasma membrane, Golgi apparatus, and ER by the markers alkaline phosphatase, α-mannosidase, and α-glucosidase II, respectively. An equal sample of each fraction was separated by SDS-PAGE and blotted with a monoclonal anti-GFP Ab. In cells expressing either LacZ or cPLA₂, GFP-AQP2(NT) was localized principally within the intracellular fractions as expected, although some plasma-membrane localization was apparent (Figure 4a). In contrast, AQP2-GFP(CT) was expressed mainly in the plasma-membrane fractions in cells expressing LacZ, whereas in cells expressing cPLA₂, AQP2-GFP(CT) was shifted markedly to the ER fractions (Figure 4a). Marker-enzyme activities of each of the fractions are presented in Figure 4b.

Figure 4

Alteration in AQP2-GFP(CT) localization with cPLA₂. (a) Immunoblot of GFP-AQP2(NT) and AQP2-GFP(CT) in cellular-membrane fractions. Two days after infection with either AdLacZ or AdcPLA₂, cellular membranes were separated by ultracentrifugation on a continuous iodixanol density gradient. Each fraction was assayed for marker-enzyme activities of plasma membrane (alkaline phosphatase), Golgi apparatus (α-mannosidase II), and ER (α-glucosidase II). Proteins were separated by SDS-PAGE and detected with 1:1000 anti-GFP Ab. GFP-AQP2(NT) is expressed principally in the ER fractions in cells infected with AdLacZ or AdcPLA₂. AQP2-GFP(CT) is expressed principally in the plasma-membrane fractions in cells infected with AdLacZ, whereas this protein is principally detected in the ER fractions of cells expressing cPLA₂. (b) Analysis of marker-enzyme activities of each fraction of the density gradient. Cellular-membrane fractions from LLC-PK₁ cells expressing GFP-AQP2(NT) or AQP2-GFP(CT) and infected with either AdLacZ or AdcPLA₂ were separated as described in Methods. Gradients were fractionated from the top to the bottom (left to right in the figure) and 50 μl of each fraction was assayed for marker-enzyme analysis. Enzymatic activities were determined as described in Methods.

Effect of cPLA₂ on membrane localization of Na⁺-K⁺-ATPase and AE 1. To test whether cPLA₂ had an effect on the trafficking of other constitutive membrane proteins, LLC-PK₁ cells infected with either AdLacZ or AdcPLA₂ were stained with Ab’s against the α-subunit of Na⁺-K⁺-ATPase or AE1/2. In cells expressing cPLA₂, the normal Na⁺-K⁺-ATPase basolateral membrane localization (Figure 5a) was almost completely abolished, and it was difficult to tell where the protein was relocated (Figure 5b). However, AE1/2 basolateral localization was unaffected (Figure 5, c and d).

Figure 5

Immunofluorescence localization of Na⁺-K⁺-ATPase α-subunit and AE1/2 in AQP2-GFP(CT) LLC-PK₁ cells. Cells grown on coverslips were incubated with either AdLacZ or AdcPLA₂ at a moi of 50 pfu/cell. Two days later, cells were fixed and permeabilized with 0.1% Triton, followed by indirect-immunofluorescence staining with an Ab against either the α-subunit of Na⁺-K⁺-ATPase or AE1/2. The secondary goat anti-rabbit IgG Ab was coupled to Cy3. In cells infected with AdLacZ, as in control cells, Na⁺-K⁺-ATPase (a) and AE1/2 (c) are localized to the basolateral membrane. However, in cells expressing cPLA₂, the basolateral-membrane staining of Na⁺-K⁺-ATPase (b) was almost completely lost, whereas there was no change in the basolateral localization of AE1/2 (d). Bar, 3μm.

To confirm that the effect of cPLA₂ on trafficking of the Na⁺-K⁺-ATPase and the anion exchanger was not related to a difference in the stability of the proteins at the plasma membrane, cells expressing either LacZ or cPLA₂ were labeled with [³⁵S]-methionine. Two days after infection with either AdLacZ or AdcPLA₂, specific activities of postnuclear supernatants of LLC-PK₁ cells (Figure 6a), SDS-PAGE profiles of the metabolically labeled proteins (Figure 6b), and total amount of cellular Na⁺-K⁺-ATPase (Figure 6c) are not affected by expression of cPLA₂. Expression of cPLA₂ results in a marked inhibition of trafficking of Na⁺-K⁺-ATPase to the plasma membrane of LLC-PK₁ cells (Figure 6d). The trafficking of AE 1 protein to the plasma membrane is not affected in cPLA₂-expressing cells.

Figure 6

Inhibition of Na⁺-K⁺-ATPase trafficking in cPLA₂-expressing cells. Two days after infection with either AdLacZ or AdcPLA₂, LLC-PK₁ cells were metabolically labeled with 0.5 mCi [³⁵S]-methionine (6 hours of pulse followed by 1 hour of chase). (a) Specific activities of postnuclear supernatants of LLC-PK₁ cells, as well as (b) SDS-PAGE profiles of the metabolically labeled proteins, are not affected by expression of cPLA₂. (c) Western blot analysis indicates that total amount of cellular Na⁺-K⁺-ATPase is also unaffected after expression of cPLA₂. (d) Using plasma-membrane biotinylation and immunoprecipitation (see Methods), we see that expression of cPLA₂ results in marked inhibition of trafficking of [³⁵S]Na⁺-K⁺-ATPase to the plasma membrane of LLC-PK₁ cells (upper panel). The trafficking of [³⁵S]AE 1 protein to the plasma membrane is not affected in cPLA₂-expressing cells (lower panel).

Trafficking in cPLA₂-expressing cells treated with inhibitors of AA metabolism. To determine if the effect of cPLA₂ on protein trafficking was due to an increase of synthesis of one of the AA metabolites, 24 hours after infection with AdLacZ or AdcPLA₂ the cells were incubated with either indomethacin (2.5 × 10^–5 M), NDGA (10^–5 M), or SKF 525A (2.5 × 10^–5 M), cyclooxygenases, lipoxygenases, or cytochrome P-450 monooxygenase inhibitors, respectively. After 24 hours, the localization of AQP2-GFP(CT) and Na⁺-K⁺-ATPase was examined in controls and in cells expressing cPLA₂. Indomethacin, NDGA, and SKF 525A did not change the abnormal protein distribution of either AQP2-GFP(CT) or Na⁺-K⁺-ATPase α-subunit in the cells expressing cPLA₂ (data not shown). In addition, incubation of the cells with 5 μM AA for 2 or 4 days had no effect on the localization of either AQP2-GFP(CT) or AE1. To stimulate AA release, cells expressing cPLA₂ or β-galactosidase were incubated for 2 hours with 10 μM of A23187, a calcium ionophore. A23187 did not modify the localization of GFP-AQP2(NT), AQP2-GFP(CT), Na⁺-K⁺-ATPase, or AE1 in cells expressing β-galactosidase or cPLA₂ (data not shown).

Protein synthesis in cells expressing cPLA₂. To determine if the effect of cPLA₂ on protein trafficking was related to an inhibition of total protein synthesis, we studied [³H]-leucine incorporation 48 hours after infection of LLC-PK₁ cells with either AdLacZ or AdcPLA₂. There was no significant difference between control cells and cells expressing β-galactosidase or cPLA₂ (368 ± 40, 502 ± 49, and 471 ± 31 cpm/10³ cells, respectively). In addition, cycloheximide treatment (25 μg/ml) for 1–3 hours had no detectable effect on the distribution of GFP-AQP2(NT) and AQP2-GFP(CT) (data not shown).

Cell morphology and Golgi apparatus organization. The actin cytoskeleton and microtubule network appeared unaffected by cPLA₂, as determined by anti-tubulin and phalloidin labeling (data not shown). To examine the structure of the Golgi apparatus, anti–β-COP Ab was used to label coat proteins on transporting vesicles within the cis-medial Golgi apparatus, and anti-giantin Ab was used to label the Golgi cisternae. In LacZ control cells, β-COP and giantin are present in condensed, perinuclear patterns, as expected (Figure 7, a and c, respectively, for β-COP and giantin). In contrast, in cells expressing cPLA₂, β-COP and giantin localization is dramatically dispersed in a manner that suggests disruption of the Golgi organization (Figure 7, b and d). After transfection with a construct encoding a GFP-cPLA₂ fusion protein, there is marked disruption of the Golgi apparatus of cells that express the protein, whereas a mutant of cPLA₂, which is catalytically inactive, has no effect on β-COP distribution (Figure 8).

Figure 7

Immunofluorescence localization of β-COP and giantin distribution in AQP2-GFP(CT) LLC-PK₁ cells. (a) In cells infected with AdLacZ, β-COP localizes to a condensed Golgi region, whereas cPLA₂-expressing cells (b) exhibit a dramatic dispersion of β-COP. (c) In cells infected with AdLacZ, giantin localizes to the Golgi cisternae. (d) By contrast, giantin localization is completely dispersed within the cytoplasm in cPLA₂-expressing cells. Bar, 5 μm.

Figure 8

Catalytically deficient mutant of cPLA₂ has no effect on β-COP localization. Kidney LLC-PK₁ epithelial cells were transfected with pEGFP only (a), pEGFP/cPLA₂ (wt) (b), or with pEGFP/cPLA₂ (mut), encoding a fusion protein of GFP with the calcium lipid-binding region of cPLA₂ (c) and stained with Ab’s against β-COP. Left column represents GFP-expressing, GFP-cPLA₂ (wt)–expressing, and GFP-cPLA₂ (mut)–expressing cells, middle column shows endogenous β-COP distribution, and right column shows the merging of the two images in each row. Asterisks indicate transfected cells. Arrowheads indicate Golgi complex of nontransfected cells, and arrows indicate the Golgi structure in transfected cells in the same cell culture. Bar, 5 μm. wt, wild-type.

Using electron microscopy, the morphology of the Golgi apparatus was analyzed in LacZ- and cPLA₂-expressing cells. The Golgi stacks were well-formed cisternae in the β-galactosidase–expressing cells (Figure 9a, arrows), whereas they were scattered and vesiculated in cells expressing cPLA₂ (Figure 9b, arrows). Nuclear membranes and ER are intact.

Figure 9

Electron microscopy examination of cellular morphology AQP2-GFP(CT) LLC-PK₁ cells. (a) The Golgi complex (arrow) appears normal in cells infected with AdLacZ. (b) In contrast, the Golgi complex in cPLA₂-expressing cells was markedly disrupted and vesiculated. ER and nuclear membranes appear normal. Bar, 1 μM.

cPLA₂ is localized to Golgi apparatus of kidney cortex and noninfected LLC-PK₁ cells. Given the dramatic effects of cPLA₂ on Golgi structure, it is of importance to evaluate whether cPLA₂ is associated with the Golgi apparatus in normal renal epithelium. cPLA₂ colocalizes with postnuclear fractions that stain positively with anti–GP-58, a marker of the Golgi complex (Figure 10, a and b) in LLC-PK₁ cells. CA-2 was used as cytosolic marker and Rab-11 as recycling endosome marker. In addition, cPLA₂ is associated with Golgi complex isolated from the rat kidney cortex, which consists primarily of proximal epithelial cells (data not shown).

Figure 10

Localization of cPLA₂ in subcellular fractions of noninfected LLC-PK₁ cells. cPLA₂ colocalizes with GP-58, a marker of the Golgi complex. (a) Western blot analyses of LLC-PK₁ cell postnuclear iodixanol density-gradient fractions with cPLA₂, GP-58 (Golgi marker), CA-2 (cytosolic marker), and Rab-11 (recycling endosomes marker) Ab’s. (b) Densitometry analysis of distribution and colocalization of cPLA₂ with LLC-PK₁ cell Golgi complex.

Cell volume and endocytosis in cells expressing cPLA₂. The size of the cells increased after cPLA₂ expression. To quantitate this effect, 50 β-galactosidase–expressing cells and 50 cPLA₂-expressing cells were measured. By determining the two-dimensional size of cells expressing cPLA₂, the estimated volume was eightfold larger than that of cells infected with AdLacZ. Rhodamine conjugated to the cell-impermeant agent dextran was used to evaluate endocytosis. Normal endocytosis is found in cells expressing cPLA₂ (Figure 11). The glycoprotein characteristics of the apical surface of β-galactosidase– and cPLA₂-expressing cells are similar as evaluated by staining with wheat germ agglutinin conjugated to rhodamine (data not shown).

Figure 11

Double-labeling immunofluorescence demonstrating endocytosis of RITC-dextran in AQP2-GFP(CT) LLC-PK₁ cells. (a) AQP2-GFP(CT) localizes on the basolateral membrane and in a perinuclear vesicular pattern in AdLacZ-infected cells. (b) RITC-dextran is endocytosed into a perinuclear location in the same cells. (c) In cPLA₂-expressing cells, the basolateral membrane is devoid of AQP2-GFP(CT). (d) In the same cells as shown in (c), RITC-dextran is endocytosed into perinuclear vesicles. No consistent difference between AdLacZ and AdcPLA₂-infected cells was observed with regard to RITC-dextran endocytosis. Bar, 5 μM.

The polarized renal epithelial cell is an excellent model for studying differential targeting of proteins to cell membranes. Normal cell-membrane trafficking is necessary for the maintenance of transepithelial transport and other cellular functions such as volume regulation. Abnormal cell trafficking, including abnormal localization of the Na⁺-K⁺-ATPase, is characteristic of several renal diseases, including polycystic kidney disease (25) and pathophysiological states, such as ischemia/reperfusion injury (26).

Replication-deficient recombinant adenoviral vectors facilitate the titrated expression of a transgene in a high percentage of cells (> 90% in these studies) so that “physiological” levels of a protein are produced (27, 28). Using stable transfectants of GFP-AQP2 fusion proteins (14), we have found that cPLA₂ alters constitutive [AQP2-GFP(CT)] but not VP-regulated [GFP-AQP2(NT)] intracellular trafficking. To confirm that this effect was also relevant to endogenous proteins, it was demonstrated that basolateral membrane localization of Na⁺-K⁺-ATPase was almost completely abolished, whereas localization of the Cl^–/HCO₃^– exchanger (AE1/2) was unaffected. [³⁵S]-methionine labeling experiments confirm that trafficking of the α-subunit of Na⁺-K⁺-ATPase to the cell membrane is altered, whereas AE1 trafficking is intact, confirming that there is a level of specificity to the effects of cPLA₂ on constitutive trafficking pathways. The absence of an effect of cPLA₂ on VP-induced membrane expression of GFP-AQP2(NT) may be explained by the localization of this protein in preformed post-Golgi vesicles that traffic to the membrane upon addition to VP. This process is unaffected by cPLA₂ activity.

While the redistribution of basolateral Na⁺-K⁺-ATPase in the proximal tubule with ischemia has been attributed to disruption of epithelial-cell cytoskeletal structural organization (29, 30), our data suggest that cPLA₂-induced changes in intracellular vesicle–trafficking processes may also account for changes in distribution of Na⁺-K⁺-ATPase. The cPLA₂ activity is significantly increased in rat proximal tubule-enriched kidney homogenates after ischemia and reperfusion (31). The decrease in plasma membrane Na⁺-K⁺-ATPase cannot be due to proteolysis, since equivalent amounts of the protein are found in control and cPLA₂–expressing cells. The α-subunit is localized to intracellular membranes in cPLA₂-expresssing cells (data not shown). The decrease in Na⁺-K⁺-ATPase activity likely accounts for the increase in cell volume measured in the absence of morphological changes in actin cytoskeleton and microtubule network (32).

The cPLA₂–induced changes in trafficking are possibly due to a direct effect of the enzyme on lipid composition of the Golgi membrane or may be an effect of AA itself. The asymmetry of the lipid bilayer, with choline-containing phospholipids located preferentially in the external leaflet and the aminophospholipids in the cytoplasmic leaflet, is necessary to maintain the fundamental properties of the different intracellular compartments within the cell (33). Hydrolysis of membrane phospholipid into lysophospholipid and free fatty acid by cPLA₂ may modify the composition of the two leaflets of the phospholipid bilayer. In addition, the modulation of the lipid content induced by cPLA₂ could alter the substrate of phospholipase D, thus affecting phosphatidic acid production locally, with secondary effects on trafficking due to the change in this important lipid (33, 34).

The complex organization and dynamic nature of the Golgi subcompartments make it very difficult to study their lipid composition (35). Phospholipase D is present in Golgi membranes (36), is activated by ADP-ribosylation factor (37), and is responsible for catalyzing the transphosphatidylation reaction of phosphatidic acid and diacylglycerol to form bisphosphatidic acid (38). The removal of an acyl chain from bisphosphatidic acid leads to the production of semilysobisphosphatidic acid, which has been localized to the most dynamic pleiomorphic regions of the Golgi complex, the transvesicular networks on both sides of the Golgi stacks (35). Thus, cPLA₂ activity might alter the lipid balance of the Golgi apparatus in such a way that stack formation or steady-state integrity is not favored. Since inhibitors of cyclooxgenases, lipoxygenases, and P-450 monooxgenase, or incubation of the cells with AA for 2–4 days, did not modify or mimic the effects of cPLA₂, it is unlikely that an AA product, or AA itself, is responsible for the effects observed.

Membrane-associated or secreted proteins contain structural information that is recognized by the intracellular sorting machinery (39). Usually, this information is conveyed by short hydrophobic amino acid sequences in the cytoplasmic tail or phosphorylation of sugar residues or specific patterns of glycosylation, mainly in the rough ER and the Golgi apparatus (40). It is possible that cPLA₂ could act on enzymes involved in posttranslational glycosylation within the Golgi apparatus. In addition, by modifying Golgi lipid composition, cPLA₂ could modify the conformational environment of Golgi membranes, which may alter critical membrane-fission events (41) or composition of lipid rafts (42).

The cPLA₂-induced modification of the Golgi apparatus is similar to the Golgi complex fragmentation that occurs in mitotic cells (43). Since cPLA₂ is constitutively phosphorylated in mitotic cells (44), and cPLA₂ activity increases in proliferating cells (data not shown), it is possible that cPLA₂ is responsible, at least in part, for the Golgi complex fragmentation observed in mitosis.

As reported by de Figueiredo et al., the ability of brefeldin A to stimulate Golgi and trans-Golgi network tubulation was inhibited by a number of antagonists of PLA₂ (45). However, because some of the antagonists used in this work are not specific and could have exerted their effects through changes in phosphatidic acid metabolism and on more than one form of PLA₂ (46), it is not possible to attribute the changes observed to a particular enzymatic activity.

In conclusion, our studies demonstrate the potential importance of intracellular group IV cPLA₂ activity for Golgi apparatus structure and function. Our results also demonstrate the selectivity of these effects of cPLA₂ for the constitutive pathway of membrane-protein trafficking and for a subset of polypeptides. These effects of cPLA₂ are likely mediated by changes in Golgi membrane lipid composition and may also explain changes observed in the Golgi structure during the cell cycle.

This work was supported by grants from the NIH: DK-39773, DK-38452, and NS-10828 (to J.V. Bonventre); HL-50361 and HL-57623 (to R.J. Hajjar); HL-54202, HL-59521, and AI-40970 (to A. Rosenzweig); and DK-38452 (to D. Brown). G. Choukroun was supported, in part, by a Lavoisier grant of the French Government and by the International Study and Research grant of the French Lilly Institute. C. Gustafson was supported by an award from Astra Research Center Boston and by a fellowship from the Jack C. Kent/National Kidney Foundation. We thank Adam Sapirstein and Nicole Lemieux for making the pEGFP/cPLA₂ and pEGFP/cPLA₂(mut) constructs.

Gabriel J. Choukroun’s present address is: Service de Néphrologie, Hôpital Necker, Enfants Malades, Paris, France.

1. Brown, D, Slow, JL. Protein trafficking and polarity in kidney epithelium: from cell biology to physiology. Physiol Rev 1996. 76:245-297.
   View this article via: PubMed Google Scholar
2. Wandinger-Ness, A., and Simons, K. 1990. The polarized transport of surface proteins and lipids in epithelial cells. In Intracellular trafficking of proteins. J. Hanover and S. Steer, editors. Cambridge University Press. Cambridge, United Kingdom. 575–612.
   View this article via: PubMed Google Scholar
3. Kelly, RB. Pathways of protein secretion in eukaryotes. Science 1985. 230:25-32.
   View this article via: PubMed CrossRef Google Scholar
4. Schatz, G, Dobberstein, B. Common principles of protein translocation across membranes. Science 1996. 271:1519-1526.
   View this article via: PubMed CrossRef Google Scholar
5. Bonventre, JV. Phospholipase A₂ and signal transduction. J Am Soc Nephrol 1992. 3:128-150.
   View this article via: PubMed Google Scholar
6. De Camilli, P, Emr, SD, McPherson, PS, Novick, P. Phosphoinositides as regulators in membrane traffic. Science 1996. 271:1533-1539.
   View this article via: PubMed CrossRef Google Scholar
7. Ktistakis, NT, Brown, HT, Sternweis, PC, Roth, MG. Phospholipase D is present on Golgi-enriched membranes and its activation by ADP ribosylation factor is sensitive to brefeldin A. Proc Natl Acad Sci USA 1995. 92:4952-4956.
   View this article via: PubMed CrossRef Google Scholar
8. Liscovitch, M. Phospholipase D: role in signal transduction and membrane traffic. J Lipid Mediat Cell Signal 1996. 14:215-221.
   View this article via: PubMed CrossRef Google Scholar
9. Bonventre, JV, et al. Reduced fertility and postischaemic brain injury in mice deficient in cytosolic phospholipase A₂. Nature 1997. 390:622-625.
   View this article via: PubMed CrossRef Google Scholar
10. Gronich, JH, Bonventre, JV, Nemenoff, RA. Purification of a high-molecular-mass phospholipase A₂ from rat kidney activated at physiological calcium concentrations. Biochem J 1990. 271:37-43.
    View this article via: PubMed Google Scholar
11. Bonventre, JV, Nemenoff, R. Renal tubular arachidonic acid metabolism. Kidney Int 1991. 39:438-449.
    View this article via: PubMed CrossRef Google Scholar
12. Slomiany, A, et al. Function of intracellular phospholipase A₂ in vectorial transport of apoproteins from ER to Golgi. Int J Biochem 1992. 24:1397-1406.
    View this article via: PubMed CrossRef Google Scholar
13. Tagaya, M, et al. Correlation between phospholipase A₂ activity and intra-Golgi protein transport reconstituted in a cell-free system. FEBS Lett 1993. 324:201-204.
    View this article via: PubMed CrossRef Google Scholar
14. Gustafson, CE, et al. Vasopressin regulated trafficking of a green fluorescent protein-aquaporin 2 chimera in LLC-PK₁ cells. Histochem Cell Biol 1998. 110:377-386.
    View this article via: PubMed CrossRef Google Scholar
15. Chibalin, AV, Katz, AI, Berggren, PO, Bertorello, AM. Receptor-mediated inhibition of renal Na(+)-K(+)-ATPase is associated with endocytosis of its alpha- and beta-subunits. Am J Physiol 1997. 273:C1458-C1465.
    View this article via: PubMed Google Scholar
16. Choukroun, G, et al. Role of the stress-activated protein kinases in endothelin-induced cardiomyocyte hypertrophy. J Clin Invest 1998. 102:1311-1320.
    View this article via: JCI PubMed CrossRef Google Scholar
17. Graham, F.L., and Preyec, L. 1991. Manipulation of adenovirus vectors. In Methods in molecular biology: gene transfer and expression protocols. Volume 7. E.J. Murray, editor. The Humana Press Inc. Clifton, New Jersey, USA. 109–128.
    View this article via: PubMed Google Scholar
18. MacGregor, G.R., Nolan, G.P., Fiering, S., Roederer, M., and Herzenberg, L.A. 1991. Use of E. coli LacZ (β-galactosidase) as a reporter gene. In Methods in molecular biology. Volume 7. E.J. Murray, editor. The Humana Press Inc. Clifton, New Jersey, USA. 217–235.
    View this article via: PubMed Google Scholar
19. Brown, D, et al. Antigen retrieval in cryostat tissue sections and cultured cells by treatment with sodium dodecyl sulfate (SDS). J Histochem Cell Biol 1996. 105:261-267.
    View this article via: CrossRef Google Scholar
20. van’t Hof, W, Resh, MD. Rapid plasma membrane anchoring of newly synthesized p59fyn: selective requirement for NH₂-terminal myristoylation and palmitoylation at cysteine-3. J Cell Biol 1997. 136:1023-1035.
    View this article via: PubMed CrossRef Google Scholar
21. Wang, KKW, Porter, A, Hajimohamadreza, I. Total protein extraction from cultured cells for use in electrophoresis and Western blotting. BioTechniques 1996. 20:662-668.
    View this article via: PubMed Google Scholar
22. Slusarewicz, P., Hui, N., and Warren, G. 1994. Purification of rat liver Golgi stacks. In Cell biology: a laboratory handbook. Volume 1. Academic Press. San Diego, California, USA. 509–516.
    View this article via: PubMed Google Scholar
23. Laemmli, UK. Cleavage of structure proteins during assembly of the head of bacteriophage T4. Nature 1970. 227:680-685.
    View this article via: PubMed CrossRef Google Scholar
24. Marshansky, V, Bourgoin, S, Londono, I, Bendayan, M, Vinay, P. Identification of ADP-ribosylation factor-6 in brush-border membrane and early endosomes of human kidney proximal tubules. Electrophoresis 1997. 18:538-547.
    View this article via: PubMed CrossRef Google Scholar
25. Wilson, PD, et al. Reversed polarity of Na(+)-K(+)-ATPase: mislocation to apical plasma membranes in polycystic kidney disease epithelia. Am J Physiol 1991. 260:F420-F430.
    View this article via: PubMed Google Scholar
26. Molitoris, BA, Dahl, R, Geerdes, A. Cytoskeleton disruption and apical redistribution of proximal tubule Na(+)-K(+)-ATPase during ischemia. Am J Physiol 1992. 263:F488-F495.
    View this article via: PubMed Google Scholar
27. Berkner, KL. Expression of heterologous sequences in adenoviral vectors. Curr Top Microbiol Immunol 1992. 158:39-66.
    View this article via: PubMed Google Scholar
28. Mulligan, RC. The basic science of gene therapy. Science 1993. 260:926-932.
    View this article via: PubMed CrossRef Google Scholar
29. Abbate, M, Bonventre, JV, Brown, D. The microtubule network of renal epithelial cells is disrupted by ischemia and reperfusion. Am J Physiol 1994. 267:F971-F978.
    View this article via: PubMed Google Scholar
30. Brown, D, Lee, R, Bonventre, JV. Redistribution of villin to proximal tubule basolateral membranes after ischemia and reperfusion. Am J Physiol 1997. 273:F1003-F1012.
    View this article via: PubMed Google Scholar
31. Nakamura, H, Nemenoff, RA, Gronich, JH, Bonventre, JV. Subcellular characteristics of phospholipase A₂ activity in rat kidney. Enhanced cytosolic, mitochondrial, and microsomal phospholipase A₂ enzymatic activity after renal ischemia and reperfusion. J Clin Invest 1991. 87:1810-1818.
    View this article via: JCI PubMed CrossRef Google Scholar
32. Leaf, A. On the mechanism of fluid exchange of tissues in vitro. Biochem J 1956. 62:241-248.
    View this article via: PubMed Google Scholar
33. Verkade, P, Simons, K. Robert Feulgen Lecture 1997. Lipid microdomains and membrane trafficking in mammalian cells. Histochem Cell Biol 1997. 108:211-220.
    View this article via: PubMed CrossRef Google Scholar
34. Harder, T, Simons, K. Caveolae, DIGs, and the dynamics of sphingolipid-cholesterol microdomains. Curr Opin Cell Biol 1997. 9:534-542.
    View this article via: PubMed CrossRef Google Scholar
35. Cluett, EB, Kuismanen, E, Machamer, CE. Heterogeneous distribution of the unusual phospholipid semilysobisphosphatidic acid through the Golgi complex. Mol Biol Cell 1997. 8:2233-2240.
    View this article via: PubMed Google Scholar
36. Ktistakis, NT, Brown, HA, Sternweis, PC, Roth, MG. Phospholipase D is present on Golgi-enriched membranes and its activation by ADP ribosylation factor is sensitive to brefeldin A. Proc Natl Acad Sci USA 1995. 92:4952-4956.
    View this article via: PubMed CrossRef Google Scholar
37. Brown, HA, Gutowski, S, Moomaw, CR, Slaughter, C, Sternweis, PC. ADP-ribosylation factor, a small GTP-dependent regulatory protein, stimulates phospholipase D activity. Cell 1993. 75:1137-1144.
    View this article via: PubMed CrossRef Google Scholar
38. van Blitterswijk, WJ, Hilkmann, H. Rapid attenuation of receptor-induced diacylglycerol and phosphatidic acid by phospholipase D-mediated transphosphatidylation: formation of bisphosphatidic acid. EMBO J 1993. 12:2655-2662.
    View this article via: PubMed Google Scholar
39. Brown, D, Breton, S. Sorting proteins to their target membranes. Kidney Int 2000. 57:816-824.
    View this article via: PubMed CrossRef Google Scholar
40. Kornfeld, S, Li, E, Tabas, I. The synthesis of complex-type oligosaccharides. II. Characterization of the processing intermediates in the synthesis of the complex oligosaccharide units of the vesicular stomatitis virus G protein. J Biol Chem 1978. 253:7771-7778.
    View this article via: PubMed Google Scholar
41. Weigert, R, et al. CtBP/BARS induces fission of Golgi membranes by acylating lysophosphatidic acid. Nature 1999. 402:429-433.
    View this article via: PubMed CrossRef Google Scholar
42. Brown, DA, London, E. Functions of lipid rafts in biological membranes. Annu Rev Cell Dev Biol 1998. 14:111-136.
    View this article via: PubMed CrossRef Google Scholar
43. Nelson, WJ. W(h)ither the Golgi during mitosis? J Cell Biol 2000. 149:243-248.
    View this article via: PubMed CrossRef Google Scholar
44. Berlin, RD, Preston, SF. Arachidonic acid mobilization is suppressed during mitosis: role of cytosolic phospholipase A₂ activation. Biochem J 1995. 309:91-97.
    View this article via: PubMed Google Scholar
45. de Figueiredo, P, Drecktrah, D, Katzenellenbogen, JA, Strang, M, Brown, WJ. Evidence that phospholipase A2 activity is required for Golgi complex and trans Golgi network membrane tubulation. Proc Natl Acad Sci USA 1998. 95:8642-8647.
    View this article via: PubMed CrossRef Google Scholar
46. Balsinde, J, Dennis, EA. Bromoenol lactone inhibits magnesium-dependent phosphatidate phosphohydrolase and blocks triacylglycerol biosynthesis in mouse P388D1 macrophages. J Biol Chem 1996. 271:31937-31941.
    View this article via: PubMed CrossRef Google Scholar