Regulation of Phospholipase D2 Activity by Protein Kinase Cα*

It has been well documented that protein kinase C (PKC) plays an important role in regulation of phospholipase D (PLD) activity. Although PKC regulation of PLD1 activity has been studied extensively, the role of PKC in PLD2 regulation remains to be established. In the present study it was demonstrated that phorbol 12-myristate 13-acetate (PMA) induced PLD2 activation in COS-7 cells. PLD2 was also phosphorylated on both serine and threonine residues after PMA treatment. PKC inhibitors Ro-31-8220 and bisindolylmaleimide I inhibited both PMA-induced PLD2 phosphorylation and activation. However, Gö 6976, a PKC inhibitor relatively specific for conventional PKC isoforms, almost completely abolished PLD2 phosphorylation by PMA but only slightly inhibited PLD2 activation. Furthermore, time course studies showed that phosphorylation of PLD2 lagged behind its activation by PMA. Concentration curves for PMA action on PLD2 phosphorylation and activation also showed that PLD2 was activated by PMA at concentrations at which PMA didn't induce phosphorylation. A kinase-deficient mutant of PKCα stimulated PLD2 activity to an even higher level than wild type PKCα. Co-expression of wild type PKCα, but not PKCδ, greatly enhanced both basal and PMA-induced PLD2 phosphorylation. A PKCδ-specific inhibitor, rottlerin, failed to inhibit PMA-induced PLD2 phosphorylation and activation. Co-immunoprecipitation studies indicated an association between PLD2 and PKCα under basal conditions that was further enhanced by PMA. Time course studies of the effects of PKCα on PLD2 showed that as the phosphorylation of PLD2 increased, its activity declined. In summary, the data demonstrated that PLD2 is activated and phosphorylated by PMA and PKCα in COS-7 cells. However, the phosphorylation is not required for PKCα to activate PLD2. It is suggested that interaction rather than phosphorylation underscores the activation of PLD2 by PKC in vivo and that phosphorylation may contribute to the inactivation of the enzyme.

Phospholipase D (PLD) 1 catalyzes the hydrolysis of phosphatidylcholine (PC) to phosphatidic acid (PA), and choline (1). PA is an important "second messenger" in several physiological processes (2). PA can be further metabolized to diacylglycerol by PA phosphohydrolase or to lysophosphatidic acid by phospholipase A2. Diacylglycerol is a well known activator of pro-tein kinase C (PKC), whereas lysophosphatidic acid mediates many important physiological functions via its G protein-coupled receptors (3). Thus, PLD influences many important intracellular events via producing these downstream products. PLD activity is regulated by many stimuli such as growth factors, cytokines, hormones, neurotransmitters, and other molecules involved in extracellular communication (1). PLD is thought to play an important role in secretion, membrane trafficking, cytoskeleton reorganization, and apoptosis (1).
PLD1 and PLD2 are two isoforms of mammalian PLD that share about 50% amino acid similarity. However, their regulatory properties are quite different, e.g. that PLD1 has a low basal activity and is regulated by PKC and members of the Rho, adenosine diphosphate ribosylation factor, and Ras/Ral families of small G proteins (4 -9). In contrast, PLD2 exhibits a high basal activity when assayed in vitro and shows little or no in vitro response to the activators of PLD1 (10 -12). However, other regulators of PLD2 activity have been reported. For example, PLD2 activity may be modulated in vivo by inhibition (11). The reported inhibitors include a heat-stable 18-kDa protein (11), adolase (13), ceramide (14), ␣-actinin (15), and ␣and ␤-synucleins (16). On the other hand, a lot of activators of PLD2 activity have also been found. These activators include phosphatidylinositol 4,5-bisphosphate (10), unsaturated fatty acid (17), caveolin (18), calcium (19), G M2 activator (20), and adenosine diphosphate ribosylation factor (12). Although the regulation of PLD1 by PKC is universally accepted, it is controversial on whether or not PLD2 is also regulated by PKC. The initial report showed that PLD2 was largely refractory to PKC activation in vitro and in vivo (11). However, recent reports demonstrate that PMA can activate PLD2 in vivo, suggesting the possible regulation of PLD2 activity by PKC (19,21,22). Hence, the role of PKC in PLD2 activation remains to be defined.
Regarding the PKC regulation of PLD1 activity, both phosphorylation-dependent (23,24) and -independent mechanisms (5,25,26) have been proposed. A protein-protein interaction rather than phosphorylation has been proposed as the main mechanism for PKC␣ to activate PLD1 (26). Compared with PLD1, only one group has explored the phosphorylation of PLD2 by PKC and its effect on activity (22). This group reported that PKC␦ is involved in PMA-induced PLD2 phosphorylation and activation in PC12 cells (22), whereas another group reported that PLD2 activity is enhanced by dephosphorylation of Ser/Thr residues (27). Thus, the role of phosphorylation in the activation of PLD2 by PKC requires further work.
We report in the present study that PMA induces PLD2 activation via the involvement of PKC␣ and that PLD2 becomes phosphorylated on both Ser and Thr residues. In addition, we demonstrate PKC␣ and PLD2 interact in vivo and that the phosphorylation is not required for the activation of PLD2. Evidence is presented that phosphorylation more likely contributes to deactivation of PLD2.
Cell Culture and Transfection-COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 100 units/ml penicillin, 100 g/ml streptomycin, and 10% fetal bovine serum in humidified 10% CO 2 . Six-well plates were seeded with 2 ϫ 10 5 cells per well, and 10-cm dishes were seeded with 8 ϫ 10 5 cells and transfected with FuGENE 6 according to the manufacturer's instructions.
In Vivo PLD Assay-After a 6-h transfection with FuGENE 6, COS-7 cells in 6-well plates were serum-starved (0.5% fetal bovine serum in Dulbecco's modified Eagle's medium in the presence of 1 Ci/ml [ 3 H]myristic acid). After overnight starvation, PLD activity was assayed as described (28). Briefly, cells were preincubated in 0.3% 1-butanol and then treated as indicated in the figure legends. Cells were then washed with ice-cold phosphate-buffered saline and stopped with methanol. Lipids were extracted, and the PtdBut product was resolved by thin layer chromatography. Bands co-migrating with a PtdBut standard were scraped and quantitated by liquid scintillation counting.
Subcellular Fractionation-COS-7 cells in 100-mm plates were harvested after transfection with PLD2 constructs and starved overnight as described above. After treatment with the appropriate reagents the cells were washed and harvested with ice-cold phosphate-buffered saline buffer. Cells were then centrifuged and resuspended in 500 l of ice-cold lysis buffer (25 mM Hepes, pH 7.2, 10% glycerol, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture). After two bursts (10 s) of sonication, the cell lysate was centrifuged at 500 ϫ g for 10 min to remove unbroken cells. The supernatant was then centrifuged at 120,000 ϫ g for 45 min at 4°C to separate the cytosolic and crude membrane fractions. The particulate fraction was washed twice with the lysis buffer and resuspended in the same buffer by passing through a 27-gauge needle until the pellet was resuspended.
In Vitro PLD Assay-The PLD activity was measured by the formation of [ 3 H]PtdBut in vitro as described previously (28) with modifications. The composition of phospholipid vesicles (11.5 M PC, 2 M phosphatidylinositol 4,5-bisphosphate, and 97.75 M phosphatidylethanolamine) was optimized for PLD2 activity. [Palmitoyl-3 H]PC (0.5 Ci/reaction) was used with 1-butanol (0.3%) as substrate. Where applicable, 100 M ATP, 200 ng of PKC␣ (PanVera), or 500 nM PMA was also included in the reaction system. The reaction mixtures were then incubated at 37°C for 30 min and stopped with chloroform/methanol/ HCl (50:98:2, v/v). The lipids were extracted from the organic phase and resolved by thin-layer chromatography. Bands co-migrating with a PtdBut standard were quantitated by liquid scintillation counting.
Western Blotting-Protein samples were analyzed by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Immobilon-P, Millipore Corp.). The blots were then blocked with 1% bovine serum albumin and incubated with appropriate primary antibodies (1 mg/ml) (1:5000) followed by incubation with horseradish peroxidase-conjugated secondary antibody. Immunoreactive bands were detected using enhanced chemiluminescence (ECL).
Immunoprecipitation-COS-7 cells cultured in 100-mm dishes were transfected and starved as described above. After treatment, the cells were washed twice with ice-cold phosphate-buffered saline and then harvested in 600 l of immunoprecipitation buffer containing 25 mM Hepes, pH 7.2, 10% glycerol, 1 mM EDTA, 1 mM EGTA, 50 mM KC1, 1% Triton X-100, 0.1% SDS, 10 mM NaF, 10 mM Na 4 P 2 O 7 , 1.2 mM Na 3 VO 4 , and protease inhibitors mixture. The cell suspension was then sonicated for 2 ϫ 5 s, and the resulting cell lysate was centrifuged at 120,000 ϫ g for 30 min to remove the detergent-insoluble fraction. The supernatant was then precleared by mixing it with 1 g of affinitypurified mouse IgG and 20 l of 1:1 slurry of protein G-agarose beads for 1 h at 4°C. The mixture was then spun, and the supernatant was incubated with 2 l of anti-Xpress mouse antibody (1 mg/ml) and 20 l of protein G-agarose beads overnight. The immunoprecipitates were washed twice with the lysis buffer and then resuspended in SDS sample buffer and analyzed by Western blotting.
In Vitro Phosphorylation of PLD2 by PKC␣-Immunoprecipitated PLD2 from PLD2-overexpressing COS-7 cells was washed with phosphorylation buffer (30 mM Tris/HCl, pH 7.0, 6 mM MgCl 2 , 0.25 mM EGTA, 0.4 mM CaCl 2 , 0.1% Triton X-100) four times and resuspended in the phosphorylation buffer. Aliquots of beads containing immunoprecipitated PLD2 were then mixed with 0.12 mM ATP, 10 Ci of [␥-32 P]ATP, 100 ng of PKC␣, and 100 nM PMA. The mixture was incubated at 30°C for 30 min and then electrophoresed through an 8% SDS-polyacrylamide gel and transferred onto a polyvinylidene difluoride membrane, and the membrane was exposed to a photographic film for autoradiography.

RESULTS
PMA Activates PLD2 Activity-As seen in Fig. 1, the transphosphatidylation product of PLD action, PtdBut, increased as early as 1 min after PMA treatment of COS-7 cells and accumulated in a time-dependent manner. PtdBut levels reached a maximum at around 15 min after PMA treatment. The data indicated that PLD2 was rapidly activated by PMA and that the activity of PLD2 (formation of PtdBut) dropped to the basal level after 15 min, indicating a short-duration response. The data demonstrated that PLD2, like PLD1, is also activated by PMA when overexpressed in COS-7 cells.
PKC Mediates PMA-induced PLD2 Activation-Because it is well established that PMA activates conventional and novel PKC isozymes, we used PKC inhibitors to explore the involvement of PKC in the PLD2 activation. Fig. 2 shows the effects of the PKC inhibitors on both basal and PMA-stimulated PLD activity in vector-and PLD2-transfected COS-7 cells. As a control, the effects of the PKC inhibitors on PLD1 activity were also examined. General PKC inhibitors such as Ro-31-8220 and bisindolylmaleimide I (Bis) showed strong inhibition of PMAinduced activation of control (vector) PLD activity and of PLD1 and PLD2 activity (Fig. 2). 2 However, the conventional PKC inhibitor Gö 6976 showed much less inhibition of PMA-induced PLD1 and PLD2 activation. Rottlerin, a PKC␦-specific PKC inhibitor, showed marked inhibition of the basal activity of both PLD1 and PLD2 but no inhibition of the effect of PMA itself. These data indicate that PKC is involved in PMA-induced activation of both PLD1 and PLD2.
PLD2 Undergoes Serine and Threonine Phosphorylation after PMA Treatment-The possible Ser/Thr phosphorylation of PLD2 after PMA treatment was examined using anti-phospho-Ser and anti-phospho-Thr antibodies. As indicated in Fig. 3A, PLD2 was basally phosphorylated on Thr but not on Ser residues. Upon PMA treatment, PLD2 became phosphorylated on Ser and Thr in a time-dependent manner. However, the phosphorylation on Ser became detectable only after 5 min of PMA treatment. The increase of phosphorylation was further demonstrated by the slower migration of PLD2 on the gel upon PMA treatment for longer times, as indicated by the Western blotting with Xpress antibodies in Fig. 3A.
Next, the effect of the PKC inhibitors on PMA-induced PLD2 phosphorylation was determined. As indicated in Fig. 3B, Ro-31-8220 and Bis completely inhibited the phosphorylation induced on both residues by treatment with PMA. Gö 6976 also showed a significant inhibition, whereas rottlerin had no effect at all. These data demonstrate that PKC, possibly the conventional isoforms rather than PKC␦, are involved in the phosphorylation of PLD2 induced by PMA.
PKC␣ Mediates PMA-induced PLD2 Phosphorylation-To obtain further information about which PKC isoform mediates the phosphorylation, the expression of different PKC isoforms in COS-7 cells was examined by Western blotting. Fig. 4A shows that PKC␣, PKC␦, PKC⑀, PKC, and PKC were detected, whereas PKC␤, PKC␥, PKC, and PKC were not even though the antibodies were shown to be effective (data not shown). Among the isoforms, PKC␣ showed a complete translocation from cytosol to membrane fractions after PMA treatment. In sharp contrast to PKC␣, PKC␦ and PKC⑀ were located mainly in the membrane fraction and showed little or no translocation when stimulated. These data imply that PKC␣ is more likely than other PKC isoforms to be involved in PMA-induced PLD2 phosphorylation.
To further explore this issue, PKC␣ or PKC␦ was overexpressed together with PLD2 in COS-7 cells, and PMA-induced PLD2 phosphorylation was examined. As shown in Fig. 4B, co-expression of PKC␣ greatly enhanced PMA-induced PLD2 phosphorylation on both Ser and Thr residues. In contrast, co-expression of PKC␦ had minimal effect. The lower panel of Fig. 4B shows that PKC␣ and PKC␦ were well expressed. The data provide further support that PKC␣, rather than PKC␦, is involved in PMA-induced PLD2 phosphorylation. To test if PKC␣ can directly phosphorylate PLD2, in vitro phosphorylation of PLD2 by PKC␣ was conducted. As shown in Fig. 4C, both PLD1 and PLD2 were phosphorylated by PKC␣ in the presence of PMA, although the phosphorylation of PLD1 was greater.
Phosphorylation Is Not Required for PLD2 Activation by PKC␣-The requirement of phosphorylation for PLD2 activation by PMA is next explored. First, the dose response of PMA on PLD2 phosphorylation was compared with that for PMA on PLD2 activation. Fig. 5A shows that PMA induced PLD2 phosphorylation by a dose-dependent manner. The phosphorylation was detectable at a PMA concentration of 2.5 nM and higher. Fig. 5B shows the dose dependence for PMA activation of PLD2. In contrast to the phosphorylation pattern, activation was evident at PMA concentrations as low as 0.5 nM and was maximal at 5 nM. The difference between these two patterns indicates that activation of PLD2 is not dependent on its phosphorylation.
To further confirm that phosphorylation is not required for PKC␣-mediated PLD2 activation, a kinase-dead mutant of PKC␣ (KN-PKC␣) was tested for its effect on the activation of PLD2. As shown in Fig. 6A, wild type PKC␣ enhanced PLD2 activity when co-transfected with PLD2 in COS-7 cells. Surprisingly, the kinase-dead PKC␣ mutant displayed an even stronger stimulation than wild type PKC␣ on basal and PLD2 activity. As a control, PKC␦ and PKC⑀ were also co-expressed with PLD2 and showed no significant stimulation, showing the specificity for PKC␣ activation of PLD2. Fig. 6B confirms that, although wild type PKC␣ enhanced both basal and PMA-in-duced PLD2 phosphorylation, the kinase-dead mutant was unable to phosphorylate PLD2. Fig. 6C further shows that in contrast to wild type, kinase-dead PKC␣ does not exhibit autophosphorylation after PMA treatment. These data strongly support the conclusion that phosphorylation is not required for PKC␣ to activate PLD2 activity.
PLD2 Binds to PKC␣ after PMA Treatment-Consistent with the fast activation of PLD2 by PKC␣ in vivo, PMA induced a rapid translocation of this PKC isoform from cytosol to the membrane fraction (Fig. 7A). Because PLD2 localizes to membranes (11,29), this raises the possibility that the translocation of PKC␣ makes it more accessible to interact with PLD2. To explore this, the co-immunoprecipitation of PLD2 and PKC␣ was examined in COS-7 cells. Fig. 7B shows that there is a basal association between PLD2 and PKC␣, and the binding is greatly enhanced by PMA treatment. Fig. 7C shows that, consistent with the results in Fig. 7B, PLD2 was present in PKC␣ immunoprecipitates and that PMA potentiated the association in a time-dependent manner. These data demonstrate that PLD2 can interact with PKC␣. To examine whether the interaction between PLD2 and PKC␣ is specific, the presence of endogenous PKC␣, PKC␦, or PKC⑀ in PLD2 immunoprecipitates was determined. As shown in Fig. 7D, endogenous PKC␣ is associated with PLD2, and the association is increased after PMA treatment in a time-dependent manner. In contrast, no PKC␦ or PKC⑀ was detected in PLD2 immunoprecipitates before and after PMA treatment. These data therefore show that PKC␣ rather than PKC␦ or PKC⑀ mediates PMA-induced PLD2 activation. The effect of PKC inhibitors on the PLD2-FIG. 4. PKC␣ is responsible for the phosphorylation. A, COS-7 cells were treated with or without 100 nM PMA for 15 min, and subcellular fractionation was conducted as described under "Experimental Procedures" to obtain the membrane (M) and cytosol (C) fractions. Total lysate was also prepared. The samples were analyzed by SDS-PAGE and subsequent Western blotting with antibodies to specific PKC isoforms. B, COS-7 cells were transfected with PLD2 alone or co-transfected with PLD2 plus increasing quantities of cDNAs for PKC␣ or PKC␦ as indicated. After starving overnight, the cells were treated with or without 100 nM PMA as indicated for 30 min. Immunoprecipitation (IP) and Western blotting (IB) were conducted as described under "Experimental Procedures." p-, phospho-. C, COS-7 cells were transfected with Xpress vector, Xpress-tagged PLD1, or Xpress-tagged PLD2. Then the Xpress immunoprecipitates were collected, and the in vitro phosphorylation assays were conducted as described under "Ex- PKC␣ interaction was also examined. As shown in Fig. 7E, Ro-31-8220, Bis, and Gö 6976 completely or almost completely inhibited PMA-induced PLD2-PKC␣ interaction, whereas rottlerin showed no inhibition at all. The large inhibitory effect of Gö 6976 on the interaction seemed inconsistent with its minimal inhibition of PLD2 activity. We hypothesized that the relationship between the interaction between PKC␣ and PLD2 and the activation of PLD2 was not linear. In other words, only a small amount of PKC␣ binding was required for the full activation of PLD2. To prove this hypothesis, we expressed increasing amounts of PKC␣ together with PLD2 and examined the effects on the PLD2-PKC␣ interaction and PLD2 activity. As shown in Fig. 7F, the addition of increasing of PKC␣ cDNA from 0.08 to 0.48 g led to a proportional increase in the PLD2-PKC␣ association. However, 0.48 g of PKC␣ cDNA produced no further increase in PLD2 activity than did 0.08 g. In other words, the relationship between PLD2-PKC␣ association and PLD2 activation was nonlinear.
Phosphorylation Suppresses PLD2 Activity-The above data indicated that protein-protein interaction is more likely than phosphorylation to be responsible for PLD2 activation by PMA/ PKC␣. The question then is, What is the role of phosphorylation in the control of PLD2 activity? Fig. 8A shows the changes of PLD2 activity after PMA treatment of COS-7 cells in the absence or presence of overexpressed PKC␣. Without PKC␣ expression, the PLD product PtdBut accumulated linearly over 10 min in the presence of PMA, showing that PLD2 remained active during this period. However, in the presence of overexpressed PKC␣, there was an increase in PtdBut before PMA treatment (0 min) followed by a slow increase with PMA. The latter change indicates that PLD2 is deactivated after several minutes of PMA treatment under the condition of PKC␣ coexpression. In support of a role for phosphorylation, Fig. 8B showed that PLD2 was phosphorylated much earlier and stronger in the presence of overexpressed PKC␣ than in its absence. This phosphorylation pattern is highly correlated with the deactivation of PLD2 activity. The in vitro assay of PLD2 in Fig.  8C shows that the inclusion of ATP in the reaction with PKC␣ and PMA inhibited PLD2 activity and that the inhibition was abolished by the PKC inhibitor Ro-31-8220. These data indicate that phosphorylation can contribute to the suppression of PLD2 activity. DISCUSSION Previously our group showed that PLD2, like PLD1, exhibited a large response to PMA when overexpressed in COS-7 cells (21). This paper represents a follow-up of this observation. The present results show that PMA stimulates PLD2 activity up to 6-fold when overexpressed in COS-7 cells (Fig. 1), in contrast to a previous report (11) but in agreement with studies in HEK293 and Sf9 cells, where PLD2 also showed large responses to PMA or PKC␣ when overexpressed (19,30).
A major issue in exploring the mechanism by which PMA/ PKC activates PLD1 or PLD2 is the role of phosphorylation. It has been reported that PLD2 is Ser/Thr phosphorylated by PKC␦ after PMA treatment of rat pheochromocytoma PC12 cells (22). In this case the Ser/Thr phosphorylation was revealed by phosphoamino acid analysis. In another paper, the Ser/Thr phosphorylation of PLD2 was implicated by using okadaic acid, an inhibitor of Ser/Thr protein phosphatases 1 and 2A (27). In the present study, using phospho-specific antibodies, we clearly demonstrate that PLD2 undergoes phosphorylation on both serine and threonine residues upon PMA treatment of COS-7 cells (Fig. 3A). Furthermore, we provide evidence that PKC␣ mediates the effect of PMA. This conclusion is based on the following points. First, Gö 6976, a PKC inhibitor that is relatively specific for PKC␣ and PKC␤1, abol- ished PMA-induced PLD2 phosphorylation (Fig. 3B). In contrast, the relatively PKC␦-specific inhibitor rottlerin did not have any inhibitory effect (Fig. 3B). A possible role for PKC␤1 was eliminated by the fact that PKC␤ was not detected in COS-7 cells (data not shown). Second, PKC␣ but not PKC␦ greatly potentiated both basal and PMA-induced phosphorylation of PLD2 (Fig. 4B). Third, the overexpression of PKC␣ but not PKC␦ or PKC⑀ stimulated the activity of PLD2 in COS-7 cells (Fig. 6A). Consistent with our results, Siddiqi et al. (19) demonstrate that in Sf9 cells, where both PKC␣ and PKC␦ expression are undetectable, PMA barely activated PLD2. However, overexpression of PKC␣, but not PKC␦, significantly enhanced both basal and PMA-stimulated PLD2 activity. In contrast, it has been reported that PKC␦ mediates PMA-induced PLD2 phosphorylation and activation in PC12 cells (22). Therefore, it is possible that different PKC isoforms are used in different cells or tissues to mediate PLD2 activation.
The issue of whether or not phosphorylation is required for PLD activation still remains controversial. This issue has mostly involved studies of PLD1. A recent paper from our lab provides evidence that phosphorylation is not required for PLD1 activation (26), but there are reports to the contrary (23,24). The requirement of phosphorylation for PKC␣-mediated PLD2 activation by PMA was therefore examined in this study. Although our findings indicate that during PMA-induced activation PLD2 becomes heavily phosphorylated on both Ser and Thr residues (Fig. 3A), there is much evidence that this phosphorylation is not required for PLD2 activation. First, although Gö 6976 eliminated the ability of PMA to phosphorylate PLD2 (Fig. 3B), it only exerted a slight inhibition on the activation of the enzyme (Fig. 2). Second, KN-PKC␣, a PKC␣ kinase-dead mutant, still activated PLD2 robustly (Fig. 6A). Third, comparison between the time courses of the effects of PMA on PLD2 phosphorylation (Fig. 3A) and activation (Fig. 1A) showed that the activation of PLD2 by PMA occurred earlier than did phosphorylation. Fourth, by comparing the concentration curve of PMA-induced PLD2 activation (Fig. 5B) and the dose-response pattern of PMA-induced PLD2 phosphorylation (Fig. 5A) we found that PLD2 was activated by PMA at concentrations (0.5 and 1 nM) where phosphorylation of PLD2 didn't occur yet. Based on the above findings, we have concluded that phosphorylation is not required for the activation of PLD2 by PMA.
Furthermore, our data suggest that the phosphorylation of PLD2 actually suppresses its enzyme activity. First, overexpression of the kinase-dead PKC␣ mutant showed a bigger stimulation of PLD2 activity than did overexpression of wildtype PKC␣ at the same level (Fig. 6A), implying that phosphorylation decreases PLD2 activity. Second, after the initial activation by PMA treatment of COS-7 cells, PLD2 lost activity faster in the presence of co-expression of PKC␣ than in its absence (Fig. 8A), and this is highly correlated with the fact that PLD2 undergoes phosphorylation much faster in the former case than in the latter (Fig. 8B). Third, in vitro PLD assay also showed that when ATP was included in the assay system, PLD2 activity decreased, and this decrease was abolished by the PKC inhibitor Ro-31-8220 (Fig. 8C). Consistent with our results, Watanabe and Kanaho (27) also demonstrate that phosphorylation leads to the inhibition of PLD2 activity.
PLD2 has been reported to bind to a lot of intracellular proteins. These proteins include the -opioid receptor (31), phospholipase ␥1 (32), c-Src (33), epidermal growth factor receptor (34), and PKC␦ (22). The binding is either ligand-dependent or -independent. PLD2 has also been reported to associate with PKC␣ in HEK-293 cells when both proteins were overexpressed, but this association was ligand-independent (30). Siddiqi et al. (19) report that when PLD2 and PKC␣ were both overexpressed in Sf9 cells, an association between them was observed and was increased by PMA. It is, therefore, possible that the protein-protein interaction between PLD2 and PKC␣ contributes to the activation of PLD2. In the present study, we showed that in COS-7 cells, an association between PLD2 and PKC␣ is also present, and this association was greatly enhanced by PMA (Fig. 7B). In addition, in the absence of PMA, KN-PKC␣ showed a stronger association with PLD2 than wild type PKC␣ (Fig. 6B). This is correlated with the fact that KN-PKC␣ induced a larger stimulation of PLD2 activity than did wild type PKC␣ (Fig. 6A). Furthermore, the effects of PKC inhibitors on PMA-induced PLD2 activation were correlated with their effects on PMA-induced PLD2-PKC␣ interaction. These observations support the conclusion that the PLD2-PKC␣ interaction is responsible for PLD2 activation. In contrast to PKC␣, interaction of PKC␦ and PKC⑀, the other two FIG. 8. Phosphorylation of PLD2 contributes to the suppression of its activity. A, COS-7 cells were transfected with PLD2 together with PKC␣ or vector control. Cells were then treated with 100 nM PMA as indicated. The PLD2 activities shown are those after subtraction of endogenous PLD activity. B, COS-7 cells were transfected with PLD2 together with PKC␣ or vector control. Cells were then treated with 100 nM PMA for the times indicated. PLD2 phosphorylation was determined as described under "Experimental Procedures." IP, immunoprecipitation; IB, Western blotting. C, membranes from PLD2 transfected COS-7 cells were obtained for in vitro PLD assay. ATP, PKC␣, PMA, and Ro-31-8220 were included in the assay as indicated. In vitro PLD assays were conducted as described under "Experimental Procedures." Data are representative of at least three independent experiments. p-, phospho-. PMA-responsive PKC isoforms, with PLD2 was not detected (Fig. 7D). Also, overexpression of PKC␦ or PKC⑀ caused no stimulation of PLD2 activity (Fig. 6A). These data demonstrate that PMA induces a specific interaction between PKC␣ and PLD2 and that PKC␣ subsequently activates PLD2 by proteinprotein interaction.
In conclusion, our findings show that PLD2 becomes Ser/ Thr-phosphorylated during its activation by PMA and that PKC␣ mediates the effect. However, the phosphorylation of PLD2 is not required for its activation. Rather, the phosphorylation is related to a slow deactivation of PLD2. We suggest that a protein-protein interaction between PLD2 and PKC␣ is responsible for PLD2 activation by PMA in COS-7 cells.