Analysis of Platelet-derived Growth Factor-induced Phospholipase D Activation in Mouse Embryo Fibroblasts Lacking Phospholipase C-γ1*

Platelet-derived growth factor (PDGF) activates phospholipase D (PLD) in mouse embryo fibroblasts (MEFs). In order to investigate a role for phospholipase C-γ1 (PLC-γ1), we used targeted disruption of the Plcg1 gene in the mouse to develop Plcg1 +/+ andPlcg1 −/− cell lines.Plcg1 +/+ MEFs treated with PDGF showed a time- and dose-dependent increase in the production of total inositol phosphates that was substantially reduced inPlcg1 −/− cells.Plcg1 +/+ cells also showed a PDGF-induced increase in PLD activity that had a similar dose dependence to the PLC response but was down-regulated after 15 min. Phospholipase D activity, however, was markedly reduced in Plcg1 −/−cells. The PDGF-induced inositol phosphate formation and the PLD activity that remained in the Plcg1 −/− cells could be attributed to the presence of phospholipase C-γ2 (PLC-γ2) in the Plcg1 −/− cells. The PLC-γ2 expressed in the Plcg1 −/− cells was phosphorylated on tyrosine in response to PDGF treatment, and a small but significant fraction of the Plcg1 −/− cells showed Ca2+ mobilization in response to PDGF, suggesting that the PLC-γ2 expressed in the Plcg1 −/− cells was activated in response to PDGF. The inhibition of PDGF-induced phospholipid hydrolysis in Plcg1 −/− cells was not due to differences in the level of PDGF receptor or in the ability of PDGF to cause autophosphorylation of the receptor. Upon treatment of the Plcg1 −/− cells with oleoylacetylglycerol and the Ca2+ ionophore ionomycin to mimic the effect of PLC-γ1, PLD activity was restored. The targeted disruption ofPlcg1 did not result in universal changes in the cell signaling pathways of Plcg1 −/− cells, because the phosphorylation of mitogen-activated protein kinase was similar inPlcg1 +/+ and Plcg1 −/−cells. Because increased plasma membrane ruffles occurred in bothPlcg1 +/+ and Plcg1 −/−cells following PDGF treatment, it is possible neither PLC nor PLD are necessary for this growth factor response. In summary, these data indicate that PLC-γ is required for growth factor-induced activation of PLD in MEFs.

Phospholipase D hydrolyzes phosphatidylcholine, generating choline and phosphatidic acid (PA) (8,9). Phosphatidic acid exerts many effects in vitro, including the stimulation of PLC-␥, phosphatidylinositol-4-phosphate kinase, and protein kinases (10). In addition, through the actions of PA phosphohydrolase and a specific phospholipase A 2 , PA can be converted to DAG and the signaling molecule lysophosphatidic acid, respectively (11).
There is conflicting evidence about whether the activation of PLC-␥ and the subsequent activation of PKC are necessary for agonist stimulation of PLD. Although there are many studies reporting the involvement of PKC in the activation of PLD by agonists (12)(13)(14)(15), there also are reports that PKC is not involved (16 -19). Furthermore, some studies have indicated that PLD can be activated by certain agonists in the absence of detectable PIP 2 hydrolysis. For example, in Madin-Darby canine kidney cells, studies with neomycin indicate that activation of PLD by purinergic agonists is independent of PLC-␥ activity (20). In certain fibroblasts, PLD activation by epidermal growth factor (EGF) has been reported to occur in the absence of measurable PIP 2 breakdown (21). However, in Swiss 3T3 fibroblasts and TRMP cells, activation of PLC-␥1 is necessary for stimulation of PLD activity by PDGF (22). Previously, we used homologous recombination to selectively disrupt the Plcg1 gene encoding PLC-␥1 in mice (23). Although this mutation was lethal, immortal mouse embryo fibroblast (MEF) cell lines were produced from Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ embryos. We have now used these cells to study PDGFinduced PLD activity. The results indicate that PLC-␥1 activity is required for PDGF-induced PLD activation.
Cell Culture-MEFs were prepared from embryonic day 9.5 embryos with (Plcg1 Ϫ/Ϫ ) and without (Plcg1 ϩ/ϩ ) targeted disruption of the Plcg1 gene, using the targeting vectors TVI and TVII (23). Wild type and null cells from the same litter were established in culture according to standard methods and maintained as immortalized non-transformed cell lines. Plcg1 gene disruption by targeting vector TVI disrupted exons encoding the X domain and the two SH2 domains of PLC-␥1, whereas targeting vector TVII replaces at the genomic level exons encoding the X domain and two SH2 domains of PLC-␥1 with a LacZ sequence. Cells targeted with TVII produce a fusion protein of the PLC-␥1 N terminus and ␤-galactosidase (23). The fibroblasts were passaged in DMEM containing 10% fetal bovine serum at 37°C in a humidified atmosphere with 5% CO 2 .
Phospholipase D Assay-Cells were plated in 60-mm tissue-culture plates. The cells were serum-starved in DMEM (Life Technologies, Inc.) containing 0.5% bovine serum albumin for 16 h prior to the start of the assay. During this time, they were labeled with 1 Ci/ml [9,10-3 H]myristic acid. At the start of the experiment, the cells were washed three times with 5 ml of phosphate-buffered saline (PBS) and pre-equilibrated at 37°C in serum-free DMEM for 1 h. For the final 10 min of preincubation, 0.3% butan-1-ol was included. At the end of the preincubation, cells were treated with the indicated agonist for 10 min or the times indicated. Incubations were terminated by removing the medium, washing on ice with 5 ml of ice-cold PBS, and adding 1.5 ml of ice-cold methanol. Cells were scraped off the plates, and the lipids were extracted and separated with methanol/chloroform/0.1 N HCl (1:1:1) according to the method of Bligh and Dyer (24). The lower phase was dried under N 2 , resuspended in 30 l of chloroform/methanol (2:1), and spotted onto silica gel 60A thin layer chromatography plates (Whatman). The plates were developed in the upper phase of the solvent system of ethyl acetate/iso-octane/H 2 O/acetic acid (55:25:50:10) and stained with iodine. A PtdBut standard (Avanti Polar Lipids) was used to locate the bands, which were scraped into scintillation vials containing 500 l of methanol and 7.5 ml of Ready Organic scintillation mixture (Beckman). Radioactivity incorporated into total phospholipids was measured, and the results were presented as percentage of total lipid cpm incorporated into PtdBut.
Measurement of Total Inositol Phosphates-Inositol phosphates were measured as described by Yeo and Exton (12). Briefly, cells were plated on 6-well tissue culture plates and labeled for 16 h with 1 Ci/ml myo-[2-3 H]inositol in inositol-free, serum-free DMEM supplemented with 0.5% bovine serum albumin. At the start of the experiment, the cells were preincubated for 1 h at 37°C. During the last 10 min of preincubation, 20 mM LiCl was included. The cells were treated with the indicated growth factors for 15 min. The experiment was terminated by removing the medium, washing with 5 ml of ice-cold PBS, and adding 750 l of ice-cold 20 mM formic acid. After 30 min of incubation on ice, the cells were neutralized with 250 l of 50 mM NH 4 OH. The cells were scraped on ice, transferred to a 1.5-ml Eppendorf tube, and spun at maximum speed for 10 min at 4°C. The supernatant (900 l) was loaded onto Bio-Rad 10-ml columns containing 1 ml of AG1-8X resin. Inositol (Ins) and inositol phosphates (InsP x ) were eluted according to Simpson et al. (25). Five hundred l of each eluted fraction was counted in 15 ml of Ready Safe (Beckman). The results are expressed as ((InsP x )/ (InsP x ϩ Ins)) ϫ 1000.
Immunoprecipitation of PLC-␥2-Subconfluent cells on 100-mm dishes were lysed in radioimmune precipitation assay buffer (50 mM Tris, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EGTA, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM Na 3 VO 4 , 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin) and precleared for 1 h with 1 g of anti-rabbit IgG and 20 l of protein A/G PLUS-agarose (Santa Cruz) at 4°C with rocking. A protein assay was performed on the precleared cell lysate, which was then diluted to 1 mg/ml in 500 l and immunoprecipitated with 3 g of rabbit polyclonal PLC-␥2 antibody for 1 h at 4°C. Twenty l of protein A/G PLUS-agarose was added overnight at 4°C with rocking. In the morning, the protein A/G PLUS-agarose beads were collected with centrifugation, washed three times with 1 ml of radioimmune precipitation buffer, and resuspended in 20 l of 2ϫ SDS sample buffer. The samples were boiled, and the immunoprecipitated proteins were separated on a 6% SDS-PAGE gel. The proteins were transferred to Immobilon-P, blocked in 3% dry milk in PBS, and probed with an antibody to P-Tyr at a concentration of 1 g/ml.
Western Blotting-Subconfluent cells in 100-mm dishes that were serum-starved in serum-free DMEM containing 0.5% bovine serum albumin for 24 h were washed once with 5 ml of PBS and scraped directly in 300 l of PBS containing 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 1 g/ml aprotinin, 1 g/ml leupeptin, and 1 g/ml pepstatin. A protein assay using the BCA method (Pierce) was performed on the samples, and cells on a duplicate 100-mm dish were counted using a hemocytometer. Proteins were separated by SDS-PAGE on a 6% (PDGFR and phosphotyrosine) or a 4 -20% gradient gel (MAP kinase and phospho-MAP kinase), transferred to Immobilon-P, blocked in 5% dry milk (MAP kinase) or 3% dry milk/PBS (P-Tyr), and probed with the antibody to PDGFR, P-Tyr, p44/42 MAP kinase, or phospho p44/42 MAP kinase.
Mobilization of Intracellular Ca 2ϩ -Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were plated on coverslips and grown to 80 -90% confluence before the addition of Earle's modified Eagle's medium plus 0.5% fetal calf serum overnight. The coverslips were then washed twice with wash buffer (10 mM Hepes, pH 7.4, 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 0.55 mM glucose), and the cells were loaded with 1 mM fluo-3 AM for 45 min at room temperature. The coverslips were washed and placed into the microscope chamber, followed by the addition of 1 ml of wash buffer containing an additional 1 mM CaCl 2 . This was followed by the addition of 25 ng/ml PDGF or 1% fresh fetal calf serum. The number of cells emitting fluorescence at a wavelength of 488 nm were counted using a Zeiss Axiovert 135 confocal microscope.
PDGF-induced Membrane Ruffling-Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were plated on glass coverslips in 6-well plates. Subconfluent cells were serum-starved for 24 h, at which time they were treated with 5 ng/ml PDGF or serum-free DMEM containing 0.5% fatty acid-free bovine serum albumin for 10 min at 37°C. The medium containing agonist was removed, and the cells were washed once with 1 ml of PBS at room temperature and fixed by adding 1 ml 3.7% formaldehyde in PBS to each well for 10 min with rocking. The fixed cells were washed twice with 5 ml of PBS at room temperature and permeabilized with 1 ml 0.2% Triton X-100 in PBS/well for 5 min with rocking. The cells were washed twice with 1 ml of PBS, and the actin stained with 3.75 units/ml BODIPY 558/568 phalloidin for 20 min in the dark with rocking. Continuing in the dark, the cells were washed three times with 1 ml of PBS, and the coverslips removed from the dish and allowed to dry in air overnight. The coverslips were sealed on a slide using clear nail polish, and fluorescence was observed with a ϫ40 objective using a Leica DMRB microscope.

RESULTS
Growth Factor-induced PLD Activity-In order to test for a role for PLC-␥1 in growth factor-induced PLD activity, we treated TVI-and TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ MEFs with EGF or PDGF. TVI-targeted Plcg1 ϩ/ϩ MEFs have approximately 2.8 ϫ 10 4 EGF receptors/cell, whereas the TVI Plcg1 Ϫ/Ϫ MEFs have approximately 5.2 ϫ 10 4 EGF receptors/ cell (26). In TVI-targeted Plcg1 ϩ/ϩ MEFs, 3 nM EGF caused a 2-fold increase in PLD activity but no increase in TVI-targeted Plcg1 Ϫ/Ϫ cells. In TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ MEFs, EGF did not induce significant PLD activity (Fig. 1A). In TVItargeted Plcg1 ϩ/ϩ MEFs, 50 ng/ml PDGF elicited a 3-fold increase in PLD activity that was inhibited in TVI-targeted Plcg1 Ϫ/Ϫ MEFs. Most strikingly, PDGF produced a robust PLD response in TVII-targeted Plcg1 ϩ/ϩ MEFs that was greatly inhibited in TVII-targeted Plcg1 Ϫ/Ϫ MEFs (Fig. 1B). Thus, both the TVI-and TVII-targeted Plcg1 ϩ/ϩ MEFs showed PDGFinduced PLD activity that was inhibited in Plcg1 Ϫ/Ϫ MEFs. Because the response to PDGF was so much greater in the TVII-targeted Plcg1 ϩ/ϩ MEFs, we selected this cell line to further investigate the role of PLC-␥1 in the PDGF-induced PLD response. InsP x Formation and PtdBut Formation in Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ MEFs-In order to investigate the extent to which disruption of PLC-␥1 eliminated PDGF-induced inositol phosphate formation, we measured total InsP x in the Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ MEFs in response to PDGF. Platelet-derived growth factor increased the production of InsP x in Plcg1 ϩ/ϩ cells in a dose-and time-dependent fashion (Fig. 2). In Plcg1 Ϫ/Ϫ MEFs, PDGF caused a smaller increase in InsP x formation. As shown in Fig. 3, PDGF treatment of Plcg1 ϩ/ϩ MEFs resulted in a doseand time-dependent increase in PtdBut, an unambiguous marker of PLD activity when cells are treated with agonist in the presence of butan-1-ol (27). However, PDGF-induced PLD activity was inhibited in Plcg1 Ϫ/Ϫ MEFs. Whereas the doseresponse curves for InsP x were similar to those for PtdBut ( Fig.  3A; cf. Fig. 2A), the time course for PtdBut indicated no further production after 15 min despite further increases in InsP x ( Fig.  3B; cf. Fig. 2B) These results are consistent with PLC-␥1 acting upstream of PLD in growth factor-induced activation, but they do not explain the cessation of PLD activation at 15 min.
Expression of PLC-␥2 and Ca 2ϩ Mobilization in Plcg1 Ϫ/Ϫ Cells-In Plcg1 Ϫ/Ϫ cells, PDGF caused a small increase in both inositol phosphate production (Fig. 2) and phospholipase D activity (Fig. 3). Furthermore, a small fraction (13.2%) of the Plcg1 Ϫ/Ϫ cells showed intracellular Ca 2ϩ mobilization upon treatment with 25 ng/ml PDGF, whereas a large majority (94%) of the Plcg1 Ϫ/Ϫ cells showed intracellular Ca 2ϩ mobilization upon treatment with 1% fetal calf serum. It has been reported that cells of hematopoietic origin express PLC-␥2, a PLC isoform that is closely related to PLC-␥1 (28). Treatment of Rat 2 cells overexpressing PLC-␥2 with PDGF causes an increase in the tyrosine phosphorylation and activation of PLC-␥2 (29). Thus, we investigated whether Plcg1 Ϫ/Ϫ cells expressed PLC-␥2. Western blot analysis of Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells with a PLC-␥2 antibody showed a large expression of PLC-␥2 in Plcg1 Ϫ/Ϫ cells when compared with Plcg1 ϩ/ϩ cells (Fig. 4A). Furthermore, upon treatment of Plcg1 Ϫ/Ϫ cells with 25 ng/ml PDGF, PLC-␥2 was phosphorylated on tyrosine and co-immunoprecipitated with the PDGF receptor. In Plcg1 ϩ/ϩ cells treated with PDGF, there was no apparent tyrosine phosphorylation of PLC-␥2 and very little co-immunoprecipitation with the PDGF receptor (Fig. 4B). These data suggest that the Plcg1 Ϫ/Ϫ cells may compensate for the disruption of Plcg1 by up-regulating PLC-␥2 and that the activation of PLC-␥2 by PDGF accounts for the small increase in inositol phosphate formation, PLD activation, and the Ca 2ϩ mobilization seen in the Plcg1 Ϫ/Ϫ cells.
PDGF Receptor Level and Autophosphorylation-To establish that the decrease in PDGF-induced production of InsP x and PtdBut in Plcg1 Ϫ/Ϫ cells was not due to a decrease in the number of PDGF receptors, Western blotting of the receptors was performed. This showed no difference in the level of PDGF receptors in Plcg1 Ϫ/Ϫ and Plcg1 ϩ/ϩ cells (data not shown). We also investigated the ability of the PDGFR to be autophosphorylated upon treatment of the Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells with 25 ng/ml PDGF. As shown in Fig. 5, phosphorylation of the PDGFR on tyrosine residues was rapid and sustained for up to 1 h in both cell types, although the effect occurred more rapidly and declined faster in the Plcg1 Ϫ/Ϫ cells. Thus, differences in the level and autophosphorylation of the PDGFR do not account for the decrease in PDGF-induced phospholipase responses in the Plcg1 Ϫ/Ϫ MEFs.  Fig. 1 were treated with the indicated concentrations of PDGF for 10 min, and the radioactivity incorporated into PtdBut was measured. The results are presented as the radioactivity incorporated into PtdBut as a percentage of the radiolabel incorporated into phospholipids. The results are plotted as the means of two separate experiments. B, TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were labeled as described in Fig. 1 and treated with 25 ng/ml PDGF for the times indicated. The data are from a single experiment representative of two.

Reconstitution of PDGF Induced PLD Activation in Plcg1
Ϫ/Ϫ Cells-In cells treated with PDGF, PLC-␥1 is recruited via its SH2 domain to the receptor, where it is activated (6). The activated PLC-␥1 hydrolyzes PIP 2 , resulting in the formation of DAG and IP 3 . Diacylglycerol activates most isozymes of PKC, whereas IP 3 promotes the release of intracellular Ca 2ϩ . It has been proposed that these two second messengers, acting alone or in combination, mediate growth factor activation of PLD (30). If this is true, then replacing these PLC-␥1 products should reconstitute PDGF-induced PLD activation in Plcg1 Ϫ/Ϫ MEFs. As shown in Fig. 6, the addition of the calcium ionophore ionomycin (5 M) plus the cell-permeable DAG analog OAG (40 M) resulted in an equal PLD response in the Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells. The response in the Plcg1 Ϫ/Ϫ cells to OAG plus ionomycin was similar to that induced by 50 ng/ml PDGF in Plcg1 ϩ/ϩ cells. These data show that the deficient PDGF-induced PLD response in Plcg1 Ϫ/Ϫ cells can be entirely restored by addition of agents that mimic the activation of PLC-␥1.
PDGF-Induced Activation of MAP Kinase-Treatment of many cell types with growth factors results in activation of the MAP kinase pathway. This activation occurs through the activation of Ras, followed by the activation of MEK kinase, MEK, and finally, MAP kinase (31). There is evidence that phospholipase C-␥1 is not involved in the growth factor-induced activation of MAP kinase (26,32). In order to confirm that the deletion of Plcg1 by targeted gene disruption did not result in secondary changes to other PDGF signaling pathways in the MEFs derived from the targeted embryos, we assessed the activation of MAP kinase in Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells using an antibody that recognizes the phosphorylated form of MAP kinase. We showed that PDGF treatment of Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells resulted in a rapid and transient increase in MAP kinase phosphorylation that was similar between the two cells (Fig. 7). This result suggests that only pathways that require the activation of PLC-␥1 are inhibited in Plcg1 Ϫ/Ϫ MEFs.
Requirement of PLD for PDGF-induced Membrane Ruffling-Two recent reports suggest that the stimulation of actin stress fiber formation by lysophosphatidic acid or ␣-thrombin is mediated by the activation of PLD (33,34). Platelet-derived growth factor has been shown to induce the formation of polymerized actin at the plasma membrane in Swiss 3T3 cells, forming membrane ruffles (35). Because PLD activation by PDGF is inhibited in Plcg Ϫ/Ϫ cells, we used these cells to test whether PLD activation was necessary for PDGF-induced membrane ruffles. Fig. 8 shows that PDGF induces membrane ruffles in Plcg1 ϩ/ϩ cells, and this is not inhibited in Plcg1 Ϫ/Ϫ MEFs. These data suggest that PLD activation may not be required for the increase of polymerized actin localized in ruffles at the plasma membrane induced by PDGF. DISCUSSION Growth factors stimulate the PLD-induced hydrolysis of phosphatidylcholine (PC) to PA and choline in a variety of cell types (30). The exact pathway by which this occurs is not clear. A number of mechanisms of activation have been proposed involving protein tyrosine kinases, PKC, Ca 2ϩ , and GTP-binding proteins (36). In an elegant study seeking to define a role for PLC-␥1 in PDGF-induced PLD activation, Yeo et al. (22) measured PLD activity in TRMP cells (a kidney epithelial cell line) expressing wild type PDGF receptors or various tyrosine mutated PDGF receptors. They reported that PDGF had no effect on PLD activity in PDGFR kinase-deficient TRMP cells, but the PDGF-induced PLD activity was restored in cells containing a mutant PDGFR that was able to bind PLC-␥1 but not other signaling proteins. Furthermore, they showed that a mutant PDGFR that could not activate PLC-␥1 was unable to activate PLD. These data suggest that PLC-␥1 is necessary and sufficient for PDGF-induced PLD activity. However, the experiments were conducted with a cell line that normally lacked the PDGFR (37). Further evidence for a role of PLC-␥ in the activation of PLD came from a study in which PLC-␥1 was overexpressed in NIH3T3 cells. Lee et al. (38) found that PDGFinduced PLD activity was directly related to the level of PLC-␥1 expressed in the cells, and that down-regulation of PKC by PMA pretreatment completely blocked PLD activation. These data again suggest PLD lies downstream of PLC-␥1 and PKC.
On the other hand, there are data that show agonist-induced PC hydrolysis or PLD activation in the absence of detectable PIP 2 breakdown (16,18,20). Cook and Wakelam (21) showed EGF stimulation of PLD activity in Swiss 3T3 cells in the absence of measurable PIP 2 hydrolysis and in the presence of a PKC inhibitor, although it was later found that EGF induced a FIG. 4. PLC-␥2 Western blot and PLC-␥2 immunoprecipitation and Western blot with P-Tyr. A, TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were lysed in 300 l of PBS containing 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 1 g/ml aprotinin, 1 g/ml leupeptin, and 1 g/ml pepstatin. Ten g of each cell lysate was separated by SDS-PAGE on a 6% gel, and the proteins were transferred to Immobilon P and probed with an antibody to PLC-␥2 as described under "Experimental Procedures." HL-60 cells were used as the source of the PLC-␥2 standard. The results are representative of three separate experiments. B, TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were treated with 25 ng/ml PDGF or serum-free medium for 3 min, lysed in radioimmune precipitation buffer, and immunoprecipitated overnight with an antibody to PLC-␥2. The immunoprecipitated proteins were separated on a 6% SDS-PAGE gel and probed with an antibody to P-Tyr (4G10) as described under "Experimental Procedures." The results are representative of two separate experiments.
FIG. 5. Western blot of platelet-derived growth factor receptor with P-Tyr antibody. TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were lysed in 300 l of PBS containing 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 1 g/ml aprotinin, 1 g/ml leupeptin, and 1 g/ml pepstatin. Ten g of each cell lysate was separated by SDS-PAGE on a 6% gel, and the proteins were transferred to Immobilon P and probed with a mouse monoclonal anti-P-Tyr antibody (4G10) as described under "Experimental Procedures." The experiment is representative of three separate experiments.
small increase in InsP x in these cells and that a PKC inhibitor did decrease the PLD response (12). When these reports are coupled with the frequent finding that PKC inhibitors produce only partial inhibition of the actions of growth factors and other agonists on PLD (10,30), questions remain about the extent of the contribution of PIP 2 hydrolysis and PKC to the regulation of PLD.
The data presented strongly suggest that PLC-␥1 is required for PLD activation by PDGF. The PLD response of TVI-targeted Plcg1 ϩ/ϩ cells to EGF and PDGF is small, and the response is inhibited in TVI-targeted Plcg1 Ϫ/Ϫ cells (Fig. 1). Furthermore, the PDGF-induced PLD response of TVII-targeted Plcg1 ϩ/ϩ cells is robust and is greatly inhibited in TVII-targeted Plcg1 Ϫ/Ϫ cells (Fig. 1). A small but reproducible, increase in PLD activity with EGF and a contrasting robust PLD response to PDGF have been seen in a variety of cell types, including Rat1 and Swiss 3T3 fibroblasts (12,39). The present study thus adds to earlier data reporting differences in the signal transduction pathways for the two growth factors (40 -42).
Phospholipase C is responsible for the hydrolysis of PIP 2 to IP 3 and DAG (6). In Plcg1 ϩ/ϩ cells, PDGF elicits an increase in the production of total inositol phosphates in a dose-and time-dependent manner (Fig. 2). In Plcg1 Ϫ/Ϫ cells, PDGF caused a small increase in inositol phosphate production (Fig. 2). Our data suggest that the small increase in inositol phosphate production in Plcg1 Ϫ/Ϫ cells is due to the expression of PLC-␥2 in the Plcg1 Ϫ/Ϫ cells (Fig. 4A). Treatment of Plcg1 Ϫ/Ϫ cells with PDGF resulted in tyrosine phosphorylation of PLC-␥2 (Fig.  4B). PDGF treatment of rat-2 cells overexpressing PLC-␥2 increases the tyrosine phosphorylation and the in vivo activity of PLC-␥2 (29). Moreover, treatment of Plcg1 Ϫ/Ϫ cells with PDGF resulted in the mobilization of intracellular Ca 2ϩ in a small population of the cells. Over-expression of PLC-␥2 in NIH3T3 cells also enhances PDGF-induced mobilization of intracellular Ca 2ϩ (43). Thus, our data suggest that the disruption of Plcg1 resulted in a compensatory up-regulation of PLC-␥2 and that this isoform, which is closely related to PLC-␥1, is responsible for the increase in InsP x formation (Fig. 2), PLD activity (Fig. 3), and intracellular Ca 2ϩ mobilization seen in Plcg1 Ϫ/Ϫ cells upon treatment with PDGF.
If PLC-␥1 acts upstream of PLD in the PDGF-induced PLDactivation pathway, then PLD activity should be inhibited in Plcg1 Ϫ/Ϫ cells. Furthermore, there should be a correlation between PLC and PLD activities. Treatment of Plcg1 ϩ/ϩ cells with PDGF results in a dose-dependent increase in PtdBut formation that mirrors the dose-response curve for PDGF-induced InsP x production in Plcg1 ϩ/ϩ (Fig. 3A; cf. Fig. 2A). The PLD response to PDGF treatment in Plcg1 Ϫ/Ϫ cells is inhibited in parallel with the decrease in the PLC response ( Fig. 3A; cf. Fig. 2A). However, the PDGF-induced PLD response reached a maximum at 15 min in the Plcg1 ϩ/ϩ or the Plcg1 Ϫ/Ϫ cells, at which time the production of PtdBut ceased (Fig. 3B). This is in contrast to the PDGF-induced InsP x production, which was still increasing at 60 min. This same pattern of phosphatidylalcohol and inositol phosphate production was reported in NIH3T3␥-1 cells, which overexpress PLC-␥1 and in which phosphatidylethanol production reached a maximum at 10 min in response to PDGF, whereas InsP x production was still increasing at 30 min (44). Exploration of the reasons for the cessation of PtdBut formation is outside the scope of the present study, but it is possible that activation of PLC and the consequent acti- FIG. 6. PLD activation by OAG and ionomycin. TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were labeled with [ 3 H]myristic acid as described in Fig. 1 and treated with 50 ng/ml PDGF, 5 M ionomycin, 40 M OAG, or ionomycin and OAG in combination for 10 min. The radioactivity incorporated into PtdBut was measured as described under "Experimental Procedures," and the results are presented as the radioactivity incorporated into PtdBut as a percentage of total radioactivity incorporated into phospholipids. The results are plotted as the means Ϯ S.E. of three separate experiments.
FIG. 7. Western blot of MAP kinase. TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells were lysed as described in Fig. 4, and equal proteins were separated by SDS-PAGE on a 4 -20% gradient gel. The proteins were transferred to Immobilon P and probed with either the rabbit polyclonal anti-p44/42 MAP kinase or the rabbit polyclonal anti-phospho-specific p44/42 MAP kinase antibody as described under "Experimental Procedures." This experiment is representative of two. vation of PKC and mobilization of Ca 2ϩ could have an initial stimulatory effect on PLD followed by an inhibitory action, due perhaps to phosphorylation of PLD or an inhibitory protein.
Phosphorylation of PLD by PKC has recently been reported to inhibit its activity (45).
The inhibition of PDGF-induced PLD activity in Plcg1 Ϫ/Ϫ cannot be attributed to a decreased level of PDGF receptors in the Plcg1 Ϫ/Ϫ cells (data not shown) or to a defect in PDGFinduced autophosphorylation (Fig. 5). In fact, the autophosphorylation of the receptor occurred more rapidly in the Plcg1 Ϫ/Ϫ cells as compared with the Plcg1 ϩ/ϩ cells, reaching a maximum level at 5 min and decreasing toward basal level by 60 min, but this difference in autophosphorylation cannot account for the decreased PDGF-induced PLD activity seen in the Plcg1 Ϫ/Ϫ cells. Another possible explanation is that PKC is deficient in the Plcg1 Ϫ/Ϫ cells. However, this does not seem to be the case because treatment of Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells with phorbol ester, an activator of PKC, results in a similar dose-dependent activation of PLD (data not shown).
In cells treated with growth factors, activated PLC-␥1 hydrolyzes PIP 2 to form IP 3 and DAG, resulting in an increase in intracellular Ca 2ϩ and the activation of PKC. Treatment of Plcg1 Ϫ/Ϫ cells with the Ca 2ϩ ionophore ionomycin and the cell-permeable DAG analogue OAG resulted in a PLD response that was similar to that in Plcg1 ϩ/ϩ cells and slightly greater than the PLD response induced by PDGF (Fig. 6). Thus, the addition of PLC-␥1 activation products to the Plcg1 Ϫ/Ϫ cells reconstituted the PDGF-induced PLD response in the Plcg1 Ϫ/Ϫ cells to the level seen in the wild type cells. These data suggest that the PLC-␥1 activation products are sufficient to completely restore the PDGF-induced PLD response lost in the Plcg1 Ϫ/Ϫ upon disruption of PLC-␥1 in these cells. Furthermore, these results and those with phorbol ester prove that PLD is not deficient in the Plcg1 Ϫ/Ϫ cells.
Growth factor treatment of cells results in a mitogenic response that is mediated by the MAP kinase. The growth factorinduced activation of MAP kinase involves the sequential activation of Ras, MEK kinase, and MEK (31). In a 3T3 cell line derivative, NR6 cells, EGF-stimulated MAP kinase activity was not affected by the inhibition of PLC with U73122 (32), and data from Ji et al. (26) showed that EGF-induced activation of MAP kinase in the TVI-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells was similar. Thus, these findings indicate that PLC-␥1 is not involved in the phosphorylation and activation of MAP kinase by growth factors. We also observed PDGF-induced phosphorylation of MAP kinase in Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells (Fig. 7), indicating that the targeted gene disruption of PLC-␥1 did not result in global changes to PDGF-signaling pathways.
Data from two recent reports suggest a role for PA in the polymerization of actin stress fibers (33,34). Actin stress fibers are a major component of the cytoskeleton in fibroblasts, where actin filaments can exist in three types of structures, including actin stress fibers, the cortical actin network, and cell surface protrusions, such as membrane ruffles and filopodia (35). Ha and Exton (34) reported that treatment of IIC9 fibroblasts with thrombin, PLD from Streptomyces chromofuscus, or exogenous PA resulted in actin stress fiber formation. In porcine aortic endothelial cells, lysophosphatidic acid treatment activated PLD, resulting in the formation of PA, in the apparent absence of the formation of other lipid second messengers (33). Lysophosphatidic acid, like exogenously added PA, also stimulated the formation of actin stress fibers (33). Although these observations generally support a role for PLD in stress fiber formation, it is possible that signals evoked by the exogenous PLD and PA are different from those elicited by activation of endogenous PLD (46). For example, they could generate lysophosphatidic acid, which could induce actin polymerization by a different mechanism.
In Swiss 3T3 cells, PDGF has been shown to induce the formation of membrane ruffles (35), and we therefore utilized the Plcg1 Ϫ/Ϫ cells to examine the role of PLD in this effect. In Plcg1 Ϫ/Ϫ cells, 5 ng/ml PDGF induced membrane ruffles similar to those induced in Plcg1 ϩ/ϩ cells (Fig. 8), even though activation of PLD was significantly inhibited (Fig. 2B). Thus, it appears that PLD and PLC activity may not be necessary for the PDGF-induced formation of membrane ruffles. It seems unlikely that the small level of PLC-␥2, inositol phosphate formation, and Ca 2ϩ mobilization would be sufficient to provoke maximal ruffling response.
In summary, the present data suggest that in mouse embryo fibroblasts, PLC-␥1 activation is necessary for the PDGF-induced activation of PLD. However, caution should be exercised in extrapolating the findings to other agonists or cell types. We are currently investigating a role for PLC-␥1 in the activation of PLD by various other agonists, including those that activate heterotrimeric G-proteins. FIG. 8. PDGF-induced membrane ruffling. TVII-targeted Plcg1 ϩ/ϩ and Plcg1 Ϫ/Ϫ cells plated on glass coverslips were treated with 5 ng/ml PDGF for 10 min. The cells were fixed and permeabilized, and the actin was stained with BODIPY phalloidin as described under "Experimental Procedures." Membrane ruffles were visualized with a ϫ40 objective using a Leica DMRB microscope. A, Plcg1 ϩ/ϩ cells treated with vehicle for 10 min. B, Plcg1 ϩ/ϩ cells treated with 5 ng/ml PDGF for 10 min. C, Plcg1 Ϫ/Ϫ cells treated with vehicle for 10 min. D, Plcg1 Ϫ/Ϫ cells treated with 5 ng/ml PDGF for 10 min.