Synergistic activation of rat brain phospholipase D by ADP-ribosylation factor and rhoA p21, and its inhibition by Clostridium botulinum C3 exoenzyme.

An activator of rat brain phospholipase D (PLD) that is distinct from the already identified PLD activator, ADP-ribosylation factor (ARF), was partially purified from bovine brain cytosol by a series of chromatographic steps. The partially purified preparation contained a 22-kDa substrate for Clostridium botulinum C3 exoenzyme ADP-ribosyltransferase, which strongly reacted with anti-rhoA p21 antibody, but not with anti-rac1 p21 or anti-cdc42Hs p21 antibody. Treatment of the partially purified PLD-activating factor with both C3 exoenzyme and NAD significantly inhibited the PLD-stimulating activity. These results suggest that rhoA p21 is, at least in part, responsible for the PLD-stimulating activity in the preparation. Recombinant isoprenylated rhoA p21 expressed in and purified from Sf9 cells activated rat brain PLD in a concentration- and GTPS (guanosine 5′-O-(3-thiotriphosphate))-dependent manner. In contrast, recombinant non-isoprenylated rhoA p21 (fused to glutathione S-transferase) expressed in Escherichia coli failed to activate the PLD. This difference cannot be explained by a lower affinity of non-isoprenylated rhoA p21 for GTPS, as the rates of [S]GTPS binding were very similar for both recombinant preparations and the GTPS-bound form of non-isoprenylated rhoA p21 did not induce PLD activation. Interestingly, recombinant isoprenylated rhoA p21 and ARF synergistically activated rat brain PLD; a similar pattern was seen with the partially purified PLD-activating factor. The synergistic activation was inhibited by C3 exoenzyme-catalyzed ADP-ribosylation of recombinant isoprenylated rhoA p21 in a NAD-dependent manner. Inhibition correlated with the extent of ADP-ribosylation. These findings suggest that rhoA p21 regulates rat brain PLD in concert with ARF, and that isoprenylation of rhoA p21 is essential for PLD regulation in vitro.

Phospholipase D (PLD) 1 catalyzes the hydrolysis primarily of phosphatidylcholine (PC) to produce choline and phosphatidic acid (PA) (1)(2)(3). In the presence of ethanol, however, PLD produces phosphatidylethanol (PEt) by a transphosphatidylation reaction. The latter product provides an unambiguous marker for assay of PLD (1)(2)(3). Using this assay, PLD activation upon agonist stimulation has been demonstrated in many types of mammalian cells and tissues (see Refs. 4 and 5 and references therein).
It has been suggested that protein kinase C, calcium ion, pertussis toxin-sensitive and -insensitive GTP-binding proteins, and protein tyrosine kinase are involved in agoniststimulated PLD activation (see Refs. 4 and 5 and references therein; see also Refs. 6 and 7). Of these putative PLD regulators, a member of the family of low molecular weight GTPbinding proteins (small G proteins), ADP-ribosylation factor (ARF), has been identified as a PLD activator by two independent research groups (8,9). Brown et al. (8) have reconstituted PLD partially purified from plasma membranes of HL-60 cells with purified ARF. Cockcroft et al. (9) have employed streptolysin O-permeabilized HL-60 cells as a PLD source for reconstitution with purified ARF. In addition to the PLD of HL-60 cells, activation by ARF has also been documented with partially purified rat brain PLD (10). Thus, PLDs of HL-60 cells and of rat brain are directly or indirectly regulated by ARF.
Another member of the small G protein family, rhoA p21, has been identified by Malcolm et al. (11) as a PLD activator in studies on rat liver membranes. Rat liver PLD, however, was found not to be activated by ARF (11). These and the preceding results suggest that there are different isoforms of PLDs and that they are regulated by distinct mechanisms. Massenburg et al. (10) have provided support for this concept, demonstrating that rat brain membranes contain two isoforms of PLD, one of which is ARF-sensitive and another oleate-sensitive but ARFinsensitive. These reports raised the question whether rat brain PLD, like rat liver PLD, might also be activated by rhoA p21.
In the process of confirming ARF activation of partially purified rat brain PLD, we have found that bovine brain cytosol contains a PLD-activating factor that is distinct from ARF and may be rhoA p21. In the present study, we show directly that rhoA p21 can also activate rat brain PLD. Inconsistent with the observations of Malcolm et al. (11), however, ADP-ribosylation of rhoA p21 by Clostridium botulinum C3 exoenzyme was found to interfere with its activation of rat brain PLD. Furthermore, it was demonstrated that rat brain PLD is synergistically activated by ARF and rhoA p21.
Partial Purification of PLD and ARF-ARF-sensitive PLD was partially purified from rat brain membranes by a method based on that of Brown et al. (8). Rat brain was homogenized in TEDP (20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol (DTT), and 0.3 mM phenylmethylsulfonyl fluoride (PMSF)), filtered through two layers of cheesecloth, and centrifuged at 20,000 ϫ g for 30 min. To extract the PLD, the crude membranes (at 5 mg of protein/ml) were incubated for 1 h at 4°C with 0.6 M NaCl in 20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 10 mM EGTA, 5 mM DTT, and 1 mM PMSF. After centrifugation at 100,000 ϫ g for 30 min, the clear supernatant was diluted 3-fold with 20 mM Na-Hepes, pH 7.5, 10 mM EGTA, 1 mM DTT, and 0.1 mM PMSF (HEDP), and subjected to heparin-Sepharose CL-6B column chromatography. The column was washed, and then PLD was eluted with a linear gradient of 0.3-2 M NaCl in HEDP. The ARF-sensitive PLD activity was eluted as a single peak at 1.1 M NaCl. This partially purified PLD preparation was used as rat brain PLD in this study.
ARF was partially purified from bovine brain cytosol. Briefly, bovine brain was homogenized, filtered, and centrifuged as described above with rat brain. Proteins from the cytosol fraction were precipitated with ammonium sulfate at 25-70% saturation and dissolved in TEDP. After dialysis against TEDP, the proteins were subjected to DEAE-Sephacel column chromatography. ARF fractions eluted at 60 mM NaCl from the column were concentrated by Mini module (Asahi Kasei, Tokyo, Japan), and then applied onto a Sephacryl S-200 column that had been equilibrated with TEDP containing 50 mM NaCl and 1 M GDP. The eluted ARF fractions were combined and concentrated by Amicon Centriprep.
Partial Purification of a PLD-activating Factor-A PLD-activating factor was partially purified from bovine brain cytosol. This factor was assayed by reconstitution with rat brain PLD. Proteins of bovine brain cytosol (as described above) were precipitated with ammonium sulfate at 55-75% saturation, and dissolved in 20 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM DTT, 1 mM EDTA, and 1 M GDP (TMDEG). After dialysis against TMDEG, the sample was loaded onto a DEAE-Toyopearl 650S column that had been equilibrated with TMDEG. The column was washed, and the activity was eluted with a linear gradient of 0 -0.2 M NaCl in TMDEG. The flow-through fractions contained ARF, and they activated rat brain PLD (data not shown). Fractions 2-18, which were devoid of ARF, also contained a PLD-stimulating activity. These fractions eluted at 60 mM NaCl as a single peak (data not shown); they were combined, and ammonium sulfate was added to 40% saturation. The sample was applied onto a phenyl-Sepharose CL-4B column that had been equilibrated with TMDEG containing 40% ammonium sulfate. The column was washed, and then the activity was eluted with a reverse gradient of 40 -0% ammonium sulfate and 0 -50% ethylene glycol in TMDEG. PLD-stimulating activity was eluted as two peaks at 25% and 2% of ammonium sulfate (data not shown). Fractions from the first peak were concentrated by Amicon Centriprep, and subjected to gel filtration chromatography on a TSK-GEL 3000 SW XL column that had been equilibrated with TMDEG containing 0.5 M KCl and 5% ethylene glycol. The PLD-stimulating activity eluted as a single peak at ϳ45 kDa (data not shown). This material was applied, after buffer exchange on a PD 10 column (Pharmacia Biotech Inc.), onto a Mono Q column that had been equilibrated with TMDEG containing 5% ethylene glycol. The column was washed, and the activity was eluted with a linear gradient of 0 -0.25 M NaCl in TMDEG containing 5% ethylene glycol (see Fig. 1). (8,10). Partially purified rat brain PLD was reconstituted in mixed phospholipid vesicles with the column fractions or recombinant rhoA p21s and/or ARF. The phospholipid vesicles (158.64 M phospholipids) comprised PE, PIP 2 , and [choline-3 H]DPPC or [2-palmitoyl-9,10-3 H]DPPC in a molar ratio of 16:1.4:1 (total volume of 50 l), and the reconstitution buffer was 50 mM Na-Hepes, pH 7.5, 3 mM EGTA, 80 mM KCl, 1 mM DTT, 3 mM MgCl 2 , and 2 mM CaCl 2 . For assay of [ 3 H]PEt formation, the reaction included 1% ethanol. After incubation at 37°C for the indicated time in the presence or absence of 40 M GTP␥S, the reaction was terminated by successive addition of 400 l of CHCl 3 /MeOH (1:2), 100 l of CHCl 3 , and 100 l of 0.2 M KCl/5 mM EDTA. Samples were centrifuged at 12,000 rpm for 2 min to separate the aqueous and organic phases. In some studies, PLD-mediated production of [ 3 H]choline was assessed by liquid scintillation spectroscopy of the aqueous extract. In other studies, [ 3 H]PEt formation was assayed by thin layer chromatography of the lipid extract, as described previously (6). In Figs. 3 and 7, the partially purified PLD-stimulating factor and recombinant isoprenylated rhoA p21 were pretreated with 1 g of C3 exoenzyme/ml and/or the indicated concentrations of NAD for the indicated time at 30°C before PLD assay. In Fig. 5B, PLD assay was carried out after recombinant rhoA p21s were loaded with GTP␥S.

PLD Assay-PLD activity was assessed by determining the production of [ 3 H]choline or of [ 3 H]PEt
[ 32 P]ADP-ribosylation-ARF was detected through its ability to stimulate auto-ADP-ribosylation of cholera toxin (15). Samples were incubated with 1 g of cholera toxin/ml in the presence of 1 M [ 32 P]NAD as already reported (16). [ 32 P]ADP-ribosylation of the column fractions by C3 exoenzyme was assessed in the presence of 400 g/ml PI as previously reported (17). PI was excluded during [ 32 P]ADP-ribosylation of recombinant rhoA p21 shown in the studies shown in Fig. 7. In this experiment, [ 32 P]NAD with a specific activity of 4,000 cpm/pmol was employed as the substrate. Incorporation of [ 32 P]ADP-ribose into the cholera toxin A 1 peptide and C3 exoenzyme substrate were analyzed by autoradiography and Bioimage BAS2000 (Fujix, Tokyo, Japan) analysis after separation of proteins on 15% gels by SDS-PAGE.
Immunoblotting-Proteins were separated by SDS-PAGE (15% gels) and transferred to nitrocellulose paper. The paper was treated for 1 h at 37°C with 50 mM Tris-HCl, pH 8.0, 80 mM NaCl, and 2 mM CaCl 2 containing 5% nonfat dry milk and 0.2% Nonidet P-40, then incubated for 1 h at 37°C with anti-rhoA p21, anti-rac1 p21, or anti-cdc42Hs p21 antibody in the same buffer. After washing with the same buffer, the paper was incubated for 1 h at 37°C with peroxidase-conjugated goat anti-rabbit IgG. The immunoreactive proteins were detected by an enhanced chemiluminescence detection system (Amersham Corp.).
[ 35 S]GTP␥S binding to recombinant rhoA p21s-Recombinant isoprenylated or non-isoprenylated rhoA p21 (10 nM) was incubated for the indicated time at 37°C with 1 M [ 35 S]GTP␥S in 20 mM Hepes-NaOH, pH 7.5, 100 mM NaCl, 1 mM MgCl 2 , 2 mM EDTA, and 10 mM DTT. After the reaction was terminated by adding 2.5 ml of buffer consisting of 20 mM Tris-HCl, pH 7.5, 3 mM MgCl 2 , 50 mM NaCl, and 1 M GTP␥S, samples were transferred to nitrocellulose filters. Filters were washed thoroughly with the same buffer, and then the bound [ 35 S]GTP␥S was quantified by liquid scintillation spectroscopy.

RESULTS
An activator of rat brain PLD that is distinct from the already identified PLD activator ARF was partially purified from bovine brain cytosol by a series of chromatographic steps (see "Experimental Procedures"). The PLD-stimulating activity and C3 exoenzyme substrate of 22 kDa apparent molecular mass were co-purified during each step (data not shown). The final chromatography on Mono Q resulted in elution of both the PLD-stimulating activity and the C3 exoenzyme substrate as two peaks (Fig. 1). The major, second peak of PLD-stimulating activity from the Mono Q column contained a 22-kDa protein as assessed by SDS-PAGE ( Fig. 2A). This protein strongly reacted with anti-rhoA p21 antibody, but not with anti-rac1 p21 or anti-cdc42Hs p21 antibody (Fig. 2B). The PLD-stimulating activity of the partially purified factor was suppressed upon incubation with both NAD and C3 exoenzyme, which specifically ADP-ribosylates and inhibits rho proteins (Fig. 3); no suppression occurred with either agent alone. These results, taken together, imply that rhoA p21 is an activator of rat brain PLD, in agreement with the finding of Malcolm et al. (11).
To confirm this hypothesis, isoprenylated rhoA p21 was expressed in and purified from Sf9 cells (13), and reconstituted with partially purified rat brain PLD. As shown in Fig. 4A, recombinant isoprenylated rhoA p21 activated in a concentration-dependent manner the rat brain PLD in the presence of GTP␥S, but not in its absence. In contrast, non-isoprenylated rhoA p21, expressed in and purified from E. coli, failed to activate PLD (Fig. 4B). Recombinant isoprenylated and nonisoprenylated rhoA p21s both bound [ 35 S]GTP␥S with similar kinetics (Fig. 5A). Furthermore, recombinant non-isoprenylated rhoA that had been preloaded with GTP␥S was also unable to effect PLD activation (Fig. 5B). These findings rule out a difference in binding of GTP␥S as the explanation for the difference in PLD activation between the two recombinant forms of rhoA p21. The results further indicate that rhoA p21 directly or indirectly activates rat brain PLD, and that posttranslational modification of rhoA p21, like ARF (8,9), is essential for it to function as a PLD activator.
Whereas the PLD in rat liver membranes is activated by rhoA p21 but insensitive to ARF (11), rat brain PLD was activated by ARF (Fig. 6) as well as by rhoA p21 (Figs. 4 and 6). When the partially purified PLD-activating factor or recombinant isoprenylated rhoA p21 was reconstituted together with ARF, synergistic activation of rat brain PLD was observed (Fig.  6). Consistent with the result illustrated in Fig. 3, incubation of recombinant isoprenylated rhoA p21 with C3 exoenzyme plus a series of concentrations of NAD attenuated the synergistic activation of PLD in a NAD concentration-dependent manner (Fig. 7). The degree of inhibition correlated well with the extent of ADP-ribosylation of recombinant isoprenylated rhoA p21 by the C3 exoenzyme (Fig. 7). DISCUSSION Malcolm et al. (11) have recently reported that rhoA p21 activates rat liver PLD. We have shown in the present study that rat brain PLD is also activated by rhoA p21 (Fig. 4A). The mechanism for activation of rat brain PLD by rhoA p21, however, seems to be different from that of rat liver PLD. In FIG. 1. Mono Q column chromatography of the PLD-activating factor of bovine brain cytosol. The PLD-activating factor that is distinct from ARF was partially purified by DEAE-Toyopearl 650S, phenyl-Sepharose CL-4B, and TSK-GEL 3000 SW XL column chromatography (as described under "Experimental Procedures"). Fractions containing the PLD-stimulating activity eluted from TSK-GEL 3000 SW XL column were subjected to Mono Q column chromatography. Each fraction was reconstituted with the partially purified rat brain PLD in the presence of 40 M GTP␥S. After incubation at 37°C for 1 h, the PLD activity (q) was determined by measuring [ 3 H]choline formation (as described under "Experimental Procedures"). C3 exoenzyme substrate in the fractions was analyzed by [ 32 P]ADP-ribosylation with C3 exoenzyme and [ 32 P]NAD (inset), as described under "Experimental Procedures." FIG. 2. Protein staining and immunoblotting of the partially purified preparation of PLD-activating factor probed with anti-rhoA p21, anti-rac1 p21, and anti-cdc42Hs p21 antibodies. Proteins in fraction 34 of the Mono Q column eluate containing the PLDstimulating activity were analyzed by SDS-PAGE on 15% gels and Coomassie Brilliant Blue (CBB) staining (A). Immunoblots of the partially purified preparation of PLD-activating factor were probed with anti-rhoA p21, anti-rac1 p21, and anti-cdc42Hs p 21 antibodies (B), as described under "Experimental Procedures. "   FIG. 3. Effects of C3 exoenzyme and/or NAD on the PLD-stimulating activity of partially purified PLD-activating factor. The partially purified PLD-activating factor was incubated for 20 min at 30°C without or with 1 g/ml C3 exoenzyme and/or 20 M NAD, reconstituted with partially purified rat brain PLD, and the PLD activity was determined by measuring [ 3 H]choline formation (as described under "Experimental Procedures"). Bars represent differences between duplicate determinations in a typical study.

FIG. 4. Effects of isoprenylated and non-isoprenylated rhoA
p21s on rat brain PLD activity. Partially purified rat brain PLD was reconstituted with isoprenylated (A) and non-isoprenylated (B) rhoA p21s, which were expressed in and purified from Sf9 cells and E. coli, respectively. The reconstituted proteins were incubated at 37°C without (E) or with 40 M GTP␥S (ql) in the presence of 1% ethanol. [ 3 H]PEt production over 1 h was analyzed as described under "Experimental Procedures." Bars represent differences between duplicate determinations in a typical study. particular, ADP-ribosylation of rhoA p21 by C3 exoenzyme perturbs the ability of rhoA p21 to activate rat brain PLD (Figs.  3 and 7), whereas rat liver PLD is still activated by ADPribosylated rhoA p21 (11).
Another difference between rat brain PLD (our observations) and rat liver PLD (data from Malcolm et al.) is the sensitivity of the PLD(s) to ARF; we found rat brain PLD to be activated by ARF (Fig. 6), whereas rat liver PLD is insensitive to ARF (11). This supports the concept that the isoforms of PLD in rat liver and brain are distinct. A possible alternative is that rat brain contains two isoforms of PLD, one of which is ARF-sensitive and another rhoA p21-sensitive, whereas rat liver contains only the rhoA p21-sensitive PLD. However, this is unlikely. If there were two isoforms of PLD in rat brain, the activation of rat brain PLDs by ARF plus rhoA p21 would be expected to be additive. On the other hand, if rat brain PLD (considered a distinct isoform from that in rat liver) were regulated by both ARF and rhoA p21 by different mechanisms, activation by these agents might be synergistic. The results of the present study provide strong support for the latter possibility, as recombinant isoprenylated rhoA p21 and ARF activated rat brain PLD in a synergistic manner (Fig. 6). Thus, it is likely that rat brain PLD differs from rat liver PLD. Very recently, Siddiqi et al. (18) and Singer et al. (19) have reported that PLDs in membranes of HL-60 cells and porcine brain, respectively, are activated synergistically by ARF and rhoA p21. Their data suggest that the PLDs in membranes of HL-60 cells and porcine brain are analogous to that in rat brain membranes.
The PLD-stimulating activity partially purified from bovine brain cytosol was inhibited only incompletely (by about 50%) on ADP-ribosylation of rhoA p21 by C3 exoenzyme and NAD (Fig.  3). Although rac1 p21 and cdc42Hs p21 were not detected in this preparation (Fig. 2B), it is still possible that another C3 exoenzyme-insensitive PLD activator(s) contaminated the preparation. It has not proven practicable to assess whether exhaustive ADP-ribosylation of the preparation results in complete inhibition, because prolonged incubation (more than 20 min at 30°C) of the partially purified PLD-activating factor drastically diminishes its activity even in the absence of C3 exoenzyme and NAD. Nevertheless, the results obtained (Fig.  3) imply that rhoA p21 is responsible, at least in part, for the PLD-stimulating activity of the preparation. Perturbation by C3 exoenzyme-catalyzed ADP-ribosylation of the ability of rhoA p21 to modulate rat brain PLD activity was clearly demonstrated in the studies employing recombinant rhoA p21 (Fig.  7). Although the rate of ADP-ribosylation of partially purified bovine brain rhoA was very slow (reaching a plateau after 60 -120 min of incubation), ADP-ribosylation of the recombinant isoprenylated rhoA reached a maximum within 5 min under the conditions employed in Fig. 7 (data not shown). Taking advantage of this characteristic of the recombinant rhoA, we were able to demonstrate that the extent of ADPribosylation of rhoA correlated well with the degree of inhibition of the synergistic activation of PLD by rhoA in concert with ARF (Fig. 7).
Synergistic activation of PLD by rhoA p21 and ARF has also been found with permeabilized, cytosol-depleted rabbit neutro- FIG. 6. Synergistic activation of rat brain PLD by ARF and the partially purified PLD-activating factor or isoprenylated rhoA p21. Partially purified rat brain PLD was reconstituted without (E) or with (q) 1 l of the PLD-activating fraction obtained from the DEAE-Toyopearl 650S column (A) or 10 nM recombinant isoprenylated rhoA p21 (B) in the presence of the indicated concentrations of ARF. After incubation for 1 h at 37°C in the presence of 40 M GTP␥S and 1% ethanol, [ 3 H]PEt formation was assessed (as described under "Experimental Procedures"). Bars represent differences between in duplicate determinations in a typical study.
FIG. 7. C3 exoenzyme-catalyzed ADP-ribosylation of recombinant rhoA p21 inhibits its ability to activate rat brain PLD synergistically with ARF. Recombinant isoprenylated rhoA p21 (25 nM) was incubated for 5 min at 30°C without (E) or with 1 g/ml C3 exoenzyme (q) in the presence of 25 g/ml ARF. ARF alone (25 g/ml) was also incubated without (Ⅺ) or with C3 exoenzyme (s) under the same conditions. The reaction mixtures included the indicated concentrations of NAD. The pretreated small G proteins were reconstituted with partially purified rat brain PLD, incubated for 10  phils, 2 suggesting that rabbit neutrophil PLD is regulated by the same (or similar) mechanism as rat brain PLD. Lambeth et al. and Bourgoin et al. have recently reported that ARF functions in concert with a 50-kDa cytosolic factor in the activation of PLDs in human neutrophils and HL-60 cells, respectively (20,21). In their studies, the molecular weight of the cytosolic factor was estimated by gel filtration chromatography. In the present study, the PLD-stimulating activity and C3 exoenzyme substrate in the bovine brain cytosol co-chromatographed on gel filtration at a position corresponding to ϳ45 kDa (data not shown). It has been established that rhoA p21 associates with a Rho GTP/GDP exchange inhibitor, Rho GDI, with a molecular mass of 27 kDa (22). Thus the complex of rhoA p21 with Rho GDI should elute from the chromatograph with an apparent molecular mass of ϳ50 kDa. Bourgoin et al. (21) found that rhoA p21 eluted with an apparent molecular mass of ϳ45 kDa, but concluded that it is not the 50-kDa PLD-activating factor of HL-60 cells as they failed to detect rhoA p21 in several PLDactivating fractions. Although the ϳ50-kDa PLD-activating factor(s) in HL-60 cells and human neutrophils remains to be identified, PLD in these cells may be regulated differently from that in rat brain. Thus, the mechanism of PLD activation may vary between one isoform of PLD and another.
Another finding in this study was that, like ARF, rhoA p21 required post-translational modification to effect its regulation of PLD: non-isoprenylated rhoA p21 failed to activate rat brain PLD, even in its GTP␥S-bound state (Figs. 4 and 5B). It has been reported that GTP-bound isoprenylated rhoA p21, but not non-isoprenylated rhoA p21, translocates to plasma membranes (23). Consistent with the latter report, rhoA p21 (isoprenylated) in the partially purified preparation of PLD-activating factor bound to phospholipids vesicles containing the PLD substrate, PC, only in the presence of GTP␥S, whereas non-isoprenylated rhoA p21 did not bind under any circumstances (data not shown). This result leads us to speculate that the GTP␥S-bound isoprenylated rhoA p21 also translocates to the phospholipid vesicles, and thereby attracts PLD to the vesicles, making the enzyme more accessible to its substrate and increasing hydrolysis. Alternatively, PLD might interact specifically with PIP 2 in the vesicles and the GTP␥S-bound isoprenylated rhoA p21 translocate to the vesicles to interact with PLD. The latter notion is based on the finding that translocation of isoprenylated rhoA p21 to the vesicles does not require the presence of PIP 2 , whereas PIP 2 is essential for PLD activity (8). 2 We are currently investigating whether PLD interacts specifically with PIP 2 .