ADP-ribosylation Factor and Rho Proteins Mediate fMLP-dependent Activation of Phospholipase D in Human Neutrophils*

Activation of intact human neutrophils by fMLP stimulates phospholipase D (PLD) by an unknown signaling pathway. The small GTPase, ADP-ribosylation factor (ARF), and Rho proteins regulate the activity of PLD1 directly. Cell permeabilization with streptolysin O leads to loss of cytosolic proteins including ARF but not Rho proteins from the human neutrophils. PLD activation by fMLP is refractory in these cytosol-depleted cells. Readdition of myr-ARF1 but not non-myr-ARF1 restores fMLP-stimulated PLD activity. C3 toxin, which inactivates Rho proteins, reduces the ARF-reconstituted PLD activity, illustrating that although Rho alone does not stimulate PLD activity, it synergizes with ARF. To identify the signaling pathway to ARF and Rho activation by fMLP, we used pertussis toxin and wortmannin to examine the requirement for heterotrimeric G proteins of the Gi family and for phosphoinositide 3-kinase, respectively. PLD activity in both intact cells and the ARF-restored response in cytosol-depleted cells is inhibited by pertussis toxin, indicating a requirement for Gi2/Gi3 protein. In contrast, wortmannin inhibited only fMLP-stimulated PLD activity in intact neutrophils, but it has no effect on myr-ARF1-reconstituted activity. fMLP-stimulated translocation of ARF and Rho proteins to membranes is not inhibited by wortmannin. It is concluded that activation of Gi proteins is obligatory for ARF/Rho activation by fMLP, but activation of phosphoinositide 3-kinase is not required.

Phospholipase D (PLD) 1 is an important signal-transducing enzyme in a wide variety of cells and catalyzes the hydrolysis of phosphatidylcholine (PC) to produce the potential second messenger phosphatidic acid (PA) (1)(2)(3)(4). Studies in neutrophils and HL60 cells have identified a requirement for both cytosolic and membrane components for PLD activation when stimulated with the nonhydrolyzable analog of GTP, GTP␥S. Through the use of reconstitution studies utilizing HL60 membranes (5) or cytosol-depleted cells (6), ARF1 and ARF3 have been identified as activators of GTP␥S-stimulated PLD activity. The requirement for Rho in PLD activation was identified separately because RhoGDI, which extracts Rho from membranes, inhibited neutrophil PLD (7). Subsequent studies utilizing purified protein demonstrated that Rho proteins stimulated PLD directly (8 -10).
Although both ARF and Rho alone can stimulate PLD, other cytosolic factors have been shown to augment these responses in human neutrophils. A 50-kDa factor has been shown to act synergistically with both ARF (11,12) and Rho in stimulating PLD (13). Similarly, both G proteins also act synergistically with protein kinase C␣ to activate PLD. Both synergistic (14,15) and additive (16,17) responses have been reported for the simultaneous addition of ARF and Rho. A single PLD that is synergistically responsive to ARF/Rho/protein kinase C␣ has been cloned and is now referred to as hPLD 1 (10).
The roles of ARF and Rho are well documented with regard to PLD activation by GTP␥S, but their roles in G proteincoupled receptors and receptor tyrosine kinases are not well defined and may be different for distinct receptors. In the case of platelet-derived growth factor and epidermal growth factor, PLD activity is reported to be downstream of phospholipase C activation (18 -20). However, in other cases, including fMLPstimulated neutrophils (21), protein kinase C plays, if any, a minor role. Using Clostridium toxins, Rho proteins have been implicated in the activation of PLD in human embryonic kidney cells overexpressing the m3 muscarinic acetylcholine receptor (22), and in Rat 1 fibroblasts stimulated with lyso-PA and endothelin (23). Insulin-mediated PLD activation is dependent on ARF and Rho proteins, and an involvement of PI 3-kinase has been implicated (24,25).
In this study we have examined the recruitment of ARF and Rho proteins to membranes by fMLP and the ability of these proteins to regulate receptor-controlled PLD in human neutrophils. We had reported previously that fMLP-dependent activation was compromised in differentiated HL60 cells depleted of their cytosolic contents (26). We report that both ARF and Rho proteins are regulated by the fMLP receptor via a pertussis toxin-sensitive heterotrimeric G protein. Although activation of PLD in intact neutrophils is inhibited by wortmannin, a relatively selective inhibitor of PI 3-kinase, wortmannin treatment is not inhibitory to the fMLP-stimulated recruitment of ARF and Rho to membranes. We conclude that the activation of PLD by the fMLP receptor is dependent on receptor-activated ARF and Rho proteins in human neutrophils coupled via G i proteins and is not obligatorily dependent on PI 3-kinase activation.
Preparation of Neutrophils-Neutrophils were prepared according to established procedures (28). 50 ml of anti-coagulated blood or a buffy coat pack was mixed with an equal volume of 2% dextran solution in phosphate-buffered saline, pH 7.2, to aggregate erythrocytes. After 20 min at room temperature, the leukocyte-rich upper layer was removed and layered onto 10 ml of Lymphoprep and centrifuged at 2,000 rpm for 20 min to separate neutrophils from other white cells. Contaminating erythrocytes were removed by hypotonic lysis.
Analysis of ARF and Rho Leakage from Permeabilized Neutrophils-Neutrophils were permeabilized for varying lengths of time with 0.4 IU/ml streptolysin O. At the required time points, 1-ml aliquots were removed and centrifuged. The proteins from the supernatants (after precipitation with trichloroacetic acid) and cell pellets were resuspended in sample buffer and analyzed for ARF and RhoA proteins by Western blot analysis using appropriate antibodies.
Assay for PLD Activity in Intact Neutrophils-Neutrophils were suspended in HEPES buffer and pretreated with 5 M cytochalasin B for 5 min. 50-l aliquots were transferred to tubes containing 2% EtOH (1% final in the assay) in the presence or absence of fMLP (1 M final). After a 10-min incubation at 37°C, assays were quenched with 700 l of CHCl 3 :MeOH (1:1). After phase separation with 350 l of water, the chloroform phase was recovered. The chloroform phase was dried under vacuum and redissolved in 50 l of chloroform. Samples were spotted onto Whatman LK6TLC silica plates. The plates were developed in chloroform:methanol:acetic acid:water (75:45:3:1), dried at room temperature, and the lipid spots localized with iodine vapors. The spots corresponding to PEt and PC were excised after iodine sublimation and put into scintillation vials. The lipids were extracted with 250 l of methanol and counted for radioactivity after the addition of scintillation fluid. Results are expressed as the percentage of dpm incorporated into PC. Lipid standard containing PA and PEt which was prepared as described previously (29) was added for localization of lipids after separation by TLC.
Reconstitution of ARF-regulated PLD in Cytosol-depleted Neutrophils-Permeabilization of cells to deplete the cytosol and reconstitution of PLD activity were essentially carried out as described previously (6,30). Briefly, labeled neutrophils were permeabilized with streptolysin O (0.4 IU/ml) for 10 min in the presence of 100 nM Ca 2ϩ in 5 ml at 37°C. The permeabilized cells were centrifuged at 4°C and resuspended in cold PIPES buffer and aliquoted (50 l) into tubes kept on ice containing appropriate reagents. The assays were carried out in the presence of 1 mM MgATP, 100 M GTP (unless indicated otherwise), 2 mM MgCl 2 , 1 M Ca 2ϩ , and 1% EtOH. ARF proteins and fMLP were added as indicated in the individual figure legends. After a 30-min incubation at 37°C, the reactions were quenched as described above.
Pertussis Toxin Treatment-Intact labeled neutrophils were pretreated with 500 ng/ml pertussis toxin for 2 h at 37°C as described previously (31). Cells were then processed as required.
C3 Transferase Treatment-Cytosol-depleted neutrophils were incubated with 1 g/ml C3 toxin in the presence of 0.5 mM NAD for 10 min at 37°C.
Wortmannin Treatment-Intact cells were incubated at 37°C for 10 min in the presence of 100 nM wortmannin.
Translocation of ARF and Rho to Membrane Compartments-1 ml of intact cells (10 7 cells/ml) was incubated in the presence or absence of 1 M fMLP for 1 or 10 min at 37°C. The cells were pretreated with 5 M cytochalasin B as indicated. At the end of the incubation, the cells were sedimented at 4°C; resuspended in buffer; and treated with diisopropyl fluorophosphate, a serine protease inhibitor, for 5 min at 4°C. After centrifugation, the cells were resuspended in 1 ml, and a mixture of protease inhibitors was added (27). After sonication, the samples were centrifuged at 100,000 ϫ g for 1 h at 4°C to obtain the membrane fraction. The membrane fractions were resuspended in sample buffer, boiled, and run on SDS-PAGE. After transfer onto polyvinylidene difluoride, blots were probed with either anti-ARF or anti-Rho antibodies. Detection was by enhanced chemiluminescence.
For translocation of ARF to membranes in permeabilized cells, 1 ml of cells (10 7 cells/ml) was incubated with streptolysin O in the presence of 1 mM MgATP, 2 mM MgCl 2 , and 1 M Ca 2ϩ . 100 M GTP, 1 M fMLP, or 10 M GTP␥S was present as indicated. After incubation for 10 min at 37°C, the samples were processed as described above for intact cells.

RESULTS
ARF Restores fMLP-dependent PLD Activity in Cytosol-depleted Human Neutrophils-To examine a requirement for ARF proteins in fMLP-stimulated PLD activity, human neutrophils were permeabilized with streptolysin O for 10 min to deplete the cells of their freely diffusable cytosolic proteins. Fig.  1A illustrates that this protocol depletes the majority of the ARF proteins from the permeabilized cells, and they are recovered in the external medium. This loss is coincident with the inability of fMLP to stimulate PLD activity in cytosol-depleted cells (Ref. 26 and Fig. 1B). fMLP regains the ability to stimulate PLD activity provided that myr-ARF1 is also added to the permeabilized cells (Fig. 1B). It was noted that adding myr-ARF1 alone raised the basal activity of PLD, and this was dependent on the presence of GTP. In the reconstituted assay, the time course of PLD activation by fMLP reached a maximum at 30 min. Fig. 1, C and D, illustrates that reconstitution of PLD activity with myr-ARF1 and fMLP is concentration-dependent. For the remainder of the experiments, fMLP was used at 1 M, and myr-ARF1 was used at 50 g/ml.
Myristoylation of ARF Is Essential for fMLP-but Not GTP␥Sdependent PLD Activity-ARF proteins are myristoylated at their NH 2 terminus, and this lipid modification is thought to be important for efficient guanine nucleotide exchange catalyzed by the ARF exchange factors (32)(33)(34). Both myristoylated (myr-ARF1) and nonmyristoylated (ARF1) ARF proteins were examined for their ability to restore fMLP-dependent PLD activity in cytosol-depleted neutrophils. Fig. 2 illustrates that myristoylation is essential for the restoration of fMLP-dependent PLD activity and also for the response observed with GTP alone. Consistent with our own observation (6) and those of others (5,35), myristoylation is not required for GTP␥S-dependent stimulation of PLD (Fig. 2). However, a 100-fold higher concentration of nonmyristoylated rARF1 is required for maximal stimulation compared with fully myristoylated ARF1 for GTP␥Sstimulated PLD activity (6). Therefore, concentrations of recombinant ARF proteins used to examine the requirement for myristoylation take this into account; for nonmyristoylated ARF, a concentration of 750 g/ml is used compared with 50 g/ml for myr-ARF1. (Effective myr-ARF1 used is approximately 5 g/ml (500 nM) because of the 10% efficiency of myristoylation in E. coli determined by mass spectroscopy analysis.) These concentrations of recombinant myr-ARF1 and nonmyr-ARF1 reflect the equivalent loading of GTP␥S when measured in vitro. Fig. 2 also illustrates that the level of PLD stimulated by these proteins is similar in magnitude when GTP␥S is the activator.
Pertussis Toxin Inhibits Myr-ARF1-restored fMLP-dependent PLD Activity-The fMLP receptor is coupled to the pertussis toxin-sensitive heterotrimeric G proteins, G i2 and G i3 (36), and ␤␥ subunits are the direct regulators of phospholipase C␤ 2 and PI 3-kinase (␥ isoform) (37,38). To address the question of whether activation of the PLD activity by the fMLP receptor in the reconstituted assay also requires a prior activation of G i proteins, we examined the influence of pertussis toxin pretreatment. Initially we confirmed that, as reported previously (39,40), pertussis toxin pretreatment led to inhibition of the fMLPstimulated PLD activity in intact cells (Fig. 3A). Pertussis toxin pretreatment also inhibits fMLP-stimulated PLD activity when the agonist and the permeabilizing agent streptolysin O are added simultaneously, conditions in which the cytosolic proteins are still present (Fig. 3B). To establish a requirement for G i proteins in the regulation of PLD activity by myr-ARF1, we examined the effect of pertussis toxin pretreatment on the myr-ARF1-restored fMLP-dependent PLD activity in cytosoldepleted neutrophils (Fig. 3C). Although the stimulation of PLD activity observed in the combined presence of GTP and myr-ARF1 was not inhibited significantly by pertussis toxin treatment, the fMLP-stimulated activity was inhibited (Fig.  3C). These results confirm that regulation of myr-ARF1 by the fMLP receptor is indirect and that one intervening component in the pathway leading to myr-ARF1 activation has to include the heterotrimeric G i proteins.
Rho Proteins Participate in PLD Activation by the fMLP Receptor-Previous studies in a variety of cells have indicated that not only ARF but also Rho can activate PLD activity (1)(2)(3)(4). It has been reported that in HL60 cells, endogenous Rho is unlikely to play a physiological role in PLD activation (16). However, these experiments were performed using GTP␥S as an activator. To examine the contribution of Rho proteins in the system analyzed here, we first established whether Rho proteins leaked out of permeabilized cells. Neutrophils were permeabilized with streptolysin O for various lengths of time, and the supernatants and the cell pellets were analyzed for Rho proteins. The majority of RhoA remains cell-associated in the permeabilized cells even after 30 min. The amount of Rho released in the supernatant was low compared with that retained in the cells (Fig. 4). Under these conditions, the majority of the ARF proteins was found to leak out of the permeabilized cells (Fig. 1A). Because Rho proteins are found mainly to be cytosolic when cells are disrupted by homogenization (27), this would suggest that RhoA proteins do not behave as freely diffusable proteins under conditions in which the cellular architecture is maintained, as is the case in permeabilized cells. In addition to RhoA, Rac proteins were also retained in the permeabilized cells (data not shown).
Despite the retention of Rho proteins, the stimulation of PLD activity by fMLP is impaired in the cytosol-depleted cells (Fig.  1B), which would suggest that Rho proteins do not play a major role in fMLP-stimulated PLD activation. However, RhoA has been shown to be a poor activator of PLD activity by itself; but when it is present together with ARF, a synergistic activation is observed (10). Thus, it was still possible that RhoA could be a contributory factor in the myr-ARF1-reconstituted response stimulated by fMLP in the permeabilized cells.
C3 transferase ADP-ribosylates Rho proteins, thereby inactivating them. To investigate a role for Rho proteins, permeabilized cells were treated with C3 transferase. The myr-ARF1restored fMLP-dependent PLD activity was partially reduced when Rho proteins were inactivated (Fig. 5). This experiment uncovers the Rho component to the PLD response observed in the presence of myr-ARF. Thus ARF and RhoA proteins act synergistically to regulate fMLP-stimulated PLD activity. ARF The concentration of myr-ARF1 was varied from 0 to 100 g/ml as indicated. The radioactivity incorporated into PEt was measured as described under "Experimental Procedures" and is expressed as a percentage of dpm present in PC. The results are from a single experiment representative of three others. Panel D, concentration dependence of fMLP to restore of myr-ARF1-dependent PLD activity in cytosol-depleted cells. Conditions are as described in the legend of panel C except that the concentration of fMLP was varied between 10 Ϫ11 and 3 ϫ 10 Ϫ6 M. The concentration of myr-ARF1 was maintained at 50 g/ml. The radioactivity incorporated into PEt was measured as described under "Experimental Procedures" and is expressed as a percentage of dpm present in PC. The results are from a single experiment representative of three others.
can stimulate PLD activity in the absence of Rho (C3-treated cells), whereas Rho requires the presence of ARF.
ARF and Rho Are Translocated to the Membrane Fraction on fMLP Stimulation-ARF and RhoA are found predominantly in the postnuclear supernatant where they are present in a GDP-bound state. Nucleotide exchange and therefore activation results in the stable interaction of ARF and Rho with membranes (41,42). The data presented indicate that ARF and Rho are required for fMLP stimulation of PLD in a reconstituted system. To verify that ARF and Rho proteins are activated in intact cells, the translocation of ARF and Rho proteins to the membranes was examined upon stimulation with fMLP. Intact neutrophils were pretreated with cytochalasin B and incubated in the presence or absence of fMLP for 1 and 10 min. At the end of the incubation, the cells were recovered and the membrane fractions prepared. The samples were run on SDS-PAGE, blotted to polyvinylidene difluoride, and probed with the appropriate antibodies. Both ARF and Rho translocated to the membrane fraction within a minute of stimulation. (Maximal activation of PLD activity by fMLP in intact cells occurs at 1 min (28,43).) ARF proteins remained membrane-associated even after 10 min, but the association of Rho was diminished at 10 min (see Fig. 6A).
PLD activation by fMLP in intact neutrophils is enhanced greatly by pretreatment with cytochalasin B (43). Several other neutrophil responses, including degranulation, respiratory burst, phospholipase A 2 activation, and protein kinase C translocation, are also potentiated greatly by cytochalasin B (44 -47). The priming of human neutrophils by cytochalasin B can also be mimicked by other physiological agonists, e.g. low concentrations of C5a, fMLP, or tumor necrosis factor, and is therefore of physiological relevance. To examine whether translocation of ARF and Rho was dependent on priming, the cells were stimulated with fMLP for 1 min with or without cytochalasin B pretreatment. The translocation of ARF was entirely dependent on cytochalasin B pretreatment, whereas some RhoA translocation could be observed in its absence. Cytochalasin B enhanced RhoA translocation (Fig. 6B). These data indicate that similar to the activation of PLD by fMLP, the translocation of the two regulators, ARF and RhoA, is more efficient in primed cells. Fig. 6C illustrates that in permeabilized cells, translocation of ARF can be observed without cytochalasin B treatment. Both fMLP and GTP␥S are effective in translocation of ARF to membranes in permeabilized cells, with GTP␥S being better than fMLP.
Potential Role of PI 3-Kinase in PLD Activation-Although we have established that ARF and Rho proteins are involved in the stimulation of PLD via the fMLP receptor, it remains to be ascertained how the monomeric GTPases are activated by the fMLP receptor. Activation of G i proteins is clearly an intermediary step as observed by the inhibition with pertussis toxin treatment. A potential candidate that could couple G i proteins to ARF and Rho activation is PIP 3 , the product of PI 3-kinase activation. Wortmannin, a direct inhibitor of PI 3-kinase (48,49), has been reported to block the activation of fMLP-stimulated PLD activity in intact human neutrophils (48,50). PIP 3 could potentially recruit the ARF and Rho exchange factors to the plasma membrane via their PH (pleckstrin homology) domains (34,(52)(53)(54)(55). The ARF exchange factors, ARNO, cytohesin, and GRP1, all contain a PH domain that can bind phosphoinositides and a SEC7 domain that is responsible for the exchange of GDP for GTP (34,(52)(53)(54). The guanine nucleotide exchange factors for Rho proteins all contain a Dbl homology domain responsible for GDP-GTP exchange and also a PH domain (55). To test for this possibility, we examined whether wortmannin could inhibit the translocation of ARF and Rho proteins in fMLP-stimulated intact cells. Contrary to our expectations, wortmannin did not block the translocation of ARF proteins to membranes and marginally inhibited RhoA translocation (Fig. 6A). We observed consistently that wortmannin alone increased the amount of Rho proteins associated in the membranes slightly.
To check that wortmannin was effective in the inhibition of PLD activity in intact cells in our hands, we confirmed that wortmannin inhibited the fMLP-stimulated PLD activation in intact cells as reported previously (48). However, only a partial inhibition by wortmannin was observed in acutely permeabilized cells examined under conditions in which cytosolic proteins are still present during the time of stimulation (data not shown). We next examined the reconstitution of PLD activity by myr-ARF1 in the cytosol-depleted neutrophils, and this was not inhibited by wortmannin (data not shown). The lack of involvement of PI 3-kinase in the activation of fMLP-stimulated PLD in the cytosol-depleted cells is also in line with our recent observations (56). We have shown that fMLP is unable to stimulate PI 3-kinase activation in cytosol-depleted cells because of the loss of PI 3-kinase (␥ isoform) and PITP (phosphatidylinositol transfer protein) from the permeabilized cells. Collectively, these results exclude the possibility that PIP 3 is an absolute requirement in the regulation of ARF and Rho proteins in human neutrophils. However, it is still possible that in intact cells PI 3-kinase has a modulatory function because wortmannin is an inhibitor in intact cells. In intact cells the activation of PLD activity is rapid and occurs within 1 min, whereas the reconstituted activity in cytosol-depleted cells is slow, and PIP 3 could potentially have a bearing on the kinetics of activation.
PMA Translocated ARF and RhoA to Membranes-To demonstrate further the lack of involvement of PI 3-kinase in the recruitment of ARF and RhoA to membranes, we took advantage of the observation that PMA stimulation of PLD activity (1-4) is not inhibited by wortmannin (57). The ability of PMA to stimulate PLD activity could be caused by the activation of protein kinase C isozymes. Alternatively, PMA could cause translocation of ARF proteins to membranes as reported previously for rat basophilic leukemic mast cells (58). In human neutrophils, we observe that PMA recruits both ARF and Rho to membranes (Fig. 7A) under conditions in which no PI 3-kinase activation occurs.
We examined the consequences of this PMA-stimulated recruitment of ARF and RhoA to membranes. Intact cells were pretreated with PMA to translocate ARF and Rho to membranes. The cells were then permeabilized with streptolysin O for 10 min and washed. The resulting cytosol-depleted cells were incubated with or without GTP␥S. In the control cells, GTP␥S stimulated a small level of PLD activity. In the PMAtreated cells, the basal activity was substantially higher compared with control cells, which most likely reflects the recruitment of protein kinase C to the membranes. The addition of GTP␥S to these cells showed an enhanced response compared with control cells (Fig. 7B). DISCUSSION Activation of many G protein-coupled receptors and receptor tyrosine kinases in intact cells leads to stimulation of PLD activity, but the intervening components that lead to PLD activation remain to be delineated. Mammalian PLD has three known members, hPLD 1a , hPLD 1b , and PLD 2 ; hPLD 1a and hPLD1 b differ by lacking a 38-amino acid insert, and both are regulated by ARF and Rho proteins. PLD 2 is constitutively active. In HL60 cells, a cell line that is related to neutrophils, the predominant isoform that is expressed is hPLD 1 (59).
ARF and Rho Proteins Are Involved in the Coupling of the fMLP Receptor to PLD via a Heterotrimeric G Protein in Human Neutrophils-In this study, we present results demonstrating that ARF and Rho proteins are involved in PLD regulation in fMLP-stimulated human neutrophils. Extended permeabilization depleted the cells of ARF proteins but not RhoA, and this compromised the ability of fMLP to stimulate PLD activity (Fig. 3). Myr-ARF1 (but not non-myr-ARF1) restores fMLP-dependent PLD activity in these cytosol-depleted cells (Fig. 1A). This result was surprising initially because we had shown previously that nonmyristoylated ARF proteins were equally competent at activating PLD when GTP␥S was used as a stimulus. Higher concentrations of nonmyristoylated ARF are required to elicit a response similar to that of myristoylated protein in GTP␥S-dependent PLD activation (5,6). Myristoylation is clearly required for efficient nucleotide exchange at physiological concentrations of Mg 2ϩ when the receptor-driven system is used (32,34,60). When GTP␥S is used, the exchange factors must be bypassed, and the limited exchange that occurs on nonmyristoylated ARF could be the result of basal turnover. The recently described guanine nucleotide exchange factor for ARF (ARNO) has been shown to promote exchange rapidly on myristoylated proteins but not nonmyristoylated ARF (34,60).
Unlike ARF (Fig. 1A), RhoA does not leak significantly from the cells upon permeabilization (Fig. 5). Although the presence of RhoA was insufficient to elicit a response to fMLP in ARFdepleted cells, in the presence of ARF a role for RhoA could be identified clearly. Treatment of cells with C3 transferase partially inhibited the reconstitution of ARF-regulated PLD activity, revealing the Rho component of fMLP-stimulated PLD activation. Synergistic activation by ARF and Rho has been noted for the expressed hPLD 1 , and the data provided here reinforce the concept that multiple inputs are involved in the activation of PLD in intact cells triggered with a physiological stimulus. From in vitro studies, ARF is more active than RhoA in activating hPLD 1 , and the data presented in this paper show that although ARF can activate PLD activity in the absence of Rho, Rho activation becomes apparent only in the presence of ARF. Further evidence that fMLP receptors can activate ARF and Rho was obtained from the observation that fMLP stimulates their translocation to the membrane fraction (Fig. 6). The conclusions drawn from our studies are at variance with those of Wakelam and co-workers (16), who reported that Rho was not a physiological activator of PLD in HL60 cells (16). However, their studies did not include an investigation with fMLP as a stimulus but instead used GTP␥S, and this could possibly account for the different conclusions.
To elucidate the pathway leading to ARF and RhoA activation by the fMLP receptor, we have considered the following sequence that could lead to ARF and RhoA activation (Fig. 8). It is established that the fMLP receptor is coupled to the pertussis toxin-sensitive G proteins G i2 and G i3 and that the pretreatment of intact cells with pertussis toxin results in the inhibition of fMLP-stimulated PLD activity (39,40). We established that the myr-ARF-restored fMLP-dependent activity was also sensitive to pertussis toxin treatment, supporting the idea that the heterotrimeric G proteins of the G i family can lead to ARF and RhoA activation.
How do activated G␣ i or ␤␥ subunits activate ARF and Rho? We tested the possibility that PI 3-kinase activation could be important in this process because it is known that wortmannin, an inhibitor of PI 3-kinase, can inhibit PLD activity in intact cells. Protein G␤␥ subunits regulate the PI 3-kinase ␥ isoform directly in these cells. This model is attractive because several exchange factors for ARF proteins have been identified recently. Proteins with SEC7 domains function as guanine nucleotide exchange factors for ARF; in mammalian cells, ARNO, GRP1, and cytohesin have been cloned recently (34,(52)(53)(54). ARNO, GRP1, and cytohesin have PH domains and have been shown to bind phosphoinositides. Rho exchange factors are proteins with PH domains and with a Dbl domain that functions as the guanine nucleotide exchange factor. Taking into consideration the observation that wortmannin inhibits activation of PLD activity in intact human neutrophils, the activation of the PI 3-kinase pathway by protein G␤␥ subunits could potentially provide the link to ARF activation. However, an alternative possibility is that these PH domains are direct targets for the ␤␥ subunits because several proteins with PH domains do bind protein G␤␥ (61).
We tested the requirement of PI 3-kinase as an intermediary step to ARF and RhoA activation in two ways. In the reconstituted assay, we found that wortmannin had no effect on the myr-ARF-reconstituted PLD activity stimulated by fMLP. This lack of inhibition is in keeping with our previous study in which we showed that in the cytosol-depleted cells, fMLP stimulation of PIP 3 production is greatly attenuated because of leakage of the PI 3-kinase ␥ isoform (56). The second approach was to examine whether the translocation of ARF and RhoA to membranes was inhibited by wortmannin in intact cells. ARF translocation was unaffected, and that of Rho was inhibited slightly. We conclude that PI 3-kinase is not involved obligatorily in the activation of ARF and Rho, and the most likely possibility is that protein G␤␥ interacts directly with the PH domains of the appropriate exchange factors. We are currently identifying the specific nature of the exchange factors that are present in neutrophils. Brefeldin A does not influence fMLP-stimulated PLD activation, indicating that the exchange factor is brefeldin A-insensitive (data not shown).
One interesting facet of neutrophil physiology is the need to prime the cells before they acquire the potential to activate many of their downstream functional responses, such as degranulation and the oxidative burst. However, PI 3-kinase and phospholipase C␤ activation by fMLP is not dependent on priming, unlike PLD and phospholipase A 2 . We have observed that recruitment of ARF to the membranes is absolutely dependent on priming and that of Rho less so. In human neutrophils, we have shown recently that the ARF-regulated PLD is localized at the secretory vesicles and that upon priming, these vesicles fuse with the plasma membrane translocating the PLD in the process (51). The requirement for neutrophils to be primed would suggest that a component, possibly PLD, has to be present at the appropriate membrane for ARF to be recruited.
The model illustrating the molecular components that participate in the activation of PLD activity is summarized in Fig.  8. fMLP activates G i proteins, and G␤␥ targets the ARF and Rho exchange factors directly for recruitment of ARF and Rho to the plasma membrane where the PLD is present. This model should be contrasted with insulin-mediated activation of PLD activity. Insulin also causes the translocation of ARF and Rho proteins, but in this case PI 3-kinase does appear to be the intermediary for the activation of the appropriate exchange factors. Because these exchange factors belong to a large family of proteins, it is likely that different cells utilize a different subset of exchange factors, and hence different intermediate components may be required. Insulin, which activates a receptor tyrosine kinase, uses PI 3-kinase to activate ARF and Rho exchange factors; and fMLP, which activates G protein-coupled receptors, uses G␤␥ to do the same. The mechanism whereby PMA causes the translocation of ARF and Rho is not known. FIG. 8. Schematic representation of the potential pathway coupling the fMLP receptor to PLD activation in human neutrophils. fMLP binds to a G protein-coupled receptor and activates G i 2/G i 3. G␤␥ subunits activate a putative ARF and Rho exchange factor via their PH domains. ARF exchange factors are characterized by the presence of a SEC7 domain and Rho exchange factors by Dbl domains that catalyze the exchange of GDP for GTP. ARF and RhoA are activated and translocate to the membranes where the PLD is present. Translocation of ARF is observed only when neutrophils are primed with cytochalasin B. Priming results in the translocation of PLD from the secretory vesicles to the plasma membrane (51).