Differential Sensitivity of P-Rex1 to Isoforms of G Protein βγ Dimers*

P-Rex1 is a specific guanine nucleotide exchange factor (GEF) for Rac, which is present in high abundance in brain and hematopoietic cells. P-Rex1 is dually regulated by phosphatidylinositol (3,4,5)-trisphosphate and the Gβγ subunits of heterotrimeric G proteins. We examined which of the multiple G protein α and βγ subunits activate P-Rex1-mediated Rac guanine nucleotide exchange using pure, recombinant proteins reconstituted into synthetic lipid vesicles. AlF–4 activated Gs,Gi,Gq,G12, or G13 α subunits were unable to activate P-Rex1. Gβγ dimers containing Gβ1–4 complexed with γ2 stimulated P-Rex1 activity with EC50 values ranging from 10 to 20 nm. Gβ5γ2 was not able to stimulate P-Rex1 GEF activity. Dimers containing the β1 subunit complexed with a panel of different Gγ subunits varied in their ability to stimulate P-Rex1. The β1γ3, β1γ7, β1γ10, and β1γ13HA dimers all activated P-Rex1 with EC50 values ranging from 20 to 38 nm. Dimers composed of β1γ12 had lower EC50 values (∼112 nm). The farnesylated γ11 subunit is highly expressed in hematopoietic cells; surprisingly, dimers containing this subunit (β1γ11) were also less effective at activating P-Rex1. These findings suggest that the composition of the Gβγ dimer released by receptor activation may differentially activate P-Rex1.

dissociation inhibitors and/or GTPase activating proteins, the stimulation of Rac-GEFs is thought to be the most biologically important mode of Rac activation (1).
The P-Rex family of proteins is a recently identified group of Rac-GEFs that has been shown to be dually modulated by the G protein ␤ 1 ␥ 2 subunit and the lipid messenger, phosphatidylinositol 3,4,5-trisphosphate (PIP 3 ) (7-9). The P-Rex family consists of three members: P-Rex1, P-Rex2a, and P-Rex2b (7)(8)(9). P-Rex1 is the only member of the P-Rex family that is expressed in hematopoietic cells, whereas P-Rex2a and P-Rex2b are found mainly in skeletal muscle and heart, respectively (7)(8)(9). All three members of the P-Rex family share a similar domain topology and exhibit GEF activity. They are all multimodular proteins containing an N-terminal tandem Dbl homology and pleckstrin homology domains followed by two DEP and two PDZ domains (7)(8)(9). P-Rex1 and P-Rex2a also contain an inositol polyphosphate 4-phosphatase (InsP x 4-phosphatase) domain that is not present in P-Rex2b; however, neither of these proteins demonstrates inositol polyphosphate 4-phosphatase activity (7,8). Thus the function of this latter domain still remains elusive. P-Rex proteins apparently serve as a link between phospholipid and G protein-coupled receptor (GPCR) signaling and act as coincidence detectors for signals arising from these two pathways (10). The activation of GPCRs leads to the release of G␤␥ subunits, which have been shown to activate the p110-␤ and p110-␥ isoforms of PtdIns 3-kinase leading to the production of PIP 3 in hematopoietic cells such as neutrophils (11)(12)(13). Therefore, the stimulation of GPCRs in hematopoietic cells leads to the production of both activators of P-Rex1. Although the specific GPCR(s) responsible for P-Rex1 activation in hematopoietic cells has not been identified, the stimulation of G i -coupled receptors expressed in neutrophils, such as the fMet-Leu-Phe receptor, is likely to lead to the activation of P-Rex1 via the liberation of G␤␥ subunits. However, an unanswered question is: does the composition of the G␤␥ dimer released upon receptor activation differentially stimulate P-Rex1?
This issue is of central importance as it has become apparent that not all G␤␥ subunits are equivalent at interacting with receptors or effectors. For example, there are large differences in the ability of certain isoforms of the G␤␥ dimer to couple the ␤-adrenergic and the adenosine A2a receptors to the G s ␣ subunit (14), the ␣ 2 -adrenergic receptor to the G i ␣ subunit (15), and the adenosine A1 or 5-HT 1A receptors to the G i ␣ subunit (16). There are also clear differences in the ability of dimers to regulate effectors (14,(17)(18)(19)(20)(21)(22)(23). Surprisingly, only the ␤ 2 ␥ 2 dimer is able to inhibit the T-type calcium channel (18) and the ␤ 5 ␥ 2 dimer inhibits the muscarinic K ϩ channel (Kir3), whereas dimers containing the ␤ 1-4 subunits activate the K ϩ channel (19). In addition, we have demonstrated that the p110␥ isoform of PtdIns 3-kinase is differentially regulated by certain isozymes of G␤␥ (20). Given that both PtdIns 3-kinase and P-Rex1 are found in cells of hematopoietic origin (7,11) and have been shown to be regulated by G protein ␤ 1 ␥ 2 subunits, we hypothesized that P-Rex1 might have a ␤␥ sensitivity profile similar to that of PtdIns 3-kinase. Therefore, we undertook a study to test the 15-30 min. About 3.5 ml of swollen resin was washed with 10 -20 ml of borate saline prior to the addition of 3 ml of concentrated EE-antibody. The resin and antibody were incubated overnight at 4°C with constant agitation. After the incubation, the resin was collected by centrifugation at 8 -10 ϫ g and the supernatant discarded. The resin was then incubated for 2 h in 0.1 M Tris, pH 8.0 (2 ml per ml of resin), washed twice with 10 -20 ml of borate saline, and stored in borate saline containing 0.02% sodium azide at 4°C until further use.
Purification of P-Rex1-Sf9 cells were infected for 60 -80 h with the P-Rex1 virus and harvested via low speed centrifugation. The cells were lysed by nitrogen cavitation at 700 p.s.i. for 20 -30 min in a buffer containing 25 mM HEPES, pH 7.5, 150 mM NaCl, 2 mM EGTA, and protease inhibitors. The cell lysate (60 ml) was centrifuged at 1,000 ϫ g and 100,000 ϫ g prior to the cytosol (60 ml) being loaded onto a 1-ml EE antibody-agarose column by gravity flow. To purify the protein, the EE column was washed with 10 ml of 25 mM HEPES, pH 7.5, 150 mM NaCl, 2 mM EGTA, and 1% Triton, followed by a 10-ml wash with 25 mM HEPES, pH 7.5, 500 mM NaCl, 2 mM EGTA. P-Rex1 was eluted in 1-ml fractions with 25 mM HEPES, pH 7.5, 150 mM NaCl, 2 mM EGTA, and 0.4 mg/ml EE peptide (CEEEEYMPME). The EE-P-Rex1 prepared using this protocol is highly pure as demonstrated in Fig. 1A. Approximately 2-3 ml of 0.4 -0.8 g/ml P-Rex1 was obtained from 500 ml of Sf9 culture.
Purification of Rac-The recombinant GST-Rac1 (Rac) protein was purified using a GST resin per the manufacturer's protocols (Amersham Biosciences). One liter of DH5␣ bacterial cell culture expressing Rac was harvested and lysed using a French Press in 30 ml of buffer containing 50 mM Tris-HCl, pH 8.0, 50 mM NaCl, 1 mM MgCl 2 , 0.1 mM EGTA, 1 mM DTT, and protease inhibitors. To extract the Rac that was membrane associated, 1% Triton X-100 was added to the cell lysate and incubated on ice for 30 min. The cell lysate was then centrifuged at 10,000 ϫ g for 30 min at 4°C. The supernatant (30 ml) was collected and incubated with 1 ml of GST resin for at least an hour at 4°C with constant mixing. After the incubation, the beads were washed twice with 10 ml of phosphate-buffered saline containing 1% Triton X-100 and three times with 10 ml of phosphate-buffered saline alone, using low speed centrifugation (3,000 ϫ g for 2 min) and gentle resuspension. The beads containing the last wash were poured into a column and allowed to drain by gravity flow. Rac was eluted off the beads in 1-ml fractions with the following buffer: 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM MgCl 2 , 1 mM DTT, and 5 mM reduced glutathione. The high purity of the Rac obtained using this method can be seen in Fig. 1B. Approximately, 3 ml of 2-3 mg/ml Rac was obtained from 1 liter of DH5␣ bacterial culture.
Purification of G Protein ␣ Subunits-The procedure for purifying the ␣ subunits of G i , G q , and G s has been described (14,20,31). Viruses encoding either G␣ 12 or G␣ 13 , and ␤ 1 and ␥ 2 subunits engineered to have a hexahistidine and FLAG tag (34) at their N termini (␤ 1HF ␥ 2HF ), were used to express specific heterotrimeric G proteins in Sf9 cells. Purification of G␣ 12 and G␣ 13 was accomplished using minor variations of the methods of Kozasa (39). For G␣ 12 , the Ni 2ϩ column was equilibrated and washed with G12 buffer containing 20 mM Tris, pH 8.0, 100 mM NaCl, 1 mM MgCl 2 , 0.5% (v/v) Genapol, 10 mM ␤-mercaptoethanol, 10 M GDP, 15 mM imidazole, and protease inhibitors. For the elution step, G12 buffer was modified by lowering the NaCl and imidazole to 50 and 10 mM, respectively; MgCl 2 and GDP were raised to 50 mM and 20 M, respectively, and 10 mM NaF and 30 M AlCl 3 were added. The eluted fractions were pooled and further purified as described (39).
All G␣ subunits were eluted with AlCl 3 and NaF (AlF 4 Ϫ ) ensuring that the proteins were properly folded and functional, as determined by their ability to be activated by AlF 4 Ϫ . The G␣ subunits were stored in the Preparation of Liposomes-Pilot experiments assayed liposomes containing various lipids and found that liposomes containing 200 M phosphatidylcholine, 200 M phosphatidylserine, 200 M phosphatidylinositol, and 10 M PIP 3 provided the highest P-Rex1 activity, as published (7). Liposomes were prepared as follows: all four lipid constituents of the vesicles were mixed at the desired concentrations, dried under argon, and the lipid cake re-hydrated with buffer containing 20 mM HEPES, pH 7.5, 100 mM NaCl, and 1 mM EGTA. The re-hydrated lipids were then sonicated on ice using the microtip of a 550 Sonic Dismembrator (Fisher Scientific) at setting 3 for 4 -8 min (20 s on/5 s off) or until the solution became translucent. The vesicles were made as a ϫ10 stock and stored under argon after each use. Liposomes devoid of PIP 3 were prepared similarly. P-Rex1 Activity Assay (P-Rex1-mediated Rac Guanine Nucleotide Exchange)-We used bacterial GST-Rac1 (Rac) in our assays because Welch et al. (7) have shown that P-Rex1 was slightly more effective at stimulating nucleotide exchange on bacterial GST-Rac1 as opposed to Rac1 proteins produced in Sf9 cells. Pilot experiments optimized the time, temperature, and concentration dependence of guanine nucleotide exchange of the various reaction components. Based on these findings, our assays were performed in the presence of 100 nM Rac (7), 30 nM P-Rex1, and varying concentrations of ␤␥ dimers in a buffer containing 20 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM EGTA, 1 mg/ml bovine serum albumin, 5 mM DTT, 5 mM MgCl 2 , and 0.001% (w/v) CHAPS (final). The activity reactions were performed either at room temperature for 4 min or at 30°C for 1 min. Testing the detergent susceptibility of the assay revealed that P-Rex1 activity was largely unaffected between 0.001 and 0.005% (w/v) CHAPS or cholate and was diminished at higher concentrations (data not shown). Unless otherwise specified, all assays were performed in the presence of 0.001% CHAPS. The total reaction volume was 25 l and the ␤␥ dimers of interest were incubated with the phospholipid vesicles on ice for 30 min prior to the addition of the other reaction components (1 mg/ml bovine serum albumin, 5 mM DTT, 5 mM MgCl 2 and P-Rex1). Rac was added at 20-s intervals and incubated with the vesicles for 10 min prior to the addition of [ 35 S]GTP␥S to initiate the reaction. The reactions were terminated by dilution with 1.5 ml of wash buffer containing 20 mM HEPES, pH 7.5, 100 mM NaCl, and 10 mM MgCl 2 and filtration over HAWP nitrocellulose filters in a Millipore 1225 Sampling Manifold filtration apparatus. The filters were washed three times with a total of 7.5 ml of wash buffer and the amount of [ 35 S]GTP␥S-Rac bound to the filter was counted in a liquid scintillation counter.
Activation of G Protein ␣ Subunits-The G␣ subunits were activated prior to reconstitution into the phospholipid vesicles to monitor their effect on P-Rex1-mediated Rac exchange activity. Each G␣ subunit was diluted 10-fold into activation buffer containing 20 mM HEPES, pH 7.5, 100 mM NaCl, and 1 mM EGTA, 5 mM MgCl 2 , 10 mM NaF, and 30 M AlCl 3 and incubated at room temperature for 30 min. The activated G␣ subunits were not directly added to the lipid vesicles, as were the G␤␥ subunits, because the AlF 4 Ϫ had a detrimental effect on the liposomes.
Instead, the activated G␣ subunits were diluted 10-fold more (a final dilution of 100-fold) into the assay and added to the reaction at 20-s intervals ϳ2 min after the addition of Rac. The ␣ subunits (G s , G i1 , G q , G 12 , and G 13 ) were diluted a 100-fold to maintain optimal detergent levels between 0.001 and 0.004% CHAPS or cholate and were tested at concentrations Ն50 nM (Fig. 3). Gel Electrophoresis and Protein Concentration Determination-The identity and purity of protein samples were confirmed by gel electrophoresis on 8 or 12% SDS gels followed by silver staining and/or by Coomassie Blue staining. Protein concentrations were determined using densitometric analysis of Simply Blue-stained gels using phosphorylase b and ovalbumin to generate standard curves for P-Rex1 and the G proteins (Rac, ␣ and ␤␥), respectively.
Calculations, Statistics, and Expression of Results-The data presented under "Results" are representative of 3 or more experiments. The data representing the effect of the various G␤␥s on P-Rex1-mediated Rac guanine nucleotide exchange activity were fitted to sigmoidal curves using the routines provided in GraphPad Prism. The EC 50 values presented in Table 1 were obtained by fitting normalized values to sigmoidal dose-response curves using the routines provided in GraphPad Prism. The data were normalized to percent of maximal Rac GTP␥S binding. Three G␤␥s (␤ 5 ␥ 2 , ␤ 4 ␥ 11 and ␤ 1 ␥ 1 ) were very poor stimulators of P-Rex1 GEF activity. These G␤␥s were not normalized, rather they were plotted in comparison to an active ␤␥ run in the same experiment. Thus EC 50 values are not reported for these G␤␥s in Table 1. The statistical significance between the fits of different data sets was determined on normalized data using the F statistic (40). Linear regression routines in GraphPad Prism were used to plot the straight lines in Figs. 2, B and C, and 4 (␤ 5 ␥ 2 ).

RESULTS
Preparation of Purified P-Rex1, Rac, and G Proteins-We prepared highly purified recombinant proteins to test the specificity of interaction of G protein ␣ and ␤␥ subunits with P-Rex1. The purity of the Rac, P-Rex1, and the G␣ 12 and G␣ 13 subunits prepared for this study are shown in Fig. 1. Note that the EE-antibody column purified recombinant P-Rex1 essentially to homogeneity from the supernatant fraction of Sf9 cells in a single step (Fig. 1A). The bacterial Rac was also purified to homogeneity using its GST tag in a single step (Fig. 1B). Fig. 1C shows that the G 12 and G␣ 13 subunits prepared by elution from a ␤␥ column are also highly pure. The G␣ s , G␣ q , G␣ i subunits and the G␤␥ dimers used in this study are of equivalent purity, as described (14,20,31,33,35).
P-Rex1 Activity-To determine the ability of various G protein ␣ and ␤␥ isoforms to activate P-Rex1, we performed P-Rex1 activity assays in synthetic lipid vesicles containing phosphatidylserine, phosphatidylcholine, and phosphatidylinositol with/without phosphatidylinositol (3,4,5)-trisphosphate as described under "Experimental Procedures." In liposomes devoid of PIP 3 , neither P-Rex1 nor ␤ 1 ␥ 2 alone stimulated Rac guanine nucleotide exchange very well ( Fig. 2A, white bars). Welch et al. (7) have shown that P-Rex1 combined with G␤␥ can stimulate Rac guanine nucleotide exchange in the absence of PIP 3 . Using our preparations of P-Rex1 and GST-Rac, we did not observe much stimulation in the absence of PIP 3 . The reason for this discrepancy is not clear. One difference between our assay conditions and the previous studies (7,8,41) is that we used GST-Rac as opposed to EE-tagged Rac as the target of P-Rex1. Thus, one potential explanation for the differences observed may be that the larger GST tag is interfering with the PIP 3 independent activation of Rac by P-Rex and the G␤␥ dimer. This possibility can be explored in future studies of the interactions between Rac, P-Rex1, and G␤␥.
In the presence of liposomes containing 10 M PIP 3 , Rac guanine nucleotide exchange was stimulated by P-Rex1 alone and was further FIGURE 1. Purification of EE-tagged P-Rex1, GST-Rac, and G protein ␣ subunits. A, recombinant, EE-tagged P-Rex1 was isolated from Sf9 cells as described under "Experimental Procedures." The purified protein was separated on an 8% SDS-PAGE and visualized by silver staining. S2 represents the 100,000 ϫ g Sf9 cell supernatant from which P-Rex1 was purified, E1-4 represents sequential elutions from the EE column with 0.4 mg/ml EE-peptide. One l of each fraction (ϳ200 ng of total protein) was run on the gel. B, recombinant, GST-tagged Rac (48 kDa) was isolated from bacterial cells as described under "Experimental Procedures." The purified GST-Rac1 protein was separated on a 12% SDS-PAGE and visualized by silver staining. The band represents 1 l of eluate (ϳ500 ng of total protein) from the GST column. C, recombinant G␣ 12 and G␣ 13 (ϳ40 kDa) were purified as described under "Experimental Procedures." The purified proteins (400 -600 ng of total protein) were resolved on a 12% SDS-PAGE and visualized with Coomassie staining. " Rac exchange activity was tested with no additions (f), with 50 nM ␤ 1 ␥ 2 (OE); with 30 nM P-Rex1 (); and with 30 nM P-Rex1 and 50 nM G␤ 1 ␥ 2 (ࡗ). C, effect of increasing concentrations of ␤ 1 ␥ 2 on P-Rex1 activity. Rac guanine nucleotide exchange activity of 100 nM Rac was measured in the presence of increasing concentrations of ␤ 1 ␥ 2 in the presence or absence of 30 nM P-Rex1, as described under "Experimental Procedures." To calculate the percent increase in Rac GTP␥S binding upon addition of ␤ 1 ␥ 2 , Rac GTP␥S binding in the absence of any additions and presence of 30 nM P-Rex1 were subtracted from the ␤ 1 ␥ 2 (F) and the P-Rex1ϩ␤ 1 ␥ 2 (Ⅺ) curves, respectively. The P-Rex1 ϩ ␤ 1 ␥ 2 data represents the mean Ϯ S.D. of 10 experiments. enhanced by addition of ␤ 1 ␥ 2 ( Fig. 2A, black bars). Having established assay conditions, which allowed a marked effect of G␤␥ on P-Rex activity, all future experiments were performed in the presence of PIP 3 containing liposomes. Fig. 2B shows the time dependence of Rac guanine nucleotide exchange activity. Note that the Rac guanine nucleotide exchange activity of pure Rac alone was low in the absence of P-Rex or G␤␥ (closed squares). Addition of 50 nM ␤ 1 ␥ 2 to Rac had no effect on the guanine nucleotide exchange activity of Rac (closed triangles). This was the case for all 13 G␤␥ dimers tested in this study (data not shown). As anticipated, addition of 30 nM P-Rex1 alone increased the amount GTP␥S bound to Rac (inverted closed triangles). The addition of 50 nM ␤ 1 ␥ 2 further stimulated the activity of P-Rex1 (closed diamonds). As expected, in the presence of both P-Rex1 (30 nM) and ␤ 1 ␥ 2 (50 nM), Rac guanine nucleotide exchange was stimulated to a much greater extent than in the presence of either P-Rex1 or ␤ 1 ␥ 2 separately, illustrating the synergistic interaction between P-Rex1 and ␤ 1 ␥ 2 (7). Fig. 2C demonstrates that ␤ 1 ␥ 2 was able to modulate P-Rex1 activity in a concentration-dependent manner in the presence of 30 nM P-Rex1 (open squares). The dimer had little effect in the absence of P-Rex1 (closed circles). The EC 50 of the ␤ 1 ␥ 2 -stimulated P-Rex1 activity was 20 nM, which is consistent with the EC 50 values noted for the effect of ␤␥ on other effectors, such as type II adenyl cyclase, phospholipase C-␤, and PtdIns 3-kinase (14,20) and more potent than previously reported for P-Rex1 (7). PIP 3 -dependent P-Rex1 Activity in the Presence of Various G␣ Subunits-We tested the ability of representatives of the different G␣ families to modulate P-Rex1 activity using five distinct G␣ subunits (G s , G i , G q , G 12 , and G 13 ). Although we used concentrations of ␣ subunits that are supermaximal (Ͼ50 nM) in other assays, such as activation of adenyl cyclase or phospholipase C-␤ (14,20), none of the G␣ subunits tested were able to stimulate Rac guanine nucleotide exchange activity above that seen with P-Rex1 alone (Fig. 3, A-E). It should be noted that the G␣ activation buffer containing AlF 4 Ϫ was inhibitory to the P-Rex1 activity assay (Fig. 3F). However, under these conditions, P-Rex1 alone was still able to catalyze nucleotide exchange on Rac and this effect was further enhanced in the presence of 100 nM G␤ 1 ␥ 2 , albeit to a lesser degree compared with the reaction in the absence of AlF 4 Ϫ (Fig. 3F).
Finally, as described under "Experimental Procedures," all ␣ subunits showed activity in at least two other assays, the ability to dissociate from the ␤␥ dimer in the presence of aluminum fluoride and the ability to couple to receptors or activate adenyl cyclase (G␣ s ) or phospholipase C-␤ (G␣ q ). Thus, the current data suggest that P-Rex1 responds selec-

The G␤␥ Sensitivity of P-Rex1
tively to the ␤␥ dimer, a property shared with ion channels and PtdIns 3-kinase (17)(18)(19)(20). (7) and G␤␥ subunits play a central role in the activation of these cells (20), especially by receptors such as the fMet-Leu-Phe, which couples to the G␣ i subunit (42). In keeping with this mechanism, P-Rex1 is markedly activated by G␤ 1 ␥ 2. However, this isoform represents only one of 84 possible G␤␥ dimer combinations (43). As most G␤␥ subunits are ubiquitously expressed and numerous studies have demonstrated that certain G␤␥ isoforms selectively regulate effectors (14,17,19,44), we tested a panel of 12 different ␤␥ dimers to determine which G␤␥ isoforms could stimulate the activity of P-Rex1 in vitro.

PIP 3 -dependent P-Rex1 Activity in the Presence of Various G␤␥ Dimers-P-Rex1 is highly expressed in hematopoietic cells
The panel of G protein ␤␥ isoforms tested was based on our experience with PtdIns 3-kinase, another G␤␥-sensitive effector in hematopoietic cells (20). The effect of the 5 different G␤ subunits on P-Rex1 activity was determined using a series of pure G␤ x ␥ 2 dimers. Fig. 4 shows that ␤ 1 ␥ 2 (open squares), ␤ 2 ␥ 2 (closed triangles), and ␤ 3 ␥ 2 (closed inverted triangles) all stimulated P-Rex1 activity with similar potencies and efficacies. Interestingly, although ␤ 4 ␥ 2 (closed diamonds) was equally as potent as ␤ 1 ␥ 2 , it was not as efficacious ( Fig. 4 and Table 1). ␤ 5 ␥ 2HF (closed circles) was not an effective stimulator of P-Rex1 activity (Fig. 4). The data in Fig. 4 was normalized and re-plotted as the percent of the maximal effect of G␤␥ and fit to sigmodal curves using the routines in GraphPad Prism and the EC 50 values are presented in Table 1. The EC 50 values shown in Table 1 for the ␤ 1-4 ␥ 2 dimers range from 10 to 20 nM; values that agree well with the results obtained in a study of the sensitivity of PtdIns 3-kinase to this panel of G␤␥ dimers (20). ␤ 5 ␥ 2HF (closed circles) was not an effective stimulator of P-Rex1 activity. Overall, the data in Fig. 4 and Table 1 indicate that, with the exception of dimers containing the ␤ 5 , varying the composition of the ␤ subunit does not have a great effect on the ability of the G␤␥ dimer to activate P-Rex1. The inability of dimers containing the ␤ 5 subunit to activate P-Rex1 is in keeping with its lack of effect on type I and type II adenyl cyclase (14) and PtdIns 3-kinase (20).
We also determined the effect of seven different G␥ subunits using a series of G␤ 1 ␥ x dimers, because ␤ 1 forms dimers with most G␥s (14,20). Fig. 5A shows that four of the different G␤ 1 ␥ x dimers tested were equally as effective as ␤ 1 ␥ 2 in promoting P-Rex1-mediated Rac guanine nucleotide exchange. As with the ␤ series of dimers, each curve was normalized and fit to a sigmoidal curve to generate EC 50 values and the data presented in Table 1. The ␤ 1 ␥ 3 , ␤ 1 ␥ 7 , ␤ 1 ␥ 10 , and ␤ 1 ␥ 13HA dimers activate P-Rex1 with EC 50 values ranging from 20 to 38 nM, values similar to those seen with the ␤-series. Fig. 5B shows that ␤ 1 ␥ 12 dimer-stimulated P-Rex1 GEF activity to a lesser extent with an EC 50 of 112 nM. Dimers containing ␤ 1 ␥ 11 were also less effective and were 3-fold less potent than ␤ 1 ␥ 2 in stimulating P-Rex1 GEF activity (Fig. 5C). Interestingly, even though the ␥ 11 subunit is highly expressed in hematopoietic cells, the ␤ 1 ␥ 11 dimer is not as effective at stimulating P-Rex1 or PtdIns 3-kinase (20). The ␥ 1 subunit is similar to ␥ 11 in its primary amino acid sequence and also contains the farnesyl lipid at its C terminus (43). Accordingly, the ␤ 1 ␥ 1 dimer was tested as another representative of a farnesylated G␥. The data in Fig. 5C show that the ␤ 1 ␥ 1 dimer was also ineffective at stimulating P-Rex1 activity. A similar lack of effect was observed with the ␤ 4 ␥ 11 dimer (Table 1).

DISCUSSION
The P-Rex proteins (P-Rex1, P-Rex2a, and P-Rex2b) have been shown to function as GEFs for the Rho subfamily of monomeric G-proteins (7-9, 41). These small GTP-binding proteins have been implicated in cytoskeletal reorganization and gene expression (5,6). Of the six members of the Rho family of GTPases (5), P-Rex1 has been shown to stimulate nucleotide exchange on Rac1, Rac2, and Cdc42 and to a lesser extent on Rho, in vitro (7). However, in cells P-Rex1 has been shown to be a Rac-specific GEF (7). P-Rex1 is a PIP 3 -dependent Rac-GEF that can be synergistically activated by G␤␥ subunits and is highly expressed in cells of hematopoietic lineage such as neutrophils (7). It is a multimodular protein consisting of several functional domains: one tandem Dbl homology/pleckstrin homology domain, two DEP and two PDZ   Table 1. All the concentration-response curves are presented in comparison to the ␤ 1 ␥ 2 -P-Rex1 activity curve (Ⅺ) presented in Fig. 2C. The data presented are averages of at least three separate experiments for each isoform of ␤␥.
domains, and an InsP x 4-phosphatase domain. The specific regions of P-Rex1 involved in mediating its PIP 3 and G␤␥ dependence have not been determined. However, studies using deletion mutations suggest that the tandem Dbl homology/pleckstrin homology domain is necessary for P-Rex1 to be modulated by both PIP 3 and G␤␥ (41). PIP 3 is a lipid messenger produced in cells upon stimulation of PtdIns 3-kinases (45). Interestingly, similar to P-Rex1, the p110␥ isoform of PtdIns 3-kinases is highly abundant in hematopoietic cells and is stimulated by G␤␥ subunits (11, 20, 46 -48). The activation of GPCRs leads to the release of G␤␥ subunits that can stimulate PtdIns 3-kinase leading to the production of PIP 3 , which can activate P-Rex1 along with the G␤␥ dimers that were initially released upon agonist binding of the receptor. Thus both co-activators of P-Rex1 are produced in cells upon GPCR activation.
There are 22 different ␣ subunits that can be expressed from 17 genes in cells (49) and seven isoforms of G␤s and 12 isoforms of G␥s that can constitute a specific G␤␥ dimer (43). The primary goal of this study was to determine whether P-Rex1 was differentially regulated by the large number of G protein isoforms known to exist in cells expressing this exchanger. Using purified recombinant proteins reconstituted into synthetic lipid vesicles, the results demonstrate a number of important points. The purified, activated G s , G i , G q , G 12 , and G 13 ␣ subunits were unable to modulate P-Rex1 activity. It is likely that each ␣ subunit was a functional and properly folded protein because: (a) each G␣ subunit was purified from an immobilized ␤␥ column and eluted via activation with AlCl 3 and NaF, which only releases active protein; (b) each of the G␣ subunits tested was able to bind GTP␥S; and (c) the ␣ subunits were effective in receptor coupling assays or able to activate effectors (14,20,31). Thus, the current data suggest that P-Rex1 responds selectively to the ␤␥ dimer, a property shared with ion channels and PtdIns 3-kinase (17)(18)(19)(20). However, it would be useful to overexpress constitutively active G␣ subunits in cells to examine their potential for regulating P-Rex1 activity in a cellular context.
Interestingly, the profile of G␤␥ dimers able to activate P-Rex1 mirrors that of PtdIns 3-kinase (20). Like PtdIns 3-kinase, P-Rex1, was markedly activated by G␤␥ subunits containing the ␤ 1-4 subunits complexed with the ␥ 2 subunit with relatively similar EC 50 values. As expected (20), the ␤ 5 ␥ 2 dimer did not activate P-Rex1. G␤ 5 is the most divergent member of the G␤ family of proteins, shows only 50% similarity to ␤ 1 (43), and is unable to regulate adenyl cyclase I, adenyl cyclase II, and PtdIns 3-kinase (14,20,50,51). However, even though both the ␤ 5 subunit and P-Rex1 are highly expressed in the brain (7,52,53), the observations that ␤ 5 can bind to certain RGS proteins with high affinity (54,55) make it difficult to speculate about the potential roles of ␤ 5 ␥ 2 in these tissues. Lastly, as was the case with PtdIns 3-kinase, the isoform of the G␥ subunit in the ␤␥ dimer had a major effect on the ability of the dimer to activate P-Rex1. Specifically, dimers containing ␥ 11 were less effective at activating P-Rex1 (Fig. 5) and the ␤ 1 ␥ 12 dimer had a lower EC 50 than dimers containing ␥ 2 , ␥ 3 , ␥ 5 , ␥ 7 , ␥ 10 , and ␥ 13 . Interestingly, the ␤ 1 ␥ 11 dimer has also been shown to be less effective at activating adenyl cyclase II, phospholipase C-␤ (35), and PtdIns 3-kinase (20). Similarly, G␥ 1 , a G␥ subunit with a similar amino acid sequence and lipid modification as G␥ 11 (43,56,57), was also less effective at stimulating P-Rex1 (Fig. 5C). Both of these G␥ subunits, ␥ 11 and ␥ 1 , are farnesylated, suggesting that the isoprenoid moiety may play a role in determining the activity of the ␤␥-effector complex at this effector. These results add to the growing body of evidence arguing that the identity of the G␤␥ subunit released upon receptor activation plays a major role in selecting their downstream effectors (14,15,18,19).
With the discovery of the P-Rex family of proteins, a direct link has been established between the activation of heterotrimeric G proteins and the stimulation of PtdIns 3-kinase and Rac (7-9). As noted earlier, the activation of G protein-coupled receptors in hematopoietic cells leads to the release/production of both co-activators of the Rac-specific GEF (7), P-Rex1, one important step in neutrophil activation and chemotaxis. Interestingly, the stimulation of all receptors does not lead to Rac activation. In neutrophils and other cells of hematopoietic origin, the activation of G i -coupled receptors, such as the fMet-Leu-Phe or the C5a receptors leads to neutrophil chemotaxis (58). In contrast, the stimulation of G s -coupled receptors, such as the A2a receptor, inhibit neu- FIGURE 5. PIP 3 -dependent activation of P-Rex1 by the different G␥ subunits. The ability of a panel of G␤ 1 ␥ x dimers to stimulate P-Rex1 GEF activity. The guanine nucleotide exchange activity of 100 nM Rac-GST was assayed in the presence and absence of 30 nM P-Rex1 with increasing concentrations of recombinant purified ␤ 1 ␥ x , as described under "Experimental Procedures." All the concentration-response curves are presented in comparison to the ␤ 1 ␥ 2 -P-Rex1 activity curve (Ⅺ) as presented in Fig. 2C. A, effect of ␤ 1 ␥ 3 (OE), ␤ 1 ␥ 7 (), ␤ 1 ␥ 10 (ࡗ), and ␤ 1 ␥ 13HA (F) on P-Rex1 activity. B, comparison of ␤ 1 ␥ 2 and ␤ 1 ␥ 12 on P-Rex1 activity; C, comparison of ␤ 1 ␥ 2 , ␤ 1 ␥ 11 , and ␤ 1 ␥ 1 on P-Rex1 activity. Data were normalized, where applicable, as described under "Experimental Procedures" and the EC 50 values are presented in Table 1. The data presented are averages of at least three separate experiments for each isoform of ␤␥.
trophil chemotaxis (59,60). Thus, it is clear that activation of heterotrimeric G proteins can both stimulate and inhibit neutrophil function. Intriguingly, activation of both G i and G s ␣ subunits will release G␤␥ subunits that potentially activate two important steps in neutrophil chemotaxis: PtdIns 3-kinase and P-Rex1. Based on our findings, we hypothesize that G␤␥ dimer combinations that do not stimulate P-Rex1 may be coupled to receptors such as the adenosine A2a receptor, which inhibit neutrophil function. The finding that both P-Rex1 and PtdIns 3-kinase respond to a similar profile of G␤␥ dimers strengthens this notion. Naturally, release of G␤␥ dimers of differing composition is just one of several possibilities to explain why activation of receptors coupled to G s may not be able to stimulate neutrophil function. Another possibility is that the dimers are released in too low a concentration to be effective or that they are compartmentalized and not able to reach certain effectors. Negative regulation of P-Rex1 or other components of the G s signaling pathway via phosphorylation by the cAMP-dependent protein kinase is another intriguing possibility that is discussed in the accompanying article (61).