Specificity and Structural Requirements of Phospholipase C-β Stimulation by Rho GTPases Versus G Protein βγ Dimers*

Phospholipase C-β2(PLCβ2) is activated both by heterotrimeric G protein α- and βγ- subunits and by Rho GTPases. In this study, activated Rho GTPases are shown to stimulate PLCβ isozymes with the rank order of PLCβ2 > PLCβ3 ≥ PLCβ1. The sensitivity of PLCβ isozymes to Rho GTPases was clearly different from that observed for G protein βγ dimers, which decreased in the following order: PLCβ3 > PLCβ2 > PLCβ1 for β1γ1/2 and PLCβ2 > PLCβ1 >>> PLCβ3 for β5γ2. Rac1 and Rac2 were found to be more potent and efficacious activators of PLCβ2 than was Cdc42Hs. The stimulation of PLCβ2 by Rho GTPases and G protein βγ dimers was additive, suggesting that PLCβ2 activation can be augmented by independent regulation of the enzyme by the two stimuli. Using chimeric PLCβ1-PLCβ2 enzymes, βγ dimers, and Rho GTPases are shown to require different regions of PLCβ2 to mediate efficient stimulation of the enzyme. Although the catalytic subdomains X and Y of PLCβ2 were sufficient for efficient stimulation by βγ, the presence of the putative pleckstrin homology domain of PLCβ2 was absolutely required for the stimulation of the enzyme by Rho GTPases. Taken together, these results identify Rho GTPases as novel PLCβ regulators, which mediate PLCβ isozyme-specific stimulation and are potentially involved in coordinating the activation of PLCβ2 by extracellular mediators in intact cells.

Mammalian PLC␤ isoforms are differentially expressed in various tissues and cell types (5,6). At the protein level, PLC␤ 1 is highly expressed in the central nervous system but is also present in several other tissues, e.g. adrenal gland, parotid gland, lung, and kidney (7)(8)(9). The PLC␤ 2 polypeptide is present at high levels in neutrophils and cultured myeloid cells but has also been detected in other cells types and tissues, including platelets (10), T lymphocytes (11), osteoblasts (12), vascular and tracheal smooth muscle cells (13,14), cerebellum (12), spleen, and thymus (15). Myeloid cells have been found to contain both PLC␤ 2 and PLC␤ 3 (9,16). However, PLC␤ 2 appears to be predominantly important in these cells, because inactivation of the PLC␤ 2 gene caused an almost complete loss of formyl peptide receptor-stimulated inositol phosphate formation in mouse neutrophils (15). Furthermore, PLC␤ 2 , but not PLC␤ 3 , was activated by complement C5a and formyl peptide receptors in transfected cells (17). The latter findings are intriguing in light of the fact that PLC␤ 3 has been shown to be stimulated to a remarkable extent by G protein ␤␥ dimers in cell-free assays (9). In addition to myeloid cells, various other cell types and tissues contain the PLC␤ 3 isoform (8,9). In contrast, the PLC␤ 4 protein shows a more limited tissue distribution and is primarily found in the retina and in specific regions of the brain (18,19).
We have previously reported the identification of a PLC␤ 2stimulating GTP-binding protein present in cytosolic fractions of bovine neutrophils (20). This cytosolic protein was shown to be a member of the Rho subfamily of GTPases, Cdc42Hs and/or Rac, associated with the Rho GDP dissociation inhibitor LyGDI (21). Rho GTPases form a subgroup of the Ras superfamily of GTP-binding proteins that have been shown to regulate a wide spectrum of cellular functions, including gene expression, cell cycle progression, and reorganization of the actin cytoskeleton (22)(23)(24)(25). The activity of the Rho GTPases is determined by the ratio of their GTP/GDP-bound forms, which is regulated by at least three regulatory proteins: guanine nucleotide dissociation inhibitors, guanine nucleotide exchange factors, and GTPaseactivating proteins (25).
Using purified proteins, we have previously demonstrated that PLC␤ 2 is activated by GTP␥S-liganded Cdc42Hs and Rac1, but not by RhoA, through direct protein-protein interaction (21). This stimulation has been shown to be independent of LyGDI but to require both C-terminal processing of the Rho * This work was supported by the Deutsche Forschungsgemeinschaft (Grant SFB 497). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
GTPases and the integrity of their effector regulating domain (21). Like G protein ␤␥ dimers, activated Rho GTPases stimulated a deletion mutant of PLC␤ 2 , PLC␤ 2 ⌬, lacking a C-terminal region necessary for stimulation by G protein ␣ q subunits (21). The specificity of PLC␤ stimulation by Rho GTPases and the mechanisms of enzyme activation by Rho GTPases remained unknown. The goal of the study was to elucidate the sensitivity of PLC␤ isozymes to stimulation by Rho GTPases versus G protein ␤␥ dimers, known activators of these enzymes, and to identify and to compare the structural requirements of PLC␤ stimulation by these two regulatory proteins. The results show that the three PLC␤ isozymes tested, PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 , are differentially sensitive to stimulation by activated Rho GTPases and ␤␥ dimers. The specificity of Rho GTPasemediated PLC␤ stimulation also differs from that reported for ␣ q -mediated PLC␤ stimulation, with PLC␤ 2 being considerably less sensitive than PLC␤ 1 and PLC␤ 3 (1)(2)(3)(4). This study identifies PLC␤ 2 as the PLC␤ isozyme most sensitive to stimulation by Rho GTPases, especially Rac1 and Rac2. Using PLC␤ 1 -PLC␤ 2 chimeras constructed on the basis of structural domains predicted by the crystal structure of PLC␦ 1 , we demonstrate that the presence of the catalytic subdomains X and Y of PLC␤ 2 is sufficient for ␤␥-dimer stimulation. In contrast, the presence of the putative pleckstrin homology (PH) domain of PLC␤ 2 is absolutely required for stimulation of the enzyme by activated Rho GTPases. Taken together, the results demonstrate, for the first time, the unique regulation of the activity of PLC␤ 2 by monomeric GTPases and G protein ␤␥ dimers requiring different structural elements of this enzyme.

EXPERIMENTAL PROCEDURES
Recombinant PLC␤ Isoforms, C-terminally Deleted PLC␤s Mutants, and PLC␤ 1 -PLC␤ 2 Chimeras-Construction of recombinant baculoviruses for expression of the bovine PLC␤ 1 and human PLC␤ 2 have been described (26,20). The cDNA of human PLC␤ 3 (27) was cloned into the EcoRI site of the baculovirus transfer vector pVL1393 (Invitrogen, Carlsbad, CA).
The PLC␤ 1 -PLC␤ 2 chimeras were generated by the PCR overlap extension method. In chimera A, the N-terminal amino acids of PLC␤ 2 (residues 1-138) were replaced by the corresponding residues of PLC␤ 1 (residues 1-142), so that the putative PH domain was substituted. Substitution of the PH domain and the putative four EF-hand motives of PLC␤ 2 (residues 1-303) for the corresponding residues of PLC␤ 1 (residues 1-307) resulted in chimera B. Chimera A was constructed from the vectors pVL1392-PLC␤ 1 and pVL1393-PLC␤ 2 ⌬ by PCR amplification using the following two pairs of primers: 5Ј-GAATTCATGGC-CGGGGCACAGC-3Ј (upstream, sense), 5Ј-GGGAGGCGTTGGCCGTC-AGCAGGTTTGTTGCCAAA-3Ј (internal, antisense), 5Ј-TTTGGCAAC-AAACCTGCTGACGGCCAACGCCTCCC-3Ј (internal, sense), 5Ј-GGAT-CCTCAGAGGCGGCTCTCCT-3Ј (downstream, antisense). The two amplified fragments were re-annealed and re-amplified using the upstream and the downstream primers, which introduced an EcoRI and a BamHI recognition site, respectively. Chimera B was constructed using the same upstream and downstream primers and the following two internal primers: 5Ј-GAAGAAAATGGAGTCGTTGCCCAGGACAAGC-TGCT-3Ј (internal, antisense) and 5Ј-CAGCTTGTCCTGGGCAACGAC-TCCATTTTCTTCTC-3Ј (internal, sense). The final fragments were then ligated at their EcoRI and BamHI sites, and the resulting constructs were inserted into the baculovirus transfer vector pVL1392.
The entire PCR-amplified regions were sequenced and found to be identical to the expected sequences. Production of recombinant baculoviruses, expression, and isolation of PLC␤ isoforms were carried out according to published protocols (26).
Recombinant Rho GTPases-The production of recombinant baculoviruses using BaculoGold TM DNA (BD Pharmingen, San Diego, CA) and pVL1393 transfer vectors containing the respective Rho GTPase cDNA ligated into the BamHI/EcoRI site has been described previously (21). Isoprenylated membrane-bound Rho GTPases were solubilized by extracting the membranes with buffer containing 23 mM sodium cholate as described previously (21). Rac2-LyGDI heterodimers were purified from cytosolic fractions of baculovirus-infected insect cells (31).

S]GTP␥S Binding-Binding of [ 35 S]GTP␥S
to Rho GTPases was assayed as described previously (21) with minor modifications. Briefly, samples (0.05-0.5 g of protein extracted from membranes of baculovirus-infected insect cells with buffer containing sodium cholate) were incubated at 30°C in an incubation mixture (40 l) containing 25 mM Hepes-NaOH, pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 20 mM MgCl 2 , 100 mM NaCl, 0.1% (v/v) GENAPOL C-100 (Calbiochem, La Jolla, CA, HPLC grade), and 100 nM [ 35 S]GTP␥S (296 GBq/mmol). The incubation was terminated after 6 h as described earlier (31). We also used [ 35  Phospholipase C Assay-Phospholipase C activity was determined as described (21) with minor modifications. In brief, 5 l of detergentsolubilized Rho GTPase and/or 5 l of purified ␤␥ were supplemented with 10 l of soluble fraction of PLC␤-baculovirus-infected insect cells and incubated at 25°C for time periods as indicated in the figure legends in a volume of 60 l containing 50 mM HEPES-NaOH, pH 7.2, 70 mM KCl, 3 mM EGTA, 2 mM dithiothreitol, 33 M [ 3 H]PI-4,5-P 2 (185 GBq/mol), 536 M phosphatidylethanolamine, and 150 nm free Ca 2ϩ . In all experiments comparing the effects of Rac2 and ␤␥, the final concentration of sodium cholate was 2 mM. Only ␤␥-mediated stimulation of wild-type versus C-terminally deleted PLC␤ isozymes ( Fig. 6) was measured in the absence of sodium cholate. Results obtained in control experiments using purified PLC␤ 2 ⌬ (31) were indistinguishable from that obtained with soluble fractions of PLC␤ 2 ⌬-baculovirus-infected insect cells.
Miscellaneous-Purification of ␤ 1 ␥ 1 isolated from bovine retinal rod outer segment membranes has been described elsewhere (32). The method used to purify membrane-associated recombinant ␤ 1,5 ␥ 2-His dimers is described in a previous study (33). SDS-PAGE, immunoblotting, and determination of protein concentrations were performed as described previously (21). Polyclonal antibodies against Rho GTPases were from Santa Cruz Biotechnology (Santa Cruz, CA). Specific antisera against PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 were a kind gift from Dr. P. J. Parker (28). The data are presented as means Ϯ S.D. of triplicate determinations.

Stimulation of Recombinant PLC␤ Isozymes by ␤␥
Dimers-To determine the specificity of PLC␤ stimulation by G protein ␤␥ dimers, the PLC␤ isozymes PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 were produced as recombinant proteins in Sf9 insect cells. Fig. 1A shows that all three isozymes were expressed as soluble proteins in infected insect cells and migrated at molecular masses of ϳ150, 140, and 160 kDa, respectively, corre-sponding to the masses described for the native PLC␤ isozymes (34,35,28). When the amounts of the three PLC␤ isozymes were normalized according to their maximal activity at 1 mM free Ca 2ϩ and 3.3 mM sodium deoxycholate (not shown) and then assayed at a more physiological Ca 2ϩ concentration of 150 nM, recombinant PLC␤ 2 displayed an ϳ3-fold higher basal activity than recombinant PLC␤ 1 and PLC␤ 3 (Fig. 1B). In all experiments shown in this study, the amounts of the three recombinant PLC␤ preparations were adjusted to equal basal phospholipase C activity at 150 nM free Ca 2ϩ in the presence of 2 mM sodium cholate. Under these conditions, purified ␤␥ dimers of bovine retinal transducin (␤ 1 ␥ 1 , 300 nM) activated PLC␤ 1 and PLC␤ 2 only slightly (1.6-fold and 3.0-fold, respectively), but caused a marked (ϳ20-fold) stimulation of PLC␤ 3 (Fig. 2).
Stimulation of Recombinant PLC␤ Isozymes by Rho GTPases-To investigate the specificity of PLC␤ 2 stimulation by Rho family members, the recombinant Rho GTPases Rac1, Rac2, and Cdc42Hs were produced in baculovirus-infected insect cells, extracted from the membrane of infected cells with detergent-containing buffer, and reconstituted with a recombinant C-terminal deletion mutant of PLC␤ 2 , PLC␤ 2 ⌬, in the presence of 100 M GTP␥S. Fig. 3 shows that both Rac1 and Rac2 caused a marked (ϳ13-fold) stimulation of PLC␤ 2 ⌬. Rac2 was slightly more potent than Rac1. Thus, half-maximal stimulation was observed at ϳ40 nM Rac2 and 100 nM Rac1. Cdc42Hs was a less potent (EC 50 : 400 nM) and less efficacious (ϳ8-fold) stimulator of PLC␤ 2 ⌬ than Rac1 and Rac2. In additional experiments, we observed a similar rank order of potency of Rho GTPases (Rac2 Ն Rac1 Ͼ Cdc42Hs) to stimulate PLC␤ 2 when full-length enzyme rather than PLC␤ 2 ⌬ was used (not shown). RhoA had no effect on full-length PLC␤ 2 or PLC␤ 2 ⌬ when tested under the same conditions (not shown).
Next, the most potent Rho GTPase, Rac2, was incubated at increasing concentrations in the presence of either GDP or GTP␥S with soluble preparations of recombinant wild-type PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 . Fig. 4 shows that PLC␤ 2 was clearly the PLC␤ isoform most sensitive to stimulation by GTP␥S-activated Rac2. Thus, half-maximal and maximal (ϳ4fold) stimulation was observed at approximately 80 and 500 nM Rac2, respectively. Rac2 also appeared to stimulate PLC␤ 2 in the presence of GDP, albeit to a much lesser (ϳ1.6-fold) extent. At high concentrations of Rac2, a reduction of PLC␤ 2 stimulation was observed both in the presence of GDP and GTP␥S, suggesting an inhibitory effect of the membrane extracts on PLC activity. PLC␤ 1 and PLC␤ 3 were also activated by Rac2, but to a much lower extent and only at much higher concentrations of GTP␥S-activated Rac2 (Ն1 M). Additional measurements (not shown) of PLC␤ stimulation by C-terminally modified Cdc42Hs and Rac1 revealed an at least 10-fold higher were subjected to SDS-PAGE and immunoblotting using PLC␤ subtype-specific antisera. The apparent molecular weights of the marker proteins are indicated. The PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 migrated at ϳ150, 140, and 160 kDa, respectively. Additional immunoreactive proteins of lower molecular masses were detected in all three preparations. Because soluble fractions of noninfected insect cells did not contain proteins reactive with the antisera used in this experiment (not shown), these proteins most likely correspond to proteolytic fragments of the full-length PLC isozymes. B, aliquots of the three soluble fractions were incubated in the presence of 150 nM free Ca 2ϩ for the times indicated at the abscissa with phospholipid vesicles containing PI-4,5-P 2 . The reaction was terminated by the addition of chloroform/methanol/concentrated HCl, and the mixture was analyzed for inositol phosphates. See "Experimental Procedures" for details. Prior to this experiment, the amounts of the samples containing soluble PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 were adjusted to give equal maximal PLC activities in the presence of 1 mM free Ca 2ϩ and 3.3 mM sodium cholate (42) (not shown) (PLC␤ 1 , 0.4 g of protein/sample; PLC␤ 2 , 1.5 g of protein/sample; PLC␤ 3 , 5.6 g of protein/sample). potency of these GTPases toward PLC␤ 2 than toward PLC␤ 1 or PLC␤ 3 . RhoA, a Rho GTPase incapable of stimulating PLC␤ 2 (21), affected neither PLC␤ 1 nor PLC␤ 3 activity under the same conditions. These data show that Rho GTPases stimulate PLC␤ 2 with a rank order of potency of Rac2 Ն Rac1 Ͼ Cdc42Hs and that, among the PLC␤ isoforms tested, PLC␤ 2 is most sensitive to stimulation by activated Rac2.
Comparison of Full-length and C-terminally Deleted PLC␤ Isozymes-A truncated PLC␤ isozyme related to PLC␤ 3 has previously been reported to be remarkably sensitive to activation by ␤␥ dimers (36). Similarly, C-terminal truncation of PLC␤ 3 from human platelets by calpain has been shown to result in a marked augmentation of ␤␥ stimulation (37). PLC␤ 1 has been shown to be cleaved by calpain between residues 880 and 881, generating two fragments of 100 and 45 kDa, respectively (38). The presence of ϳ45to 50-kDa proteins in the preparations of recombinant PLC␤ isozymes, which are likely to represent C-terminal proteolytic fragments of the enzymes (cf. Fig. 1), raised the possibility that the observed order of PLC␤ stimulation by Rac2 was a consequence of C-terminal proteolysis of the PLC␤ isozymes. To challenge this hypothesis, we examined and compared the ability of Rac2 to stimulate the activity of C-terminal deletion mutants of PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 . A schematic representation of the wild-type and mutant PLC␤ isozymes is shown in Fig. 5A. As shown in Fig. 5B, all three mutants were expressed in baculovirus-infected insect cells as soluble proteins and migrated on SDS-polyacrylamide gels at the expected molecular weights. The deletion mutants displayed Ca 2ϩ sensitivities indistinguishable from those of their wild-type counterparts (not shown). Reconstitution of the wild-type and mutant PLC␤ isoforms with GTP␥S-activated Rac2 (500 nM) or ␤ 1 ␥ 1 (300 nM) showed that, surprisingly, the deletion of the C-terminal regions of each PLC␤ isozyme did not change the efficiency of ␤ 1 ␥ 1 to cause PLC␤ stimulation (Fig. 6). In agreement with earlier findings (20), the removal of the C-terminal part of PLC␤ 2 enhanced the extent of Rac2-mediated stimulation of the enzyme. Importantly, however, no enhancement of Rac2-mediated stimulation was observed for the C-terminally deleted variants of PLC␤ 1 and PLC␤ 3 . Therefore, removal of the C-terminal regions of PLC␤ isozymes does not change the rank order of specificity of PLC␤ stimulation by both Rac2 and G protein ␤␥ subunits.
Simultaneous Stimulation of PLC␤ 2 by Rac2 and ␤␥ Subunits-The next experiments were designed to examine whether the different specificities of PLC␤ isozyme stimulation by Rho GTPases and ␤␥ dimers may reflect independent regulation of PLC␤ 2 by both stimulators. We have previously shown that ␤ 5 ␥ 2-His is a more potent stimulator of PLC␤ 2 ⌬ than is ␤ 1 ␥ 2-His and that the latter, but not the former ␤␥ dimer, activates PLC␤ 3 (33). A comparison of the effects of ␤ 5 ␥ 2-His and ␤ 1 ␥ 2-His on the activity of wild-type PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 is shown in Fig. 7. Among the PLC␤ isoforms tested, PLC␤ 3 was the isoform most sensitive to stimulation by ␤ 1 ␥ 2-His , followed by PLC␤ 2 and PLC␤ 1 , which was hardly activated by this ␤␥ preparation. In marked contrast, ␤ 5 ␥ 2-His proved to be a potent and efficacious activator of full-length PLC␤ 2 , but it did not affect full-length PLC␤ 1 or PLC␤ 3 . Because activation of PLC␤ 2 by ␤ 5 ␥ 2 occurred at low concentrations (EC 50 : ϳ10 nM) and reached saturation within the range of ␤␥ dimer concentrations tested, ␤ 5 ␥ 2-His was chosen as the ␤␥ dimer to compare the effects of Rac2 and ␤␥ dimers on PLC␤ 2 activity when added alone or in combination at maximally effective concentrations (Fig. 8). Because stimulation of PLC␤ 2 by ␤␥ dimers is sensitive to high concentrations of detergent (32), detergentfree Rac2-LyGDI heterodimers were used as a source of Rac2. In our hands, Rac2-LyGDI heterodimers and monomeric Rac2 stimulated PLC␤ 2 with equal potency and efficacy in the presence of phospholipid vesicles containing the phospholipase C substrate PI-4,5-P 2 (31). Fig. 8 shows that addition of GTP␥S allowed activation of Rac2 from the heterodimer (EC 50 : ϳ20 nM) leading to an 3-fold stimulation of PLC␤ 2 ⌬ and that stimulation of the enzyme was observed both in the absence and in the presence of ␤ 5 ␥ 2-His (EC 50 : ϳ20 nM). The extent of stimulation by GTP␥S-activated Rac2-LyGDI was additive with that . Phospholipase C activity was measured at 150 nm free Ca 2ϩ in the presence of phospholipid vesicles containing PI-4,5-P 2 . The results are given as the percentage of the basal activities, which were set to 100%. For any given pair of wild-type and mutant PLC␤ isozyme, the basal activities varied by less than a factor of two (not shown). obtained by ␤ 5 ␥ 2-His , suggesting independent stimulation of PLC␤ 2 by Rho GTPases and ␤␥ subunits. Interestingly, partial inhibition of ␤ 5 ␥ 2-His -mediated PLC␤ 2 stimulation by Rac2-LyGDI was measured in the presence of GDP. Because there was no effect of Rac2-LyGDI on basal PLC␤ 2 activity, this result may indicate that inactive Rac2-LyGDI may interact with ␤ 5 ␥ 2-His and/or noncompetitively interfere with ␤ 5 ␥ 2-Hismediated PLC␤ 2 activation.
Structural Requirements of PLC␤ Stimulation by Rho GTPases versus ␤␥ Dimers-The independent stimulation of PLC␤ 2 by ␤␥ subunits and Rho GTPases prompted us to delineate the structural elements of PLC␤ 2 required for enzyme activation. The site of interaction of PLC␤ 2 with ␤␥ was localized by others to the region between the catalytic subdomain Y residues Glu 574 and Lys 583 (39,40). However, ␤␥ dimers have also been reported to bind to the isolated PH domains of PLC␤ 1 and PLC␤ 2 (41). To identify the sites on PLC␤ 2 relevant for activation by Rac2 and ␤␥, we took advantage of the fact that both stimulators elicited only a slight effect on PLC␤ 1 ⌬, but markedly activated PLC␤ 2 ⌬. The effects of the two stimulators were examined on the activity of PLC␤ 1 -PLC␤ 2 chimera in which N-terminal portions of PLC␤ 2 ⌬ had been replaced by the corresponding regions of PLC␤ 1 ⌬ (Fig. 9). Two chimera, designated A and B, carrying the putative PH domain and the PH domain together with the four EF-hand motifs of PLC␤ 1 ⌬, respectively, were analyzed. A third chimera, comprising the PH domain, the EF-hand motifs, the catalytic subdomain X of PLC␤ 1 ⌬, and the catalytic subdomain Y of PLC␤ 2 ⌬, was catalytically inactive and hence not used for further analysis. Although chimera A was expressed in baculovirus-infected insect cells at levels considerable lower than were PLC␤ 2 ⌬ and chimera B, the two chimeras and PLC␤ 2 ⌬ were indistinguishable in terms of the dependence of their catalytic activity on Ca 2ϩ (not shown). Fig. 10 shows that substitution of the N-terminal portions of PLC␤ 2 ⌬ for the corresponding regions of PLC␤ 1 ⌬ led to reduction of the degree of ␤ 1 ␥ 1 -mediated stimulation. Specifically, at the highest concentration of ␤ 1 ␥ 1 tested (4 M), the degrees of stimulation were 129-, 88-, 66-, and 8-fold for PLC␤ 2 ⌬, chimera A, chimera B, and PLC␤ 1 ⌬ (Fig. 10A). Because the effects of ␤ 1 ␥ 1 did not reach saturation within the range of concentrations tested, it is currently unclear whether the reduced stimulation is due to a decrease in the potency or efficacy of ␤ 1 ␥ 1 . Very interestingly, substitution of the putative PH domain of PLC␤ 2 ⌬ for its counterpart of PLC␤ 1 ⌬ caused an almost complete (Ͼ95%) loss of stimulation by GTP␥S-activated Rac2 (Fig. 10B). In additional experiments (not shown), we found that chimeras A and B were also resistant to stimulation by GTP␥S-activated Cdc42Hs. Taken together, these results not only show that the structural requirements of PLC␤ 2 stimulation by ␤␥ dimers and by the Rho GTPases Rac2 and Cdc42Hs are distinct but also suggest that the putative PH domain of PLC␤ 2 is critically involved in mediating its activation by Rho GTPases.

Specificity of PLC␤ Stimulation by Rho GTPases-
We have previously shown that the Rho GTPases Rac1 and Cdc42Hs, but not RhoA, stimulate the activity of PLC␤ 2 (21,31). In this study, we demonstrate that the PLC␤ isozymes PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 are differentially sensitive to stimulation by Rho GTPases and G protein ␤␥ dimers. Activated Rac2 is shown to stimulate PLC␤ isozymes with the rank order of potency and efficacy of PLC␤ 2 Ͼ PLC␤ 3 Ն PLC␤ 1 . This rank order is clearly different from the order observed for G protein ␤␥ dimers, which is PLC␤ 3 Ͼ PLC␤ 2 Ͼ PLC␤ 1 for most ␤␥ dimers, e.g. ␤ 1 ␥ 1 or ␤ 1 ␥ 2-His (cf. Figs. 2 and 7 and Refs. 9, 42, and 43) and PLC␤ 2 Ͼ PLC␤ 1 ϾϾϾ PLC␤ 3 for ␤ 5 ␥ 2-His (cf. Fig. 7). Furthermore, the results reported here show that both Rac1 and Rac2 are more FIG. 7. Stimulation of PLC␤ isozymes by ␤ 5 ␥ 2 and ␤ 1 ␥ 2 dimers. Soluble proteins of insect cells expressing PLC␤ 1 (0.3 g of protein/ sample), PLC␤ 2 (0.5 g of protein/sample), or PLC␤ 3 (4.0 g of protein/ sample) were reconstituted with increasing amounts of purified recombinant ␤ 1 ␥ 2-His (filled triangles) and ␤ 5 ␥ 2-His (filled squares). PLC activity was measured for 30 min at 150 nM free Ca 2ϩ in the presence of phospholipid vesicles containing PI-4,5-P 2 . potent activators of PLC␤ 2 than Cdc42Hs. All three PLC␤stimulating Rho GTPases, Rac2, Rac1, and Cdc42Hs, preferentially activate PLC␤ 2 (PLC␤ 2 Ͼ PLC␤ 3 Ն PLC␤ 1 ) (not shown). RhoA did not stimulate any of the three PLC␤ isozymes tested under the same conditions (not shown). Notably, the specificity of PLC␤ activation by Rho GTPases described here did not depend on the presence of detergents, which could conceivably differentially affect the activities of the PLC␤ isozymes, because purified detergent-free Rac2-LyGDI heterodimers (31) stimulated the PLC␤ isozymes studied here with the same rank order as did detergent-solubilized monomeric, C-terminally processed Rac2 (not shown). Interestingly, the rank order of PLC␤ stimulation by Rho GTPases also differs from that reported for activation of these enzymes by members of the ␣ q subfamily of G protein ␣ subunits: PLC␤ 1 Ն PLC␤ 3 Ͼ PLC␤ 2 (3,4). Together with the observation that the C-terminal region of PLC␤ isozymes is required for the stimulation by ␣ q subunits (43,44), but not for stimulation by Rho GTPases (20 and Fig. 6), this finding suggests that the stimulation of PLC␤ by ␣ subunits of heterotrimeric G proteins differs mechanistically from stimulation of this enzyme by Rho GTPases. Thus, PLC␤ isozymes can be activated, in the nanomolar concentration range, by both subunits of G proteins and by Rho GTPases but with distinct PLC␤ isozyme specificities.
The Role of the C-terminal Regions of PLC␤s in Their Regulation by Rho GTPases and ␤␥-Proteolytic cleavage of PLC␤ 3 by calpain at a site upstream of the C2 domain has been suggested to enhance ␤␥-mediated stimulation (36,37). Our data show, however, that truncation of recombinant PLC␤ 1 , PLC␤ 2 , and PLC␤ 3 at a site corresponding to the calpain cleavage site in PLC␤ 1 had no effect on their sensitivity to ␤␥ stimulation. The finding that the deletion of the C-terminal region of PLC␤ 3 did not enhance ␤␥ stimulation suggests that the increased PLC␤ 3 stimulation following calpain cleavage reported by Banno and coworkers (37) may not simply result from the removal of an inhibitory constraint built up by the C-terminal amino acids of PLC␤ 3 . Instead, it is more likely that the C-terminal region generated by treatment of PLC␤ 3 with calpain still interacts with the remaining part of the enzyme, to inhibit its activity, whereas it is absent from the deletion mutant studied here. Consistent with earlier results (20), Rac2 stimulation of PLC␤ 2 was enhanced in the absence of the C-terminal region of the enzyme. However, because this effect was not observed in the case of PLC␤ 1 and PLC␤ 3 , the PLC␤ 2 specificity described here is clearly not a consequence of Cterminal proteolysis of the enzymes.

Simultaneous Stimulation of PLC␤ Isozymes by Rho
GTPases and G Proteins-The fact that the stimulatory effects of ␤ 5 ␥ 2 and GTP␥S-activated Rac2 on PLC␤ 2 were strictly additive at saturating concentrations of the two activators suggests that there are separate sites on PLC␤ 2 for the interaction with ␤␥ and Rac2. Similar observations have been made previously for the activation of PLC␤ 2 and PLC␤ 3 by ␣ q and ␤␥ (9,45). In additional experiments, we have found that neither ␤␥-mediated activations nor ␣ q -mediated activations of PLC␤ 3 or PLC␤ 1 were influenced by GTP␥S-activated Rac2. Collectively, these results suggest that PLC␤ isozymes can be isozyme-specifically activated by three different stimulators, G protein ␣ subunits, G protein ␤␥ dimers, and Rho GTPases, via independent regulatory sites. Structural Requirements of PLC␤ Stimulation by Rho GTPases versus ␤␥ Dimers-The generation of chimeric PLC␤ enzymes made up of portions from an isoform stimulated only poorly by ␤␥ dimers and Rho GTPases, PLC␤ 1 ⌬, together with the remaining portions of PLC␤ 2 ⌬, an isoform markedly sensitive to both activators, allowed to delineate the structural elements of PLC␤ 2 required for the regulation by ␤␥ and Rac2. The fact that chimeras A and B were still markedly activated by ␤␥ dimers suggests that the catalytic subdomains of PLC␤ 2 are both necessary and sufficient for ␤␥ stimulation. This is Prior to this experiment, the amounts of the samples containing soluble PLC isoforms had been adjusted to give similar basal activities in the presence of 150 nM free Ca 2ϩ and 2 mM sodium cholate (not shown) (PLC␤ 2 ⌬, 2 g of protein/sample; chimera A, 60 g of protein/ sample; chimera B, 3 g of protein/sample; PLC␤ 1 ⌬, 0.3 g of protein/ sample). There was no effect of ␤ 1 ␥ 1 (1 M) or Rac2 (1 M) on phospholipase C activity of soluble fractions (60 g of protein/sample) of noninfected insect cells or insect cells infected with baculovirus encoding E. coli ␤-galactosidase.
FIG. 9. Linear representation of PLC␤ 2 ⌬, chimera A, chimera B, and PLC␤ 1 ⌬. The amino acid sequences were aligned using the ClustalW program contained in the PC/GENE software package (Intelligenetics, Mountain View, CA). Chimeras A and B are composed of the first 142 and 307 residues, respectively, of PLC␤ 1 , followed by the 695 and 530 C-terminal residues, respectively, of PLC␤ 2 ⌬. AB, antibody recognition site. consistent with a previous report showing that a region within the catalytic Y domain of PLC␤ 2 contains the stimulatory ␤␥ interaction site (40). The lower extent of stimulation of chimeras A and B by ␤␥ relative to PLC␤ 2 ⌬ is in line with the suggestion that an additional binding site for ␤␥ dimers may exist in the putative PH domain of PLC␤ 2 (41), which, albeit not absolutely required for ␤␥ stimulation, may increase the affinity of the ␤␥-PLC␤ 2 interaction. In addition, our results suggest that the region corresponding to the putative PH domain of PLC␤ 1 , an isozyme barely activated by ␤␥, is capable of substitute for the corresponding region of PLC␤ 2 . This is consistent with the recent report describing interaction of an isolated PH domain of PLC␤ 1 with ␤␥ dimers (41). Interestingly, construction of a chimera consisting of the putative PH domain of PLC␤ 2 and the catalytic subdomains X and Y of PLC␦, an enzyme that is not regulated by ␤␥ dimers, resulted in a ␤␥regulated enzyme (46). This suggests that ␤␥ may interact with multiple sites in phospholipase C isozymes and that these sites can be provided even by isozymes that are poorly (e.g. PLC␤ 1 ) or not at all (e.g. PLC␦ 1 ) sensitive to ␤␥ stimulation. An important outcome of our experiments on chimeric PLC␤ 1 -PLC␤ 2 enzymes is the observation that the substitution of the putative PH domain of PLC␤ 2 by the corresponding region of PLC␤ 1 abolished Rac2-mediated stimulation of the enzyme. This result not only demonstrates different structural requirements of PLC␤ 2 stimulation by ␤␥ and Rac2, but also shows, for the first time, that the putative PH domain of PLC␤ 2 is specifically and critically involved in mediating the regulation of the activity of the PLC␤ isozyme. It is currently unknown whether Rac2 directly binds to the PH domain of PLC␤ 2 or whether Rac2 induces a conformational change of the enzyme involving the PH domain. Interestingly, the PH domain of PLC␤ 2 was found to be required for both membrane targeting and catalytic activity of recombinant PLC␤ 2 in transfected COS-7 cells (44). The functional role of the region corresponding to the putative PH domain in PLC␤ isozymes is poorly understood. The isolated PH domain of PLC␤ 1 has been reported to bind inositol phospholipids (PI-3-P Ͼ PI-4,5-P 2 , PI-3,4,5-P 3 ) (47). A cooperative mechanism involving phosphatidylinositol 3-phosphate and ␤␥ subunits has been proposed to regulate plasma membrane localization and activation of PLC␤ 1 through the putative PH domain of this enzyme. A similar scenario involving Rac2 and ␤␥ could be depicted in the case of PLC␤ 2 . Thus, PLC␤ isozymes seem to act as a point of convergence of transmembrane signaling: PLC␤ 1 integrating signals emanating from inositol phospholipid 3-kinase and G protein ␣ q subunits, PLC␤ 3 those from G protein ␣ q and ␤␥ subunits, and PLC␤ 2 those from G protein ␤␥ subunits and Rac/Cdc42Hs.
The finding that the activity of PLC␤ isozymes can be specifically regulated by three different stimulators, G protein ␣ subunits, G protein ␤␥ dimers, and the Rho GTPases Rac and Cdc42Hs also enhances the cellular repertoire to coordinate, both spatially and temporally, responses to extracellular signals acting through stimulation of PLC␤ isozymes and thereby to enhance the degree of signal specificity. An intriguing possibility is that the Rho GTPases Rac and Cdc42Hs act as organizers to mediate recruitment of PLC␤ 2 to allow activation by G protein-coupled receptors only at specific sites within the cell and/or only within a specific time frame during or after receptor activation. The mechanisms involved in the recruitment of PLC␤ isozymes to their phospholipid are still not known. It seems clear, however, that the known activators, ␣ q and ␤␥, are not involved in this process (48,49). Interestingly, although PLC␤ 3 is strongly activated by ␤␥ dimers in cell-free systems, this isoform is, in marked contrast to PLC␤ 2 , not stimulated by chemoattractant receptor activation or exoge-nous ␤␥ dimers in transiently transfected COS-7 cells (17). These results support the notion that each PLC␤ isoform may require distinct additional components either for recruitment or stimulation of the catalytic activity. For example, activated members of the ␣ q subfamily of G protein ␣ subunits have been shown to permit PLC␤ stimulation by receptors acting through ␤␥ subunits of pertussis toxin-sensitive G proteins (50). The results presented herein suggest that Rac1, Rac2, and Cdc42Hs may contribute to the specificity and/or efficacy of PLC␤ signaling. Interestingly, both PLC␤ 2 and Rac2 are the major representatives of the PLC␤ and Rac GTPase subfamilies, respectively, in myeloid cells (16,51). Both proteins are activated in response to activation of chemoattractant receptors in a pertussis-toxin sensitive manner (17,52,53,54). The recent finding, that expression of a dominant-interfering form of Cdc42Hs in myeloid-differentiated HL-60 cells drastically reduced the formyl peptide receptor mediation of (a) formation of inositol 1,4,5-trisphosphate, (b) increase in the concentration of intracellular Ca 2ϩ , and (c) rapid activation of Rac2 (55), supports our hypothesis that Rac2 and possibly Rac1 and Cdc42Hs are critically involved in receptor-mediated regulation of PLC␤ 2 activity in intact cells. Although the mechanisms by which chemoattractant receptors stimulate Rac/Cdc42 are not defined and the functional relevance of phosphatidylinositol 3-kinase to this process is a subject of debate, the recently identified ␤␥and PI-3,4,5-P 3 -dependent Rac exchanger P-Rex 1 appears to fill the gap between chemoattractant receptors and Rac GTPases in leukocytes (56).
In conclusion, our results demonstrate that the specificity of PLC␤ stimulation by the Rho GTPases Rac and Cdc42Hs differs from the specificity observed for ␣ q/11 subunits and ␤␥ dimers of heterotrimeric G proteins and that PLC␤ 2 represents the PLC␤ isozyme most sensitive to stimulation by the Rho GTPases Rac and Cdc42 among the PLC␤ isozymes investigated in this study. Moreover, there are separate sites on PLC␤ isozymes for the regulation by Rho GTPases, G protein ␣ subunits, and ␤␥ subunits allowing independent activation by these stimulators.