Activation of Protein Kinase D by Signaling through Rho and the α Subunit of the Heterotrimeric G Protein G13 *

Protein kinase D (PKD/PKCμ) immunoprecipitated from COS-7 cells transiently transfected with either a constitutively active mutant of Rho (RhoQ63L) or the Rho-specific guanine nucleotide exchange factor pOnco-Lbc (Lbc) exhibited a marked increase in basal activity. Addition of aluminum fluoride to cells co-transfected with PKD and wild type Gα13 also induced PKD activation. Co-transfection of Clostridium botulinum C3 toxin blocked activation of PKD by RhoQ63L, Lbc, or aluminum fluoride-stimulated Gα13. Treatment with the protein kinase C inhibitors GF I or Ro 31-8220 prevented the increase in PKD activity induced by RhoQ63L, Lbc, or aluminum fluoride-stimulated Gα13. PKD activation in response to Gα13 signaling was also completely prevented by mutation of Ser-744 and Ser-748 to Ala in the kinase activation loop of PKD. Co-expression of C. botulinum C3 toxin and a COOH-terminal fragment of Gαq that acts in a dominant-negative fashion blocked PKD activation in response to agonist stimulation of bombesin receptor. Expression of the COOH-terminal region of Gα13 also attenuated PKD activation in response to bombesin receptor stimulation. Our results show that Gα13 contributes to PKD activation through a Rho- and protein kinase C-dependent signaling pathway and indicate that PKD activation is mediated by both Gαq and Gα13 in response to bombesin receptor stimulation.

PKD/PKC is a serine/threonine protein kinase (6,7) with distinct structural, enzymological, and regulatory properties (8). In particular, PKD is rapidly activated in intact cells through a mechanism that involves phosphorylation (8). Specifically, exposure of intact cells to phorbol esters, cell-permeant DAGs, or bryostatin induces rapid PKD phosphorylation and activation, which is maintained during cell lysis and immunoprecipitation (8 -13). Several lines of evidence, including the use of PKC-specific inhibitors and co-transfection of PKD with constitutively active PKC mutants, indicate that PKD is activated through a novel PKC-dependent signal transduction pathway in vivo (9 -11). The residues Ser-744 and Ser-748 in the activation loop of PKD have been identified as critical phosphorylation sites in PKD activation induced by phorbol esters (14). Taken together, these results suggest an important connection between PKCs and PKD and indicate that PKD can function downstream of PKC in a novel signal transduction pathway.
Heterotrimeric guanine nucleotide-binding regulatory proteins (G proteins) are composed of ␣, ␤, and ␥ subunits and transduce external signals from heptahelical receptors to intracellular effectors (15). Mammalian G protein ␣ subunits are classified into four subfamilies: G s , G i , G q , and G 12 . The ␣ subunit of G q stimulates the ␤ isoforms of phospholipase C (PLC) that catalyze the production of inositol 1,4,5-trisphosphate that triggers the release of Ca 2ϩ from internal stores and diacylglycerol (DAG) that activates the classical and novel isoforms of PKC (reviewed in Ref. 16). We reported that a variety of neuropeptide agonists that signal through heptahelical receptors and couple to heterotrimeric G proteins, including bombesin, bradykinin, endothelin, and vasopressin, induce rapid PKD activation in normal and neoplastic cells (11,13,17,18). Although each of these receptors activates G q and G␣ q signaling stimulates PKD activity (19), the kinetics and degree of PKD activation differ among the different receptors even in the same cell line (11). Furthermore, expression of a COOHterminal fragment of G␣ q that acts in a dominant-negative fashion attenuated (but did not eliminate) PKD activation in response to agonist stimulation of bombesin receptor (19). Taken together, these results prompted us to consider the possibility that G protein-coupled receptors (GPCRs) stimulate PKD activation not only via G␣ q but also through other, as yet unidentified, G protein-mediated signaling pathways.
Many G q -coupled receptors also interact with other heterotrimeric G proteins including members of the G 12 family which mediate activation of the low molecular weight G proteins of the Rho subfamily (20 -25) via guanine nucleotide exchange factors that directly link the G␣ subunits to Rho (26 -28). Rho plays a major role in promoting cytoskeletal responses including formation of actin stress fibers, assembly of focal adhesions, and tyrosine phosphorylation of focal adhesion proteins and has been implicated in gene expression, cell migration, prolif-eration, and transformation (29 -31). Interestingly, a number of recent studies have suggested a convergence between Rhoand PKC-mediated signaling in yeast and mammalian cells (32)(33)(34)(35)(36)(37). In the budding yeast, the homologues of mammalian PKC and RhoA, Pkc1p and Rho 1p, respectively, act through a common mechanism that appears to involve a direct interaction between these proteins (32,33). In epithelial and endothelial cells, treatment with Clostridium difficile toxin, which inactivates all members of the Rho subfamily, prevented PKC translocation and activation (34). Rho signaling has been shown to enhance AP-1 transcription in T lymphocytes, and a molecular association between Rho and PKC has been demonstrated in these cells (35). Recently, Slater et al. (36) have demonstrated that Rho-GTP potently stimulates PKC␣ activity in vitro using recombinant proteins, and Sagi et al. (37) reported that G␣ q and PLC signaling are synergistic with Rho. In addition, Rho has been implicated in the stimulation of pathways leading to lipid-derived second messenger synthesis and PKC activation (38,39) including phosphatidylinositol 4,5bisphosphate (PtdIns(4,5)P 2 ) generation (40,41), PLD activation (42,43), and inhibition of DAG kinase isoforms (44). These considerations prompted us to examine whether, in addition to G␣ q , Rho-and G␣ 13 -mediated signaling can promote PKD activation in intact cells and whether endogenous Rho and G␣ 13 could contribute to PKD activation in response to bombesin receptor stimulation.
The results presented here demonstrate that co-expression of PKD with constitutively activated Rho, the Rho-specific guanine nucleotide exchange factor pOnco-Lbc (45,46), or G␣ 13 stimulated by aluminum fluoride induced a marked increase in PKD activity. The Clostridium botulinum C3 toxin, which inactivates Rho, selectively prevented PKD activation in response to Rho, Lbc, and G␣ 13 and attenuated PKD activation in response to bombesin GPCR activation. PKD activation induced by G␣ 13 /Rho signaling was prevented by PKC inhibitors and by mutation of Ser-744 and Ser-748 in the kinase activation loop of PKD to alanine, a non-phosphorylatable residue. Expression of a COOH-terminal fragment of G␣ q that acts in a dominant-negative fashion together with the C3 toxin virtually abolished PKD activation induced by agonist stimulation of the bombesin receptor. Thus, our results identify PKD as a novel downstream target in G␣ 13 and Rho signaling and indicate that bombesin GPCR stimulation promotes PKD activation via both G q -and G␣ 13 /Rho-dependent pathways.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-COS-7 cells were maintained by subculture in 10-cm tissue culture plates every 3-4 days in Dulbecco's modified Eagle's medium supplemented with 10% FBS at 37°C in a humidified atmosphere containing 10% CO 2 . For experimental dishes, cells were subcultured at 6 ϫ 10 4 cells/ml in 6-(5 ml) or 10-cm (10 ml) dishes on the day prior to transfections. All transfections and co-transfections were carried out with equivalent amounts of DNA (6 g/6-cm dish and 12 g/10-cm dish). Transfections were carried out in Opti-MEM (Life Technologies, Inc.) using Lipofectin (Life Technologies, Inc.) at 10 l/6-cm dish or 20 l/10-cm dish, added to cells in a final volume of 2.5 ml/6-cm dish or 5 ml/10-cm dish, following formation of DNA-Lipofectin complexes according to the protocol provided by the manufacturer. Cells were allowed to take up complexes in the absence of FBS for 5-6 h or overnight and then FBS (10% final concentration) in Opti-MEM was added to the dishes to yield a final volume of 5 ml/6-cm dish or 10 ml/10-cm dish. Cells were used for experiments after a further 48 -72 h of incubation.
cDNA Constructs Used in Transfections-The wild type G␣ q subunit cDNAs in the eukaryotic expression vector pcDNA-1 (Invitrogen) (47) were obtained from the American Type Tissue Collection (Manassas, VA). The murine wild type G␣ 13 subunit cDNAs in pcDNA-1 were gifts from Dr. H. R. Bourne (University of California, San Francisco) (48). The constructs pcDNA3-PKD encoding PKD (49) and pcDNA3-PKD mutants encoding PKD with site-specific mutations within the activa-tion loop in the catalytic domain including the single mutants (PKD-S744A and PKD-S748A) and double mutants (PKD-S744A/S748A and PKD-S744E/S748E) have been described previously (14). BNR-pCD2 containing the cDNA encoding the bombesin/gastrin-releasing peptide receptor was kindly provided by Dr. J. F. Battey, Jr. (NIDCD, National Institutes of Health, Bethesda). Expression vectors for Rho, RhoQ63L, Ras, RasV12, Rac, and RacV12 in pcDNA3 were provided by Dr. D. Chang (UCLA). The production of the vector GFP-␣ q CT encoding a fusion protein of GFP containing G␣ q (residues 305-359) at its carboxyl terminus have been described previously (14). The plasmids pEF-LacZ and pEF-C3 were provided by Dr. R. Treisman (Imperial Cancer Research Fund, London, UK). The pOnco-Lbc was provided by Dr. D. Toksoz (Tufts University, Boston) and was described previously (50).
Polymerase chain reaction was used to generate DNA encoding for the carboxyl-terminal region of G␣ 13 (residues 333-377) using the murine G␣ 13 cDNA as a template with sense (5Ј-GCTCAAGCTTC-GAAACGCCGGGACCAGCAGCAG-3Ј) and antisense (5Ј-GGTGGATC-CTCACTGCAGCATGAGCTGCTTCAG-3Ј) primers. The resulting DNA fragment was subcloned into the BamHI and HindIII restriction sites of pcDNA-3. The fidelity of the polymerase chain reaction was confirmed by DNA sequencing. The BamHI/HindIII DNA fragment was cloned in p⑀GFP-C1 (CLONTECH, Inc., La Jolla, CA) such that the resulting fusion protein produced by this plasmid would be a hybrid ⑀GFP containing G␣ 13 (residues 333-377) at its carboxyl terminus.
Immunoprecipitations-Transfected COS-7 cells were washed twice with Dulbecco's modified Eagle's medium and equilibrated in 5 ml of the same medium at 37°C for 1-2 h. Some dishes were treated with various pharmacological agents during this equilibration period or with agonists for 10 min at the end of this period, as indicated in the corresponding figure legends. Cells were lysed in buffer A (50 mM Tris-HCl, pH 7.6, 2 mM EGTA, 2 mM EDTA, 1 mM dithiothreitol, 100 g/ml leupeptin, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, hydrochloride (Pefabloc), and 1% Triton X-100). PKD was immunoprecipitated at 4°C for 3 h with the PA-1 antiserum (1:50 dilution) raised against the synthetic peptide EEREMKALSERVSIL that corresponds to the COOH-terminal region of PKD as described previously (6,49). The immune complexes were recovered using protein-A coupled to agarose.
In Vitro Kinase Assays-Immune complexes were washed twice with lysis buffer and then twice with kinase buffer consisting of 30 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol. Autophosphorylation reactions were initiated by combining 20 l of immune complexes with 5 l of a phosphorylation mixture containing 100 M [␥-32 P]ATP (specific activity, 400 -600 cpm/pmol) in kinase buffer. Following incubation at 30°C for 10 min, the reactions were terminated by addition of 1 ml of ice-cold kinase buffer and placed on ice. Immune complexes were recovered by centrifugation, and the proteins were extracted for SDS-PAGE analysis by addition of 2ϫ SDS-PAGE sample buffer (200 mM Tris-HCl, pH 6.8, 0.1 mM sodium orthovanadate, 1 mM EDTA, 6% SDS, 2 mM EDTA, 4% 2-mercaptoethanol, 10% glycerol). Dried SDS-PAGE gels were subjected to autoradiography to visualize radiolabeled protein bands.
For assays of exogenous substrate phosphorylation, immune complexes were processed as for autophosphorylation reactions, and then substrate (syntide-2, final concentration 2.5 mg/ml) was added in the presence of 100 M [␥-32 P]ATP (400 -600 cpm/pmol) in kinase buffer (final reaction volume, 30 l). After incubation at 30°C for 10 min, the reactions were terminated by adding 100 l of 75 mM H 3 PO 4 , and 75 l of the mixed supernatant was spotted to Whatman P-81 phosphocellulose paper. Papers were washed thoroughly in 75 mM H 3 PO 4 and dried, and radioactivity incorporated into syntide-2 was determined by detection of Cerenkov radiation in a scintillation counter.
Western Blot Analysis-Samples of cell lysates were directly solubilized by boiling in SDS-PAGE sample buffer. Following SDS-PAGE on 8% gels (for PKD) or 10% gel (for G proteins), proteins were transferred to Immobilon-P membranes (Millipore), as described previously (9,23) and blocked by overnight incubation with 5% non-fat dried milk in PBS, pH 7.2. Membranes were incubated at room temperature for 3 h with antisera specifically recognizing either PKD, the different G proteins (G␣ q or G␣ 13 ), or GFP at 1:250 -1:500 dilution in phosphate-buffered saline containing 3% non-fat dried milk. Immunoreactive bands were visualized using either horseradish peroxidase-conjugated anti-rabbit IgG and subsequent enhanced chemiluminescence detection or 125 Ilabeled protein A followed by autoradiography. The G␣ q antiserum was raised against a synthetic peptide corresponding to the COOH-terminal decapeptide of this G protein that was cross-linked to keyhole limpet hemocyanin with glutaraldehyde. The G␣ 13 antiserum was raised against the synthetic peptide CLHDNLKQLMLQ (which corresponds to the carboxyl-terminal peptide 367-377 of murine G␣ 13 with an amino-terminal cysteine added for coupling) cross-linked to keyhole limpet hemocyanin with the hetero-bifunctional reagent sulfosuccinimidyl 4-(p-maleimidophenyl) butyrate, as described (52). The antibody for GFP was raised in rabbits to a glutathione S-transferase-GFP fusion protein, as described recently (53).
Materials-[␥-32 P]ATP (370 MBq/ml), 125 I-labeled protein A (15 mCi/ ml), horseradish peroxidase-conjugated donkey anti-rabbit IgG, and enhanced chemiluminescence reagents were from Amersham Pharmacia Biotech. Protein-A agarose and Pefabloc were from Roche Molecular Biochemicals. Opti-MEM and Lipofectin were from Life Technologies, Inc. GF I and cytochalasin D were obtained from Sigma. Ro 31-8220 and 1-(5-isoquinolinesulfonyl)-homopiperazine (HA1077) were from Calbiochem. All other reagents were from standard suppliers or as described in the text and were the highest grade commercially available.

RESULTS AND DISCUSSION
Expression of Constitutively Activated Rho Induces PKD Activation-To test whether PKD activation can be induced by signaling pathway(s) initiated by the low molecular weight G proteins of the Rho subfamily, COS-7 cells were co-transfected with PKD and expression vectors encoding wild type Rho, Ras, Rac or constitutively active forms of the three proteins, RhoQ63L, RasVal12, or RacVal12. PKD was immunoprecipitated from the lysates of transfected cells with PA-1 antiserum, and the immune complexes were incubated with [␥-32 P]ATP, subjected to SDS-PAGE, and analyzed by autoradiography to detect the prominent 110-kDa band corresponding to autophosphorylated PKD.
The results presented in Fig. 1A show for the first time that cells co-transfected with RhoQ63L and PKD exhibit a marked increase in PKD activity compared with cells transfected with either PKD and wild type Rho or with PKD alone. Similar results were obtained when PKD activity in immunocomplexes was determined by phosphorylation of syntide-2 (54, 55), a synthetic peptide demonstrated previously (6) to be an excellent substrate for PKD. We verified that the level of PKD expression in cells co-transfected with Rho (either wild type or QL) and PKD was similar to those transfected with PKD ( Fig.  1C). In contrast, PKD activity was increased only slightly by overexpression of wild type or constitutively activated mutants of Rac or Ras (Fig. 1, A and B).
Expression of the C. botulinum C3 toxin, which specifically ADP-ribosylates Rho at residue 41 and impairs its function (46), markedly attenuated PKD activation induced by Rho QL (Fig. 1, A and B, right). Expression of the C3 toxin did not interfere with the expression of PKD ( Fig. 1, C, right). The results presented in Fig. 1 indicate that Rho activation is a potential pathway leading to PKD activation.
Low molecular weight G proteins of the Rho subfamily cycle between an inactive form bound to GDP and an active GTPbound state. Guanine nucleotide exchange factors positively modulate the small GTPases by catalyzing the dissociation of the bound GDP to allow the association of GTP. The lbc oncogene encodes a specific guanine nucleotide exchange factor for Rho and causes cellular transformation through activation of the Rho signaling pathway (45,46). In order to substantiate the conclusions drawn from data in Fig. 1, we transfected COS-7 cells with PKD either with or without constitutively activated Lbc (pOnco-Lbc), and we examined the effects of Lbc-mediated activation of endogenous Rho on PKD activity in immunoprecipitates. As shown in Fig. 2, expression of pOnco-Lbc (Lbc) induced a significant increase in PKD activity, as shown by assays of either autophosphorylation ( Fig. 2A) or syntide-2 phosphorylation (Fig. 2B). PKD activation induced by Lbc was virtually abrogated by co-expression of C. botulinum C3 toxin, confirming that Lbc leads to PKD activation via Rho. Endogenous Rho was sufficient to mediate the effects of Lbc since the expression of wild type Rho with Lbc did not produce a significant further increase in PKD activity. The expression of Lbc with or without C3 toxin did not affect the level of PKD expression (Fig. 2C).
Aluminum Fluoride Stimulates PKD Activation in COS-7 Cells Transfected with G␣ 13 -Recently, the G 12 subfamily has been implicated in pathways leading to activation of the low molecular weight G proteins of the Rho subfamily (20 -25). Specifically, G␣ 13 stimulates the nucleotide exchange activity of p115 GEF for Rho thereby leading to Rho activation (26 -28). In order to examine the effect of G␣ 13 signaling on PKD activity, we transiently transfected COS-7 cells with vector or wild type G␣ 13 , as well as G␣ q , and then stimulated the cells with either 10 M aluminum fluoride or PDB as a positive control. Aluminum fluoride activates heterotrimeric G proteins due to its ability to mimic the ␥-phosphoryl group of GTP when com- plexed with the GDP-bound ␣ subunit (56). PKD activity in immunocomplexes was determined by autophosphorylation or by phosphorylation of syntide-2.
As shown in Fig. 3, PKD isolated from unstimulated COS-7 cells had low catalytic activity that was markedly activated by PDB stimulation of intact cells (ϳ10-fold increase). Overexpression of G␣ q or G␣ 13 in COS-7 cells did not induce any detectable PKD activation. Addition of aluminum fluoride to cells transfected with PKD alone failed to induce any significant increase in PKD activity, whereas addition of aluminum fluoride to cells co-transfected with PKD and wild type G␣ q induced a marked increase in PKD activity, in agreement with our recent results (19). A salient feature of the results shown in Fig. 3 is that aluminum fluoride stimulation of cells transfected with G␣ 13 also induced a significant increase in PKD activity in immunocomplexes as measured by autophosphorylation (Fig.   3A) or syntide-2 phosphorylation assays (Fig. 3B). Western blot analysis confirmed that the cells transfected with the G␣ q or G␣ 13 expression plasmids overexpressed these G␣ subunits (Fig. 3C).
Next, we used C. botulinum C3 toxin to determine whether

FIG. 2. pOnco-Lbc induces PKD activation in COS-7 cells, and this induction is blocked by the Rho inhibitor C3 toxin.
Exponentially growing COS-7 cells were co-transfected with pcDNA3-PKD (PKD) and pSR␣ or pSR␣ encoding pOnco-Lbc (Lbc) or wild type Rho (Rho). Two of these cell dishes were also co-transfected with pEF-LacZ (Ϫ) or pEF-C3 (ϩ), as indicated in this figure. Three days after transfection, the cultures were lysed. A, the lysates were immunoprecipitated with PA-1 antiserum, and PKD activity in the immunocomplexes was determined by an in vitro kinase assay (IVK) as described under "Experimental Procedures," followed by SDS-PAGE and autoradiography. A representative autoradiogram is shown. The position of autophosphorylated PKD at apparent M r 110,000 is indicated by the arrow to the left. Similar results were obtained in four independent experiments. The bar graph shows the quantification of the level of PKD autophosphorylation in these experiments performed by scanning densitometry. The results expressed as an increased fold over control in phosphorylation are means Ϯ S.E. (n ϭ 4). B, syntide-2 phosphorylation in immune complexes. The results expressed as an increased fold over control in phosphorylation represent the mean Ϯ S.E. obtained from three independent experiments, each performed in duplicate. C, levels of expression of PKD in each transfection were analyzed by Western blotting (W. Blot) aliquots of total cell lysates with antiserum against PKD. PKD activation in response to G␣ 13 signaling proceeds through a Rho-dependent pathway. As illustrated in Fig. 3, expression of C3 toxin blocked the increase in PKD activity produced by aluminum fluoride stimulation for the cells co-expressing PKD with G␣ 13 . In contrast, C3 toxin did not prevent the increase in PKD activity induced by treatment with aluminum fluoride of cells co-expressing PKD with G␣ q . These results strongly indicate that Rho mediates PKD activation induced by aluminum fluoride in G␣ 13 -transfected cells.
We reported previously (19) (and confirmed in this study) that PKD isolated from COS-7 expressing constitutively activated G␣ 12 or G␣ 13 did not exhibit increased catalytic activity. In view of the results shown in Fig. 3, we considered the possibility that chronic signaling leading to PKD activation by constitutively activated G␣ 13 could be impaired. For example, G proteins including G␣ 12 or G␣ 13 have been shown to be phosphorylated by phorbol ester-sensitive PKC isoforms (57)(58)(59) leading to their desensitization (60). To determine whether PKD is activated in response to acute rather than chronic signaling through G␣ 13 , COS-7 cells were transfected with vectors encoding wild type G␣ 13 or G␣ q , and after 72 h of incubation, the cultures were challenged with aluminum fluoride for various times (1-24 h). As shown in Fig. 3D, the increase in PKD activity induced by aluminum fluoride in cells transfected with G␣ 13 declined gradually reaching almost baseline values after 24 h of continued exposure to aluminum fluoride. In contrast, PKD activation in response to aluminum fluoride-stimulated G␣ q remained undiminished even after 24 h of treatment. These results indicate that chronic stimulation of G␣ 13 signaling results in desensitization of PKD activation mediated by this pathway.

The PKC Inhibitors GF I and Ro 31-8220 Prevent PKD Activation by Aluminum Fluoride in COS-7 Cells
Transfected with G␣ 13 -A number of recent studies have suggested a convergence between Rho-and PKC-mediated signaling in a variety of cell types (32)(33)(34)(35)(36)(37). Consequently, we determined whether PKCs mediate PKD activation induced by G␣ 13 activation, using inhibitors that discriminate between PKCs and PKD. COS-7 cells transiently transfected with wild type G␣ 13 were treated for 1 h with the potent inhibitors of phorbol estersensitive isoforms of PKC GF I (also known as GF 109203X or bisindolylmaleimide I) or Ro 31-8220 (61, 62), prior to stimulation with 10 M aluminum fluoride. As shown in Fig. 4A, exposure to either GF I or Ro 31-8220 potently blocked PKD activation induced by aluminum fluoride in G␣ 13 -transfected cells. In contrast, the compound GF V, which is structurally related to GF I but biologically inactive, did not affect PKD activation in response to aluminum fluoride in these cells. Previously, we demonstrated (9, 11) that either GF I or Ro 31-8220 does not inhibit PKD activity when added directly to the in vitro kinase assay at concentrations identical to those required to block PKD activation by aluminum fluoride in G␣ 13 -transfected COS-7 cells. Thus, the results shown in Fig. 4 imply that GF I and Ro 31-8220 do not inhibit PKD activity directly but interfere with G␣ 13 -mediated PKD activation in intact COS-7 cells by blocking PKC.
To substantiate further that PKD activation induced through the G␣ 13 /Rho signaling pathway is mediated by PKC, we tested whether PKC inhibitors also inhibit PKD activation induced by either expression of pOnco-Lbc (Lbc) or RhoQL. COS-7 cells transiently transfected with wild type PKD and either Lbc or RhoQL were incubated with GF I/Ro 31-8220 or GF V. As shown in Fig. 4B, exposure to either GF I or Ro 31-8220 potently blocked PKD activation induced by expression of either Lbc or RhoQL. In contrast, addition of GF V did not affect PKD activation in response to Lbc or RhoQL.
The Rho-associated coiled-coil forming protein serine/threonine kinase (ROCK) family composed of ROCK (also known as Rho kinase or ROCK-II) and the closely related p160ROCK (also known as ROCK-I) has been identified as one of the downstream targets of Rho-GTP (63-65) that transduce Rho activation into cytoskeletal responses (66,67). The ROCKs have also been implicated in PtdIns(4,5)P 2 generation (41) and PLD activation (43) which potentially could lead to Rhodependent PKC activation. The protein kinase inhibitor 1-(5isoquinolinesulfonyl)-homopiperazine (HA-1077) has been recently identified as a potent inhibitor of ROCK (68). As shown in Fig. 4, A and B, treatment with 20 M HA-1077 did not prevent PKD activation in response to aluminum fluoride-stimulated G␣ 13 , RhoQL, or Lbc. Similarly, disruption of the actin cytoskeletal organization in response to cytochalasin D did not interfere with PKD activation produced by aluminum fluoridestimulated G␣ 13 , RhoQL, or Lbc. All these results indicate that stimulation of G␣ 13 /Rho signaling promotes PKD activation through a PKC-dependent but ROCK-independent pathway. Ϫ ) for 30 min and lysed. The lysates were immunoprecipitated with PA-1 antiserum, and PKD activity in the immunocomplexes was determined by autophosphorylation (in vitro kinase assay (IVK)) or by phosphorylation of syntide-2, as described under "Experimental Procedures." A and B, upper panels, the autoradiograms shown are representative of at least three independent in vitro kinase assay experiments with similar results; lower panels, syntide-2 phosphorylation in immune complexes. Results represent the mean Ϯ S.E. from three independent experiments, each performed in duplicate. 13 -The residues Ser-744 and Ser-748 in the activation loop of PKD are required for PKC-dependent PKD activation induced by phorbol esters (14), oxidative stress (69), and bombesin (19). If Ser-744 and Ser-748 are target sites for activating phosphorylation events in response to G␣ 13 signaling, their conversion to Ala should reduce or eliminate G␣ 13 -mediated activation of PKD. To test this possibility we used PKD mutants with single or double substitutions of these residues cloned in the expression vector pcDNA3 (i.e. PKD-S744A, PKD-S748A, or PKD-S744A/S748A). COS-7 cells, co-transfected with wild type PKD or PKD mutants and G␣ 13 , were treated with or without aluminum fluoride. PKD kinase activity was measured by either autophosphorylation or syntide-2 phosphorylation.

Substitution of Ser-744 and Ser-748 by Alanine Prevents PKD Activation in Response to G␣
As shown in Fig. 5, PKD isolated from unstimulated cells had low catalytic activity that was markedly activated by aluminum fluoride-stimulated G␣ 13 . Single substitutions of either Ser-744 or Ser-748 by Ala resulted in PKD mutants that displayed reduced activity after stimulation (50% decrease in both single Ala mutants compared with stimulated wild type PKD). In contrast, substitution of both Ser-744 and Ser-748 for Ala in PKD completely blocked kinase activation induced by aluminum fluoride-stimulated G␣ 13 (Fig. 5, A and B). In all cases, the protein expression levels of the transfected PKD mutants were comparable to that of wild type PKD, as shown by Western blot analysis (Fig. 5C). Thus, substitution of Ser-744 and Ser-748 in the activation loop of PKD by neutral non-phosphorylatable residues prevents the activation of this enzyme by aluminum fluoride-stimulated G␣ 13 in vivo.
As shown in Fig. 5, A and B, and in agreement with previous results (14), replacement of both serine residues with glutamic acid (PKD-S744E/S748E) markedly increased PKD basal activity. Interestingly, the activity of the PKD-S744E/S748E mutant was not significantly further enhanced by aluminum fluoride-stimulated G␣ 13 , suggesting that phosphorylation of these two sites induces maximal PKD activation in response to these pathways.
Role of Endogenous G␣ 13 in Mediating PKD Activation in Response to Bombesin Receptor Activation-The COOH terminus of G proteins plays a key role in their interaction with cognate receptors (70). Recently, peptides corresponding to this region of G␣ q or G␣ i have been shown to target the receptor-G protein interface in a selective manner and thereby block receptor-mediated PLC activation (71) and inwardly rectifying K ϩ channel activity (72,73), respectively. For example, transient transfection of COS-7 cells with ␣ 1B -adrenergic receptors or M 1 muscarinic receptors and the COOH-terminal region of G␣ q attenuated inositol phosphate production in response to receptor activation (71).
In the present study, a dominant-negative strategy was also used to test the role of endogenous G␣ 13 in bombesin receptormediated PKD activation. We generated chimeric fusion proteins between the COOH-terminal region of G␣ 13 (referred as G␣ 13 CT) and GFP from Aequorea victoria, which forms an independent 30-kDa domain with inherent fluorescence (51). Initially, we verified that the GFP-G␣ 13 CT chimera is expressed in transiently transfected COS-7 cells as judged by Western blot analysis using antibodies directed against either GFP or the COOH-terminal region of G␣ 13 (Fig. 6A). In addition, we also visualized the expression of the GFP-G␣ 13 CT chimera by examining GFP fluorescence in individual COS-7 cells (results not shown).
Next, we determined whether expression of GFP-G␣ 13 CT interferes with PKD activation via the bombesin receptor. COS-7 cells were co-transfected with PKD, bombesin receptor, and either GFP-G␣ 13 CT or GFP. After 72 h, the cells were challenged with either bombesin or PDB for 10 min and then lysed. PKD activity, after immunoprecipitation, was assayed by autophosphorylation. The results illustrated in Fig. 6B demonstrate that expression of GFP-G␣ 13 CT markedly attenuated the increase of PKD activity induced by bombesin. In contrast, expression of GFP-G␣ 13 CT did not interfere with PKD activation in response to PDB which directly stimulates PKC leading to PKD activation and therefore bypasses the receptor/G protein interaction. These results indicate that endogenous G␣ 13 contributes to PKD activation in response to bombesin receptor activation.
Role of Endogenous G␣ q and Rho in Mediating PKD Activation in Response to Bombesin Receptor Activation-In order to determine the contribution of Rho-dependent pathways to PKD Ϫ ) for 30 min and lysed. The lysates were immunoprecipitated with PA-1 antiserum, and PKD activity was determined by in vitro kinase assay (IVK) as described under "Experimental Procedures." A, the autoradiogram of in vitro kinase assay shown is representative of at least three independent experiments. The position of autophosphorylated PKD at apparent M r 110,000 is indicated by the arrow to the left; the bar graph shows the quantification of the level of PKD autophosphorylation in these experiments performed by scanning densitometry. The results expressed as a percentage of the maximum increase in phosphorylation are means Ϯ S.E. of three independent experiments. B, syntide-2 phosphorylation in immune complexes. The results expressed as an increased fold over control in phosphorylation represent the mean Ϯ S.E. obtained from three independent experiments, each performed in duplicate. C, levels of expression of wild type PKD and the different PKD mutants were analyzed by Western blotting (W. Blot) aliquots of total cell lysates with PKD antiserum. activation in response to bombesin GPCR stimulation, we cotransfected COS-7 cells with expression vectors encoding bombesin GPCR and PKD with or without a C3 toxin expression vector. After 72 h, the cells were challenged with either bombesin or PDB for 10 min and lysed, and PKD activity, after immunoprecipitation, was assayed by autophosphorylation or syntide-2 phosphorylation. As illustrated in Fig. 7 (A and B), expression of C3 toxin attenuated the increase in PKD activity produced by bombesin but did not interfere with PKD activation promoted by PDB. These results indicate that maximal bombesin-induced PKD activation requires functional Rho.
Recently, we reported (19) that expression of a chimeric fusion protein between the COOH-terminal region of G␣ q (referred as G␣ q CT) and GFP attenuated bombesin-induced PKD activation. Here, we determined whether expression of C3 toxin together with GFP-G␣ q CT interferes with PKD activation induced by bombesin in an additive manner. COS-7 cells were co-transfected with PKD, bombesin receptor, and GFP-G␣ q CT with or without C3 toxin. In agreement with previous results, expression of GFP-G␣ q CT (but not GFP) attenuated the increase of PKD activity induced by bombesin (Fig. 7, A and B).
The salient feature of the results presented in Fig. 7 is that co-transfection of C3 toxin together with GFP-G␣ q CT almost abolished bombesin-induced PKD activation. In contrast, expression of GFP-G␣ q CT together with C3 toxin did not interfere with PKD activation in response to PDB which directly stimulates PKC. We verified that the GFP-G␣ q CT chimera is expressed in transiently transfected COS-7 cells as well as in cells co-transfected with GFP-G␣ q CT and C3 toxin, as judged by Western blot analysis using antibodies directed against either GFP or the COOH-terminal region of G␣ q (Fig. 7C, upper  panel). In addition, we also confirmed that protein expression levels of PKD were comparable under all the experimental conditions used (Fig. 7C, lower panel). Taken together, the results presented in Figs. 6 and 7 indicate that endogenous G 13 , G q , and Rho mediate PKD activation in response to bombesin receptor activation.
Concluding Remarks-Activation of a number of receptors that couple to heterotrimeric G proteins including bombesin, bradykinin, endothelin, vasopressin, and lysophosphatidic acid have been shown to stimulate PKD activation in a variety of Exponentially growing COS-7 cells were co-transfected with pcDNA3-PKD (PKD), BNR-pCD2 (BR) containing the cDNA encoding the bombesin/gastrin-releasing peptide receptor, pEF-LacZ (pEF), or pEF-C3 (C3), and pcDNA3 encoding GFP (GFP) or GFP-␣ q CT (GFP-␣ q CT), as indicated. Three days after transfection, the cultures were left unstimulated (Ϫ) or stimulated (ϩ) with either 200 nM PDB or 10 nM bombesin (Bom) for 10 min and lysed. The lysates were immunoprecipitated with PA-1 antiserum, and PKD activity in the immunocomplexes was determined by autophosphorylation (in vitro kinase assay (IVK), A) or by phosphorylation of syntide-2 (B) as described under "Experimental Procedures." A, the autoradiogram of in vitro kinase assay shown is representative of at least three independent experiments with similar results. The bar graph shows the quantification of the level of PKD autophosphorylation in these experiments performed by scanning densitometry. The results expressed as a percentage of the maximum increase in phosphorylation are mean Ϯ S.E. of three independent experiments. B, syntide-2 phosphorylation in immune complexes. Results expressed as an increased fold in phosphorylation represent the mean Ϯ S.E. from three experiments, each performed in duplicate. C, upper panel, levels of expression of GFP-␣ q CT were analyzed by Western blotting (W. Blot) aliquots of transfected cell lysates with either G␣ q or GFP antibodies, as indicated; bottom panel, levels of expression of PKD were analyzed by Western blotting aliquots of total cell lysates with PKD antiserum. cell types. Although each of these receptors activates G q and G␣ q signaling stimulates PKD activity (Ref. 19 and results presented here), the kinetics and degree of PKD activation differ among the different receptors even in the same cell line. Some of the variation may be the result of receptors coupling to more than one G protein initiating additional signaling pathways that can also contribute to PKD activation. In this context, it is also relevant that expression of a COOH-terminal fragment of G␣ q that acts in a dominant-negative fashion attenuated (but did not eliminate) PKD activation in response to agonist stimulation of bombesin receptor (Ref. 19 and results presented here). These considerations raised the possibility that GPCRs promote PKD activation not only via G␣ q but also through other, as yet unidentified, G protein signaling pathways.
Many G q -coupled receptors also interact with heterotrimeric G proteins of the G 12 family that are known to promote Rho activation via the novel guanine nucleotide exchange factor p115 GEF which directly links G␣ 13 to Rho (26,27). It is increasingly recognized that Rho and PKC signaling interact in a variety of cell types (see Introduction for references). In particular, recent studies demonstrated that Rho-GTP potently stimulates PKC␣ activity in vitro (36) and that G␣ q and phospholipase C signaling are synergistic with Rho in vivo (37). Consequently, we examined in this study whether Rho-and G␣ 13 -mediated signaling can promote PKD activation in intact cells and whether endogenous Rho could mediate PKD activation in response to bombesin receptor stimulation.
Our results demonstrate that PKD immunoprecipitated from COS-7 cells transiently transfected with either a constitutively active mutant of Rho (RhoQ63L) or the Rho-specific guanine nucleotide exchange factor Lbc exhibited a marked increase in basal activity. These findings demonstrate, for the first time, that Rho-regulated signaling leads to PKD activation. Further-more, addition of aluminum fluoride to cells co-transfected with PKD and wild type G␣ 13 also induced a marked increase in PKD activity. Transfection of C. botulinum C3 toxin specifically blocked activation of PKD by RhoQ63L, Lbc, or by aluminum fluoride-stimulated G␣ 13 . In contrast, expression of the C3 toxin did not interfere with PKD activation induced through G␣ q . These results imply that, under our experimental conditions, the expression of the C3 toxin did not restrict the supply of PtdIns(4,5)P 2 necessary for PLC-mediated production of DAG. These results indicate that G␣ 13 leading to Rho activation is a potential signaling pathway that mediates PKD activation.
PKD activation in response to either G␣ 13 or RhoQL and pOnco-Lbc signaling is prevented by treatment with selective PKC inhibitors. PKD activation in response to G␣ 13 signaling is also prevented by mutation of Ser-744 and Ser-748 to Ala in the kinase activation loop of PKD. Furthermore, aluminum fluoride-stimulated G␣ 13 did not induce a further increase in PKD activity when Ser-744 and Ser-748 were mutated to Glu to mimic the phosphorylated residues. These data suggest that G␣ 13 signaling, like bombesin receptor activation, G␣ q , and PDB, leads to PKD activation through phosphorylation of Ser-744/748 in the activation loop of PKD through a PKC-dependent pathway.
Dominant-negative strategies to uncouple heptahelical receptors from their cognate G proteins have received much attention, but only recently has it been shown that expression of the COOH-terminal region of G proteins can competitively inhibit receptor-G protein interaction (71,73). For example, expression of the last 55 amino acids of G␣ q has been shown to target the receptor-G q interface in a selective manner and thereby block receptor-mediated PLC activation in cultured cells and in transgenic mice (71). In agreement with these recent results, we reported previously (19) that FIG. 8. Signal transduction pathways involved in PKD activation in response to bombesin stimulation. The scheme illustrates the molecular and pharmacological approaches used in this study. PKD activation in response to acute bombesin receptor signaling is mediated by both G␣ q and G␣ 13 through PKC. As shown in this pathway, PKD can be activated (3) by aluminum fluoride (AlF 4 Ϫ , stimulator of G 13 or G q ) or pOnco-Lbc (Lbc, active form of GEF) or RhoQL (constitutively active mutant of Rho); PKD activation could be blocked (-͉) by COOH-terminal fragments of G 13 (␣ 13 CT) or G q (␣ q CT) (competitive inhibitors of receptor G protein interaction), by C3 toxin (C3, specific inhibitor of Rho) or GF1 or Ro 31-8220 (Ro) (selective inhibitors of PKC). Mutation of Ser-744 and Ser-748 to Ala in the kinase activation loop of PKD prevents PKD activation through the G 12 and G q pathways. Rho kinase (ROCK) inhibitor, HA1077 (HA), and disruption of the actin cytoskeleton with cytochalasin D (CytD) do not interfere with PKD activation in response to bombesin. (Also see the text for details and abbreviations.) expression of a COOH-terminal fragment of G␣ q that acts in a dominant-negative fashion attenuated (but did not eliminate) PKD activation in response to bombesin. In the present study, we demonstrate, for the first time, that expression of the COOH-terminal region of G␣ 13 markedly reduces PKD activation in response to bombesin receptor stimulation. Furthermore, expression of C. botulinum C3 toxin also attenuated PKD activation in response to bombesin, implying that G␣ 13 /Rho signaling contributes to GPCR-induced PKD activation. Consistent with this notion, expression of the COOHterminal region of G␣ q together with the C. botulinum C3 toxin, which inactivates Rho, virtually abolished PKD activation in response to bombesin receptor stimulation. In contrast, expression of the COOH-terminal region of G␣ q together with the C3 toxin did not interfere with PKD activation in response to PDB which bypasses receptor pathways and directly activates classic and novel PKCs. These results support a model in which bombesin GPCR stimulation induces PKD activation through both G␣ q and G␣ 13 /Rho signaling pathways.
In conclusion, our results demonstrate that G␣ 13 contributes to PKD activation through a Rho-and PKC-dependent signaling pathway and indicate that PKD activation is mediated by both G␣ q and G␣ 13 in response to acute bombesin receptor stimulation, as summarized in the scheme shown in Fig. 8. The findings presented here identify PKD as a novel downstream target in G␣ 13 and Rho signaling.