Parallel Activation of Phosphatidylinositol 4-Kinase and Phospholipase C by the Extracellular Calcium-sensing Receptor*

The calcium-sensing receptor (CaR) is a G protein-coupled receptor that regulates physiological processes including Ca2+ metabolism, Na+, Cl−, K+, and H20 balance, and the growth of some epithelial cells through diverse signaling pathways. Although many effects of CaR are mediated by the heterotrimeric G proteins Gαq and Gαi, not all signaling pathways regulated by CaR have been identified. We used human embryonic kidney (HEK)-293 cells that stably express human CaR to study the regulation of inositol lipid metabolism by CaR. The nonfunctional mutant CaRR796W was used as a negative control. We found that CaR regulates phosphatidylinositol (PI) 4-kinase, the first step in inositol lipid biosynthesis. In cells pretreated with U73122 to inhibit phospholipase C activation and to block the degradation of PI 4,5-bisphosphate to form [3H]inositol trisphosphate (IP3), CaR stimulated the accumulation of [3H]PI monophosphate (PIP). Additionally, wortmannin, an inhibitor of both PI 3-kinase and type III PI 4-kinase, blocked CaR-stimulated accumulation of [3H]PIP and inhibited [3H]IP3 production. CaR-stimulated inositol lipid synthesis was attributable to PI 4-kinase and not PI 3-kinase because CaR did not activate Akt, a downstream target of PI 3-kinase. CaR associates with PI 4-kinase based on the findings that CaR and the 110-kDa PI 4-kinase β can be co-immunoprecipitated with antibodies against either CaR or PI 4-kinase. The PI-4 kinase in co-immunoprecipitates with anti-CaR antibody was activated in Ca2+-stimulated HEK-293 cells, which stably express the wild type CaR. Pertussis toxin did not affect the formation of [3H]IP3 or the rise in intracellular Ca2+ (Handlogten, M. E., Huang, C. F., Shiraishi, N., Awata, H., and Miller, R. T. (2001) J. Biol. Chem.276, 13941–13948). RGS4, an accelerator of GTPase activity of members of the Gαi and Gαq families, attenuated the CaR-stimulated PLC activation and IP3 accumulation, which is mediated by Gαq, but did not inhibit CaR-stimulated [3H]PIP formation. In HEK-293 cells, which express wild type CaR, Rho was enriched in immune complexes co-immunoprecipitated with the anti-CaR antibody. C3 toxin, an inhibitor of Rho, also inhibited the CaR-stimulated [3H]IP3production but did not lead to CaR-stimulated [3H]PIP formation, reflecting inhibition of PI 4-kinase. Taken together, our data demonstrate that CaR stimulates PI 4-kinase, the first step in inositol lipid biosynthesis conversion of PI to PI 4-P by Rho-dependent and Gαq- and Gαi-independent pathways.

The calcium-sensing receptor (CaR) 1 is a member of the G protein-coupled receptor superfamily. It has a characteristic seven transmembrane domain structure and contains a long extracellular amino terminus that is 19 -25% identical to the ␥-aminobutyric acid type B (GABA B ) and metabotropic glutamate receptors (1). CaR is expressed in a number of tissues, including parathyroid glands (2), kidney (3), and the nervous system (4) and is also expressed at the mRNA or protein levels in many other tissues and cells, including the large and small intestines (5), osteoblasts (6), keratinocytes (7), the hippocampus (8), renal tubular cells (9), human breast (10), endocrine and exocrine pancreas (11), and fibroblasts (12). In addition to Ca 2ϩ and Mg 2ϩ , ions whose metabolism is regulated by CaR, it is activated by a wide range of divalent cations, including some that are toxic such as Pb 2ϩ , Cd 2ϩ , and Fe 2ϩ (11,13). CaR inhibits parathyroid hormone secretion in response to increases in extracellular Ca 2ϩ (Ca 2ϩ o ) and also inhibits parathyroid cell proliferation. In the kidney, activation of CaR inhibits reabsorption of Na ϩ , Cl Ϫ , Mg 2ϩ , Ca 2ϩ , and H 2 O (14). The precise function of CaR in other tissues that are not directly involved in Ca 2ϩ metabolism is not fully understood, but it may participate in paracrine signaling (15). Loss and gain of function mutations lead to several human diseases, familial hypocalciuric hypercalcemia, neonatal severe hyperparathyroidism, and autosomal dominant hypocalcemia (16,17).
Despite work from many laboratories, the signaling pathways regulated by CaR are not fully defined. CaR can stimulate G␣ i subunits to inhibit the activity of adenylyl cyclase (18) and activate extracellular signal-regulated protein kinase (ERK) (19). It activates members of the G␣ q subfamily to stimulate phospholipase C (PLC) that hydrolyzes phosphatidylinositol 4,5-bisphosphate (PI 4,5-P 2 ) to generate diacylglycerol (DAG) and inositol trisphosphate (IP 3 ) or increase intracellular Ca 2ϩ (Ca 2ϩ i ) (20). CaR activates a number of intracellular enzymes and kinases, including Src kinase, phospholipase A 2 , PLC, and phospholipase D in parathyroid and human embryonic kidney * This work was supported by grants from the American Heart Association (to C. (HEK)-293 cells that stably express CaR (12,21). Activation of CaR also inhibits the apical 70-pS K ϩ channels in cells from the medullary thick ascending limb of Henle via a cytochrome P450-dependent metabolite of arachidonic acid (22). The molecular mechanism(s) of activation or inhibition of these signaling pathways and ion transport systems has not been fully defined. CaR inhibits cell proliferation, an activity that is uncharacteristic of G␣ i -and G␣ q -coupled receptors, suggesting that it may act via additional pathways (23,24). CaR is distributed in caveolin-rich plasma membrane domains (25) suggesting that it may interact with multiple signaling proteins, including some that may act independently of heterotrimeric G proteins.
In the last two decades, IP 3 and DAG, both products of hydrolysis of PI 4,5-P 2 by PLC have been defined as second messengers of fundamental importance. Recently individual polyphosphoinositides, originally thought to be simply metabolic intermediates or products, have been recognized as having many specific roles in cell biology (26). In the biosynthetic pathway leading from PI to PI 4,5-P 2 , PI 4-kinase carries out the first step resulting in the formation of PI 4-P. PI 4-P is then converted to PI 4,5-P 2 by PI 4-P 5-kinase (38). Besides acting as a substrate for PLC, PI 4,5-P 2 plays an important role in the regulation of cellular organelle trafficking (27), cytoskeleton rearrangement (28), phospholipase D (29), ion transporters (30), and G protein-gated inward rectifying K ϩ channels (31). In addition, PI 4,5-P 2 inhibits caspases (32) and gelsolin, a caspase substrate (33), to regulate apoptosis. PI 3-P, PI 3,4-P 2 , and PI 3,4,5-P 3 can stimulate protein kinase B (Akt), which regulates cell survival through the phosphorylation and inactivation of the Bcl-2 homolog Bad and caspase-9 (34,35). These lipids also activate other kinases with 3-phosphoinositide-binding pleckstrin homology (PH) domains, including Bruton's tyrosine kinase (Btk) and inducible T-cell kinase (Itk), which activate Src family kinases (36,37).
Although the breakdown of polyphosphoinositides by G protein-coupled receptor-activated PLC has been well established, little attention has been paid to regulation of the earlier steps in polyphosphoinositide synthesis. In the following studies, we investigated the role of CaR in activation of PI 4-kinase using HEK-293 cells that stably express the wild type CaR. We found that CaR activates PI 4-kinase independently of PLC, demonstrating regulation of the first step in polyphosphoinositide synthesis by a G protein-coupled receptor. Additionally, we found that CaR stimulated PI 4-kinase independently of G␣ i and G␣ q , by a mechanism involving Rho.  (20,39). Myc-C 3 toxin pEF was generously provided by Dr. Sternweis (U.T. Southwestern Medical Center). The rabbit polyclonal anti-PI 4-kinase ␤ and anti-Rho (-A, -B, -C) antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). The monoclonal antiphospho-Akt (4E2) and the polyclonal anti-phospho-ERK antibodies were obtained from New England Biolabs (Beverly, MA), and the polyclonal anti-c-Myc antibody (A-14) was supplied by Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Dynabeads protein A and Dynal MPC were purchased from Dynal (Lake Success, NY). The FuGENE 6 transfection reagent was from Roche Diagnostics Corp.

Materials-All
Cell Culture, Prelabeling, and Transfection-HEK-293 cell lines that express pcDNA3, CaR-HApcDNA3, or CaR R796W -HApcDNA3 were maintained in DMEM supplemented with 10% fetal bovine serum, 5 units/ml penicillin, 5 g/ml streptomycin, and 0.2 mg/ml G418 (13). Experiments were performed in multiple clones with similar results. Rat-1 cells were cultured in the same medium without G418. HEK-293 cells that stably express either the wild type CaR or the mutant CaR R796W were cultured in 12-well plates and prelabeled with 5-10 Ci/ml myo-[ 3 H]inositol in 0.5 ml of inositol-free DMEM containing 10% fetal bovine serum for 48 -50 h. In some experiments, after 24 h of prelabeling, the cells were transiently transfected with LacZpCMV5, Myc-C 3 toxin pEF, or MycRGS4pCMV5 using the FuGENE 6 transfection reagent for 24 h before experiments. Rat-1 cells were transfected with CaR-HApcDNA3 using the FuGENE 6 reagent. Twenty-four hours after transfection, the cultures were starved with serum-free DMEM for 24 h and then incubated at 37°C for 5 min in the presence or absence of either 100 ng/ml epidermal growth factor (EGF) or 5 mM CaCl 2 . In some experiments, cells were pretreated with 15 M wortmannin for 30 min and then incubated with agonist, and the samples were harvested for immunoblotting.

Measurement of [ 3 H]IP 3 and [ 3 H]PIP-
After equilibration for 30 min in inositol-free DMEM containing 20 mM Hepes (pH 7.4) and 20 mM LiCl in the presence or absence of different concentrations of U73122 or wortmannin, the prelabeled cells in 12-well plates were incubated at 37°C for 5 min in 0.4 ml of equilibration medium in the presence or absence of either 5 mM CaCl 2 or 150 M ATP. The reactions were terminated by adding 0.4 ml of 10% perchloric acid to each well. The plates were stored at 4°C for 3-5 h, the extracts were transferred to microcentrifuge tubes, neutralized by addition of ϳ0.32 ml of 1 M Hepes (pH 7.4) containing 1 M KOH and 1 mM EDTA, and then centrifuged at 12,000 rpm for 5 min. The supernatants were applied to AG1-X8 anion exchange columns, and tritiated inositol-containing compounds were separated as described earlier (39). The total cellular lipids in the pellets were extracted with chloroform, 1% HCl in methanol, and water (6:6:5.5, v/v/v). The chloroform phase was transferred to glass tubes and evaporated under a N 2 steam. The total cellular lipids were separated by TLC and identified by comigration with commercial standards in a solvent system containing chloroform, methanol, 20% methylamide, and 0.1 M ammonium hydroxide (40:28:5:2, v/v/v/v) (Ref. 40 and Fig.  2B). The standards were visualized with iodine vapor, and the areas corresponding to PI and PIP were scraped into scintillation vials and quantitated by liquid scintillation spectrometry. In some experiments in which LacZ, C 3 toxin, or RGS4 were transiently transfected into the cells that stably expressed the wild type CaR or mutant CaR R796W , the expression of the transiently expressed protein was documented by immunoblotting of the same samples from which the lipids were extracted. After transfer of the organic phase to glass tubes, the aqueous phase was carefully removed, the protein pellets were washed with acetone once, the pellet was dried with air, 40 -50 l of SDS-PAGE sample buffer were added, and the samples were adjusted to a neutral pH. The samples were boiled for 5 min, subjected to SDS-PAGE, and processed for immunoblotting with the antibodies indicated.
Immunoblotting-HEK-293 cells were homogenized in a buffer containing 20 mM Hepes, pH 8.0, 2 mM MgCl 2 , 1 mM EDTA, and protease inhibitors and centrifuged at 2000 rpm for 10 min. The postnuclear supernatant was centrifuged at 13,500 rpm for 60 min, and the resulting pellet was resuspended in the same buffer. The protein concentration was determined and adjusted to the same concentration by adding SDS-PAGE sample buffer. The samples were subjected to SDS-PAGE and processed for immunoblotting with the antibodies indicated and visualized with enhanced chemiluminescence.
Immunoprecipitation-The stable cells were treated with or without 5 mM CaCl 2 for 5 min and then lysed by buffer A containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% C 12 E 10 (lubrol), 1 mM EDTA, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, and 50 g/ml leupeptin, benzamidine, and aprotinin. The samples were transferred to microcentrifuge tubes and centrifuged at 13,500 rpm for 30 min. The protein concentration in cell lysates was determined. Equal amounts of protein from cell lysates were precipitated with antibody against CaR using Dynabeads protein A and Dynal MPC, according to the manufacturer's instructions. For PI-4 kinase immunoprecipitation, the crude membrane fraction described above was extracted on ice for 60 min in buffer A, and the extracts were centrifuged at 13,500 rpm for 60 min. The resultant supernatant was processed for immunoprecipitation using antibody against PI-4 kinase ␤. To attach each of the antibodies to the Dynabeads protein A, 200 l of Dynabeads protein A were transferred to microcentrifuge tubes, the tubes were placed in Dynal MPC for 2 min, and the fluid was removed. 600 l of monoclonal anti-CaR antibody was added to one tube, and 80 l of 0.1 M sodium phosphate buffer, pH 8.0 with 20 l of polyclonal anti-PI 4-kinase ␤ antibody was added to the other. Samples were resuspended with Dynabeads protein A at 4°C with slow rotation mixing for 30 min. To capture the antibody-loaded Dynabeads protein A, the tubes were placed in Dynal MPC to remove the fluid, and the complexes were rinsed once with 0.1 M sodium phosphate buffer and resuspended in 210 l of the same buffer. Each 50-l aliquot of antibody-loaded Dynabeads protein A was incubated with 1 mg of protein from the cell lysates or membrane extract for 30 min, washed twice to limit nonspecific binding, and the fluid removed by placing the samples in Dynal MPC and aspiration. SDS sample buffer was added to tubes, which were vortexed for 30 s and placed in Dynal MPC. The sample buffer was collected and cooked for 3 min, and then the samples were subjected to SDS-PAGE and immunoblotted using the antibodies indicated.

Natural [ 3 H]Inositol-PI Preparation and in Vitro PI 4-Kinase
Assay-Natural [ 3 H]inositol-PI was prepared by labeling HEK-293 cells at 2-days preconfluence with myo-[ 3 H]inositol (10 Ci/ml in 10% fetal bovine serum and inositol-free DMEM). Total cellular lipids were extracted, and the radiolabeled cellular lipids were separated by TLC. The [ 3 H]inositol-PI fraction was eluted from the silica gel as described previously (41). The [ 3 H]inositol-PI was Ͼ97% pure. The enzymes for the in vitro PI 4-kinase assay were obtained by co-immunoprecipitation using the anti-CaR antibody. The activity of PI 4-kinase was assayed essentially as described by Zhao et al. (42) but modified for the natural [ 3 H]inositol-PI substrate. Briefly, the 200-l reaction mixture contained 50 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 , 1 mM EGTA, 200,000 cpm of natural [ 3 H]inositol-PI, 0.4% Triton X-100, 0.5 mg/ml bovine serum albumin, 3 mM ATP, and the immunoprecipitated proteins. The reaction was started by the addition of the reaction mixture to the immunoprecipitation tubes, vortexing the samples, and incubating them at 30°C for 15 min. Reactions were then terminated by the addition of 0.5 ml of 1% HCl in methanol. The samples were extracted and separated by TLC, and the radioactivity was measured by liquid scintillation counting.

RESULTS
To investigate the regulation of the phosphoinositide cycle by CaR, we stably expressed the wild type CaR or the nonfunctional mutant CaR R796W in HEK-293 cells (Fig. 1A) and meas- Our recent studies showed no effect of pertussis toxin on CaR-stimulated IP 3

formation or the rise in Ca 2ϩ
i (20). This indicates that the regulation of CaR-coupled inositol lipid metabolism is G i -independent. However, the role of G␣ q in the stimulation of IP 3 formation by CaR was demonstrated using expression of RGS4 (a Regulator of G protein Signaling protein for G␣ q family members) to attenuate G␣ q activation (20).
PLC, which is activated by many signaling pathways, is a key enzyme in the catabolism of PI 4,5-P 2 to produce IP 3 and DAG. The level of PI 4,5-P 2 and consequently its availability as a substrate for PLC, depends on its rate of synthesis by PI 4-kinase and PI 4-P 5-kinase as well as its rate of degradation by PLC. To determine whether a signal originating from CaR regulates PI kinases and consequently the supply of PI 4,5-P 2 , we used U73122, a specific inhibitor of PLC, to block catabolism of PI 4,5-P 2 . The cells were pretreated with different concentrations of U73122 (0 -5 M) and then incubated with 5 mM Ca 2ϩ o for 5 min. Fig. 2A shows that U73122 inhibits [ 3 H]IP 3 formation in a dose-dependent manner, with nearly complete inhibition at 5 M, the maximum concentration. To demonstrate activation of PI kinase activity by CaR, the total cellular lipids from the same samples shown in Fig. 2A  Genetic and biochemical studies have led to the classification of PI kinases in three general families, PI 3-kinase (type I) and PI 4-kinase (types II and III). Wortmannin, a fungal metabolite, inhibits type III PI 4-kinase and PI 3-kinase but not type II PI 4-kinase (38,43). We used wortmannin to determine which PI kinases were activated by CaR. Preincubation of the cells with wortmannin resulted in nearly complete inhibition of CaR-stimulated [ 3 H]IP 3 formation (Fig. 3A). The usual series of enzymes that leads to IP 3 accumulation is PI 4-kinase and PI 4-P 5-kinase, which synthesize PI 4,5-P 2 from PI, and PLC, which hydrolyzes PI 4,5-P 2 . Using the same samples as in  (Fig. 3A).
In order to exclude the possibility that CaR activates PI 3-kinase, a possible source of PIP, we measured the activation of protein kinase B (Akt) a downstream effector of PI 3-kinase. The PI 3-P, PI 3,4-P 2 , and PI 3,4,5-P 3 produced by PI 3-kinase activate protein kinase B (Akt) and Bruton's family tyrosine kinases, which contain 3-phosphoinositide binding pleckstrin homology (PH) domains (34,35). HEK-293 cells express Akt at low levels, and activation of it by EGF, normally a potent stimulus, is difficult to detect. To circumvent this problem, we transiently expressed the wild type CaR in Rat-1 cells, and measured Akt activation in response to 100 ng/ml EGF and 5.0 mM Ca 2ϩ o . Fig. 4A shows that in Rat-1 cells EGF stimulates phosphorylation of Akt and that this effect is inhibited by pretreatment of the cells with 15 M wortmannin. The wild type CaR was expressed at similar levels in the different groups of transfected cells as shown in Fig. 4B. Fig. 4C shows that 100 ng/ml EGF (5 min) stimulates phosphorylation of both Akt and ERK, while 5 mM Ca 2ϩ (5 min) stimulates phospho-rylation of ERK, but not Akt. Activation of ERK by CaR indicates that it is expressed and functional. These results demonstrate that activation of CaR does not result in activation of PI

FIG. 2. Effect of U73122 on Ca 2؉ o -induced [ 3 H]IP 3 and [ 3 H]PIP formation in HEK-293 cells. HEK-293 cells that express wild type CaR (WT) or mutant CaR R796W (Mut) were prelabeled with [ 3 H]inositol
for 48 h, were pretreated with different concentration of U73122 for 30 min to inhibit PLC activation, and were then incubated in the presence of 5 mM CaCl 2 at 37°C for 5 min.

FIG. 4. Effect of EGF and Ca 2؉
o on PI 3-kinase activity as measured by the level of phospho-Akt in Rat-1 cells. Rat-1 cells were serum-starved for 24 h and incubated in the presence or absence of 15 M wortmannin (Wort) for 30 min and then treated with or without 100 ng/ml EGF for 5 min. The samples were harvested, the protein concentration was measured, and equal amounts of cellular protein were processed for immunoblotting using an anti-phospho-Akt antibody (A). Rat-1 cells were transiently transfected with the wild type CaR. After 24 h of transfection, the cells were serum-starved for 24 h and incubated in the presence or absence of 100 ng/ml EGF or 5 mM CaCl 2 for 5 min. The samples were harvested, the protein concentration was determined, and the equal amounts of cellular protein were processed for immunoblotting using the anti-CaR antibody (B) and the antiphospho-Akt antibody or anti-phospho-ERK antibody (C). Two separate experiments yielded similar results.
3-kinase but indicate that it does activate the wortmanninsensitive type III PI 4-kinase.
The preceding results indicate that CaR activates PI 4-kinase. To identify which form of PI 4-kinase is expressed in our cell system, crude membrane and cytosol fractions were prepared from HEK-293 cells and processed for immunoblotting using an anti-PI 4-kinase ␤ antibody. A single 110-kDa band representing PI 4-kinase ␤ was detected in both crude membranes and cytosol but was predominantly expressed in the cytosolic fractions (Fig. 5A). To determine whether CaR and PI 4-kinase associate, we co-immunoprecipitated these proteins from cell lysates (using the anti-CaR antibody) or crude membrane extracts (using the anti-PI 4-kinase antibody) of HEK-293 cells that express either the wild type CaR (WT) or control (V) G418-resistant cells. Reciprocal co-immunoprecipitation using the anti-PI 4-kinase ␤ and anti-CaR antibodies was performed, and the samples were immunoblotted with either the anti-CaR or anti-PI 4-kinase antibodies (Fig. 5, B and C). In Fig. 5B, a band representing PI 4-kinase ␤ was present in the cells that express CaR as well as the control cells as would be expected. In the right panel, a single band of ϳ130 kDa representing CaR was co-immunoprecipitated by the anti-PI 4-kinase ␤ antibody. Similarly, in Fig. 5C, where the immunoprecipitation was performed with the anti-CaR antibody, CaR was found in the cells that express it (left panel). In the right panel, PI 4-kinase ␤ was strongly present in the cells that express CaR. The mutant CaR R796W also appears to co-immunoprecipitate with PI 4-kinase but less reliably than the wild type CaR. Additionally, two bands corresponding to Rho (24 and 29 kDa) were also co-immunoprecipitated with CaR (Fig. 5D). To directly verify that CaR stimulates PI 4-kinase ␤, we performed in vitro assays of PI 4-kinase activity in the immune complexes immunoprecipitated by the anti-CaR antibody. Using natural [ 3 H]inositol-PI as a substrate, PI 4-kinase activity was assayed directly for the production of [ 3 H]PIP by thin layer chromatography. As shown in Fig. 6, [ 3 H]PIP formation was significantly increased in the immunoprecipitates from the cells that expressed wild type CaR and that were treated with 5 mM Ca 2ϩ o . These results demonstrate that immune complexes precipitated by the anti-CaR antibody contain PI 4-kinase activity and that CaR stimulates PI 4-kinase. Rho, a potential PI 4-kinase activator, is also found in the complex immunoprecipitated by the anti-CaR antibody. These results indicate that CaR, Rho, and PI 4-kinase are present in the same microdomain of the cell membrane and that they interact, possibly directly.
Based on our whole cell (Figs. 2 and 3) and in vitro data (Figs. 5 and 6), PI 4-kinase is a downstream effector of CaR. However, the mechanism by which CaR activates PI 4-kinase is not defined. To test for roles for G␣ i , G␣ q , and Rho, we transiently expressed RGS4 and C 3 toxin in HEK-293 cells that stably express the wild type CaR or mutant CaR R796W and that had been prelabeled with myo-[ 3 H]inositol. C 3 toxin specifically ADP-ribosylates and inactivates Rho, but not Rac or Cdc42, and RGS4 attenuates G␣ i and G␣ q activation blocking PLC activation. Expression of RGS4 was documented by immunoblotting (Fig. 7A), but the signal for C 3 toxin was much weaker (data not shown). The cells were incubated in the presence or absence of 5 mM CaCl 2 for 5 min, and [ 3 H]IP 3 release and [ 3 H]PIP formation were measured. Fig. 7B shows that both C 3 toxin and RGS4 inhibited ϳ30% of the CaR-stimulated [ 3 H]IP 3 release. This degree of inhibition is consistent with the transient transfection efficiency in our system. Inhibition of IP 3 release is indicative of inhibition of PLC activation or inhibition of an earlier step in inositol lipid biosynthesis, PI 4 kinase or PI 4-P 5-kinase. CaR stimulates PLC via G␣ q , so expression of RGS4 leads to a reduction in [ 3 H]IP 3 release and a CaR-dependent increase in [ 3 H]PIP because its catabolic metabolism is blocked. Expression of C 3 toxin leads to reduced [ 3 H]IP 3 production but not accumulation of [ 3 H]PIP. C 3 toxin has no effect on PLC activation but can reduce PLC substrate supply (44) at earlier steps, the synthesis of PIP and PIP 2 , which are catalyzed by PI 4-kinase and PIP 5-kinase. Additionally, if C 3 toxin inhibited only PIP 5-kinase, accumulation of PIP would have been observed. However, PIP did not accumulate, so C 3 toxin must have inhibited PI 4-kinase, the enzyme that con- verts PI to PIP. Work by others (45) has shown that Rho regulates PI 4-kinase, so it is most likely that CaR activates PI 4-kinase via Rho. These results demonstrate that CaR stimulates inositol lipid metabolism at a minimum of two sequential steps, activation of PI 4-kinase by Rho and activation of PLC by G␣ q .
To determine whether this regulation of sequential steps in PI metabolism is a general phenomenon for G q -coupled receptors, we chose ATP, an agonist for purinergic receptors, to stimulate PLC and PI 4-kinase activities in HEK-293 cells. The cells, prelabeled with myo-[ 3 H]inositol, were pretreated with or without 5 mM U73122 and then incubated with 150 M ATP or 5 mM CaCl 2 . The results in Fig. 8 show a similar pattern of [ 3 H]IP 3 release and [ 3 H]PIP formation as was found with CaR.

DISCUSSION
The importance of inositol lipids in cell signaling became clear with the discovery that PLC leads to the breakdown of PI 4,5-P 2 and the generation of two important second messengers, IP 3 and DAG. IP 3 binds to IP 3 receptors on the endoplasmic reticulum to release Ca 2ϩ i , and DAG is a physiologic activator of protein kinase C. In addition, PI 4,5-P 2 has recently been found to play important and direct roles in the regulation of the activity of phospholipase D (29), ion transporters (30), ion channels (31), vesicle trafficking (27), and the state of the actin cytoskeleton (28). Additional studies have shown that the family of inositol lipid molecules has effects on cell proliferation and apoptosis (26,32,33). A recent report indicated that PI 3-kinase and PI 4-kinase, the enzymes involved in inositol lipid synthesis, may be involved in Alzheimer's disease because the specific activities of these enzymes were reduced by 43-59% in brains of patients with Alzheimer's disease (46).
The levels of the different inositol lipids in cells are controlled at the level of synthesis by specific lipid kinases and at the level of hydrolysis by PLC. The synthesis of PI 4,5-P 2 begins with the conversion of PI to PI 4-P by PI 4-kinase and then PI 4-P to PI 4,5-P 2 by PI 4-P 5-kinase. PI 4-P 5-kinase is regulated by Rho kinase (47), but little information is available about the regulation of PI 4-kinase by extracellular signals. EGF stimulates PI 4-kinase activity by a mechanism that involves phosphorylation of PI 4-kinase and the formation of a multiprotein complex including the EGF receptor, a 55-kDa wortmannininsensitive PI 4-kinase, possibly PI 4-P 5-kinase and PLC (48). Regulation of this first committed step in inositol lipid synthesis, PI 4-kinase activity, by G protein-coupled receptors has not been described, although this enzyme activity is required for sustained angiotensin II-stimulated IP 3 production (49).
Recently, we demonstrated that CaR-induced formation of IP 3 and the resultant rise of Ca 2ϩ i are not affected by pertussis toxin (20). HEK-293 cells treated with U73122 (a specific inhibitor of PLC or expressed RGS4 that attenuates G␣ q -stimulated PLC) activation of the CaR-stimulated PIP accumulation occurs, demonstrating activation of PI 4-kinase, PI 4-P 5-kinase, or both by CaR. The lack of effect of pertussis toxin and RGS4 on CaR-stimulated PIP accumulation indicates that activation of PI 4-kinase is G␣ i -and G␣ q -independent. Addition- ally, because pertussis toxin did not affect CaR-stimulated IP 3 production, it cannot inhibit PI 4-kinase or PIP 5-kinase activities (20). PI 4-kinase interacts with CaR, based on the findings that CaR and the 110-kDa PI 4-kinase ␤ were co-immunoprecipitated reciprocally, and because immune complexes precipitated with the anti-CaR antibody contained activated PI 4-kinase when the cells were pretreated with Ca 2ϩ o . The fact that Rho was enriched in the immunoprecipitates of CaR-expressing cells indicates that Rho may participate in CaR-stimulated PI 4-kinase activity. C 3 toxin, which inactivates Rho proteins by ADP-ribosylation, inhibits CaR-stimulated IP 3 release by reducing PLC substrate supply (44) but not PLC activity, because it does not result in an increase in PIP synthesis. Rho must stimulate this signal cascade at earlier steps such as conversion of PI to PI 4-P by PI 4-kinase or conversion of PI 4-P to PI 4,5-P 2 by PIP 5-kinase. If C 3 toxin blocked PIP 5-kinase and not PI 4-kinase, expression of C 3 toxin would have resulted in an increase in PIP accumulation. However, this was not observed (Fig 7C). Therefore, C 3 toxin must inhibit PI 4-kinase activity. In our system, the blockade of PI 4-P synthesis by C 3 toxin demonstrates CaR-activated PI 4-kinase via the Rho-dependent pathway.

FIG. 7. The influence of C 3 toxin and RGS4 on [ 3 H]IP 3 and [ 3 H]PIP formation in Ca 2؉
CaR stimulated the synthesis of two different forms of PIP through either PI 3-kinase or PI 4-kinase. The products of these two enzymes differ in the position of phosphate on the inositol ring and are difficult to separate by thin layer chromatography. PI 4-P, a product of PI 4-kinase, can be converted to PI 4,5-P 2 by PI 4-P 5-kinase, which also converts PI 3-P, a product of PI 3-kinase, to PI 3,5-P 2 . Serunian et al. (50) reported that PLC from many different sources hydrolyzes only PI 4,5-P 2 , the product of PI 4-P but not PI 3,4-P 2 , PI 3,5-P 2 , or PI 3,4,5-P 3 , the products of PI 3-P. Consequently, because CaR stimulates the production of IP 3 , the PIP we measured must be derived from PI 4-kinase and must be PI 4-P. To confirm that PI 4-kinase and not PI 3-kinase was activated by CaR, we transiently transfected the wild type CaR into Rat-1 cells where PI 3-kinase and its signaling cascade could be studied. The ability of EGF to activate Akt coupled with the failure of CaR to do so, demonstrates that CaR stimulates PI 4-kinase and not PI 3-kinase.
Mammalian PI 4-kinase enzymes are classified as type II or III depending on their sensitivity to inhibition by wortmannin. Type II PI 4-kinase is a 50 -56-kDa protein, is tightly membrane-bound, and is not inhibited by wortmannin. The wortmannin-sensitive Type III PI 4-kinase activity is attributed to two proteins, a 110-kDa ␤-form and a 230-kDa ␣-form, both of which are located in the endoplasmic reticulum, Golgi, and cytosol. The type II and type III PI 4-kinases are all involved in agonist-induced polyphosphoinositide synthesis in mammalian cells (38,43). We used several approaches to determine which type or form of the enzyme is activated by CaR. First, wortmannin inhibited the formation [ 3

H]IP 3 and [ 3 H]PIP in Ca 2ϩ
ostimulated cells that expressed the wild type CaR, indicating that CaR stimulates a wortmannin-sensitive PI 4-kinase (Fig.  3B). Second, our cells express a 110-kDa protein that is recognized by an anti-PI 4-kinase ␤ antibody in the appropriate cell fractions (Fig. 5A). Reciprocal co-immunoprecipitation of CaR and the 110-kDa PI 4-kinase ␤ demonstrates that these two signaling proteins are co-localized in the cell and further indicates that they interact, possibly directly (Fig. 5, B and C). Finally, the immune complex assay using an antibody against CaR demonstrated that the PI 4-kinase activity associates with CaR activation (Fig. 6). Taken together, our data demonstrate that CaR stimulates PI 4-kinase ␤ in mammalian cells.
Nakanishi et al. (49) reported that PI 4-kinase activity was required for sustained angiotensin II -stimulated IP 3 produc-tion but did not specifically demonstrate stimulation of its activity by angiotensin II. Using wortmannin or LY-294002, several other groups showed that PI 4-kinase is necessary for agonist-induced muscarinic cholinergic receptor endocytosis in SH-SY5Y neuroblastoma cells (51), for agonist-stimulated Ca 2ϩ extrusion in human platelets (52), and for selecting apical delivery vesicles in polarized MDCK cells (53). Gasman et al. (45) determined that G␣ o -mediated exocytosis was associated with the RhoA-and PI 4-kinase-dependent pathway. In HEK-293 cells, Rho kinase stimulates PI 4-P 5-kinase as demonstrated by an increase or decrease of cellular PI 4,5-P 2 levels with transfection of RhoA, Rho kinase, or C 3 toxin treatment (47).
Although these data suggest that RhoA participates in the regulation of inositol lipid metabolism, its precise role and the extracellular signals that modulate its activity are not defined. Our results demonstrate that the activities of PI 4-kinase and PLC are regulated in parallel by distinct mechanisms but that the signal originates at CaR (Fig. 9). PI 4-kinase is activated by a Rho-dependent mechanism that is independent of G␣ i and G␣ q , and PLC is regulated by G␣ q (20). The dual regulation is also observed in the regulation of IP 3 release and PIP formation by ATP receptors and may also be found in many other signaling systems that are regulated by G␣ q -coupled receptors.
The nonfunctional mutant CaR R796W , which contains a mutation in the third cytoplasmic loop, a receptor region that characteristically contacts G proteins, does not activate PI 4-kinase. This same mutant receptor also fails to activate other signaling pathways that presumably involve G proteins including activation of ERK, cytosolic phospholipase A 2 , and PLC (13,19,54). The third cytoplasmic loop may have functions in addition to activating heterotrimeric G proteins, because this region of the D 2 dopamine receptor interacts with the cytoskeletal protein, filamin (55).
Our results demonstrate that the levels of inositol lipids are regulated at multiple steps by CaR, a G protein-coupled receptor. CaR stimulates PI 4-kinase, the first committed step in inositol lipid biosynthesis via Rho-dependent and G␣ i -and G␣ q -independent mechanisms, and it stimulates PLC via a G␣ q -dependent mechanism. Regulation of the sequential steps in inositol lipid metabolism by different mechanisms could be important because the intermediate products have biological activity. In addition to serving as a substrate for PLC to produce IP 3 and DAG, PIP 2 , the product of PI 4-kinase and PI 4-P 5-kinase, has many roles in the cell including regulation of ion channels, transporters, and the state of the actin cytoskeleton (27)(28)(29)(30)(31). Situations could arise where an increase in PIP 2 is required without the production of IP 3 and DAG. Alternatively, if PLC is activated to produce IP 3 and DAG, the increased supply of PIP 2 might be coordinated with activation of PLC to avoid levels of PIP 2 in excess of what could be metabolized by PLC. The production of all of these lipids may be coordinated through a multiprotein complex, which is composed of CaR and PI 4-kinase, and may also include other PI kinases such as Rho, Rho kinase, and phospholipases.