Stimulation of phosphatidylinositol-4-phosphate 5-kinase by Rho-kinase.

The serine/threonine kinase Rho-kinase was recently identified as a downstream effector of the small GTPase Rho, mediating effects of Rho on the actin cytoskeleton. Also phosphatidylinositol 4,5-bisphosphate (PI(4,5)P(2)) has been implicated in the regulation of actin polymerization. As the synthesis of PI(4,5)P(2) has been suggested to be affected by Rho proteins, we investigated whether Rho-kinase is involved in the control of PI(4,5)P(2) levels. Overexpression of RhoA in HEK-293 cells increased phosphatidylinositol 4-phosphate (PI4P) 5-kinase activity and concomitantly enhanced cellular PI(4,5)P(2) levels, whereas overexpression of the Rho-inactivating C3 transferase decreased both PI4P 5-kinase activity and PI(4,5)P(2) levels. These effects of RhoA could be mimicked by overexpression of wild-type Rho-kinase and of the constitutively active catalytic domain of Rho-kinase, Rho-kinase-CAT. In contrast, a kinase-deficient mutant of Rho-kinase had no effect on PI4P 5-kinase activity. Importantly, the increase in PI4P 5-kinase activity and PI(4,5)P(2) levels by wild-type Rho-kinase, but not by Rho-kinase-CAT, was completely prevented by coexpression of C3 transferase, indicating that the effect of Rho-kinase was under the control of endogenous Rho. In cell lysates, addition of recombinant RhoA and Rho-kinase-CAT stimulated PI4P 5-kinase activity. Finally, the increase in PI(4,5)P(2) levels induced by both Rho-kinase-CAT and RhoA was reversed by the Rho-kinase inhibitor HA-1077. Our data suggest that Rho-kinase is involved in the Rho-controlled synthesis of PI(4,5)P(2) by PI4P 5-kinase.

Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P 2 ) 1 is a versatile phospholipid with pleiotropic cellular functions (1). PI(4,5)P 2 is not only a substrate for phospholipase C (PLC) and phosphatidylinositol 3-kinases (PI 3-kinases), it also serves as a membrane docking site for pleckstrin homology-containing proteins (2). Furthermore, PI(4,5)P 2 has been reported to interact with several proteins/enzymes involved in vesicle trafficking, including phospholipase D (PLD) and ADP-ribosylation factor (3,4). The binding and displacement of various actin regulatory proteins from actin filaments by PI(4,5)P 2 allows the polymerization of these filaments, indicating an important role for PI(4,5)P 2 in the organization of the actin cytoskeleton (5). Also small GTPases of the Rho family have been implicated in the regulation of the cytoskeleton. Thus, Rho, Rac, and Cdc42 exert distinct effects on cytoskeletal arrangement, thereby controlling cell morphology and essential physiological functions as adhesion, cell division, cell motility, and cytokinesis (6,7). Evidence has been gathered that Rho family GTPases and PI(4,5)P 2 may be connected in common signaling pathways, and that the major PI(4,5)P 2 -generating enzyme, phosphatidylinositol-4-phosphate 5-kinase (PI4P 5-kinase) is a downstream target of Rho proteins. In mouse fibroblast lysates, addition of recombinant RhoA stimulated PI4P 5-kinase activity (8), and a physical association between RhoA and PI4P 5-kinase was observed (9). Tolias et al. (10) demonstrated an association between PI4P 5-kinase and Rac1, but not with RhoA. In permeabilized platelets, a constitutively actice Rac mutant was able to uncap actin filaments by a PI(4,5)P 2 -dependent mechanism and to induce PI(4,5)P 2 synthesis (11). Recently, we have shown that treatment of cells with toxin B from Clostridium difficile, that specifically inactivates Rho family GTPases (12), resulted in the reduction of cellular PI(4,5)P 2 levels by 50 -90%, accompanied by profound changes in cell morphology (13,14).
Several serine/threonine kinases have recently been identified as functional effectors of Rho, including protein kinase N (15,16), citron kinases (17,18), and the family of Rho-associated kinases (19 -22). Rho-associated kinase (Rho-kinase; also designated as ROK␣ and ROCK-II) was shown to mimic the effects of RhoA on the cytoskeleton by inducing the formation of actin stress fibers and focal adhesions (23,24). Rho-kinase activated myosin, and hence stimulated the assembly of myosin-actin filaments, both by directly phosphorylating the myosin light chain itself (25) and by phosphorylating the myosinbinding subunit of myosin phosphatase (26). Also the NHE1 Na ϩ /H ϩ exchanger, another downstream target of RhoA linked with cytoskeletal rearrangement, was phosphorylated and activated by Rho-kinase (27). Recently, we observed a role for Rho-kinase in phospholipid signaling, as Rho-kinase potentiated the stimulation of PLD by the m3 muscarinic acetylcholine receptor in HEK-293 cells (28).
As Rho-kinase is mediating signals from Rho to the actin cytoskeleton and as the formation of the actin assembly inducing phospholipid PI(4,5)P 2 has been shown to be regulated by Rho family proteins, we examined whether Rho-kinase can modulate the synthesis of PI(4,5)P 2 . Here we show that RhoA stimulated PI4P 5-kinase activity, resulting in enhanced cellular PI(4,5)P 2 levels in HEK-293 cells. Furthermore, Rho-kinase was found to regulate PI4P 5-kinase in a Rho-dependent manner. These results implicate a role for Rho-kinase in the Rhocontrolled synthesis of PI(4,5)P 2 and identify a further contribution of Rho-kinase to phospholipid metabolism.
Cell Culture and Transfection-HEK-293 cells were routinely passaged in Dulbecco's modified Eagle's medium/F-12 medium with 10% fetal calf serum (29). For transfection studies, DNA encoding human RhoA was subcloned into the pRK5 expression vector. Myc-tagged C3 transferase in pEF (30) was kindly donated by Dr. A. Hall. The cDNAs encoding Myc-tagged wild-type Rho-kinase, the constitutively active catalytic domain of Rho-kinase (Rho-kinase-CAT; amino acids 6 -553), and the kinase-deficient mutant of Rho-kinase-CAT (Rho-kinase-CAT-KD; K121G) were subcloned into pEF (20,23,25). Hemagglutinintagged PI4P 5-kinase ␤ (31), generously provided by Drs. H. Ishihara and Y. Oka, was cloned into the pCMV vector. Subconfluent monolayers in 145-mm culture dishes were transfected with the indicated amounts of DNA using the calcium phosphate method (32). Transfection efficiency of the HEK-293 cells was determined by in situ screening of ␤-galactosidase-positive cells in monolayers co-transfected with constitutively active pSV␤-gal (Promega), and appeared to range between 50 and 80%. Assays were performed 48 or 72 h after transfection.
Preparation of Recombinant RhoA and Rho-kinase-CAT-The cDNA encoding human RhoA was cloned into pAcGHLT baculovirus transfer vector (Pharmingen), and the cDNA encoding the catalytic fragment of Rho-kinase was inserted into the pAcGLT vector (25). Sf9 cells were infected for 48 h at 25°C. Thereafter, the cells were lysed by sonication in 10 mM Tris/HCl, pH 7.5, 50 mM NaCl, 10 mM MgCl 2 , 1 mM dithiothreitol, and 10 M phenylmethylsulfonyl fluoride. The lysates were cleared by centrifugation for 1 h at 20,000 ϫ g, and the supernatants, containing GST-RhoA or GST-Rho-kinase-CAT, were incubated with glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) for 30 min at 4°C. After washing the beads three times, RhoA or Rho-kinase-CAT proteins were released from the bound glutathione S-transferase fusion constructs by overnight incubation with thrombin (10 units) in 50 mM Tris/HCl, pH 8.0, 150 mM NaCl, 5 mM MgCl 2 , and 1 mM dithiothreitol. The beads were removed by centrifugation, and thrombin was inactivated by treatment with p-aminobenzamidine beads. Protein concentrations were determined with the Bradford method. The recombinant protein preparations were analyzed for their homogeneity by SDS-polyacrylamide gel electrophoresis and Coomassie staining, and were found to contain essentially only one band.
Analysis of Cellular Phosphoinositides-For determination of phosphoinositide levels, cells were replated in poly-L-lysine-treated 60-mm culture dishes 24 h after transfection, and labeled with myo-[ 3 H]inositol (0.5 Ci/ml) in inositol-free medium for 24 h. The monolayers were washed once with phosphate-buffered saline and scraped in 1.5 ml of cold 2.4 N HCl. Phase separation was performed by addition of 5 ml of chloroform/methanol/concentrated HCl (200:100:0.75, v/v) and centrifugation for 10 min at 2,000 ϫ g. The lipid phase was vacuum dried, resuspended in chloroform, and spotted on Silica Gel 60 TLC plates (Merck), impregnated with 1.2% potassium oxalate. The plates were chromatographed for 2 h in chloroform, methanol, 2.5 N ammonium hydroxide (9:7:2, v/v). Authentic phosphoinositide standards were visualized by staining with iodine vapor, and the area corresponding to PI(4,5)P 2 was scraped into vials, and the radioactivity measured by liquid scintillation counting. The amount of PI(4,5)P 2 was normalized for protein content.
Determination of PI(4,5)P 2 Masses-Unlabeled cells were replated in poly-L-lysine-treated 35-mm culture dishes 24 h after transfection, and cultured for another 24 h. Monolayers were extracted with 0.5 M trichloroacetic acid to remove soluble inositol phosphates and the remaining membrane lipids were assayed for PI(4,5)P 2 as described by Chilvers et al. (33). Briefly, lipids were extracted with acidified chloroform/ methanol and hydrolyzed with 1 M KOH at 100°C for 15 min. Samples were desalted and neutralized by filtration through Dowex-50 columns (Bio-Rad), and Ins(1,4,5)P 3 in the eluate was measured by a radioreceptor assay, based on the bovine adrenal cortical Ins(1,4,5)P 3 -binding protein (33).
Samples were mixed with 5 l of a sonicated solution of phospholipid (70 M PI4P and 35 M phosphatidylserine, final concentrations) and preincubated for 15 min at 25°C with additional compounds as indicated. Reactions were started by the addition of 20 M [␥-32 P]ATP (1 Ci/assay). After 5 min (or the time indicated in the figure legend), the reactions were stopped by adding 0.3 ml of methanol, 1 N HCl (1:1, v/v) and extracted with 0.25 ml of chloroform. The organic layer was dried, resuspended in chloroform, and chromatographed on oxalate-pretreated Silica Gel 60 plates as described above. The reaction products were visualized by autoradiography and identified by comparison with unlabeled standards. The appropriate regions of the TLC plates were scraped, and the radioactivity quantitated by liquid scintillation counting.
High Pressure Liquid Chromatography Analysis-Extracted lipid products of the PI4P kinase assay were deacylated as described (34), using monomethylamine instead of ethanolamine. The products were mixed with glycerophosphoinositol phosphate standards, and analyzed by HPLC on a Mini Q PC 3.2/3 column (Amersham Pharmacia Biotech) with on-line metal-dye-detection (MDD-HPLC) as described (35).

RhoA Increases PI(4,5)P 2 Levels in HEK-293 Cells-
We previously demonstrated that intoxication of HEK-293 cells with toxin B from C. difficile resulted in a decrease of cellular PI(4,5)P 2 levels, indicating that members of the Rho GTPase family are positively involved in the regulation of PI(4,5)P 2 synthesis (13,36). In the present study, we examined whether Rho itself could mediate a change in PI(4,5)P 2 levels. Transfection of HEK-293 cells with RhoA resulted in an increase in cellular PI(4,5)P 2 levels (Fig. 1A). In cells transfected with 100 g of RhoA DNA, the increase was about 50% compared with mock-transfected cells. On the other hand, transfection of cells with C3 transferase, that specifically ADP-ribosylates and thereby inactivates Rho, but not Rac or Cdc42, resulted in a decrease of cellular PI(4,5)P 2 levels, by about 40% at 100 g of C3 DNA (Fig. 1B). No differences in total phosphoinositide labeling was observed in cells transfected with either RhoA or C3 transferase (data not shown). These data indicated that RhoA positively regulated the cellular pool of PI(4,5)P 2 .
RhoA Stimulates PI4P 5-Kinase Activity in HEK-293 Cells-The major cellular pathway for the generation of PI(4,5)P 2 is the phosphorylation of PI4P on the D-5 position of the inositol ring by PI4P 5-kinase (1). We next examined whether RhoA enhanced PI(4,5)P 2 levels in HEK-293 cells by stimulating the activity of PI4P 5-kinase. Lysates of cells transfected with either RhoA or C3 transferase were assayed for in vitro PI4P 5-kinase activity, using PI4P and [␥-32 P]ATP as substrates. PI4P 5-kinase activity was clearly elevated in lysates from RhoA-overexpressing cells ( Fig. 2A). The stimulatory effect of RhoA was even more pronounced when the lysates were preincubated with GTP␥S. In contrast, in lysates from C3 transferase-expressing cells, [ 32 P]PI(4,5)P 2 formation was reduced by about 50% (Fig. 2B). Furthermore, the stimulation of PI4P 5-kinase activity by GTP␥S was abolished in lysates from C3expressing cells. To substantiate the stimulating effect of RhoA on PI4P 5-kinase activity, we performed the in vitro kinase assay in the presence of purified recombinant RhoA. RhoA concentration dependence stimulated PI4P 5-kinase activity up to 2-fold (Fig. 3A). The formation of [ 32 P]PI(4,5)P 2 appeared to be linear for at least 20 min, both in the absence and presence of RhoA (Fig. 3B). The stimulatory effect of recombinant RhoA was judged to be specific, as a non-relevant protein produced and purified by identical procedures had no effect on PI4P 5-kinase activity.
All in vitro phosphorylation reactions were performed in the presence of 0.1% Nonidet P-40 that completely prevented the generation of the interfering product PI(3,4)P 2 by PI 3-kinases. Thus, using PI(4,5)P 2 as a substrate, no PI 3-kinase activity could be detected under these assay conditions and the RhoAstimulated production of phosphatidylinositol bisphosphate from PI4P was insensitive to the PI 3-kinase inhibitor LY294002 (data not shown). To obtain final proof, the phosphatidylinositol bisphosphate reaction product generated in the presence of RhoA was deacylated and identified by HPLC analysis with on-line metal dye-detection. The product proved to be exclusively PI(4,5)P 2 based on comigration with a glycerophosphoinositol(4,5)P 2 standard (Fig. 4).
Rho-kinase Is a Positive Effector of PI4P 5-Kinase in HEK-293 Cells-The serine/threonine kinase Rho-kinase has been recognized as an important downstream effector of RhoA. Recently, we provided evidence for the implication of Rho-kinase in the Rho-controlled PLD stimulation by the m3 muscarinic acetylcholine receptor (28). The newly identified involvement of Rho-kinase in phospholipid signaling prompted us to investigate whether Rho-kinase also takes part in the Rho-controlled stimulation of PI4P 5-kinase. Cells were transfected with different Rho-kinase constructs, and after 48 h the cells were lysed and assayed for PI4P 5-kinase activity. Overexpression of the wild-type Rho-kinase resulted in a clear increase in PI4P 5-kinase activity (Fig. 5). Overexpression of the constitutively active Rho-kinase-CAT was even more effective. Transfection of HEK-293 cells with 25 and 50 g of Rho-kinase-CAT DNA increased the PI4P 5-kinase activity by 43 and 64%, respectively (Fig. 5). No stimulating effect on PI4P 5-kinase activity was observed in lysates from cells that had been transfected with the kinase-deficient mutant, Rho-kinase-CAT-KD. To investigate whether the stimulatory effect of Rho-kinase on PI4P 5-kinase activity was under the control of Rho, cells were co-transfected with Rho-kinase constructs and C3 transferase. The stimulatory effect of wild-type Rho-kinase could be completely abolished by coexpression of C3 transferase, indicating that the transfected Rho-kinase was under the control of endogenous Rho (Fig. 6A, left panel). As Rho-kinase-CAT DNA encodes the constitutively active catalytic domain of Rho-kinase, it was anticipated that the stimulatory effect of this construct on PI4P 5-kinase was independent of endogenous Rho. Indeed, coexpression of C3 transferase did not prevent the increased [ 32 P]PI(4,5)P 2 production induced by Rho-kinase-CAT (Fig. 6A, right panel).
As commercial PI4P might be contaminated with minor amounts of PI5P, the Rho-kinase-driven generation of PI(4,5)P 2 could theoretically be mediated by a 4-kinase rather than a 5-kinase. To exclude this possibility, cells were cotransfected with the recently cloned PI4P 5-kinase ␤ (31). In cells expressing the PI4P 5-kinase isoform, stimulation of PI(4,5)P 2 generation by the Rho-kinases was dramatically (10 -50-fold over basal) potentiated (Fig. 6B). Again, the stimulating effect of wild-type Rho-kinase, but not of Rho-kinase-CAT, was largely prevented by C3 transferase. These data consolidate the identification of PI4P 5-kinase as the responsible target of Rho-kinase.
The stimulation of PI4P 5-kinase by Rho-kinase was further studied using recombinant Rho-kinase-CAT. The functionality of the purified kinase was confirmed by its ability to phosphorylate myosin light chain in a cell-free system (data not shown). Addition of purified Rho-kinase-CAT potently and concentration dependence increased the PI4P 5-kinase activity in cell lysates (Fig. 7). The stimulation was up to 3-fold at the highest Rho-kinase-CAT concentration. The effect of Rho-kinase-CAT was not restricted to HEK-293 cell lysates, as stimulation of PI4P 5-kinase was also observed in bovine brain homogenates (data not shown). No phosphorylation of PI4P was observed in the absence of cell lysates, indicating that Rho-kinase-CAT had apparently no direct lipid kinase activity (data not shown).
Rho-kinase Is Implicated in RhoA-induced Elevation of PI(4,5)P 2 Levels-As Rho-kinase was shown to enhance the in vitro PI4P 5-kinase activity, we examined whether Rho-kinase affected the cellular level of PI(4,5)P 2 . Transfection of cells with Rho-kinase constructs indeed resulted in an increase in cellular PI(4,5)P 2 (Fig. 8). Wild-type Rho-kinase elevated the PI(4,5)P 2 level by about 40% and Rho-kinase-CAT by about 60%. Very similar as shown above for the PI4P 5-kinase activity, cotransfection with C3 transferase prevented the increase in PI(4,5)P 2 induced by wild-type Rho-kinase, but did not affect the increase by Rho-kinase-CAT. To substantiate the observed changes in PI(4,5)P 2 levels in metabolically labeled cells, PI(4,5)P 2 masses were additionally determined by a radioreceptor assay. In control, mock-transfected cells PI(4,5)P 2 mass amounted to 103 Ϯ 28 pmol/mg of protein (mean Ϯ S.E. for four experiments). In cells transfected with Rho-kinase constructs, PI(4,5)P 2 mass was increased to 248 Ϯ 56 pmol/mg (100 g of Rho-kinase DNA) and 292 Ϯ 67 pmol/mg (50 g of Rho-kinase-CAT DNA).
Recently, the anti-vasospastic compound HA-1077 has been reported to be an ATP-competitive Rho-kinase inhibitor (37,38). Indeed, we found that Rho-kinase-CAT-induced protein phosphorylation in HEK-293 lysates was specifically and completely inhibited by HA-1077, without appreciably affecting the basal phosphorylation pattern (data not shown). The IC 50 for and GroPIns(4,5)P 2 (Assay ϩ GroPIns(4,5)P 2 ) standards, respectively. The upper two traces were generated by running the individual standards alone. The lower control trace was obtained with the reaction product of an heat-inactivated lysate (Control). this inhibition was determined to be about 0.6 M, which fits well with the reported K i for Rho-kinase of 0.33 M (38). Incubation of intact HEK-293 cells with 10 M HA-1077 for 60 min reversed the increase in PI(4,5)P 2 by Rho-kinase-CAT, without affecting the basal level (Fig. 9A). Most important, the increase in the PI(4,5)P 2 content induced by overexpression of RhoA was similarly completely abolished by HA-1077 (Fig. 9B). DISCUSSION In the present study we have shown that RhoA as well as the Rho effector Rho-kinase stimulate the catalytic activity of PI4P 5-kinase. The stimulation of the in vitro PI4P 5-kinase activity was reflected in enhanced PI(4,5)P 2 levels in intact cells. Stimulation of PI4P 5-kinase by RhoA was initially reported by Chong et al. (8). These authors also described a physical association between RhoA and PI4P 5-kinase (9). In their study, the binding of RhoA to PI4P 5-kinase was independent of whether RhoA was in the GTP-or GDP-bound state, whereas only GTP-bound RhoA stimulated PI4P 5-kinase. Therefore, the authors hypothesized that activation of PI4P 5-kinase might require an additional Rho effector (9). Here we show that overexpression of RhoA as well as purified recombinant RhoA stimulated PI4P 5-kinase activity, whereas expression of C3 transferase reduced the formation of PI(4,5)P 2 . Most important, Rho-kinase was recognized to act as a downstream effector of Rho in the stimulation of PI(4,5)P 2 synthesis. The catalytic activity of Rho-kinase appeared to be essential, as overexpression of a kinase-dead mutant of the catalytic domain, Rho-kinase-CAT-KD, had no effect on PI4P 5-kinase activity. The observation that coexpression of C3 transferase prevented the stimulation by wild-type Rho-kinase, but not by the constitutively active catalytic domain of Rho-kinase, Rhokinase-CAT, allowed two important conclusions. First, the stimulation of PI4P 5-kinase activity by wild-type Rho-kinase was physiologically regulated by endogenous Rho. Second, the autonomous effect of Rho-kinase-CAT confirmed the notion that the isolated catalytic domain of Rho-kinase is indeed functioning as a dominant active form of Rho-kinase. Wild-type Rho-kinase contains a catalytic, a coiled-coil, a Rho-binding, and a pleckstrin-homology domain, and the removal of the regulatory domains was shown to yield a constitutively active catalytic domain (23). The differential effects of C3 transferase on the elevated cellular PI(4,5)P 2 levels induced by wild-type Rho-kinase versus Rho-kinase-CAT gave further support for the existence of a Rho 3 Rho-kinase 3 PI4P 5-kinase signaling cascade. The significance of Rho-kinase in the Rho-controlled synthesis of PI(4,5)P 2 was finally confirmed by the Rho-kinase inhibitor HA-1077. HA-1077 fully suppressed the Rho-kinase-CAT, as well as the RhoA-induced increase in PI(4,5)P 2 content.
The canonical pathway for the synthesis of PI(4,5)P 2 , i.e. phosphorylation of PI on the D-4 position of the inositol ring by PI 4-kinase to PI4P, which is then further phosphorylated on the D-5 position by PI4P 5-kinase, is believed to represent the major mechanism by which cells generate PI(4,5)P 2 . Nevertheless, in the last few years cellular phosphoinositide metabolism was appreciated to be much more complex. The type I PI4P 5-kinases were found to phosphorylate not only PI4P, but also to accept PI, PI3P, and PI(3,4)P 2 as substrates (39,40). The former type II PI4P 5-kinases were exposed to be actually PIP 4-kinases, that phosphorylate PI3P and PI5P at the D-4 position (41). Together with the PI 3-kinases and several phosphoinositide phosphatases, a complex phosphoinositide network metabolism is emerging rather than straight, single-line pathways. In the present study, the kinase assay conditions were optimized for the measurement of PI4P 5-kinase activity. Typically, PI4P was offered as substrate at low concentrations, to permit phosphorylation on the D-5 position and to avoid the phosphorylation of possibly contaminating PI5P (41,42). To prevent the production of the interfering lipid PI(3,4)P 2 , all assays were performed in the presence of nonionic detergent, which inhibits PI 3-kinase activity (43,44). Indeed, HPLC analysis of the reaction products identified the bisphosphorylated phosphoinositide to be solely PI(4,5)P 2 .
Recently, three isoforms of type I PI4P 5-kinases were cloned, designated as ␣, ␤, and ␥ (31,45,46). The isoforms had molecular masses of 68 kDa (␣ and ␤) and 90 kDa (␥). Overexpression of all three PI4P 5-kinase isoforms induced polymerization of actin in atypical short fibers in vivo (46,47). The actin fibers induced by the ␣ isoform were insensitive to dominantnegative N19 RhoA, suggesting that RhoA was located upstream of PI4P 5-kinase (47). Actin polymerization by PI4P 5-kinase was supposed to be initiated by the displacement of actin regulatory proteins by PI(4,5)P 2 (5,11). The surprising observation that also a kinase-deficient mutant of PI4P 5-kinase induced actin polymerization suggested that also structural interactions of the enzyme molecule might be involved (46). Co-expression of the ␤ isoform with the Rho-kinase constructs dramatically potentiated the stimulating effect of Rhokinase on the production of PI(4,5)P 2 , confirming that indeed PI4P 5-kinase is modulated by Rho-kinase.
The present findings together with our previous observation that RhoA and Rho-kinase potentiates the stimulation of PLD by the m3 muscarinic acetylcholine receptor (28) establishes a role for Rho-kinase in phospholipid signaling. As PI(4,5)P 2 stimulates PLD activity (3) and conversely, the PLD product phosphatidic acid is recognized as an important positive effector of PI4P 5-kinase (48,49), the two enzyme reactions are expected to be closely interacting in vivo. Our findings that Rho-kinase stimulates the activities of both PI4P 5-kinase and PLD further supports this thesis. To exclude that the effects of Rho-kinase on PLD were indirectly mediated by increased production of PI(4,5)P 2 , PLD assays were typically performed in the presence of PI(4,5)P 2 . Conversely, the potentiating effect of Rho-kinase on PI4P 5-kinase activity was observed both in the absence and presence of phosphatidic acid (data not shown). In the intact cell, however, RhoA and Rho-kinase might effect PI4P 5-kinase and PLD in concerted actions.
Unlike PLC and PLD, where reaction products can be accumulated by inhibiting their degradation by lithium chloride or by transphosphatidylation with ethanol, respectively, the direct assessment of in vivo PI4P 5-kinase activity is complicated by the fact that PI(4,5)P 2 is substrate for several enzymes (PLCs, PI 3-kinases, and PIP phosphatases) as well as binding partner for regulatory proteins. This has hindered exact studies on potential agonist-dependent biosynthesis of PI(4,5)P 2 and regulation of PI4P 5-kinase by cell surface receptors and cell adhesion (50). Nevertheless, the stimulation of PI4P 5-kinase by RhoA and Rho-kinase in vitro was reflected by elevated PI(4,5)P 2 levels in vivo. Pertussis toxin-sensitive activation of PI4P 5-kinase activity by G protein-coupled receptors was demonstrated in permeabilized neutrophils (51). Furthermore, PI4P 5-kinase activity was found associated with the activated epidermal growth factor receptor, but tyrosine phosphorylation alone seemed insufficient to activate the enzyme (52). An intriguing role in the synthesis of PI(4,5)P 2 has been proposed for phosphatidylinositol transfer protein, namely to present phosphoinositides to their successive kinases (53,54). Recently, we reported that PI(4,5)P 2 levels in HEK-293 cells were regulated by pervanadate and tyrosine kinase inhibitors, suggesting that PI4P 5-kinase is positively regulated by tyrosine phosphorylation (36). Combination experiments with C. difficile toxin B suggested that the putative tyrosine kinase was likely to act upstream of the Rho proteins (36). Besides GTPases of the Rho family, in HL60 cells the small GTPase ADP-ribosylation factor also stimulated PI(4,5)P 2 synthesis, but it was unresolved whether this was a direct stimulation or the result of enhanced phosphatidic acid production (55,56). Very recently, ADP-ribosylation factor was shown to directly activate PI4P 5-kinase ␣ in concert with phosphatidic acid (57). Future studies should elucidate how these modulators of PI4P 5-kinase, including RhoA, Rho-kinase, and ADP-ribosylation factor, are coordinated in vivo. Furthermore, it has to be determined whether and which of the isoforms of PI4P 5-kinase is a direct target for Rho-kinase.