Association of Protein Kinase Cμ with Type II Phosphatidylinositol 4-Kinase and Type I Phosphatidylinositol-4-phosphate 5-Kinase*

Protein kinase Cμ (PKCμ), also named protein kinase D, is an unusual member of the PKC family that has a putative transmembrane domain and pleckstrin homology domain. This enzyme has a substrate specificity distinct from other PKC isoforms (Nishikawa, K., Toker, A., Johannes, F. J., Songyang, Z., and Cantley, L. C. (1997) J. Biol. Chem. 272, 952–960), and its mechanism of regulation is not yet clear. Here we show that PKCμ forms a complex in vivo with a phosphatidylinositol 4-kinase and a phosphatidylinositol-4-phosphate 5-kinase. A region of PKCμ between the amino-terminal transmembrane domain and the pleckstrin homology domain is shown to be involved in the association with the lipid kinases. Interestingly, a kinase-dead point mutant of PKCμ failed to associate with either lipid kinase activity, indicating that autophosphorylation may be required to expose the lipid kinase interaction domain. Furthermore, the subcellular distribution of the PKCμ-associated lipid kinases to the particulate fraction depends on the presence of the amino-terminal region of PKCμ including the predicted transmembrane region. These results suggest a novel model in which the non-catalytic region of PKCμ acts as a scaffold for assembly of enzymes involved in phosphoinositide synthesis at specific membrane locations.

diacylglycerol (DAG)), novel PKCs (nPKC) comprising ␦, ⑀, , and (activated by DAG and acidic phospholipid but insensitive to calcium), atypical PKCs / and (mechanism of regulation not clear) (1)(2)(3)(4)(5)(6). Another subgroup of PKCs may be defined by PKC (human)/PKD (mouse homologue of PKC) (7)(8)(9)(10)(11). PKC differs in some structural features from other PKC isozymes in that PKC contains a putative transmembrane domain and a pleckstrin homology (PH) domain in its amino-terminal regulatory region providing a mechanism for constitutive association with the membrane (7,8). Recent investigations using various approaches such as overexpression and down-regulation of specific isozymes indicate that each PKC isozyme plays a unique role in signal transduction processes (1,5). It is highly likely that the distinct functions of these isozymes are a consequence of isozyme-specific substrates and/or interacting proteins. However, information about the substrate specificities of each PKC family member has been quite limited.
Recently, we determined optimal substrate motifs for all classes of PKC isozymes using an oriented peptide library technique and found that each PKC isozyme has a unique optimal substrate sequence (12). Interestingly, PKC has very different substrate specificities from other isozymes in that PKC showed extreme selectivity for peptides with Leu at position Ϫ5 amino-terminal of the phosphorylated Ser. This finding suggested that PKC is involved in a unique signaling pathway distinct from those mediated by other PKC family members. PKC/PKD was recently shown to be activated by mitogenic regulatory peptides and platelet-derived growth factor (PDGF) by a pathway that requires cPKC or nPKC activation upstream (13). In lymphocytes, PKC is shown to regulate signaling via Syk and phospholipase C␥1 (14). However, substrate proteins or interacting proteins of PKC in intact cells are poorly understood.
Of interest, recent data indicate that PKC is located in the Golgi apparatus and is involved in basal transport processes (15). Lipid kinases, such as phosphatidylinositol 3-kinase (Pt-dIns 3-K), phosphatidylinositol 4-kinase (PtdIns 4-K), and phosphatidylinositol-4-phosphate 5-kinase (PtdIns-4P 5-K), have also been implicated in membrane trafficking (16). One of the PtdIns 4-K isozymes, PtdIns 4-K␤, is concentrated in the Golgi apparatus (17). The yeast VPS34 gene product, Vps34p, which is a PtdIns-specific 3-kinase, is required for the efficient sorting and delivery of proteins from the Golgi to the vacuole (18). Vps34p exists as a complex with Vps15p, a 160-kDa Ser/ Thr protein kinase, which recruits Vps34p to the membrane of the Golgi complex and enhances Vps34p PtdIns 3-kinase activity (19). A human homologue of Vps34p has recently been characterized as part of a complex with a human homologue of Vps15p, an adaptor protein called p150 (20). Interestingly, this * This work was supported by National Institutes of Health Grants RO1-GM36624 and by P50-HL56993 (SCOR Grant in Atherosclerosis). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
In this study we found that PKC associates with at least two lipid kinases, a type II PtdIns 4-K and type I PtdIns-4-P 5-K, in an isozyme-specific manner. Analysis using various deletion mutants of PKC suggest that these two lipid kinases effectively bind to PKC through its amino-terminal region and that the protein kinase activity of PKC is essential to this association. These results suggest that PKC plays a role in assembly of specific phosphatidylinositol-phosphorylating enzymes at the membrane.

EXPERIMENTAL PROCEDURES
Materials-Isozyme-selective antipeptide antibodies for PKC␣, -⑀, and -were purchased from Santa Cruz Biotechnology. Two anti-PKC antibodies were used as follows: one raised against amino acids 6 -25 mapping at the amino terminus of PKC (␣--N) and one against amino acids 893-912 mapping at the carboxyl terminus of PKC (␣--C). Anti-FLAG M2 monoclonal antibodies were purchased from Eastman Kodak Co. LipofectAMINE reagents were purchased from Life Technologies, Inc. P81 phosphocellulose paper was purchased from Whatman.
cDNA Expression Vectors-The PKC⑀ construct was prepared as described previously (6). FLAG M2-tagged PKC construct was provided by Dr. Matthew Coghlan and Dr. Margaret Chou. A glutathione S-transferase (GST) fusion of the cysteine-rich region (amino acids 34 -290) of PKC named GST-CYS and a GST fusion of the pleckstrin homology domain (amino acids 380 -545) named GST-PH were expressed in bacteria and purified with glutathione-Sepharose beads as described previously (7). The delta1-79 PKC mutant was constructed by digesting pBpl4 (7) with ApaI and NsiI. Overhanging 5Ј-and 3Ј-ends were filled with the Klenow enzyme and ligated in the BamHI/blunt site of pSV40, a pCDNA3-derived vector. The delta 1-340 mutant was constructed by cutting the delta 1-79/pSV construct with HindIII, isolation of the 800-base pair HindIII fragment followed by religating of the vector/PKC portion. Expression of that construct results in the translation initiation at Met-341. The expression of both deletion mutants was verified by Western blot analysis using antipeptide antibodies for PKC (epitope corresponding to amino acids 893-912 mapping at the carboxyl terminus of PKC). A kinase-deficient mutant of PKC, K612W, was generated by site-directed mutagenesis using the in vitro mutagenesis kit and was subcloned into pcDNA3.
Cell Culture and Transfections-COS-7 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum using 100-mm plastic dishes. Exponentially growing COS-7 cells were transfected with the various plasmids using LipofectAMINE reagents, according to the manufacturer's instructions. Swiss 3T3 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum. For PDGF stimulation experiments, subconfluent Swiss 3T3 cells were washed with phosphate-buffered saline (PBS), cultured in Dulbecco's modified Eagle's medium with 0.1% fetal calf serum for 24 h, and stimulated with the indicated concentration of PDGF for 5 min.
For subcellular fractionation studies, cells were homogenized in lysis buffer lacking detergent. The cytosolic and particulate fractions were separated by 105,000 ϫ g centrifugation as described previously (23). PKC was then immunoprecipitated from each fraction as above.
Association of Lipid Kinases from COS-7 Cell Lysate with GST-fused Proteins-COS-7 cells were washed with PBS and lysed in lysis buffer. Crude cell extract was clarified by centrifugation (14,000 rpm for 10 min). GST or GST-fused proteins (3 g of each) were incubated with the cell lysate at 4°C for 2 h with constant rocking. The beads were washed twice with 1-ml volume of ice-cold PBS, twice with 1-ml volume of ice-cold washing buffer (0.1 M Tris-HCl, pH 7.4, 0.1 M LiCl), and twice with 1-ml volume of TNE (10 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA). Lipid kinase assays were then performed on the beads as described below.
Western Blot Analysis-Western blot analysis was done as reported previously (24). The immunoprecipitated proteins were dissolved in 20 l of sodium dodecyl sulfate (SDS) buffer (62.5 mM Tris, 2% SDS, 5% 2-mercaptoethanol, and 5% glycerol, pH 6.8) and boiled for 5 min. This solution was subjected to 7.5% SDS-polyacrylamide gel electrophoresis. After electrophoresis, the proteins were electrophoretically transferred to nitrocellulose membrane. Isozyme-specific anti-PKC antibodies were used as the primary antibody. Anti-rabbit IgG or anti-mouse IgG conjugated to horseradish peroxidase was used as a secondary antibody. The membranes were finally visualized by chemiluminescence as described by the manufacturer (NEN Life Science Products).
PKC Assay-PKC activity was assayed in vitro essentially as described previously using the standard PKC vesicle assay (25). The reaction mixture (30 l) contained 100 M ATP with [␥-32 P]ATP (5 Ci), 1 mM dithiothreitol, 5 mM MgCl 2 , 25 mM Tris-HCl, pH 7.5, 20 g/ml PS, 10 M DAG, 200 M CaCl 2 (for PKC␣), 0.5 mM EGTA (for PKC⑀, and ), and the indicated amount of synthetic substrate peptide. Reactions were started by addition of immunoprecipitated PKC and incubated at 30°C for 10 min. Reaction mixtures were spotted onto P81 phosphocellulose paper and washed 4 times in 500 ml of 1% phosphoric acid. Incorporation of 32 P was determined by liquid scintillation counting. For each experimental condition, values for control reactions lacking substrate peptide were subtracted as blanks. Autophosphorylation of PKC was carried out essentially as described above, but no substrate was added. Proteins of the reaction mixture were separated by SDSpolyacrylamide gel electrophoresis and visualized by autoradiography.
Lipid Kinase Assay and HPLC Analysis-Lipid kinase assays were performed as described previously with a slight modification (26). Briefly, the reaction mixture (100 l) contained 50 M ATP, 20 mM HEPES, pH 7.5, 10 mM MgCl 2 , 0.2 mg/ml sonicated lipids, 20 Ci of [␥-32 P]ATP in the presence (for PtdIns 4-K and PtdIns-4-P 5-K assay) or absence (for PtdIns 3-K assay) of 0.3% Triton X-100. Assays were performed at 37°C for 20 min and then stopped with 25 l of 5 N hydrochloric acid. The lipid was extracted with 160 l of 1:1 (v/v) chloroform:methanol. The organic layer was collected, subjected to thin layer chromatography (TLC) on Silica gel 60 plate, and developed with 1-propanol/2 M acetic acid (65:35 by volume). Phosphoinositides were visualized by autoradiography and quantitated using a molecular imager (Bio-Rad). The lipid products on the TLC plate were collected, deacylated, and analyzed by HPLC as described in detail elsewhere (27). Standards were made from tritiated PtdIns-4-P and PtdIns-4,5-P 2 (NEN Life Science Products), which were deacylated and included with the 32 P-labeled product in the HPLC run. PtdIns-3-P, PtdIns-3,4-P 2 , and PtdIns-5-P standards were made as described previously (28).

Association of PtdIns 4-K and PtdIns-4-P 5-K Activities with
PKC-To examine whether PKC associates with lipid kinases, phosphoinositide kinase activities were measured in immunoprecipitates of PKC. A PKC expression vector or control vector were transiently transfected into COS-7 cells, and cell lysates were immunoprecipitated with an anti-PKCspecific antibody raised against the amino terminus of PKC. The resultant immunocomplexes were subjected to Western blot analysis and in vitro protein kinase assay to verify the expression of PKC. The immunocomplex obtained from PKC-transfected COS-7 cells contained a 110-kDa protein that immunoblotted with anti-PKC antibody (Fig. 1A). This is the expected migration position for PKC (7). The 110-kDa protein was not detected in immunoprecipitates from nontransfected cells (Fig. 1A). Three synthetic peptides, ␣, ␤I, and peptide, were used as substrates for protein kinase assays. These peptides are based on the optimal substrate motifs of PKC␣, -␤I, and -, respectively (12). The immunocomplex obtained from PKC-transfected COS-7 cells phosphorylated ␤I and peptides equally but not the ␣ peptide (Fig. 1B). The immunocomplex obtained from control vector-transfected cells showed no significant protein kinase activity (Fig. 1B). These results are consistent with our previous finding that PKC preferentially phosphorylates the ␤I and peptides compared with the ␣ peptide (12). These results confirm the overexpression of functional PKC in COS-7 cells. These immunocomplexes were subjected to a lipid kinase assay using crude brain phosphoinositides (CBP) as lipid substrates in the presence or absence of 0.3% of Triton X-100. As shown in Fig. 1C, the PKC immunoprecipitate contained activity that phosphorylated lipids to produce products that migrate on thin layer chromatography in the region of PtdIns-4-P and PtdIns-4,5-P 2 standards. This activity was enhanced in the presence of Triton X-100. Since Triton X-100 is known to inhibit PtdIns 3-kinase activity (29), these results argue against PtdIns 3-kinase being responsible for the phosphorylation detected.
We checked the possibility that PtdIns 4-K and PtdIns-4-P 5-K association is a common property of PKC family members. For this purpose, cPKC␣, nPKC⑀, and atypical PKC were used, each of which is classified into a different subfamily of PKCs from PKC. The expression vectors of PKC⑀, -, orwere transiently transfected into COS-7 cells, and the enzymes were immunoprecipitated with isozyme-specific antibody. Endogenous PKC␣ was investigated since this enzyme is highly expressed in COS-7 cells. The expression of each isozyme was confirmed by Western blot analysis (Fig. 3A) and protein kinase assays of immunoprecipitates (data not shown). Each immunocomplex was also subjected to PtdIns 4-K (Fig. 3B) and PtdIns-4-P 5-K (Fig. 3C) assays. Both lipid kinase activities were normalized to the ␤I peptide phosphorylating activity, because ␤I peptide is a relatively good substrate for all the PKC isozymes as described previously (12). Significant PtdIns 4-K activity was detected in the PKC immunoprecipitate, whereas less activity was detected in the PKC␣ immunoprecipitate, and no PtdIns 4-K activity was detected in PKC⑀ and PKC immu- noprecipitates (Fig. 3B). Furthermore, PtdIns-4-P 5-K activity was only detected in the PKC immunoprecipitate (Fig. 3C). These results indicate that PtdIns 4-K and PtdIns-4-P 5-K association is not a general property of PKC family members.
The Amino-terminal Region of the Regulatory Domain of PKC Is Involved in the Association with the Lipid Kinases-Two PKC deletion mutants, the delta 1-79 and delta 1-340, were used to determine the region of PKC involved in the association of the lipid kinases. These mutants were transfected into COS-7 cells and immunoprecipitated with anti-PKC antibody raised against the carboxyl terminus of PKC. The expression of these mutants was verified by Western blot analysis (Fig. 4B), by protein kinase activity (Fig. 4C), and by autophosphorylation (Fig. 4D) in immunoprecipitates. The delta 1-79 mutant was detected as two bands on Western blot. The upper band of the doublet may result from hyperphosphorylation of the delta 1-79; consistent with this idea, under autophosphorylation conditions more radioactivity is detected in the upper band (Fig. 4D). Consistent with the absence of region 1-79, neither band was blotted with the antibody against the amino terminus of PKC (not shown). The autophosphorylation level of the delta 1-340 mutant was reduced by 75% compared with that of wild type PKC and the delta 1-79 mutant. Since the delta 1-340 mutant has similar peptide kinase activity to wild type PKC (Fig. 4C), these results suggest that the major autophosphorylation site might be present in the region between residues 79 and 340. The immunocomplexes were subjected to PtdIns 4-K (Fig. 4D) and PtdIns-4-P 5-K (Fig. 4E) assays. Both lipid kinase activities associating with the delta 1-340 mutant were reduced by 80% compared with activities associated with wild type PKC and the delta 1-79 mutant. These results indicate that the region FIG. 3. Lipid kinases specifically associate with PKC compared with other PKC isoforms. Expression vectors of PKC⑀, -, and -or control vectors were transiently transfected into COS-7 cells and immunoprecipitated with an isozyme-specific antibody. Immunoprecipitated PKC␣ was prepared from control COS-7 cells using a specific antibody. The resultant immunocomplexes were subjected to Western blot analysis (A). A PtdIns 4-K assay using PtdIns/PtdSer (1:1) (B) or a PtdIns-4-P 5-K assay using PtdIns-4-P/PA (1:1) (C) was performed in the presence of 0.3% Triton X-100, as described under "Experimental Procedures." Phosphoinositides were visualized by autoradiography and quantitated as described under "Experimental Procedures." The data are presented as the percentage of the activity present in the immunoprecipitated PKC fraction. Both lipid kinase activities were normalized to the ␤I peptide-phosphorylating activities present in the immunocomplexes. The data are representative of three experiments. The PtdIns 4-K to protein kinase ratio for PKC was 16.9 and the PtdIns-4-P 5-K to protein kinase ratio was 18.5.

FIG. 4. The amino-terminal region of the regulatory domain of PKC is involved in the association with lipid kinases.
Schematic drawing of the structure of wild type PKC and the amino-terminal deletion mutants of PKC, delta 1-79 and delta 1-340 (A). TM, transmembrane domain; CYS, cysteine-rich domain; AD, activation domain; PH, pleckstrin homology domain. These constructs were transfected into COS-7 cells and immunoprecipitated with anti-PKC antibody raised against the carboxyl terminus of PKC. The resultant immunocomplexes were subjected to Western blot analysis (B), in vitro protein kinase assay using ␤I peptides as substrates (C), and autophosphorylation assay (D). PtdIns 4-K (E) and PtdIns-4-P 5-K (F) assays were performed as described in the legend for Fig. 3. Both lipid kinase activities were normalized to the ␤I peptide phosphorylating activities present in the immunocomplexes. Autophosphorylation of various PKC constructs and phosphorylation of phosphoinositides were visualized by autoradiography and quantitated as described under "Experimental Procedures." The data are presented as the percentage of the activity present in immunoprecipitates of full-length PKC. Error bars represent the standard error from three independent experiments. The PtdIns 4-K to protein kinase ratio for full-length PKC was 10.5, and the PtdIns-4-P 5-K to protein kinase ratio was 11.2. between residues 79 and 340 of PKC, which contains cysteine rich domain I and II, is involved in the association with both lipid kinases.
Characterization of the PKC-associated PtdIns 4-K and Pt-dIns-4-P 5-K-In order to confirm the above observations and further characterize these lipid kinases, the GST-fused aminoterminal region of PKC coding for amino acids 34 -290 (GST-CYS domain), GST-fused pleckstrin homology domain coding for amino acids 380 -545 (GST-PH domain), or GST only were incubated with COS-7 lysate and assayed for PtdIns 4-K and PtdIns-4-P 5-K activities. As shown in Fig. 5B, the GST-CYS domain associated with a PtdIns kinase activity while very little activity associated with GST alone or the GST-PH domain. This PtdIns kinase activity was inhibited by adenosine in a dose-dependent manner (85% by 500 M adenosine) (Fig. 5C) but was not inhibited by 100 nM wortmannin which caused 100% inhibition of PtdIns 3-K activity (Fig. 5D). This result indicates that this PtdIns kinase has properties similar to type II PtdIns 4-K which is markedly stimulated by detergent and inhibited by adenosine (30). Similar characteristics were obtained using wild type PKC-associated PtdIns kinase (data not shown).
The GST-CYS domain also bound PtdIns-4-P 5-K activity, whereas the GST-PH domain and GST alone had very little associated activity (Fig. 5E). In agreement with results in Fig.  1C, the GST-CYS domain-associated PtdIns-4-P 5-K activity was activated by phosphatidic acid, consistent with this being a type I PtdIns-4-P 5-K (31). All of these results suggest that PKC associates with a type II PtdIns 4-K and a type I PtdIns-4-P 5-K through its amino-terminal region (amino acids 80 -290).
Protein Kinase Activity of PKC Is Required for the Association of PKC with the Lipid Kinases-To determine whether the protein kinase activity of PKC is required for the association with the lipid kinases, a kinase-deficient mutant of PKC, K612W mutant, was transfected into COS-7 cells. The immunoprecipitated K612W mutant showed neither the ␤I peptide kinase activity nor autophosphorylation activity compared with those of wild type PKC (Fig. 6, B and C). The immunoprecipitates of wild type PKC and the K612W mutant were subjected to lipid kinase assays using crude brain phosphoinositides as lipid substrates. As shown in Fig. 6D, no lipid kinase activity was observed in the immunoprecipitate of the K612W mutant, whereas marked productions of both PtdIns-P and PtdIns-P 2 were observed in the immunoprecipitate of wild type PKC, suggesting that protein kinase activity of PKC is GST only were incubated with COS-7 lysate. After washing the beads, a PtdIns 4-K assay was performed (B). A PtdIns 4-K assay was also performed on the GST-CYS domain in the presence or absence of the indicated amount of adenosine (C) or 100 nM of wortmannin (D). The effect of 100 nM wortmannin on purified PtdIns 3-K is shown for comparison (D). The assay conditions were the same as described in the legend for Fig. 3. The data are presented as the percentage of the activity in the absence of inhibitors. The PtdIns-4-P 5-K assay was performed in the presence of PS or PA as described in the legend to Fig.  3 (E). Phosphoinositides were visualized by autoradiography and quantitated as described under "Experimental Procedures. "   FIG. 6. A kinase-deficient mutant of PKC, K612W, does not associate with lipid kinase activities. Wild type PKC or a kinasedeficient mutant of PKC, K612W, was transfected into COS-7 cells and immunoprecipitated with anti-PKC antibody raised against the amino terminus of PKC. The resultant immunocomplexes were subjected to Western blot analysis (A), a protein kinase assay using ␤I peptide as a substrate (B), and an autophosphorylation assay (C). Lipid kinase assays (D) were performed as described in the legend for Fig. 3 using crude brain phosphoinositides as lipid substrates. Lipid products were visualized by autoradiography as described under "Experimental Procedures." required for the association of PKC with the lipid kinases.

The Association of PKC with the Lipid Kinases Is Not Affected by PDGF-induced Activation of PKC in Swiss 3T3
Cells-PKD, a mouse homologue of PKC, has been shown to be activated by various mitogenic regulatory peptides and PDGF through another PKC isozyme-dependent signal transduction pathway in Swiss 3T3 cells (13). The effect of PDGF on PKC protein kinase activity and on the lipid kinase activities was investigated using Swiss 3T3 cells. As shown in Fig. 7, A and C, immunoprecipitates of endogenous PKC from Swiss 3T3 cells contained both PtdIns 4-K and PtdIns-4-P 5-K activities. PDGF caused marked activation of PKC protein kinase activity (3-and 8-fold for 3 and 16 ng/ml of PDGF, respectively) (Fig. 7B). However, PDGF did not affect the lipid kinase activities associated with PKC (Fig. 7C). Since PDGF causes the enhanced phosphorylation of PKC not by activation of autophosphorylation but by activation of another PKC isozyme (13), this observation along with results from Fig. 6 suggests that the basal protein kinase activity of PKC is essential and sufficient for the association of both lipid kinase activities.
Domains of PKC Required for Location of PtdIns 4-K and PtdIns-4-P 5-K at the Membrane-The subcellular distribution of wild type PKC and the PKC mutants (the delta 1-79, the delta 1-340, and K612W mutants) were investigated. After 105,000 ϫ g ultracentrifugation to separate cytosol and particulate fractions, the PKC proteins were immunoprecipitated and analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting (Fig. 8A) and by a protein kinase assay (Fig.  8B). Wild type PKC and the K612W mutant were mainly recovered from the particulate fraction rather than the cytosolic fraction. In contrast, the delta 1-79 mutant was primarily in the cytosolic fraction rather than the particulate fraction. The delta 1-340 was almost exclusively recovered from the cytosolic fraction. These results indicate that the amino-terminal hydrophobic sequence plays a major role in membrane binding and that the region between 79 and 340 provides some binding to the membrane.
The PKC immunocomplexes obtained from the two fractions were also subjected to PtdIns 4-K and PtdIns-4-P 5-K assays (Fig. 8C). For wild type PKC, lipid kinase activities were found primarily in protein from the particulate fraction, as expected. For the delta 1-79 mutant, lipid kinase activities were found in both the cytosolic PKC and the membranebound PKC. As expected from the results in Fig. 4, the delta 1-340 mutant had relatively little lipid kinase activity associated but the activity was confined to the cytosolic fraction. Consistent with the previous results, the K612W mutant associated with little lipid kinase activity. Thus, the PtdIns 4-K and PtdIns-4-P 5K activities can complex with PKC whether or not it is membrane-bound, and the amino-terminal regions of PKC are required for location of the lipid kinases to the membrane.

DISCUSSION
In the present study, we found that PKC associates with at least two lipid kinases, a type II PtdIns 4-K and a type I PtdIns-4-P 5-K. These lipid kinase activities could be immunoprecipitated with PKC-specific antibodies from COS-7 cells overexpressing PKC or from Swiss 3T3 cells that normally express PKC. In contrast, very little lipid kinase activity could be detected in immunoprecipitates of PKC␣, PKC⑀, or PKC. The protein kinase activity of PKC was found to be essential for the association with these lipid kinase activities, because a kinase-deficient mutant of PKC, K612W, did not associate with either lipid kinase activity. Furthermore, the subcellular distribution of the PKC-associated lipid kinases to the particulate fraction depends on the presence of the aminoterminal region of PKC, including the predicted transmembrane region. These results suggest that PKC can act as a scaffold to locate phosphoinositide kinases at the membrane and that the association or activation can be regulated by autophosphorylation of PKC.
Multiple types of PtdIns 4-K exist in mammalian cells (40). The type II PtdIns 4-Ks were initially characterized as membrane-associated 55-kDa proteins whose lipid kinase activities are inhibited by adenosine and the monoclonal antibody 4C5G (30,32). The gene for the 55-kDa PtdIns 4-K has not been cloned. However, two other mammalian PtdIns 4-K genes have been cloned (PtdIns 4-K␣ and PtdIns 4-K␤) and shown to encode proteins with high homology to PtdIns 3-K (32,33). These enzymes are not intrinsic membrane proteins, and the mechanism by which they are brought to the membrane where their substrates reside is not known (40). Interestingly, the recently cloned PtdIns 4-K␤ was shown to be concentrated in the Golgi (17). Since PKC was reported to be concentrated in the Golgi (15), we investigated the possibility that the PtdIns 4-K activity associated with PKC is due to PtdIns 4-K␤. However, we could find no evidence for association of either PtdIns 4-K␣ or PtdIns 4-K␤ with PKC using overexpression or isoform-specific antibodies (not shown). In addition, the adenosine FIG. 7. The association of PKC with lipid kinases is not affected by PDGF-induced activation of PKC in Swiss 3T3 cells. Subconfluent Swiss 3T3 cells were incubated with 0.1% fetal calf serum for 24 h and stimulated with the indicated concentration of PDGF for 5 min. Endogenous PKC was immunoprecipitated with anti-PKC antibody (anti-amino terminus) or with control antibody. The immunocomplexes were used for Western blot analysis (A), protein kinase assay using ␤I peptide as a substrate (B), and lipid kinase assays (C). PtdIns 4-K and PtdIns-4-P 5-K activities were measured using PtdIns/PtdSer (1:1) and PtdIns-4-P/PA (1:1) as substrates, respectively, as described in the legend for Fig. 3. Lipid products were visualized by autoradiography as described under "Experimental Procedures." sensitivity, detergent activation, and wortmannin resistance of the PKC-associated PtdIns 4-K are consistent with the properties of the 55-kDa type II enzyme that has not yet been cloned (34). Only approximately 3% of the total cellular PtdIns 4-K and PtdIns-4-P 5-K activities co-precipitate with PKC (not shown), indicating either that PKC associates with isoforms that are in low abundance or that protein modification is required to assemble the complex.
Two classes of phosphatidylinositol phosphate kinases have been characterized, type I and type II. The type I enzymes phosphorylate PtdIns-4-P at the 5 position to form PtdIns-4,5-P 2 . Two different genes for type I enzymes have been cloned and expressed (type I␣ and type I␤), and both enzymes are significantly activated by phosphatidic acid (31,35,36). The PtdIns-4-P 5-K activity associated with PKC is activated by phosphatidic acid, consistent with it being due to a type I enzyme. However, attempts to show association between PKC and either type I␣ or type I␤ have thus far failed. 2 The type II PtdIns-4-P 5-K was also thought to be a PtdIns-4-P 5-K; however, recent research in this laboratory showed that this enzyme phosphorylates PtdIns-5-P at the 4 position to make PtdIns-4,5-P 2 (28). The activity associated with PKC does not phosphorylate PtdIns-5-P.
We found that deletion of 79 amino acids from the amino terminus of PKC (the delta 1-79 mutant) did not affect the binding of either PtdIns 4-K or PtdIns-4-P 5-K, whereas deletion of 340 amino acids (the delta 1-340 mutant) diminished binding of both enzymes by 80%. Furthermore, the CYS domain of PKC coding for amino acids 34 -290 could effectively bind both lipid kinase activities from COS-7 cell lysates. These results suggest that PKC associates with these two lipid kinases through its amino-terminal region (amino acids 80 -290). On the other hand, a kinase-deficient PKC mutant, K612W, failed to associate with either lipid kinase, even though the K612W mutant has an intact amino-terminal region. These observations suggest that the protein kinase activity of PKC is required for inducing a conformational change of PKC that exposes the amino-terminal region which is involved in the association with lipid kinases. If so, what is the target substrate of PKC in this model? The only major phosphorylated protein present in the immunoprecipitated PKC following in vitro addition of [␥-32 P]ATP was PKC itself (Figs. 4D and 6C). This is also true for the delta 1-79 mutant (Fig.  4D). Furthermore, the autophosphorylation level of the delta 1-340 mutant was reduced by 75% compared with wild type PKC under conditions in which the protein kinase activities (using ␤I peptide) of wild type PKC and the delta 1-340 mutant were the same. All of these results suggest that the major target substrate of PKC, which is involved in the recruitment of lipid kinase activity, is PKC itself and that the major in vitro phosphorylation sites are located in the amino terminus of PKC. The possibility that the kinase-inactive PKC (K612W) associates with the lipid kinases but fails to activate them cannot be excluded by the present data. It will be necessary to clone the genes for the PKC-associated PtdIns 4-K and PtdIns-4-P 5-K to resolve further the mechanism of recruitment and activation since they do not appear to be cloned genes.
It has been shown that Vps34p, which is a phosphatidylinositol-specific 3-kinase (PtdIns 3-K) and is involved in vesicle-2 K. Nishikawa and K. F. Tolias, unpublished results. Wild type PKC, the delta 1-79 mutant, the delta 1-340 mutant, or the K612W mutant were transfected into COS-7 cells. Cytosolic and particulate fractions were prepared as described under "Experimental Procedures." PKC was precipitated from these two fractions, and the immunoprecipitates were analyzed by Western blot (A) or were assayed for the ability to phosphorylate the ␤I peptide or assayed for PtdIns 4-K and PtdIns-4-P 5K activities as described in the legend to Fig. 4. mediated transport of proteins to the vacuole, associates with Vps15p, a 160-kDa membrane-bound Ser/Thr protein kinase (19). Vps15p recruits Vps34p to the membrane of the Golgi complex and enhances Vps34p PtdIns 3-K activity, since mutational inactivation of Vps15p protein kinase activity stops its association with Vps34p and blocks activation of Vps34p lipid kinase activity (19,37). A human homologue of Vps34p has also been characterized as part of a complex with a human homologue of Vps15p, called p150 (20). This complex also has a protein kinase activity, and the association of the Vps34p homologue with p150 increases the PtdIns 3-K activity (21). However, the structure of p150 is quite different from that of PKC, in that p150 contains an amino-terminal myristoylation site, a Ser/Thr protein kinase domain, a region with homology to the 65-kDa regulatory subunit of protein phosphatase 2A, and a region containing a WD40 repeat motif (21).
It has been demonstrated that both p150 and Vps15p are modified by myristoylation at the amino-terminal region (21,38), which may allow their targeting to the cytoplasmic face of the Golgi membrane, thereby allowing PtdIns 3-K to gain access to the membrane. PKC is the only PKC isozyme that contains a putative transmembrane sequence (7). Consistent with this fact, membrane targeting of the PKC-associated lipid kinases depends on the presence of amino-terminal region of PKC, including this transmembrane sequence (Fig. 8). Furthermore, PKC was shown to be located in the Golgi membrane and involved in basal vesicle transport processes (15). The results presented here suggest that PKC may play a role in vesicle transport processes by recruiting two different types of lipid kinases to the Golgi membrane to facilitate local production of PtdIns-4-P and PtdIns-4,5-P 2 . In addition to their roles as precursors of Ins-1,4,5-P 3 and diacylglycerol, these lipids have been implicated in various vesicle transport events (16,39).
Stimulation of Swiss 3T3 cells with PDGF caused marked activation of PKC but did not affect the lipid kinase activities associated with PKC, suggesting that the basal protein kinase activity of PKC is essential and sufficient for the association of both lipid kinases. Recently, PKD, a mouse homologue of human PKC, was shown to be activated not only by PDGF but also by other cell stimuli, such as vasopressin, endothelin, and bradykinin through a pathway that requires another PKC isozyme upstream (13). Interestingly, bombesin was demonstrated to induce PKD phosphorylation in vivo in a time-and dose-dependent manner, but the major phosphorylation sites were different from those of unstimulated PKD (basal phosphorylation sites) judging from the tryptic phosphopeptide maps (13). All of these observations support a model in which PKC can be phosphorylated by at least two steps in intact cells; the first step is an autophosphorylation that is necessary for the induction of a conformational change of PKC allowing the association with lipid kinases, and the second step is a transphosphorylation step mediated by another PKC isozyme-dependent pathway that is involved in full activation of PKC.
In summary, PKC associates with type II PtdIns 4-K and type I PtdIns-4-P 5-K. This association is mediated through an amino-terminal region (amino acids 78 -290) of PKC, and the protein kinase activity of PKC is essential for the association with lipid kinases. We propose that autophosphorylation of the amino-terminal region of PKC induces a conformational change that enables both lipid kinases to associate. Thus PKC may act as a regulated scaffold for recruitment of enzymes to the membrane for the purpose of local phosphoinositide synthesis.