Cpk is a novel class of Drosophila PtdIns 3-kinase containing a C2 domain.

We report the identification of a novel class of phosphatidylinositol (PtdIns) 3-kinases whose members contain C-terminal C2 domains. We have isolated Drosophila and murine genes (termed cpk and cpk-m respectively) by polymerase chain reaction amplification of cDNA libraries with degenerate primers corresponding to conserved regions of PtdIns kinases. The amino acid sequences of Cpk and Cpk-m are most similar to that of p110, a family of PtdIns 3-kinases that mediates the responses of cells to mitogenic stimuli. The Cpk and Cpk-m sequences are similar to a large, central region of p110, but differ from p110 at their N and C termini. The N termini of the Cpk proteins do not contain any recognizable protein motif, while the C termini contain “C2 domains,” a feature unique among PtdIns kinases. Cpk has an intrinsic PtdIns kinase activity and can phosphorylate PtdIns and PtdIns-4-P, but not PtdIns(4,5)P2, at the D3 position of the inositol ring. Cpk is the first PtdIns 3-kinase identified with this particular substrate specificity. We have identified two potential Cpk-binding proteins, p90 and p190, and have determined that both Cpk and p190 may be tyrosine phosphorylated. This finding suggests that Cpk function may be regulated by tyrosine kinases.

We report the identification of a novel class of phosphatidylinositol (PtdIns) 3-kinases whose members contain C-terminal C2 domains. We have isolated Drosophila and murine genes (termed cpk and cpk-m respectively) by polymerase chain reaction amplification of cDNA libraries with degenerate primers corresponding to conserved regions of PtdIns kinases. The amino acid sequences of Cpk and Cpk-m are most similar to that of p110, a family of PtdIns 3-kinases that mediates the responses of cells to mitogenic stimuli. The Cpk and Cpk-m sequences are similar to a large, central region of p110, but differ from p110 at their N and C termini. The N termini of the Cpk proteins do not contain any recognizable protein motif, while the C termini contain "C2 domains," a feature unique among PtdIns kinases. Cpk has an intrinsic PtdIns kinase activity and can phosphorylate PtdIns and PtdIns-4-P, but not PtdIns(4,5)P 2 , at the D3 position of the inositol ring. Cpk is the first PtdIns 3-kinase identified with this particular substrate specificity. We have identified two potential Cpk-binding proteins, p90 and p190, and have determined that both Cpk and p190 may be tyrosine phosphorylated. This finding suggests that Cpk function may be regulated by tyrosine kinases.
PtdIns 1 kinases regulate diverse cellular processes, which include cell signaling, cell cycle progression, and intracellular protein sorting (reviewed in Herman et al., 1992;Kapeller and Cantley, 1994;Stephens et al., 1993;Kunz et al., 1993). PtdIns kinases phosphorylate phosphoinositol lipids at distinct positions on the inositol ring; PtdIns 3-kinases phosphorylate the D3 hydroxyl group on the inositol ring, while PtdIns 4-kinases phosphorylate the D4 hydroxyl group (reviewed in ). All PtdIns kinases contain a core region of sequence similarity, which is believed to be the catalytic domain. This domain, termed the "PtdIns kinase domain," shares limited sequence similarity with the catalytic domain of protein kinases (Hiles et al., 1992) and mutation of conserved residues results in loss of PtdIns kinase activity (Dhand et al., 1994;Schu et al., 1993). Surprisingly, the mechanism used by PtdIns kinases to modulate regulatory pathways remains unknown.
A number of receptor tyrosine kinases, src-like tyrosine kinases and viral oncoproteins bind and activate a cellular PtdIns 3-kinase (reviewed in ). Studies of mutants that abrogate the binding of this PtdIns 3-kinase to these molecules suggest that PtdIns 3-kinases can mediate mitogenic and cell motility responses of cells to growth factors and oncoproteins. Purification of the polypeptide subunits of this PtdIns 3-kinase revealed that the enzyme exists as a heterodimeric complex composed of a 110-and an 85-kDa subunit Fry et al., 1992;Morgan et al., 1990;Shibasaki et al., 1991). The 110-kDa subunit contains a C-terminal PtdIns kinase domain, as well as a small domain at its N terminus that is sufficient for binding to the 85-kDa subunit (Holt et al., 1994;Hiles et al., 1992;Klippel et al., 1994). The 85-kDa subunit serves as an adapter and binds activated growth factor receptors and other tyrosine phosphorylated molecules through two Src homology 2 (SH2) domains (Hu et al., 1992;Klippel et al., 1992;McGlade et al., 1992;Reedijk et al., 1992;Yoakim et al., 1992;Yonezawa et al., 1992). The association of the enzyme with activated growth factor receptors may localize it to the plasma membrane, where its phospholipid substrates reside.
Recently, a number of genes have been identified that are similar to a large, central portion of p110, but which differ from p110 at their N and C termini (Stoyanov et al., 1995;Zhou et al., 1995). The product of one such gene, p110␥, can be activated in vitro by either the ␣ or ␤␥ subunits of heterotrimeric G proteins (Stoyanov et al., 1995). This observation suggests that different classes of receptors may utilize distinct mechanisms to activate p110-related enzymes (Stoyanov et al., 1995).
The stimulation of cells with growth factors or tumor antigens results in a rapid increase in the cellular concentrations of PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 from a low basal level, which suggests that these molecules may represent novel second messengers. Several protein kinases have recently been described which may represent downstream effectors of the lipid products of PtdIns 3-kinases. The Akt protein kinase can be activated in vivo by treating cells with the mitogen platelet-derived growth factor, and the binding of PtdIns 3-kinase to the platelet-derived growth factor receptor is essential for Akt activation (Franke et al., 1995). Akt can be directly activated in vitro with PtdIns-3-P, although the effects of PtdIns(3,4)P 2 and PtdIns-(3,4,5)P 3 on Akt activity are not known. In addition to Akt, particular isoforms of protein kinase C can be activated by the lipid products of PtdIns 3-kinases. Protein kinase C isoforms ␦, ⑀, , and can be activated in vitro with PtdIns(3,4)P 2 or PtdIns(3,4,5)P 3 , but not with PtdIns-3-P (Nakanishi et al., 1993;Toker et al., 1994). Therefore, it is possible that Akt or isoforms of protein kinase C may directly mediate PtdIns 3-kinase signaling by responding to increases in the cellular concentration of PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 .
A PtdIns 3-kinase has also been implicated in the regulation of intracellular protein sorting. Vps34 was initially identified as Saccharomyces cerevisiae mutant defective in the trafficking of proteins to the vacuole (Herman and Emr, 1990). Vps34 can phosphorylate PtdIns, but not PtdIns-4-P or PtdIns(4,5)P 2 in vitro, which is consistent with absence of detectable PtdIns(3,4,5)P 3 in yeast . Vps34 is the major PtdIns 3-kinase in yeast; VPS34 mutant strains contain no detectable PtdIns-3-P .
There is evidence for the existence of PtdIns 3-kinase activities distinct from p110/p85 and Vps34, and for this reason we have sought to identify novel PtdIns 3-kinases (reviewed in Stephens et al. (1993)). We have isolated and characterized a novel PtdIns 3-kinase from Drosophila melanogaster that can phosphorylate both PtdIns and PtdIns-4-P in vitro. We have designated this gene cpk (cpk: C2 containing PtdIns kinase) because it contains a "C2 domain." We have also identified a mouse gene (named cpk-m) that contains a C2 domain and shares extensive sequence identity with cpk. Therefore, the cpk and cpk-m represent a new class of PtdIns 3-kinases. Polymerase chain reaction products of approximately 400 base pairs were recovered and sequencing revealed open reading frames with sequence identity to p110 PtdIns 3-kinases. These DNA fragments were then used as probes to screen cDNA libraries. Large cDNAs, which did not contain the 5Ј ends of the cpk cDNAs, were recovered from the Drosophila and murine libraries. The 5Ј ends of the cDNAs were extended using a 5Ј rapid amplification of cDNA ends kit (Life Technologies, Inc.). Construction of the Drosophila cDNA library from 4 -8 h Drosophila embryos has been previously described (Brown and Kafatos, 1988). The murine cDNA libraries used were random and oligo(dT)primed mouse brain and mouse liver libraries purchased from Clontech. Standard procedures for cloning were used (Sambrook et al., 1989). The sequence of DNA was determined using an A.L.F. DNA Sequencer (Pharmacia).
Generation of ␣-Cpk Sera-A fragment of the Drosophila Cpk protein was expressed in the Escherichia coli strain BL21DE3(lysS) as a hexahistidine fusion protein. Briefly, the Drosophila cpk cDNA was digested with NcoI (1892) and HpaI (4157) and the resulting 2265-base pair fragment was ligated into the NcoI and Ecl136II sites of pet23d (Novagen). This construct drives expression of an 85-kDa Cpk hexahistidine fusion protein, named Pet.1. This protein was found to reside com-pletely in inclusion bodies. Therefore, inclusion bodies were purified, solubilized in 1 ϫ Laemmli sample buffer, and electrophoresed on a 8% preparative gel (Sambrook et al., 1989). Pet.1 protein was eluted from a gel slice, as described previously (Jessus and Beach, 1992) and then used to immunize rabbits (Berkeley Antibody Co.). The polyclonal rabbit serum, designated ␣-Cpk, was purified on an affinity column. The affinity column was prepared by coupling 2 mg of Pet.1 protein to an Affi-Gel 10 solid support, according to the manufacturer's instructions (Bio-Rad). This antigen column was used to immunoaffinity purify ␣-Cpk serum, as described previously (Harlow and Lane, 1988). The affinity purified serum was then incubated with whole cell BL21DE3(lysS) lysates that had been immobilized to a PVDF membrane (Millipore). In this manner, antibodies to E. coli proteins that coelute with Pet.1 from the gel slice were eliminated. Preimmune serum was similarly treated, for use as a control. Polyclonal ␣-peptide serum was generated by immunizing rabbits with the P6 peptide (NH 2 CRQD-FLSQPSTSSSQY COOH), which corresponds to amino acids 419 -434 of the Cpk protein. The P6 peptide was conjugated to the carrier and then used to immunize rabbits (Berkeley Antibody Co.). The Pet.1 protein was also used to immunize mice which were subsequently used to produce the monoclonal cell line ␣-Cpk.m1 (Program of Excellence, Monoclonal Antibody Facility).
Preparation of Drosophila Lysates and Immunochemical Assays-Lysates were prepared by Dounce homogenizing 0 -12 h Drosophila embryos in lysis buffer (lysis buffer: 20 mM HEPES pH 7.5, 150 mM sodium chloride, 2 mM EDTA, 10 mM sodium fluoride, 10 mM sodium phosphate (pH 7.5), 10 mM tetrasodium pyrophosphate, 10 mM sodium orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 10% glycerol, 10 trypsin inhibiting units/ml aprotinin, and 20 M leupeptin). The lysates were then frozen in aliquots at Ϫ70°C. Immediately prior to use, an aliquot was thawed, diluted with lysis buffer containing 1% Triton X-100, and the insoluble proteins were pelleted in a microcentrifuge. Cpk protein was detected by immunoblotting in the following manner. Proteins from lysates were resolved by SDS-PAGE on a 6% gel and then transferred to a PVDF membrane (Millipore) using high molecular weight transfer buffer (Harlow and Lane, 1988). The blots were incubated with the appropriate serum diluted in TBS-T (TBS-T: 50 mM Tris pH 8.0, 150 mM sodium chloride, and 0.1% Tween 20) containing 5% dry milk and 1% ovalbumin and then processed using an enhanced chemiluminescence kit (Amersham). A silver staining kit was used to visualize proteins in immunprecipitates (Accurate Chemical and Scientific Corp.).
PtdIns Kinase Assays-Cpk protein was precipitated from either COS-7 cell or Drosophila lysates as described previously (Harlow and Lane, 1988). The precipitations were washed four times in lysis buffer containing 1.0% Triton X-100 and then twice in PtdIns kinase assay buffer (PtdIns kinase assay buffer: 30 mM HEPES pH 7.5, 30 mM magnesium chloride). PtdIns kinase assays in which PtdIns was used as the substrate were performed as described previously (Kaplan et al., 1987;Whitman et al., 1988). The PtdIns kinase assays were modified in the following manner for the determination of PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 substrate specificities. The PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 lipid substrates were mixed with an equal amount of phosphatidylserine and then sonicated to form vesicles. Preparation of vesicles with phosphatidylserine assures that the physical properties of the PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 vesicles are approximately equivalent. The products of these kinase assays were resolved by TLC using Silica Gel 60 plates (Whatman) in a buffer consisting of chloroform:actetone:methanol:acetic acid:water (80:30:26:24:14). Cpk kinase assays were further modified by the addition of 0.05% CHAPS to the vesicle substrates and the PtdIns kinase assay buffer. The addition of CHAPS was determined to stimulate Cpk PtdIns kinase activity in vitro.
Determination of the Position on the Inositol Ring Phosphorylated by Cpk-PtdIns-3-P and PtdIns-4-P were resolved using TLC with a borate buffer system that has previously been described in detail (Walsh et al., 1991). [␥-32 P]PtdIns-3-P and [␥-32 P]PtdIns-4-P standards were generated in the following manner. [␥-32 P]PtdIns-3-P was produced by phosphorylating PtdIns with [␥-32 P]ATP using a constitutively active p110 mutant protein (p110*) whose construction and expression has been previously described (Hu et al., 1995). [␥-32 P]PtdIns-4-P was produced by phosphorylating PtdIns with [␥-32 P]ATP with lysates (20 g) prepared from 0 to 12-h Drosophila embryos. PtdIns 4-kinases are generally the most abundant PtdIns kinases found in lysates. In order to verify that the major product of this reaction is indeed PtdIns-4-P, the reaction products were demonstrated to comigrate with an unlabeled PtdIns-4-P standard (Sigma), but could be resolved from the [␥-32 P]PtdIns-3-P standard. The unlabeled PtdIns-4-P standard was visualized by iodine staining.
Expression of Proteins in COS-7 Cells-A plasmid was constructed that expressed Cpk-HA fusion proteins in COS-7 cells. NotI and SmaI sites were introduced into the cpk cDNA at the position of the stop codon using the primer dPIK 34 (dPIK 34: 5Ј-CCCCGGGTCAGCGGC-CGCCGTTCCTGGACACCGCGCCCAG-3Ј), which corresponds to nucleotides 5755-5795 of the cDNA. A triple tandem copy of the HA1 epitope on a NotI DNA fragment was ligated into the NotI site (Tyers et al., 1992). An SpeI site was introduced at the position of the initiating methionine using the primer dPIK 29 (dPIK29: 5Ј-TTAGACGAGAC-TAGTATGTCAAATCAAGCG-3Ј), which corresponds to nucleotides 132-162 of the cpk cDNA. The resulting 5683-base pair SpeI/SmaI fragment was ligated into the XbaI/SmaI sites of the mammalian expression vector pCG. pCG is a derivative of pEVRF (Matthias et al., 1989) with a modified polylinker that contains the human cytomegalovirus enhancer/promoter region and the translation initiation region of the herpes simplex virus thymidine kinase gene. A kinase-deficient Cpk protein was constructed by changing lysine 1347 to arginine with the primer dPIK 27 (dPIK27: 5Ј-GTGGGACCTGATGCCGAATCTTTAC-CGGCTATCTTTAGGTGCGGA-3Ј). A constitutively active p110 mutant protein (p110*) was expressed as a control, and its construction and expression has been previously described (Hu et al., 1995).
GenBank Accession Numbers for cpk and cpk-m cDNAs-The accession numbers for the nucleotide sequence of cpk and cpk-m sequences are U52192 and U52193, respectively.

RESULTS
Isolation of the Drosophila and Murine cpk cDNAs-We identified both the Drosophila and murine cpk genes by polym-erase chain reaction amplification of cDNA libraries using degenerate primers corresponding to conserved amino acids in PtdIns kinase catalytic domains. DNA fragments of the expected size were obtained and analysis of the DNA sequences revealed open reading frames with a high degree of sequence identity to the catalytic domains of PtdIns kinases. These DNA fragments were used as probes to screen Drosophila and murine libraries and multiple large cDNA clones were obtained. These clones did not contain the 5Ј end of the Drosophila and murine cDNAs, and therefore the 5Ј ends were extended using anchored polymerase chain reaction (5Ј rapid amplification of cDNA ends). The Drosophila cpk cDNA is likely to be fulllength for the following reasons. The size of the cDNA (6.9 kilobase pairs) is consistent with the size of the mRNA as estimated by Northern blot analysis (data not shown). Conceptual translation of the cDNA revealed a large open reading frame encoding a protein with a predicted molecular mass of 210 kDa. The first methionine in this open reading frame is encoded by the first ATG in the cDNA and it is preceded by an in-frame stop codon. The Drosophila and murine Cpk proteins are 34% identical and 48% similar (Fig. 1).
Analysis of the sequence of the Drosophila and murine Cpk proteins revealed that they are significantly more similar to the p110 family of PtdIns 3-kinases than to other families of PtdIns kinases. The Cpk genes are 31% identical and 43% similar to a large central region of p110 (Hiles et al., 1992). This region includes both the PtdIns kinase domain ( Fig. 2A, black box) and an adjacent region in which p110 and Cpk can be distinguished from other PtdIns kinases ( Fig. 2A, striped box). The Cpk proteins are also more similar to p110 family of PtdIns kinases in the most conserved portion of the catalytic domain (Fig. 2C). In this region, the Cpk proteins share approximately 45-50% identity with p110␣, p110␤, or p110␥. In contrast, the Cpk proteins share 35, 29, and 26% identity in this region with the Vps34, Pik1, and Tor2 PtdIns kinases, respectively.
The Cpk proteins differ from p110 at both their N and C termini, and therefore probably do not represent p110 homologues. The N termini of the Cpk proteins do not contain any recognizable domain, including the domain at the N terminus of the p110 that is responsible for binding to the p85 adapter molecule (Holt et al., 1994;Klippel et al., 1994). The N termini of the Cpk proteins are rich in the amino acids serine, glutamine, and proline, and the significance of this is unknown. The C termini of Cpk and Cpk-m contain a C2 domain ( Fig. 2A). C2 domains are found in a diverse group of proteins and may mediate binding to phospholipids or other proteins. The C2 domains in Cpk and Cpk-m are 52% similar to each other, and approximately 38% similar to C2 domains present in protein kinase C, synaptotagmin, or rabphilin (Fig. 2B). The possible significance of the C2 domains in Cpk and Cpk-m is discussed below.
Detection of the Drosophila Cpk Protein in Lysates-Polyclonal sera that recognize the Drosophila Cpk were generated to characterize the biochemical properties of the Cpk. A portion of the Drosophila Cpk was produced as a fusion protein in E.  (Hu et al., 1993;Stoyanov et al., 1995;Volina et al., 1994), Tor2 (Kunz et al., 1993), Pik1 (Flanagan et al., 1993) and Vps34 (Herman and Emr, 1990) PtdIns kinases. The PtdIns kinase domains are depicted in black, a region in which Cpk and p110 are related is depicted with stripes, and the C2 domain and p85 binding domains are indicated. B, comparison of the amino acid sequence of the C2 domains in Drosophila and murine Cpk with the C2 domains in rabphilin (Shirataki et al., 1993), synaptotagmin II (Geppert et al., 1991), and protein kinase C (Schaeffer et al., 1989). Rabphilin, synaptotagmin II, and protein kinase C are indicated by Rab, Syt II, and Pkc, respectively. Both conserved and identical residues are shaded. The following groups of amino acids were considered to be conserved: Ala, Val, Leu, Ile, Met; Asp, Glu; Lys, Arg; Asn, Gln; Phe, Tyr; Ser, Thr. C, comparison of the amino acid sequences of a part of the catalytic domains of Cpk and Cpk-m with the catalytic domains of p110␣ (Volina et al., 1994), p110␤ (Hu et al., 1993), p110␥ (Stoyanov et al., 1995), Vps34 (Herman and Emr, 1990), Pik1 (Flanagan et al., 1993), and Tor2 (Kunz et al., 1993). Only identical amino acids are shaded. The amino acid numbers are indicated to the right of the sequence. Sequence similarity of the Cpk proteins with other proteins was obtained by searching data bases with the program FASTA (Pearson and Lipman, 1988). coli and used to immunize rabbits. The resulting polyclonal serum (named ␣-Cpk) was affinity purified and used to probe an immunoblot of lysates prepared from 0 to 12-h Drosophila embryos. The ␣-Cpk immune serum recognizes a polypeptide migrating at approximately 210 kDa, the predicted molecular mass of the Cpk protein (Fig. 3A). p210 is not recognized by ␣-Cpk preimmune serum and is likely to be the Cpk gene product.
To verify that p210 is the Cpk gene product, we have generated an independent serum that recognizes Cpk protein. Rabbits were immunized with a peptide (P6) corresponding to a region of Cpk not present in the Cpk fusion protein described above. The resulting serum, which was designated ␣-P6, was used to precipitate protein from Drosophila lysates. The precipitates were resolved by SDS-PAGE, transferred to a PVDF membrane, and probed with ␣-Cpk serum (Fig. 3B). Both the ␣-Cpk and ␣-P6 immune sera precipitated a 210-kDa polypeptide that was recognized by ␣-Cpk serum. p210 was not detected in control precipitates using preimmune sera. Therefore, it is likely that p210 represents the Drosophila Cpk gene product.

Drosophila Cpk Has PtdIns 3-Kinase Activity-Since
Cpk is related to PtdIns kinases, it was of interest to determine whether it could phosphorylate PtdIns, and whether this phosphorylation occurred on the D3 or D4 position of the inositol ring. We have determined that Cpk protein precipitated from lysates prepared from Drosophila embryos can phosphorylate PtdIns (Fig. 4A). Cpk protein was precipitated from lysates using the ␣-Cpk and ␣-P6 sera. Half of each precipitate was assayed for PtdIns kinase activity and the other half was used for the detection of Cpk protein on an immunoblot. We detected a PtdIns kinase activity in precipitates using the ␣-Cpk and ␣-P6 immune sera, but not in precipitates using preimmune sera. The PtdIns kinase activity was competed by preincubating the ␣-Cpk serum with the Cpk fusion protein or the ␣-P6 serum with the P6 peptide. Therefore, Cpk protein precipitated from Drosophila lysates has a PtdIns kinase activity.
It is possible that the activity detected in the ␣-Cpk and ␣-P6 precipitates results from a PtdIns kinase coprecipitating with Cpk, rather than from Cpk protein itself. We have determined independently that Cpk can phosphorylate PtdIns by assaying the activity of Cpk protein obtained by exogenous expression in COS-7 cells (Fig. 4B). Wild-type and kinase-deficient Cpk proteins were tagged with an HA epitope and then expressed in COS-7 cells. A kinase-deficient Cpk mutant was constructed by changing a conserved lysine in the catalytic domain to arginine. Wild type and mutant Cpk proteins were precipitated from COS-7 cell lysates. One-half of each precipitate was assayed for PtdIns kinase activity and the other half was used for the detection of Cpk protein on an immunoblot. Precipitates containing wild-type Cpk protein contained a PtdIns kinase activity, while precipitates containing an equivalent amount of the mutant Cpk protein contained approximately 6-fold less activity (Fig. 4B). These data demonstrate that Cpk protein possesses intrinsic PtdIns kinase activity. Interestingly, the specific activity of Cpk purified from Drosophila lysates was significantly higher (approximately 10-fold) than that of Cpk derived from exogenous expression in COS-7 cells (data not shown). It is possible that COS-7 cells lack regulatory molecules required for full activation of Cpk kinase activity.
We have characterized the drug sensitivity and divalent cation requirement of the Cpk PtdIns kinase activity relative to p110. p110 PtdIns 3-kinase activity is sensitive to wortmannin, a fungal metabolite that has been shown to be a selective inhibitor of PtdIns kinases. In vitro, the wortmannin sensitivity of the Cpk PtdIns activity is similar to that of p110. The IC 50 (half-maximal inhibition) value for p110 was determined to be FIG. 3. Detection of Cpk protein in lysates prepared from Drosophila embryos. A, immunoblot of lysates (30 g) prepared from 0 to 12-h Drosophila embryos probed with ␣-Cpk polyclonal serum. A polypeptide migrating at approximately 210 kDa, the predicted molecular mass of the Cpk protein, is present in blots probed with the immune serum, but absent in blots probed with the preimmune serum. B, precipitation of Cpk protein from embryo lysates. Lysates prepared from 0 to 12-h Drosophila embryos were precipitated with ␣-Cpk preimmune (lane 1), ␣-Cpk immune (lane 2), ␣-P6 preimmune (lane 3), and ␣-P6 immune sera (lane 4). The precipitates were resolved by SDS-PAGE, transferred to a PVDF membrane, and probed with the ␣-Cpk serum. Both ␣-Cpk or ␣-P6 immune sera precipitate a 210-kDa protein that represents the Cpk gene product. The ␣-Cpk serum also recognizes a 95-kDa protein on an immunoblot. We do not believe that this protein is an antigenically related PtdIns kinase, since it is not precipitated by the ␣-Cpk serum (lane 2). Preimmune and immune sera are indicated by P and I, respectively. Lysates were also precipitated with ␣-Cpk serum which had been preincubated with Cpk protein (lane 3) or ␣-P6 serum which had been preincubated with the P6 peptide (lane 6). The precipitates were split in half; half of each precipitate was assayed for PtdIns kinase activity, and the other half was used to detect Cpk protein by probing an immunoblot with ␣-Cpk serum. B, wild-type (lane 1) and kinase deficient (lane 2) Cpk proteins which had been tagged with an HA epitope were expressed in COS-7 cells. The Cpk proteins were precipitated from lysates using an ␣-HA serum. Half of each precipitate was assayed for PtdIns kinase activity and the other half was used to detect Cpk protein by probing an immunoblot with ␣-Cpk serum. P and I indicate preimmune and immune sera, respectively. I c represents either ␣-Cpk or ␣-P6 sera which had been preincubated for 10 min with 0.5 g of competitor. The ␣-Cpk immune serum had been preincubated with Cpk protein and the ␣-P6 immune serum had been preincubated with P6 peptide. 7.5 nM, which is a value consistent with previously studies (Woscholski et al., 1994). The IC 50 for wortmannin inhibition of Cpk was 11 nM (data not shown). Also, p110 requires the addition of either Mg 2ϩ or Mn 2ϩ to in vitro kinase assays, although the enzyme is more active in the presence of Mg 2ϩ . In contrast, Cpk strictly requires the presence of Mg 2ϩ in in vitro kinase assays, and the enzyme is inactive in the presence of Mn 2ϩ (data not shown).
We have examined whether Cpk phosphorylates PtdIns at the same position of the inositol ring as p110, the D3 position (Fig. 5). PtdIns-3-P and PtdIns-4-P can be resolved by TLC using a borate buffer (Walsh et al., 1991). Cpk protein was precipitated from either Drosophila (Fig. 5A) or COS-7 cell (Fig. 5B) lysates and used to phosphorylate PtdIns. The ␥-32 Plabeled products of these reactions were separated by TLC. The Cpk products migrated at the position of a ␥-32 P-labeled PtdIns-3-P standard, but not a PtdIns-4-P standard. The Cpk products were then mixed with either PtdIns-3-P or PtdIns-4-P standards and the mixtures were resolved by TLC. The lipid products of Cpk comigrated with the PtdIns-3-P standard, but not with the PtdIns-4-P standard, suggesting that the Cpk reaction products are PtdIns-3-P and that Cpk is a PtdIns 3-kinase.
Cpk Can Phosphorylate PtdIns and PtdIns-4-P, but Not PtdIns(4,5)P 2 in Vitro-PtdIns 3-kinases have distinct substrate specificities in vitro; Vps34 can only phosphorylate PtdIns, while p110/p85 can phosphorylate PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 . We determined the lipid substrate specificity of Cpk. Cpk protein was precipitated from Drosophila lysates and this protein phosphorylated PtdIns and PtdIns-4-P, but not PtdIns(4,5)P 2 using in vitro kinase assays (Fig. 6). Cpk is the first PtdIns 3-kinase identified with this particular substrate specificity. A constitutively active p110 (p110*, Hu et al. (1995)) was used as a control and it phosphorylated PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 . We have also determined that wild-type Cpk protein obtained from exogenous expression in COS-7 cells displayed the same substrate specificity as protein derived from Drosophila lysates (data not shown). A kinasedeficient Cpk protein obtained from exogenous expression in COS-7 cells served as a control and was unable to phosphorylate any of these substrates.
Identification of Potential Cpk-binding Proteins-The identity of Cpk-binding proteins may provide information concerning the role of Cpk in vivo. We have generated a monoclonal ␣-serum (␣-Cpk.m1) which can specifically recognize Cpk for the purpose of identifying putative Cpk interacting proteins. ␣-Cpk.m1 recognizes Cpk on an immunoblot of lysates prepared from 0 to 12-h Drosophila embryos (Fig. 7A). Cpk protein was precipitated from Drosophila lysates using ␣-Cpk.m1, the precipitates were resolved by SDS-PAGE, and proteins were visualized by silver staining (Fig. 7B). In addition to Cpk, two proteins with apparent molecular sizes of 90 and 190 kDa could clearly be visualized. p90 and p190 were not recognized directly by ␣-Cpk.m1 on an immunoblot, which suggests that these proteins are not fragments of Cpk (data not shown). In addition, the observation that ␣-Cpk.m1 does not recognize p90 or p190 on a immunoblot suggests that ␣-Cpk.m1 is not directly precipitating these proteins from lysates (Fig. 7A).
We investigated the possibility that either Cpk or any coprecipitating proteins are tyrosine phosphorylated. The p110/p85 PtdIns 3-kinase interacts with activated tyrosine kinase growth factor receptors, and both p85 and p110 are subsequently phosphorylated on tyrosine (reviewed in Stephens et al. (1993)). Cpk was precipitated from Drosophila lysates, the precipitated proteins were resolved by SDS-PAGE, transferred to a PVDF membrane, and probed with an ␣-phosphotyrosine antiserum (Fig. 7C). The Cpk precipitations contained two ␣-phosphotyrosine reactive polypeptides migrating at 190 and 210 kDa. This finding suggests that both Cpk and p190 may be tyrosine phosphorylated and raises the possibility that Cpk function may be regulated by tyrosine kinases. Cpk protein was precipitated from Drosophila embryo lysates and the precipitate was divided into three aliquots. One aliquot was used to phosphorylate either PtdIns (lane 2), PtdIns-4-P (lane 4), or PtdIns(4,5)P 2 (lane 6) using in vitro kinase assays. The reaction products were separated by TLC. A constitutively active p110 protein (p110*) served as a control and it was able to phosphorylate PtdIns (lane 1), PtdIns-4-P (lane 3), and PtdIns(4,5)P 2 (lane 5) substrates. PIP, PIP 2 , and PIP 3 refer to phosphatidylinositol phosphate, phosphatidylinositol bisphosphate, and phosphatidylinositol trisphosphate, respectively. DISCUSSION We report the identification of a novel class of PtdIns 3-kinases whose members contain C-terminal C2 domains. We have isolated Drosophila and murine genes (termed cpk and cpk-m, respectively) whose products share 34% sequence identity and 48% sequence similarity. The finding of a related genes in vertebrates and invertebrates suggests that Cpk may define a new class of conserved PtdIns 3-kinase and that Cpk regulatory pathways may also be conserved. In a recent independent study, MacDougall et al. (1995) also report the identification of this gene.
Cpk and Cpk-m are more closely related to the p110 family of PtdIns 3-kinases than to other families of PtdIns kinases. Amino acids 863-1587 of Cpk are approximately 31% identical and 43% similar to a large, central region of mammalian p110 (Hiles et al., 1992). Cpk and Cpk-m differ from p110 at their N and C termini and therefore probably do not represent p110 homologues. The N termini of the Cpk proteins do not contain any recognizable domain, and do not resemble a domain at the N terminus of p110 that mediates its interaction with the p85 subunit (Holt et al., 1994;Klippel et al., 1994). Indeed, we have determined that Cpk does not form a complex with p85 when both proteins are coexpressed in COS-7 cells (data not shown).
The Drosophila and murine Cpk proteins contain a C2 domain at their C termini. C2 domains are found in a diverse group of proteins, including some isoforms of protein kinase C, synaptotagmin, and rabphilin (reviewed in Nishizuka (1992), Perin et al. (1990), and Shirataki et al. (1993)). C2 domains were initially described as Ca 2ϩ -dependent phospholipid binding motifs that can mediate the translocation of proteins to membranes (reviewed in Nishizuka (1992)). Indeed, MacDougall et al. (1995) demonstrate that recombinant C2 domain derived from their PtdIns kinase clone (termed 68D) can weakly bind to phospholipids. In addition to binding phospholipids, C2 domains can also mediate interactions between proteins. A C2 domain from synaptotagmin, a synaptic vesicle protein, can bind AP-2 complexes (AP-2: adapter protein 2) with a high affinity (Zhang et al., 1994). AP-2 is a protein complex which is believed to play a role in the assembly of clathrin-coated pits and may target proteins to coated pits for recycling. Also, a C2 domain from protein kinase C can mediate its interaction with RACKs (RACKs: receptors for activated C kinases, reviewed in Liscovitch and Cantley (1994)). RACKs have been implicated in the regulation of the kinase activity and intracellular localization of protein kinase C (reviewed in Mochly-Rosen (1995)). 2 Therefore, it is possible that the C2 domain in Cpk may function as a regulatory domain by binding lipids or proteins which modulate its intracellular localization or enzymatic activity.
We have characterized the enzymatic activity of the Cpk. PtdIns kinases phosphorylate PtdIns at either the D3 or D4 hydroxyl group of the inositol ring. Cpk can phosphorylate PtdIns at the D3 position of the inositol ring, in a manner similar to p110 PtdIns 3-kinases. We have also determined the substrate specificity of the Cpk. p110 can phosphorylate PtdIns, PtdIns-4-P, and PtdIns(4,5)P 2 in vitro. In contrast to p110, Cpk can phosphorylate PtdIns and PtdIns-4-P, but not PtdIns(4,5)P 2 in vitro. Cpk is the first PtdIns 3-kinase identified with this particular substrate specificity. The kinase assays were performed using Cpk derived from two sources, native protein purified from Drosophila lysates and recombinant protein purified from COS-7 cells. Also, we have determined that the PtdIns 3-kinase activity associated with Cpk is intrinsic by demonstrating that point mutations within the catalytic domain of Cpk result in an inactive enzyme. MacDougall et al. (1995) also report a similar in vitro substrate specificity associated with recombinant 68D purified from insect cells. We do not know to what extent the substrate specificity determined using in vitro kinase assays reflects the substrate specificity in cells.
The treatment of cells with mitogenic stimuli results in the rapid accumulation of the products of PtdIns 3-kinases from a low basal level. This observation suggests that these lipids may represent novel second messengers, although candidate effectors have only recently been described. Three protein kinases whose activity can be regulated by the lipid products of a PtdIns 3-kinase have been described. The Akt protein kinase is specifically activated by PtdIns-3-P (Franke et al., 1995), while particular isoforms of protein kinase C are activated by PtdIns(3,4)P 2 or PtdIns(3,4,5)P 3 (Nakanishi et al., 1993;Toker et al., 1994). It is possible that Cpk acts in a pathway in which 2 D. Mochly-Rosen, personal communication. prepared from 0 to 12-h Drosophila embryos were probed with ␣-Cpk.m1 monoclonal serum. A polypeptide migrating at approximately 210 kDa represents cpk. B, visualization of 90-and 190-kDa proteins coprecipitating with Cpk. Drosophila lysates were precipitated with protein A beads alone (1) or beads in which ␣-Cpk.m1 had previously been immobilized (2). The precipitates were resolved on a 6% SDS-polyacrylamide gel and proteins were visualized by silver staining. In addition to Cpk, polypeptides of approximately 90 and 190 kDa can be specifically detected in ␣-Cpk precipitates. C, both Cpk and p190 may be tyrosine phosphorylated. Drosophila lysates were precipitated with protein A beads alone (1) or beads in which ␣-Cpk.m1 had previously been immobilized (2). The precipitates were resolved on a 6% SDS-polyacrylamide gel, transferred to a PVDF membrane, and probed with an ␣-phosphotyrosine serum (␣-PY). Polypeptides migrating at the molecular weight of Cpk and p190 are recognized by the ␣-phosphotyrosine serum.
its lipid products regulate a currently unidentified effector.
Several PtdIns 3-kinases are known to be associated with regulatory proteins. For example, the p110 PtdIns 3-kinase associates with an 85-kDa subunit which is essential for the proper localization and activation of p110. p85 binds to tyrosine-phosphorylated growth factor receptors, and this interaction may localize p110 to the plasma membrane, where its lipid substrates reside (Hu et al., 1992;Klippel et al., 1992;McGlade et al., 1992;Reedijk et al., 1992;Yoakim et al., 1992;Yonezawa et al., 1992). In addition to functioning as an adapter, p85 is required for the enzymatic activity of p110; p110 must be associated with p85 to be fully active (Holt et al., 1994;Klippel et al., 1994). The Vps34 PtdIns 3-kinase associates with the Vps15 protein kinase and this complex regulates intracellular protein sorting . Vps15 is required for both the membrane localization and the PtdIns kinase activity of Vps34 . Therefore, it is possible that Cpk associates with a protein that can modulate its intracellular localization or enzymatic activity. We have identified two potential Cpk interacting proteins, p90 and p190. Interestingly, both Cpk and p190 are tyrosine phosphorylated. The only other PtdIns kinase known to be tyrosine phosphorylated is p85/p110 and this phosphorylation occurs upon recruitment of the enzyme to activated growth factor receptors. It is possible that p190 represents a growth factor receptor, its substrate, or another protein that modulates the activity of Cpk.
Surprisingly little is known about the physiological role of PtdIns 3-kinases in a multicellular organism. We can now utilize the sophisticated genetic tools available in Drosophila to study Cpk. Drosophila is also an excellent system for a biochemical analysis of Cpk, since lysates that contain active Cpk enzyme are likely to contain Cpk regulators and associated proteins. The observation that Cpk may be tyrosine phosphorylated in vivo is intriguing, and the identification of the tyrosine kinase which phosphorylates it might provide information concerning the physiological role of Cpk.