Subtype-specific Translocation of Diacylglycerol Kinase α and γ and Its Correlation with Protein Kinase C*

We examined the translocation of diacylglycerol kinase (DGK) α and γ fused with green fluorescent protein in living Chinese hamster ovary K1 cells (CHO-K1) and investigated temporal and spatial correlations between DGK and protein kinase C (PKC) when both kinases are overexpressed. DGKα and γ were present throughout the cytoplasm of CHO-K1 cells. Tetradecanoylphorbol 13-acetate (TPA) induced irreversible translocation of DGKγ, but not DGKα, from the cytoplasm to the plasma membrane. The (TPA)-induced translocation of DGKγ was inhibited by the mutation of C1A but not C1B domain of DGKγ and was not inhibited by staurosporine. Arachidonic acid induced reversible translocation of DGKγ from the cytoplasm to the plasma membrane, whereas DGKα showed irreversible translocation to the plasma membrane and the Golgi network. Purinergic stimulation induced reversible translocation of both DGKγ and α to the plasma membrane. The timing of the ATP-induced translocation of DGKγ roughly coincided with that of PKCγ re-translocation from the membrane to the cytoplasm. Furthermore, re-translocation of PKCγ was obviously hastened by co-expression with DGKγ and was blocked by an inhibitor of DGK (R59022). These results indicate that DGK shows subtype-specific translocation depending on extracellular signals and suggest that PKC and DGK are orchestrated temporally and spatially in the signal transduction.

Diacylglycerol (DG) 1 is a second messenger regulating various cellular responses (1,2). One of the important roles of DG is the activation of protein kinase C (PKC) (1,3,4). Thus, DG is very important for regulation of PKC activity and cellular response. DG is produced physiologically as a result of the signal-induced hydrolysis of phosphatidylinositol by phospholipase C and also from phosphatidylcholine by phospholipase D. Generated DG is phosphorylated to phosphatidic acid by diacylglycerol kinase (DGK) or cleavage by DG lipase (2,5,6). DGK is an important enzyme for inactivating PKC by attenuation of the DG level, contributing to regulation of the cellular response. In addition, phosphatidic acid itself activates PKC and PLC␥1 (7,8) and modulates Ras GTPase-activating protein (9). DGKs have additional important functions for various cellular responses.
To date at least nine subtypes of mammalian DGKs have been cloned and divided into five groups based on structure (2). Generally, all DGKs have cysteine-rich regions homologous to the C1A and C1B motifs of PKCs in the regulatory domain at the N terminus of the protein and possess a conserved catalytic domain in the C terminus of the protein. Type I DGKs, including DGK␣, -␤, and -␥, have EF-hand motifs and two cysteinerich regions in the regulatory domain (10 -12). Type II DGKs such as DGK␦ and -, have a pleckstrin homology domain instead of the EF-hand motif in addition to two cysteine-rich regions (13,14). Interestingly, the catalytic domains of DGK␦ and -are separated. Type III, consisting of DGK⑀, has only two cysteine-rich regions in the regulatory domain. Type IV, DGK and -, has a unique motif similar to the myristoylated alanine-rich C-kinase substrate (MARCKS) phosphorylation site in the regulatory domain and four ankyrin repeats at its C terminus (15,16). The final group, type V, includes DGK, which has three cysteine-rich regions and a pleckstrin homology domain with an overlapping Ras-associating domain (17). In contrast to the accumulated knowledge of molecular structure, the functions and regulatory mechanism of each DGK subtype are still unknown.
Recently, we developed a system to monitor the translocation of PKCs fused with GFP in living cells (18). Using this system, we demonstrated that each subtype of PKC shows distinct translocation in response to various stimuli, suggesting that each subspecies has a spatially and temporally different targeting mechanism that depends on the extracellular and intracellular signals (19,20). Thus, we hypothesize that the subtype-specific targeting of PKC contributes to the subspeciesspecific functions. Interestingly, in addition to PKC, translocation of DGK by several stimuli such as phorbol ester (21,22), calcium (23), and carbachol (24) has been shown by analysis of immunoblots or of DGK activity, although previous studies concerning the translocation of DGKs have been performed using crude DGKs, and subtype-specific translocation could not be demonstrated. Furthermore, the involvement of PKC in the regulation of activity and localization of DGK␣ has been described (24 -27). These observations suggest that the translocation of DGKs is an effective process to regulate their activity and subtype-specific functions and that DGKs and PKCs are not only functionally but also temporally and spatially correlated in their signal transduction pathways. In this study, we examined translocations of DGK␣ and -␥ fused with GFP and investigated temporal and spatial correlations between DGK and PKC in addition to their functional correlation.

EXPERIMENTAL PROCEDURES
Materials-Adenosine triphosphate (ATP), uridine triphosphate (UTP), calcium ionophore A 23187, and tetradecanoylphorbol 13-acetate (TPA) were purchased from Sigma. Arachidonic acid and 1,2dioctanoylglycerol (DiC8) were purchased from Doosan Serdary Research Laboratories (Eaglewood Cliffs, NJ) and Biomol Research Laboratories (Plymouth Meeting, PA), respectively. DGK inhibitor (R59022) was a product of Research Biochemicals International. All the other chemicals used were of analytical grade. A plasmid including cDNA for a variant of blue fluorescence protein (BFP) was donated by Dr. Osumi (Himeji Institute of Technology, Japan) (28).
Cell Culture-COS-7 cells were purchased from the Riken cell bank (Tsukuba, Japan). The CHO-K1 cell strain was a gift from Dr. Nishijima (National Institute of Health, Tokyo, Japan). COS-7 cells were cultured in Dulbecco's modified Eagle's medium, and CHO-K1 cells were cultured in Ham's F-12 medium (Life Technologies, Inc.) at 37°C in a humidified atmosphere containing 5% CO 2 . Both media contained 25 mM glucose, and both were buffered with 44 mM NaHCO 3 and supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml). The fetal bovine serum used was not heat-inactivated.
Construct of Plasmids Encoding GFP-DGK␥ Fusion Protein-The plasmid bearing cDNA for pig DGK␣ (designated as BS 585) (10) was donated by Dr. Kanoh (Sapporo Medical University School of Medicine, Japan). A cDNA fragment of rat DGK␥ with a XhoI site in the 5Јterminus and an SmaI site in the 3Ј terminus was produced by a PCR with cDNA for rat DGK␥ (12) as the template. The sense and antisense primers were 5Ј-AACTCGAGAAGATGAGTGACGGGCAATGG-3Ј and 5Ј-TTCCCGGGAGTCCTTTGAACGGCTTTTCCT-3Ј, respectively. The PCR product of DGK␥ was first subcloned into a TA cloning vector, pCR TM 2.1 (Invitrogen, San Diego, CA). The plasmid was designated as BS465. The cDNA encoding DGK␥ in BS465 was digested with XhoI and SmaI then subcloned into the XhoI and SmaI site in pEGFPN1 (CLONTECH, Palo Alto, CA) or SalI and SmaI site in pEGFPC1 (CLONTECH) (BS470 and BS561, respectively). Similarly, a cDNA fragment of pig DGK␣ with an XhoI site in the 5Ј-terminus and a SmaI site in the 3Ј-terminus was produced by PCR with BS585 as the template and subcloned into a TA cloning vector, pCR TM 2.1 (BS597). Finally, the DGK␣ was subcloned into the SalI and SmaI site in pEG-FPC1 (BS606).
Site-directed Mutagenesis of GFP-DGK␥-Site-directed mutagenesis was performed according to the manufacturer's recommended protocol with ExSite PCR-based site-directed mutagenesis kit (Stratagene, La Jolla, CA), using BS465 as a template. The sense and antisense primers for producing a mutant DGK␥ whose Cys-285 in C1A domain was substituted to Gly (C1A mutant) were 5Ј-GCCACATCATGCTGAT-GGGC-3Ј and 5Ј-CAAAGTTGCAGTAAGTGGGCTT-3Ј, respectively. The primers for a mutant DGK␥, whose Cys-348 in the C1B region was substituted to Gly (C1B mutant), were 5Ј-GCCACAAAAGTATCAAGT-GCTAC-3Ј and 5Ј-CACGGTCACACTTGACGGA-3Ј. Mutagenesis was confirmed by verifying sequences. Each C1A or C1B mutant of DGK␥ cDNA was subcloned into SalI and SmaI sites in pEGFPC1, as in the case of BS561 (designated BS691 and BS692, respectively).
For immunoblotting, the samples were subjected to 7.5% SDS-polyacrylamide gel electrophoresis, followed by blotting onto a polyvinylidine difluoride membrane (Millipore, Bedford, MA). Nonspecific binding sites were blocked by incubation with 5% skim milk in 0.01 M phosphate-buffered saline, pH 7.4 (PBS) for 18 h at 4°C. The membrane was then incubated with anti-DGK␥ antibodies (diluted 1: 1000) (12) or anti-GFP antibody (CLONTECH) for 1 h at room temperature. After washing with 0.01 M PBS containing 0.03% Triton X-100 (PBS-T), the membrane was incubated with peroxidase-labeled anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) for 30 min. After three rinses with PBS-T, the immunoreactive bands were visualized using a chemiluminescence detection kit (ECL, Amersham Pharmacia Biotech).
To determine the kinase activity, appropriate volumes of the homogenate samples described above were subjected to octyl glucoside mixed-micelle assay (29). The amount of each sample subjected to kinase assay was determined based on the amount of DGKs or their fusion proteins assessed by immunoblotting. 1-Steroyl-2-arachidonoylsn-glycerol was used as a substrate. The radioactivity of phosphatidyl acid was separated on a 20-cm Silica Gel 60 (Merck) thin layer chromatography plate using a chloroform:methanol:acetic acid (65:15:5) solution and detected by BAS 2000 (Fujix, Tokyo, Japan).
Observation of Translocation of the GFP Fusion Proteins-Plasmids (approximately 5.5 g) were transfected into 1.0 ϫ 10 6 cells by lipofection using TransIT TM -20 (Mirus Co., Madison, WI) according to the manufacturer's standard protocol. After being cultured at 37°C for 16 -24 h, the cells were spread onto glass-bottom culture dishes (MatTek Corp., Ashland, MA). Experiments were performed 16 -48 h after the transfection.
The culture medium was replaced with HEPES buffer composed of 135 mM NaCl, 5.4 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM HEPES, 10 mM glucose, pH 7.3 (Ringer's solution). Translocation of the GFP fusion protein was triggered by a direct application of various stimulants at 10 times higher concentration into the Ringer's solution to obtain the appropriate final concentration.
The fluorescence of the fusion protein was monitored with a confocal laser-scanning fluorescent microscope (LSM 410 invert, Carl Zeiss, Jena, Germany) at 488-nm argon excitation using a 515-535-nm band pass barrier filter. All experiments were performed at 37°C.
For simultaneous observation of GFP-DGK␥ and PKC␥-GFP in the same field, a plasmid-bearing variant, BFP, was co-transfected with PKC␥-GFP into CHO-K1 cells as a marker for detecting PKC␥-GFPexpressing cells. The GFP-DGK␥-or PKC␥-GFP-expressing cells, which were individually transfected, were spread again and mixed into the same glass-bottomed culture dishes. The CHO-K1 cells expressing GFP-DGK␥ or PKC␥-GFP were determined based on the fluorescence of BFP. The fluorescence of the BFP was monitored with a confocal laserscanning fluorescent microscope (LSM 410 invert, Carl Zeiss, Jena, Germany) at 364-nm UV laser excitation using a 397-nm long pass barrier filter, whereas that of GFP was monitored at 488-nm argon excitation using a 515-535-nm band pass barrier filter. After confirming, translocation of GFP-DGK␥ or PKC␥-GFP was triggered and observed based on the fluorescence of GFP as described above.
Co-detection of the Golgi Network and GFP-DGK␣ Translocated by Stimulation Using Arachidonic Acid-Texas red-conjugated wheat germ agglutinin was used to monitor the Golgi network. After induction of the translocation of GFP-DGK␣ by 100 M arachidonic acid, the cells were fixed with 4% paraformaldehyde and 0.2% picric acid in 0.1 M PBS for 30 min. After washing 3 times with 0.1 M PBS, the fixed cells were treated with PBS containing 0.3% Triton X-100 and 10% normal goat serum for 20 min. The cells were then incubated with 0.5 g/ml Texas red-conjugated wheat germ agglutinin (Molecular Probes, Leiden, the Netherlands) in PBS-T for 30 min. Finally, the fluorescence of Texas red and GFP were observed under a confocal laser-scanning fluorescent microscope, the former at 588-nm argon excitation using a 590-nm long pass barrier filter and the latter at 488-nm argon excitation using a 510 -535-nm band pass barrier filter. Fig. 1A shows constructs of two fusion proteins of DGK␥ with GFP. GFP-DGK␥ possesses GFP at the N terminus of the protein, whereas GFP is located at the C terminus of DGK␥-GFP. Immunoblotting using the anti-DGK␥ antibody revealed that the molecular size of both fusion proteins is about 120 kDa, which is about 30 kDa larger than intact DGK␥ (Fig. 1B). Furthermore, no significant degraded products were detected, although nonspecific bands were observed in all lanes. Anti-GFP antibody also recognized the 120-kDa bands of DGK␥-GFP and GFP-DGK␥ but not intact DGK␥ or other proteins (data not shown). Fig. 1C shows that GFP-DGK␥ had kinase activity as significant as that of intact DGK␥, whereas no activity of DGK␥-GFP was detectable. Therefore, we produced only GFP-DGK␣. GFP-DGK␣ had reasonable molecular mass (about 110 kDa) and showed sufficient kinase activity (data not shown). In the following experiments, GFP-DGK␥ and GFP-DGK␣ were used.

Properties of the Fusion Protein of DGK␥ and -␣ with GFP-
Distinct Translocation of DGK␥ and DGK␣-When GFP-DGK␣ and -␥ were expressed in CHO-K1 cells, the fluorescence of GFP-DGK␥ and GFP-DGK␣ was observed throughout the cytoplasm and in the nucleus (Figs. 2 and 3). The expression level of GFP-DGK␥ and GFP-DGK␣ in the nucleus varied in the cells.
As a result of stimulation with 100 M ATP, DGK␥ in the cytoplasm was translocated to the plasma membrane within 0.5-1 min and returned to the cytoplasm within 5-10 min (Fig. 2, top row). As a result of stimulation with 100 M arachidonic acid, DGK␥ in the cytoplasm was translocated to the plasma membrane within 30 s and returned to cytoplasm within 12 min, which was similar to ATP-induced translocation (Fig. 2, middle row). However, these stimuli exerted no effect on DGK␥ in the nucleus. Activation by 1 M TPA induced obvious translocation of DGK␥ from the cytoplasm to the membrane (Fig. 2, bottom row). The TPA-induced translocation began within 30 s and remained on the plasma membrane for at least 60 min after the treatment. The TPA-induced translocation was not inhibited by staurosporine, an inhibitor for protein kinases (data not shown).
In contrast to DGK␥, the effects of these stimuli on the translocation of DGK␣ were quite distinct. TPA caused no translocation of DGK␣, unlike DGK␥, and ATP-induced translocation of DGK␣ occurred in 30 s and returned in 2 min, which meant that it occurred and was completed earlier than DGK␥ (Fig. 3, top and bottom rows). Arachidonic acid at 100 M translocated DGK␣ in the cytoplasm to the perinuclear area in addition to the plasma membrane (Fig. 3, middle row). Furthermore, DGK␣ in the nucleus was also affected by arachidonic acid. DGK␣ showed dot-like accumulation within the nucleus 3 min after arachidonic acid stimulation (Fig. 3, middle). Occasionally, DGK␣ was translocated from the nucleoplasm to the nuclear membrane during the late phase of the arachidonic acid-induced translocation (data not shown).
Target Site of GFP-DGK␣ on Stimulation with Arachidonic Acid-To identify the intracellular compartment in which GFP-DGK␣ accumulated in response to arachidonic acid, the Golgi network was visualized by Texas red-conjugated wheat germ agglutinin in the CHO-K1 cells expressing GFP-DGK␣ after arachidonic acid treatment. Intense GFP fluorescence was present on the plasma membrane and perinuclear area (Fig. 4, left). Fluorescence of Texas red was present around the nucleus, indicating the Golgi network and also the plasma membrane due to glycosylated transmembrane proteins (Fig. 4,  center). An overlapping image shows that the fluorescence of Texas-red and of GFP were co-localized in the perinuclear region (Fig. 4, right).
Importance of C1A Domain in the TPA-induced Translocation of DGK␥-Two mutants of DGK␥ fused with GFP were produced. In one mutant, Cys-285 in the C1A region was replaced by Gly (C1A mutant), and in the other mutant, Cys-348 in the C1B region was substituted with Gly (C1B mutant). Both mutants showed unique localization when expressed in CHO-K1 cells. Unlike intact DGK␥, in almost of the cells, intense fluorescence was present throughout the cytoplasm with faint fluorescence in the nucleus (Fig. 5). TPA at 1 M induced irreversible translocation of the C1B mutant similar to that of intact DGK␥. In contrast, the C1A mutant was not translocated in response to TPA.
Comparison of the ATP-induced Translocation of DGK␥ and PKC␥-For the purpose of clarifying spatial and temporal correlations between DGK␥ and PKC␥, their ATP-induced translocations were compared. To distinguish DGK␥ and PKC␥, both cDNA of PKC␥-GFP and variant BFP were simultaneously transfected into the CHO-K1 cells and spread into a glassbottomed dish together with the CHO-K1 cells that were separately transfected with GFP-DGK␥ alone. Accordingly, blue with slight green fluorescence and bright green fluorescence in Fig. 6 showed the CHO-K1 cells expressing PKC␥ and DGK␥, respectively. Intense fluorescence of PKC␥-GFP was observed throughout the cytoplasm with faint fluorescent in the nucleus. As a result of stimulation with 100 M ATP, PKC␥ was translocated from the cytoplasm to the plasma membrane at 10 s after the stimulation, and translocation was most significant at 30 s. Then, PKC␥ was re-translocated within 1 min. In contrast to PKC␥, the translocation of DGK␥ was initiated around 30 s after the stimulation and was seen most significantly at 2 min. Finally, DGK␥ was re-translocated at 5 min. However, there was no significant difference in the time course of ATP-induced translocations of GFP-DGK␣ and PKC␥-GFP (data not shown).
Functional Correlation between DGK␥ and PKC␥-Effects of co-expression of DGK␥ and of a DGK inhibitor on the translocation of PKC␥-GFP were investigated. Fig. 7 shows the translocation of PKC␥-GFP in different conditions. In normal condi-

FIG. 2. Translocation of GFP-DGK␥.
Top row, application of 100 M ATP induced a translocation of GFP-DGK␥ from the cytoplasm to the plasma membrane. The translocation was observed within 1 min after the stimulation. Thereafter, the GFP-DGK␥ was gradually re-translocated from the membrane to the cytoplasm and was mostly restored to a state similar to that before the stimulation within 12 min. Bar, 10 m. tions, after simultaneous application of 10 M DiC8, and 1 M calcium ionophore A23187, PKC␥-GFP was translocated from the cytoplasm to the plasma membrane within 2 min and restored within 8 min (Fig. 7B). When co-expressed with DGK␥, the same stimulation induced translocation of PKC␥-GFP within 20 s, but PKC␥-GFP was quickly re-translocated from the membrane to the cytoplasm, which was completed by 2 min (Fig. 7A). On the other hand, in the presence of R59022 (a DGK inhibitor), re-translocation of PKC␥-GFP was significantly inhibited, although the initiation of the translocation was not altered (Fig. 7C). Inhibitors for PKC did not alter the translocation of DGK␥ induced by ATP, arachidonic acid, or TPA (data not shown).

DISCUSSION
It has been shown that both PKC␥ and DGK␥ are abundantly expressed in Purkinje cells (12,30) and possess not only cysteine-rich regions but also calcium-sensitive domains (2). These findings suggest a functional correlation between this DG-dependent kinase and DG-catalyzing kinase. DGK␣ has a very similar enzymological character to that of DGK␥ but shows glial expression in the brain (11), suggesting that each DGK subtype has a specific function. Thus, we chose DGK␣ and -␥ to study the different functions among many DGK subtypes.
Although GFP is a useful tool as a marker protein, it should be verified that each GFP fusion protein has the same biological properties as its native protein. Especially, to judge the movement of the fusion protein by GFP fluorescence, it is necessary to confirm that no significant cleavage product of the fusion protein is present. We, therefore, produced two types of fusion proteins, DGK␥-GFP and GFP-DGK␥, and compared their enzymological and immunological properties with native DGK␥. Both DGK␥-GFP and GFP-DGK␥ were of appropriate sizes, since the molecular weights of DGK␥ and GFP are 88 and 27 kDa, respectively (Fig. 1B), and no significant degraded products were detected by either anti-DGK␥ or GFP antibodies, although nonspecific bands were observed with anti-DGK␥ antibody. Unexpectedly, DGK␥-GFP showed no kinase activity, whereas the activity of GFP-DGK␥ was almost the same as that of native DGK␥ (Fig. 1C). This suggests that the C terminus of DGK␥ may play an important role in its activity. Previously, we confirmed that both GFP-PKC␥ and PKC␥-GFP showed similar kinase activity to that of intact PKC␥ (18). These results indicate that it depends on the property of each protein whether GFP is attached at the N or C terminus of the protein. Based on the finding that only GFP-DGK␥ had kinase activity, we can assume that GFP was fused to the N terminus of DGK␣. GFP-DGK␣ was recognized as the 110-kDa band, which is a suitable size because DGK␣ is an 80-kDa protein and has enough kinase activity (data not shown).
To date, we have accumulated knowledge concerning translocations of PKCs in CHO-K1 cells (19,20,31). In addition, translocation of DGK by several stimuli such as phorbol ester (21,22), calcium (23), and also receptor-mediated translocation (24) have so far been reported. Therefore, the effects of TPA (a phorbol ester), calcium ionophore (A23187), and ATP as an agonist of purinergic receptors expressed in CHO-K1 cells (32) on the translocation of GFP-DGK␥ and GFP-DGK␣ were examined in CHO-K1 cells. Moreover, the effect of arachidonic acid was examined based on the finding that fatty acids induce distinct translocation among PKC subtypes (19). Calcium ionophore at 20 M induced generally similar translocations between GFP-DGK␥ and GFP-DGK␣. Namely, both DGK fusion proteins in the cytoplasm moved to the plasma membrane at 10 -15 s after the ionophore stimulation and were finally restored to the cytoplasm (data not shown). In contrast, ATP, TPA, and arachidonic acid showed distinct effects on the translocations of DGK␥ and -␣, as described above (Figs. 2 and 3). This is the first report visualizing the subtype-specific translocation of DGK␣ and -␥ in living cells.
Although DGKs have cysteine-rich regions homologous to the C1A and C1B motifs of PKCs, phorbol 12,13-dibutyrate binding in DGKs, unlike PKCs, has never been detected (2,33). Recently, Hurley et al. (34) compared amino acid sequences of C1 domains of many proteins, including PKCs (␣, ␤, ␥, ␦, ⑀, , , and ) and DGKs (␣, ␤, ␥, ␦, ⑀, and ), and divided them into two groups, typical and atypical. C1 domains in the typical group fit the profile for phorbol ester binding, whereas those in the atypical did not. According to Hurley et al. (34), DGK C1 domains have no property to bind DG and TPA except for C1A regions in DGK␥ and ␤. Therefore, to elucidate whether the C1A region is responsible for TPA-induced translocation of DGK␥, we investigated the translocation of the two mutants. Based on a previous report that mutation on the 17th cysteine in the C1B of PKC␦ completely abolished its phorbol ester binding (35), Cys-285 in the C1A region or Cys-348 in the C1B region were replaced by Gly, which corresponds to the 17th cysteine in the C1B region of PKC␦. The C1B mutant showed TPA-induced irreversible translocation, whereas the C1A mutant was not translocated in response to TPA (Fig. 5), indicating that interaction between TPA and the C1A domain of DGK␥ resulted in the translocation. Together with the results showing that DGK␣ does not respond to TPA, these results support Hurley's classification. Among the C1 domains in Hurley's typical group, those of PKCs can be further divided into two subgroups based on the data from Irie et al. (36). C1B of all conventional and novel PKCs have a high affinity binding to phorbol ester, whereas all their C1A domains except for the C1A of PKC␥ show very low affinity. Since phorbol ester binding of DGK has not been detected in vitro, the C1A domain of DGK␥ is thought to belong to the low affinity group. More interestingly, arachidonic acid induced the translocation of C1A mutant but not the C1B mutant of DGK␥ (data not shown). These results indicate not only that the loss of TPAinduced translocation observed in the C1A mutant is not due to the loss of translocation ability but also that the domain interacting with arachidonic acid differs from that interacting with TPA. Alternatively, Cys-285 in C1A domain of DGK␥ may be unnecessary for arachidonic acid binding.
When expressed in the CHO-K1 cells, both GFP-DGK␥ and GFP-DGK␣ were observed in the cytoplasm and the nucleoplasm, although the expression level of the fusion proteins in the nucleoplasm varied from cell to cell. The percentage of cells expressing the fusion protein in the nucleoplasm increased during the culture (data not shown). In contrast, both mutants of DGK␥ were detected mainly in the cytoplasm, and a significantly low amount of the mutant was expressed in the nucleus. Since neither mutant showed a significant kinase activity (data not shown), kinase activity may be necessary for expression of DGK␥ and -␣ in the nucleoplasm or for their translocation into the nucleoplasm. Alternatively, the different localization of the C1 mutants may be due to loss of interaction with an unknown protein, because it was suggested that the C1 region of DGK is important for protein-protein interaction (2). Furthermore, despite a lack of kinase activity, the mutants showed translocation, indicating that kinase activity is not necessary for the translocation of DGK, as seen in the translocations of PKCs (18,20).
It is also noteworthy that GFP-DGK␥, but not GFP-DGK␣, showed a similar translocation to that of PKC␥-GFP. For example, PKC␥-GFP and GFP-DGK␥ were translocated to the plasma membrane reversibly by arachidonic acid (Fig. 2 and  19), whereas GFP-DGK␣ was targeted to the Golgi network in addition to the plasma membrane (Fig. 4). TPA caused translocation of both GFP-DGK␥ and PKC␥-GFP but not GFP-DGK␣. Furthermore, DGK␥ and PKC␥ showed spatially similar but temporally different translocations after purinergic receptor activation (Fig. 6). This phenomenon may indicate a functional correlation of the two kinases, because PKC is probably translocated to the plasma membrane before DGK is targeted there, so PKC is activated by DG for a long time. In other words, the time lag between the translocation of DGK and PKC may regulate the duration of PKC activation. In fact, co-expression of DGK␥ hastens the re-translocation of PKC␥ from the membrane to the cytoplasma, and the inhibitor of DGK blocked the re-translocation of PKC␥ (Fig. 7), indicating clearly the functional correlation of DGK and PKC. Further experiments, however, are necessary to conclude that the temporal and spatial orchestration of PKC␥ and DGK␥ occurred in vivo, i.e. in Purkinje cells in which both kinases are abundantly expressed.
In the present study, we could not find any evidence that PKC directly regulates the translocation of DGK, although the direct phosphorylation of DGK by PKC was reported previously. For example, Kanoh et al. (25,26) report in vitro and in vivo phosphorylation of DGK by PKC, and Nobe et al. (24) also show that phosphorylation of DGK by PKC increased kinase affinity to the micell. Furthermore, Topham et al. (27) indicate that PKC␣ and -regulated nuclear localization of DGK. However, in the present study, treatment with staurosporine, a PKC inhibitor, exerted no effect on the initial localization or on any translocation of the two subtypes of DGKs (data not shown). Further experiments are required to elucidate the regulatory mechanism of DGK by PKC.
In conclusion, we demonstrated that DGK as well as PKC shows subtype-specific translocation and that translocation depends on extracellular signals and contributes to the regulation of DGK activity and its subtype-specific function, although the significance of this study may be limited because of using overexpressed enzymes. Our results suggested that functionally correlated proteins such as PKC and DGK are orchestrated temporally and spatially in the signal transduction mediated via both kinases.