Phorbol Ester-regulated Oligomerization of Diacylglycerol Kinase δ Linked to Its Phosphorylation and Translocation*

Diacylglycerol kinase (DGK) plays an important role in signal transduction through modulating the balance between two signaling lipids, diacylglycerol and phosphatidic acid. In yeast two-hybrid screening, we unexpectedly found a self-association of the C-terminal part of DGKδ containing a sterile α-motif (SAM) domain. We then bacterially expressed the SAM domain fused with maltose-binding protein and confirmed the formation of dimeric and tetrameric structures. Moreover, gel filtration and co-immunoprecipitation analyses demonstrated that DGKδ formed homo-oligomeric structures in intact cells and that the SAM domain was critically involved in the oligomerization. Interestingly, phorbol ester stimulation induced dissociation of the oligomeric structures with concomitant phosphorylation of DGKδ. Furthermore, we found that DGKδ was translocated from cytoplasmic vesicles to the plasma membrane upon phorbol ester stimulation. In this case, DGKδ mutants lacking the ability of self-association were localized at the plasma membranes even in the absence of phorbol ester. A protein kinase C inhibitor, staurosporine, blocked all of the effects of phorbol ester,i.e. oligomer dissociation, phosphorylation, and translocation. We confirmed that tumor-promoting phorbol esters did not directly bind to DGKδ. The present studies demonstrated that the formation and dissociation of oligomers serve as the regulatory mechanisms of DGKδ and that DGKδ is a novel downstream effector of phorbol ester/protein kinase C signaling pathway.

Diacylglycerol kinase (DGK) 1 phosphorylates diacylglycerol to produce phosphatidic acid (1). The roles of diacylglycerol and phosphatidic acid as lipid second messengers have been attracting much attention. Diacylglycerol is known to be an activator of conventional and novel protein kinase Cs (PKC), chimerins, Unc-13, and Ras guanyl nucleotide-releasing pro-tein (2)(3)(4). Phosphatidic acid has been reported in a number of studies to modulate phosphatidylinositol-4-phosphate kinase, Raf-1 kinase, atypical PKC, and many other important enzymes (5,6). Phosphatidic acid is also known to have mitogenic effects in a variety of cells (7). Because the cellular concentrations of these signaling lipids must be strictly regulated, DGK is thought to be one of the key enzymes involved in the cellular signal transduction.
It is now recognized that DGK represents a large gene family of isozymes differing remarkably in their structures, the modes of tissue expression, and in the enzymological properties (5,6,8). To date, nine mammalian DGK isozymes (␣, ␤, ␥, ␦, ⑀, , , , and ) containing in common the zinc finger structures and the catalytic region are classified into five subgroups according to their structural features (5,6,8). In contrast to mammals, only a few DGK isoforms have been identified in organisms such as worms (Caenorhabditis elegans) and flies (Drosophila melanogaster) (5,6). Moreover, no DGK gene has yet been detected in yeast. It is therefore likely that mammalian DGK isozymes have essential roles in biological processes specific to higher vertebrates, such as development/differentiation, neural networking, the immune system, and tumorigenesis. However, our knowledge of specific functions of the individual DGK isozymes is still limited at present.
We previously reported that DGK␦ (type II isozyme) contains a pleckstrin homology (PH) domain at the N terminus and a sterile ␣-motif (SAM) domain at the C terminus (9). Because another type II isozyme, DGK, was reported to have no SAM domain (10), DGK␦ has a unique structural feature among the DGK isozymes identified so far. Although the DGK␦ PH domain has been reported to bind to inositol 1,3,4,5-tetrakisphosphate, phosphatidylinositol 3,4,5-trisphosphate, and myosin II (11,12), functions of the SAM domain of DGK␦ are still unclear. The SAM domain, which is ϳ70 amino acids long, was first described as a module that is present in a small group of yeast sexual differentiation and Drosophila polyhomeotic proteins (13). The SAM domain was subsequently found to occur in a wide range of proteins. Proteins containing SAM domains include the Eph family of receptor tyrosine kinases (14,15), serine/threonine kinases (16), Src homology 2 domain-containing adaptor proteins (17), ETS transcription factors (18), p53 (17), and many other proteins. Such an extended occurrence suggests that SAM is an evolutionally conserved protein-protein interaction domain that is involved in the regulation of numerous developmental processes among diverse eukaryotes. Some SAM domains were reported to form homo-and heterooligomers (16 -18). However, recent studies showed that although SAM domains of EphA4 and B2 receptors formed homooligomers, oligomer formation was not detected for EphA2 and B1 receptors (19 -21). The function of the SAM domain, therefore, does not appear to be limited to oligomer formation, and can be variable depending on different proteins. The present work for the first time shows that the DGK isoform forms oligomeric structures through the SAM domain and that phorbol ester stimulation dynamically regulates the molecular assembly and intracellular localization of DGK␦.

EXPERIMENTAL PROCEDURES
Plasmid Constructs-cDNAs coding for DGK␦-CTR (amino acid residues 882-1170), DGK␦-CC (amino acid residues 882-1097), DGK␦-SAM (amino acid residues 1098 -1170), and DGK␦-⌬SAM (amino acid residues 1-1096) were generated from a human DGK␦ cDNA clone (9) by polymerase chain reaction (PCR). DGK␦-W1101G cDNA clone was generated by replacing the Trp-1101 with Gly using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). For subsequent subcloning into expression vectors, primers were designed such that the resulting DGK␦ cDNA fragments contained a 5Ј-EcoRI restriction site and a 3Ј-stop codon followed by a 3Ј-EcoRI site. The resulting DGK␦ fragments were fused in-frame with green fluorescent protein (GFP) in the pEGFP-C3 vector (CLONTECH, Tokyo, Japan) and with the 3xFLAG tag in the p3xFLAG-CMV-7.1 (Sigma, Tokyo, Japan). Plasmids encoding the SAM domain of DGK␦ fused to the maltosebinding protein (MBP, 43 kDa) were obtained by inserting the SAM cDNA fragments into the pMAL-c2X vector (New England BioLabs, Beverly, MA). The full-length DGK␣ cDNA was amplified by PCR and subcloned in-frame to the p3xFLAG-CMV-7.1 vector.
Yeast Two-hybrid Analysis and cDNA Cloning-The DGK␦ fragments were fused in-frame with the LexA DNA-binding domain in the pBTM116 vector (kindly provided by Dr. A. Yoshimura, Kyushu University, Fukuoka, Japan) and also fused with the GAL4 activation domain in the pGAD10 vector (CLONTECH). The yeast two-hybrid screen was done as described by Endo et al. (22). In brief, the C-terminal region of DGK␦ (amino acid residues 882-1170) used as a bait was fused with the LexA DNA-binding domain in pBTM116 vector (pBTM116/ DGK␦-CTR). Transformation of yeast strain L40 was performed by the lithium acetate method. Yeast transformants containing pBTM116/ DGK␦-CTR were subsequently transformed with the human skeletal muscle cDNA library in pGAD10 and selected for growth on Trp Ϫ Leu Ϫ His Ϫ plates. Two independent positive colonies (clones 18 and 21) that gave ␤-galactosidase activity were isolated from 8 ϫ 10 6 yeast transformants. An 5-bromo-4-chloro-3-indolyl-␤-D-galactoside (X-gal) filter assay was employed to measure ␤-galactosidase activity. Loss of pBTM116/DGK␦-CTR from positive colonies was induced by Leu Ϫ selection and positive library plasmids were isolated. To verify the interaction, positive plasmids were purified and retransformed into L40 cells together with or without pBTM116/DGK␦-CTR. Transformants were selected for growth on Trp Ϫ Leu Ϫ plates and assessed for ␤-galactosidase activity. Sequencing and data base analyses revealed that clone 18 encoded a DGK␦ fragment (residues 889 -1170), whereas clone 21 possessed an unidentified sequence.
Other sets of two-hybrid analysis were performed to determine selfassociation of the C-terminal parts of DGK␦, such as wild-type and point-mutated SAM domains. The L40 cells were transformed with various pBTM116-and pGAD10-DGK␦ cDNA constructs. Transformants were selected by growth on Trp Ϫ Leu Ϫ plates and assessed for ␤-galactosidase activity.
Cell Culture and Transfection-COS-7 and human embryonic kidney (HEK) 293 cells were maintained in DMEM (Sigma) containing 10% fetal bovine serum at 37°C in an atmosphere containing 5% CO 2 . Cells were transiently transfected with cDNAs using Effectene transfection reagent according to the manufacturer's instructions (Qiagen, Tokyo, Japan). After 2 days, cells were used for further analysis.
Expression and Purification of MBP Fusion Proteins-XL1-Blue cells (Stratagene) were transfected with various pMAL-c2X constructs, and MBP-fusion proteins were expressed and purified according to the procedure recommended by the manufacturer (New England BioLabs). In this case, expression of fusion proteins was induced by 0.3 mM isopropyl-1-thio-␤-D-galactoside at 37°C for 2 h. The cells were then harvested and lysed by freeze-thaw cycles in the column buffer (20 mM Tris-HCl (pH 7.4), 200 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml pepstatin A, 5 g/ml leupeptin). Fusion proteins were purified using an amylose resin affinity column.
Size Exclusion Chromatography-Gel filtration was performed on a Superose 12 HR 10/30 column (Amersham Biosciences, Tokyo, Japan) equilibrated with the column buffer supplemented with (for Triton X-100 extracts) or without (for MBP fusion proteins) 1% Triton X-100, and run with the fast protein liquid chromatography system equipped with a UV monitor (Amersham Biosciences) at a flow rate of 0.5 ml/min at room temperature. The column was calibrated with gel filtration molecular weight markers (Sigma). For analysis of the MBP fusion proteins, aliquots (100 l) of the purified samples were loaded onto the column, and 0.5-ml fractions were collected. For analysis of the Triton X-100 extracts from COS-7 cells expressing FLAG-tagged DGK␦, the cells were harvested in 20 mM Tris-HCl (pH 7.4), 1 mM EDTA, 200 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, Complete protease inhibitor mixture (Roche Molecular Biochemicals, Tokyo, Japan), and phosphatase inhibitor mixture II (Sigma). The mixture was centrifuged at 2,000 ϫ g for 3 min at 4°C, and the resultant supernatant was further centrifuged at 100,000 ϫ g for 30 min at 4°C to give the Triton X-100 extracts. Aliquots (100 l) of the extracts were loaded onto the column, and 0.5-ml fractions were collected. The fractions were analyzed by Western blotting using an anti-FLAG monoclonal antibody (M2, Sigma). In some experiments, the Triton X-100 extracts of HepG2 cells were similarly fractionated, and the fractions were analyzed for endogenous DGK␦ using anti-DGK␦ antibody as previously described (9).
Western Blot Analysis-Gel filtration fractions, precleared cell lysates, and immunoprecipitates were separated on 7.5 or 16.5% SDS-PAGE. The separated proteins were transferred to a nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany) and blocked with Block Ace (Dainippon Pharmaceutical, Tokyo, Japan). The membrane was incubated with anti-FLAG M2 monoclonal antibody or anti-GFP monoclonal antibody (B-2, Santa Cruz Biotechnology) in 10% Block Ace for 1 h. The immunoreactive bands were visualized using peroxidaseconjugated anti-mouse IgG antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) and SuperSignal (Pierce, Rockford, IL).
Phorbol Ester Binding Assay-The [ 3 H]phorbol 12,13-dibutyrate (PDBu) binding assay was performed as described (23,24). The binding was determined in duplicated incubation mixtures (250 l each), containing [ 3 H]PDBu (17.0 Ci/mmol, 40 nM) (Amersham Biosciences), 550 ϫ g supernatant of COS-7 cells (50 g of protein), 50 mM Tris-HCl, pH 7.4, 0.1 mM CaCl 2 , 100 g/ml phosphatidylserine, and bovine ␥-globulin (4 mg/ml). Incubation was carried out in 1.5-ml Eppendorf tubes for 30 min at 37°C. The samples were then chilled for 5 min at 0°C and added with 187 l of 35% (w/w) polyethylene glycol 6000 in 50 mM Tris-HCl, pH 7.4. The mixtures were then incubated for 15 min at 0°C to permit precipitation of proteins. The precipitates were spun down at 10,000 ϫ g at 4°C for 20 min. The supernatant was removed by aspiration and blotted with filter papers. After the tip of the Eppendorf tube was cut off, the radioactivity in the pellet was measured to determine the bound [ 3 H]PDBu. Nonspecific binding determined in the presence of unlabeled PDBu (40 M) was subtracted from the data.
Protein Phosphorylation Analysis in Vivo-COS-7 cells expressing 3xFLAG-tagged DGK␦ were used at 40 h post-transfection. HepG2 cells (100-mm dish) or transfected COS-7 cells (35-mm dish) were washed four times with Tris-buffered saline, incubated for 1 h in DMEM without sodium phosphate and sodium pyruvate (Invitrogen, Tokyo, Japan) containing 0.1% BSA, and then labeled for 2 h with [ 32 P]orthophosphate (0.3 mCi/ml, Amersham Biosciences) in phosphate-free DMEM, 0.1% BSA. The labeled cells were subsequently incubated in phosphate-free DMEM, 0.1% BSA containing 1 M 12-O-tetradecanoylphorbol 13-acetate (TPA) or 0.1% Me 2 SO for 1 h. Cells were then harvested in buffer 1 and then centrifuged at 10,000 ϫ g for 20 min at 4°C to give the cell lysates. Cell lysates (300 l) were precleared with Protein A/G PLUSagarose (10 l). Anti-FLAG (M2, for COS-7 cells transfected) or anti-DGK␦ (against the C-terminal peptide, amino acids 1156 -1170, of human DGK␦, for HepG2 cells) antibody (2 g) was added to precleared lysates to immunoprecipitate 3xFLAG-tagged or endogenous DGK␦ proteins. After 1 h, Protein A/G PLUS-agarose (5 l) was added followed by a 1-h incubation at 4°C. After washing the agarose beads five times with buffer 1, immunoprecipitated proteins were extracted with 50 l of SDS sample buffer and then analyzed by SDS-PAGE. The radioactive signal was visualized by phosphorimaging using a BAS1800 Bio-imaging Analyzer (Fuji Film, Tokyo, Japan).
Fluorescence Microscopy-HEK293 cells were grown on poly-L-lysine-coated glass coverslips and transiently transfected with cDNAs encoding various DGK␦ constructs N-terminally fused with GFP. After 2 days, HEK293 cells were washed four times with PBS and then cultured in DMEM, 0.1% BSA. After 3 h, the medium was exchanged with DMEM, 0.1% BSA containing 1 M TPA. After 1 h, cells were fixed with 3.7% formaldehyde in PBS for 10 min and then washed five times with PBS at room temperature. The coverslips were mounted using Vectashield (Vector Laboratories, Burlingame, CA). Cells were examined using an inverted confocal laser scanning microscopy (Zeiss LSM 510). Images were processed using Adobe Photoshop 6.0 (Adobe Systems, San Jose, CA).

Self-association of the DGK␦ SAM Domain in the Yeast Two-
hybrid System-We screened yeast two-hybrid cDNA libraries prepared from human skeletal muscle and obtained two positive clones that interacted with a C-terminal portion (amino acid residues 882-1170) of DGK␦ containing two sets of coiledcoil structures and a SAM domain (Fig. 1A). Interestingly, sequence analysis revealed that one of the positive clones (clone 18) was DGK␦ (residues 889 -1170) itself.
Because both of the coiled-coil structures and the SAM domain could participate in protein-protein interaction, we next tested which region was responsible for the self-association. The DGK␦ SAM domain interacted with each other in a yeast two-hybrid assay as demonstrated by the formation of blue colonies, whereas the DGK␦ coiled-coil structures did not interact with the counterpart (Fig. 1, B and C). The result indicates that the SAM domain is sufficient to form a dimeric structure.
An aromatic residue, Tyr-912 (corresponding to Trp-1101 in DGK␦), at the 8th position in the SAM domain of the EphB2 receptor was reported to be critical for dimer formation (20). On the other hand, the SAM domain of the EphA2 receptor, in which the critical Tyr is mutated to Gly, did not form a dimer structure (19). Trp-1101 in DGK␦ was thus substituted by Gly to see whether this Trp was critical for the interaction. As shown in Fig. 1, B and D, the SAM-W1101G mutant did not show a detectable interaction in the yeast two-hybrid assay. This indicates that Trp-1101 is a residue critical for homodimer formation, suggesting that the dimer structure of the DGK␦ SAM domain has general relevance to other SAM domain-containing proteins.
Gel Filtration Analysis of MBP-DGK␦-SAM and MBP-DGK␦-SAM-W1101G-To further study the multimeric nature of the DGK␦ SAM domain in vitro, the purified MBP-DGK␦-SAM and MBP-DGK␦-SAM-W1101G were analyzed using a Superose 12 HR column. Although MBP-DGK␦ SAM at a low loading concentration (0.6 M) eluted at the monomer position (not shown), a peak of the fusion protein at 6 M of loading concentration was at 12.2 ml (ϳ120 kDa, dimer) ( Fig. 2A, top panel). We confirmed by SDS-PAGE analysis that the appearance of the fusion protein (52 kDa) was well correlated with the UV peak pattern ( Fig. 2A, bottom panel). When the loading concentration was increased to 45 M, MBP-DGK␦ SAM eluted at the tetramer position (ϳ240 kDa) (Fig. 2B). In contrast to the wild-type SAM, MBP-DGK␦-SAM-W1101G at 6 M of loading concentration eluted at the monomer position (Fig. 2C). Moreover, even if the loading concentration was increased to 45 M, the mutant did not form a multimeric structure (Fig. 2D). These data, essentially consistent with those of the yeast twohybrid assay, indicate that the DGK␦ SAM domain is capable of forming at least tetrameric structures, and that Trp-1101 in the SAM domain is essential for the interaction. In this respect, MBP-DGK␦-SAM-W1101G eluted at the same retention volume (13.1 ml) as that of the MBP-DGK␦-SAM (wild-type) monomer form, indicating that the mutation (Trp-1101 to Gly) introduced did not significantly affect the folding of this domain.
Oligomer Formation of DGK␦ through the SAM Domain in Vivo-To assess whether intact DGK␦ formed the oligomer structure in vivo, we performed gel filtration analysis of DGK␦ expressed in COS-7 cells. In this case DGK␦ was solubilized with 1% Triton X-100 and gel filtration was done in the presence of Triton X-100, because more than 90% of wild-type DGK␦ and its mutants were recovered in particulate fractions by cell fractionation. It was assumed that monomer (140 kDa plus Triton X-100 micelle size: 90 kDa), homo-dimer (280 plus 90 kDa), and homo-tetramer (560 plus 90 kDa) of DGK␦ would elute at retention volumes 10.5-11.5, 9.5-10.0, and 7.5-8.5 ml, respectively. As shown in Fig. 3, most of the wild-type enzyme was recovered in the Նdimer fractions. We also observed that more than 80% of the endogenous DGK␦ in HepG2 cells was recovered in the Նdimer fractions, in particular the tetramer fraction, although the exact quantitation of the immunoreactive bands was hindered by the low concentration of endogenous DGK␦. This finding indicates that the oligomerization is not the artifact of enzyme overexpression. We could not exclude the possibility that the Նdimer fractions contained heterodimer/oligomers of DGK␦ with unknown partner(s). However, because the availability to the overexpressed enzyme of the endogenous oligomerization partners, if any, should be limited, the formation of hetero-dimer/oligomer was considered to be negligible. In contrast to the wild-type enzyme, most of DGK␦ lacking the SAM domain (DGK␦-⌬SAM) eluted at the monomer position. Most of DGK␦ having the mutated SAM domain (DGK␦-W1101G) also eluted at the monomer position. These results strongly suggest that DGK␦ forms oligomeric structures in vivo and that the SAM domain plays a critical role for the oligomer formation. Notably, small portions of DGK␦-⌬SAM and DGK␦-W1101G proteins were recovered in the Նdimer fractions, suggesting that in addition to the SAM domain the N-terminal region of the enzyme contributes, albeit to a lesser extent, to the formation of oligomeric structure.
To further confirm homo-oligomer formation of DGK␦ in vivo, we carried out co-immunoprecipitation analysis using the lysates of COS-7 cells co-expressing GFP-tagged and 3xFLAGtagged DGK␦ proteins. When 3xFLAG-tagged DGK␦ was immunoprecipitated with anti-FLAG antibody, GFP-tagged DGK␦ was co-immunoprecipitated (Fig. 4A). Reciprocally, when GFP-tagged DGK␦ was immunoprecipitated with anti-GFP antibody, 3xFLAG-tagged DGK␦ was co-immunoprecipitated. To confirm the specificity of co-immunoprecipitation, we performed the same experiment using COS-7 cells expressing GFP-tagged DGK␦ and another isoform, DGK␣, which was tagged with 3xFLAG. As shown in Fig. 4B, no DGK␣ band was detected in the anti-GFP immunoprecipitation fraction. Taken together, we concluded that DGK␦ specifically formed homooligomer structures in vivo.
To assess the contribution of the SAM domain toward the oligomer formation, we performed co-immunoprecipitation experiments of the SAM alone or DGK␦ mutants lacking the SAM domain or substituted at Trp-1101 by Gly in the SAM domain. As shown in Fig. 4C, 3xFLAG-SAM was detected in the anti-GFP immunoprecipitation fraction. Moreover, in comparison with wild-type DGK␦, significantly less 3xFLAG-DGK␦-⌬SAM and DGK␦-W1101G were found in the anti-GFP-immunoprecipitated complexes (Fig. 4D). We confirmed the presence of comparable amounts of the GFP fusion proteins in anti-GFP immunoprecipitates. These results indicate that the SAM domain is sufficient for self-association in vivo, contributing critically to the oligomer formation of the fulllength DGK␦. As observed in gel filtration analysis, detectable 3xFLAG-DGK␦-⌬SAM and DGK␦-W1101G bands still remained in the anti-GFP immunoprecipitated complexes (Fig.  4D). This again suggests that in addition to the SAM domain the N-terminal region of the enzyme is able to weakly contribute to the formation of oligomer structure.
Effects of Phorbol Ester Stimulation on the Oligomerization of DGK␦-After detecting the SAM domain-mediated oligomerization of DGK␦, we tested effects of cell stimulation on the physical state of DGK␦. After serum starvation of COS-7 cells expressing FLAG-tagged enzymes for 3 h, more than 90% DGK␦ eluted in Նdimer fractions in gel filtration (Fig. 5) as already noted for endogenous enzyme (Fig. 3). Interestingly, when the cells were stimulated with TPA for 1 h, we found a clear shift of the elution pattern of wild-type DGK␦ from Նdimer to monomer fractions. In this experiment, a large portion of DGK␦ lacking the SAM domain had already been monomeric and TPA did not affect the assembly of this mutant. Essentially the same result was obtained with DGK␦-W1101G. 2 The results indicated that TPA stimulation induced conversion of oligomers to a monomeric form of DGK␦.
To confirm the dissociation of the DGK␦ oligomers induced by TPA stimulation, we performed co-immunoprecipitation analysis using cells expressing differently tagged enzymes. As shown in Fig. 6A, TPA treatment markedly reduced co-precipitation of the DGK␦ molecules. Together with the gel filtration data shown in Fig. 5, it becomes clear that TPA stimulation induces dissociation of the oligomeric structures of DGK␦.
We next asked whether the SAM domain alone was involved in inducing TPA-dependent oligomer-monomer conversion. As shown in Fig. 6B, no change of the SAM domain band was observed after TPA stimulation in a co-immunoprecipitation assay. Furthermore, in gel filtration analysis, no significant change of elution pattern of the SAM domain alone was observed for cells stimulated with TPA. 2 The results suggest that although the assembly of DGK␦ is mediated by the SAM domain, the TPA-induced dissociation of the oligomers requires the participation of the N-terminal region of DGK␦ other than the SAM domain. TPA, we investigated translocation of DGK␦. As shown in Fig.  7, C and D, we found that in HEK293 cells TPA induced translocation of GFP-DGK␦ from cytoplasmic vesicles to the plasma membrane. In contrast, cytoplasmic/nucleoplasmic localization of GFP alone was not affected by TPA stimulation (Fig. 7, A and B). Interestingly, GFP-DGK␦-⌬SAM was localized at the plasma membrane even under unstimulated conditions (Fig. 7, E and F). Moreover, GFP-DGK␦-W1101G was also distributed at the plasma membrane without TPA stimulation (Fig. 7, G and H). The data strongly suggest that DGK␦ mutants lacking the ability of self-association apparently mimic TPA stimulation. Thus, the dissociation of the oligomeric structure of DGK␦ was well correlated with its plasma membrane localization. In addition, it is strongly suggested that the oligomer formation mediated by the SAM domain exerts a blocking effect on the translocation.

Role of the SAM Domain in Enzyme Translocation
Because DGK␦ has Cys-rich, zinc finger-like structures, it was possible that TPA directly interacted with the enzyme, causing its recruitment from the cytoplasm to the plasma membrane. Thus, we next tested the phorbol ester binding activity of GFP-tagged DGK␦. Because the expression level of wild-type DGK␦ was lower (30 -50%) than that of PKC␤I (Fig. 8B), we also used DGK␦ lacking the PH domain (DGK␦-⌬PH) in this experiment to obtain an improved expression level. The extract from the PKC␤I-transfected cells used as a positive control showed markedly increased PDBu binding activity (Fig. 8A). However, the extract from COS-7 cells expressing wild-type DGK␦ gave no detectable PDBu binding activity. Even though DGK␦-⌬PH whose expression level reached the same level of that of PKC␤ (Fig. 8B) was used, the PDBu binding activity failed to be detected. Phorbol ester is, therefore, unlikely to recruit DGK␦ to the plasma membrane through its direct binding. This conclusion is also supported by the fact that the enzyme oligomerization and plasma membrane translocation, both regulated by phorbol ester (Figs. 6 and 7) were mediated through the SAM domain, which is apparently not involved in the interaction with phorbol esters.
Phosphorylation of DGK␦ by TPA Stimulation-To assess the molecular basis of the effects of TPA, we next asked whether DGK␦ was phosphorylated in TPA-stimulated cells. As shown in Fig. 9A, DGK␦ expressed in COS-7 cells was phosphorylated by endogenous protein kinase(s) in the presence of 1 M TPA. We also observed that endogenous DGK␦ in HepG2 cells was phosphorylated by TPA stimulation (Fig. 9B). This finding indicates that the phosphorylation is not the artifact of protein overexpression.
To prove the involvement of PKC in the phosphorylation process, we examined whether the in vivo phosphorylation of DGK␦ could be inhibited by a PKC inhibitor, staurosporine. As shown in Fig. 9C, staurosporine substantially reduced labeling of immunoprecipitated DGK␦ from cells stimulated with TPA. Staurosporine at a concentration of 0.3 M caused complete inhibition of DGK␦ labeling. These results indicate that PKC is involved in the phosphorylation and that DGK␦ is a novel downstream effector of the phorbol ester-PKC pathway.
Effects of Staurosporine on the Oligomerization and Translocation of DGK␦-In addition to PKC, many proteins, for example, a vav oncogene product, chimerins, and Ras guanyl nucleotide-releasing proteins have been reported to be activated by tumor-promoting phorbol ester (3,4). To examine whether the oligomerization and translocation of DGK␦ caused by TPA stimulation could be correlated with PKC action, we analyzed effects of staurosporine on these events. As previously shown in Fig. 6A, compared with the control cells, significantly less 3xFLAG-DGK␦ was co-precipitated with GFP-DGK␦ when the cells were stimulated with TPA in the absence of staurosporine (Fig. 10A). Remarkably, staurosporine (0.3 M) markedly blocked the effects of TPA, and in this case the co-precipitation of the DGK␦ molecules occurred to the same extent as observed for nonstimulated cells. This indicates that dissociation of DGK␦ oligomers in TPA-treated cells is mediated by the action of PKC. Furthermore, in the presence of staurosporine, the translocation of DGK␦ from cytoplasmic vesicles to the plasma membrane caused by TPA stimulation became undetectable (Fig. 10B). These results indicate that in TPA-stimulated cells, PKC is at least in part responsible for conversion of oligomeric DGK␦ to a monomer, and also for the translocation of DGK␦. DISCUSSION We demonstrated that DGK␦ formed dimer/oligomer structures through its SAM domain in vitro and in vivo. We further showed that TPA stimulation induced disassembly of the oligomeric structures and a concomitant enzyme translocation to the plasma membranes. Although the SAM domain oligomerization has been generally thought to be constitutive and static, the present studies showed that the interaction between the DGK␦ SAM domains was dynamically regulated. The analysis of endogenous and overexpressed DGK␦ showed that most of DGK␦ in resting cells exists constitutively as oligomers, in particular as a tetramer. The cell stimulation by TPA overcomes the protein-protein interactions mediated by the SAM domain, and this effect of TPA can be accounted for by the action of PKC in view of the effects of protein kinase C inhibitor, staurosporine tested at a low concentration (0.3 M). The present work showed that DGK␦ was phosphorylated in a TPA-dependent manner. However, we do not know yet whether PKC directly phosphorylates DGK␦, or phosphorylates unknown factor(s) that subsequently participate(s) in the regulation of oligomer formation and/or enzyme translocation. Because conventional PKCs are known to be activated at and stably associated with plasma membranes in TPA-stimulated cells (27), it is also difficult to envisage how the TPA-activated PKC could control the molecular assembly and translocation of cytoplasmic DGK␦. Further work is needed to elucidate the molecular mechanisms of the effects of PKC on the physical state of DGK␦.
Reflecting the importance of DGKs in the attenuation of the diacylglycerol signal, several DGK isozymes have been described to be relocated to specific sites within the stimulated cells. DGK␣ was translocated from the cytosol to the plasma membranes in the activated T-lymphocytes (28) and in Chinese hamster ovary cells treated with arachidonic acid (25). The molecular mechanism underlying DGK␣ translocation is not well understood. DGK␥, on the other hand, has been shown to bind phorbol ester with high affinity (29) and was translocated from the cytosol to the plasma membranes in TPA-stimulated cells (25). Furthermore, in TPA-treated fibroblasts, the nuclear DGK was shown to be phosphorylated by PKC, resulting in its relocation from the nucleus to the cytosol (26). In the present work DGK␦ was translocated from the cytoplasmic vesicles to the plasma membranes in TPA-treated cells and this translocation was closely linked to disassembly of the enzyme oligomers. DGK␦ thus appears to need to be a monomer to become associated with plasma membranes. These observations indicate that different DGK isozymes are moved to specific sites under distinct mechanisms so that each DGK species may phosphorylate different diacylglycerol pools in distinct subcellular compartments on different demands.
TPA may directly recruit DGK␦ to the plasma membrane by binding to its Cys-rich, zinc finger-like structures. However, because we did not detect significant binding of phorbol ester to intact DGK␦ expressed in COS-7 cells (Fig. 8), this possibility is unlikely. Although the critical 20 residues conserved in all conventional and novel PKC C1 domains were reported to play important roles in phorbol ester binding (3,29), indeed, the zinc fingers of DGK␦ had only 60 -80% sequence homology with respect to the 20 residues, suggesting that DGK␦ zinc fingers do not serve as phorbol ester-binding domains. Although we cannot deny the possibility that the DGK␦ zinc fingers bind TPA with low affinity, the translocation was shown to be dependent on the action of PKC. Moreover, DGK␦-⌬SAM and DGK␦-W1101G, both of which still contain zinc finger structures, can be localized at the plasma membrane even in the absence of TPA. Taken together, the effects of direct binding of TPA to the enzyme are not considered to be involved in the translocation.
The DGK␦ SAM domain was shown to form at least homotetrameric structures. This indicates that the DGK␦ SAM domain has two interfaces. Thanos et al. (20) proposed based on crystallographic data that the oligomeric structure of the EphB2 receptor SAM domain is assembled by two interfaces. In this model, an aromatic residue, Tyr-912 (corresponding to Trp-1101 in DGK␦), at the 8th position in the SAM domain of the receptor is critical for the dimer/oligomer formation. Because the mutation at Trp-1101 in the DGK␦ SAM resulted in a complete loss of the ability of self-association, the mode of oligomer formation of the DGK␦ SAM domain can fit with this model. The abolishment of self-association of SAM-W1101G even at high concentrations also suggests that the dimer formation via this interface including Trp-1101 may act as a trigger and may facilitate subsequent tetramer formation through another interface.
Earlier studies suggest that SAM domains of tyrosine kinases and transcription factors form homo-as well as heterooligomers and that such oligomers have important physiological roles. For instance, it has been shown that tyrosine kinase activity of Eph receptors was activated by forming a dimer and that tetramerized ligands for these receptors triggered a variety of signaling events such as recruitment of low molecular weight protein-tyrosine phosphatase (30). SAM domains from various polycomb group proteins, which regulate homeotic gene transcription, also form specific homo-and hetero-oligomers and may be important for generating large protein complexes within the cells (18,31). In the case of DGK␦, which is neither a receptor nor a transcription factor, the enzyme assembly mediated through the SAM domain exerts blocking effects on the enzyme translocation. This effect is, to our knowledge, a novel function of SAM domain. Although we could not detect apparent formation of hetero-oligomers of DGK␦, the possibility remains that the DGK␦ SAM domain recruits signaling partners through heteromeric SAM-SAM interactions after the conversion to a monomer form. We could not detect significant changes of catalytic activity of DGK␦ obtained from the TPAtreated cells when measured in vitro. We therefore do not know whether the formation and dissociation of the enzyme oligomers have any relevance to the control of the enzyme catalytic activity.
To date, several reports have already described that DGK␣ and DGK were phosphorylated by protein kinases such as PKC, protein kinase A, and Src tyrosine kinase (26,(32)(33)(34). We demonstrated that DGK␦ is also an isozyme phosphorylated by protein kinase(s) in TPA-stimulated cells. Because a PKC inhibitor, staurosporine, inhibited the phosphorylation of DGK␦ by TPA stimulation, DGK␦ was phosphorylated directly by PKC or indirectly by other kinases that were activated by PKC. Because TPA stimulates mitogen-activated protein kinase through PKC (35,36), DGK␦ may be phosphorylated by mitogen-activated protein kinase and/or other protein kinases involved in the mitogen-activated protein kinase pathway. The oligomerization may be controlled by serine/threonine phosphorylation. Alternatively, it is possible that DGK␦ has more than two distinct phosphorylation sites, which are essential for oligomer dissociation and post-dissociation events such as membrane translocation, respectively. To examine these possibilities in future studies, identification of the phosphorylation site(s) and then the generation of DGK␦ mutants mimicking nonphosphorylated and phosphorylated enzymes are required.
We have not revealed physiological functions of DGK␦ yet. Tumor-promoting phorbol ester stimulates many kinds of physiological machineries, such as cell growth and differentiation and tumorigenesis (2)(3)(4). Moreover, most of the SAM-containing proteins are involved in the regulation of developmental processes (17). DGK␦ may thus be involved in such biologically important events. The present studies demonstrate that oligomer-monomer conversion, phosphorylation, and subcellular translocation induced by TPA stimulation are important mechanisms of the control of DGK␦ action and that DGK␦ is a novel downstream effector of phorbol ester/PKC signaling pathway.