Synthesis and Phorbol Ester Binding of the Cysteine-rich Domains of Diacylglycerol Kinase (DGK) Isozymes

Diacylglycerol kinase (DGK) and protein kinase C (PKC) are two distinct enzyme families associated with diacylglycerol. Both enzymes have cysteine-rich C1 domains (C1A, C1B, and C1C) in the regulatory region. Although most PKC C1 domains strongly bind phorbol esters, there has been no direct evidence that DGK C1 domains bind phorbol esters. We synthesized 11 cysteine-rich sequences of DGK C1 domains with good sequence homology to those of the PKC C1 domains. Among them, only DGKγ-C1A and DGKβ-C1A exhibited significant binding to phorbol 12,13-dibutyrate (PDBu). Scatchard analysis of rat-DGKγ-C1A, human-DGKγ-C1A, and human-DGKβ-C1A gaveK d values of 3.6, 2.8, and 14.6 nm, respectively, suggesting that DGKγ and DGKβ are new targets of phorbol esters. An A12T mutation of human-DGKβ-C1A enhanced the affinity to bind PDBu, indicating that the β-hydroxyl group of Thr-12 significantly contributes to the binding. TheK d value for PDBu of FLAG-tagged whole rat-DGKγ (4.4 nm) was nearly equal to that of rat-DGKγ-C1A (3.6 nm). Moreover, 12-O-tetradecanoylphorbol 13-acetate induced the irreversible translocation of whole rat-DGKγ and its C1B deletion mutant, not the C1A deletion mutant, from the cytoplasm to the plasma membrane of CHO-K1 cells. These results indicate that 12-O-tetradecanoylphorbol 13-acetate binds to C1A of DGKγ to cause its translocation.

Diacylglycerol kinase (DGK) 1 and protein kinase C (PKC) both interact with the second messenger diacylglycerol (DG) (1,2). DGK phosphorylates DG to produce phosphatidic acid, whereas PKC is allosterically activated by DG in the presence of phosphatidylserine. Therefore, DGK may inhibit the activation of PKC by attenuating DG levels, contributing to the regulation of intracellular signal transduction.
To date, nine subtypes of mammalian DGKs have been cloned (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15). All DGK isozymes consist of a conserved catalytic domain and two or three cysteine-rich C1 domains designated as C1A, C1B, and C1C (16). These isozymes are classified into five classes according to the other functional domains (Fig.  1). The class I isozymes (DGK␣, -␤, and -␥) have calcium binding domains (EF-hands). The class II isozymes (DGK␦ and -) have a pleckstrin homology domain at the N terminus, and their catalytic region is split into two domains unlike the other DGK isozymes. DGK⑀ has a simple structure and is classified as a class III isozyme. The class IV isozymes DGK and have a myristoylated alanine-rich C kinase substrate homology domain and four ankyrin repeats. DGK, which has three C1 domains unlike other DGK and PKC isozymes, is the only isozyme in class V.
The similarity between DGK and PKC isozymes in structure is in the cysteine-rich C1 domains. Recent investigations using NMR spectroscopy and x-ray crystallography have revealed the three-dimensional structure of C1B domains of PKC␣, PKC␥, and PKC␦ (17)(18)(19)(20). Each PKC C1 domain has six conserved cysteines and two histidines in the typical core structure HX 12 CX 2 CX 13-14 CX 2 CX 4 HX 2 CX 7 C (where X is any amino acid) that coordinates two atoms of zinc in a tetrahedral geometry (21,22). We previously synthesized the 50-mer core structure of the C1 domains of all PKC isozymes and showed that most of the C1 domains of conventional and novel PKC isozymes strongly bind phorbol 12,13-dibutyrate (PDBu) with dissociation constants (K d ) in the nanomolar range (23,24).
The core structure of DGK C1 domains, HX 10 -12 CX 2-6 CX 9 -19CX 2 CX 4 HX 2-4 CX 5-10 C, is slightly different from that of PKC C1 domains, but the six cysteines and two histidines are precisely conserved. In particular, DGK␤-C1A, DGK␥-C1A, DGK␦-C1A, DGK␦-C1B, DGK-C1A, and DGK-C1C have the same core structure as the PKC C1 domains (Fig. 2) and are deduced to show a significant phorbol ester binding affinity. However, there have been no reports that DGK isozymes bind phorbol esters. Ahmed et al. (25) reported that human-DGK␣ expressed in Escherichia coli or in COS-7 cells did not bind PDBu. Sakane et al. (26) also detected no specific PDBu binding in either rat-DGK␤ or human-DGK␥ expressed in COS-7 cells. On the other hand, a recent investigation by Shirai et al. (27) clearly showed that DGK␥ expressed in Chinese hamster ovary-K1 (CHO-K1) cells was translocated from the cytoplasm to the plasma membrane irreversibly following 12-O-tetradecanoylphorbol 13-acetate (TPA) treatment. They also suggested that the C1A domain of DGK␥ is responsible for this translocation based on the point mutation of each C1 domain. These results prompted us to examine the phorbol ester binding ability of each DGK␥ C1 domain, and we have reported recently (28) that rat-DGK␥ binds PDBu with high affinity as a preliminary communication. Further detailed studies on the C1 domains of several DGK isozymes other than those of rat-DGK␥ have been carried out along with the C1A peptide of human-DGK␥ which Sakane et al. (26) employed. These additional data have been compiled to give this full report.

EXPERIMENTAL PROCEDURES
General-The following spectroscopic and analytical instruments were used: matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS), Applied Biosystems Voyager-DE TM STR (20 kV); peptide synthesizer, Pioneer TM peptide synthesizer model 9030 (Applied Biosystems); and HPLC, Waters model 600E with model 2487UV detector. MALDI-TOF-MS measurements were carried out as follows; each peptide dissolved in 0.1% trifluoroacetic acid aqueous solution (50 pmol/l) was mixed with saturated ␣-cyano-4-hydroxycinnamic acid in 50% CH 3 CN containing 0.1% trifluoroacetic acid at a ratio of 1:1. One microliter of the resultant solution was subjected to the measurement. Angiotensin I and ACTH-(7-38) were used as external references. HPLC was carried out on a YMC-packed SH-342-5 (ODS, 20 mm inner diameter ϫ 150 mm) column (Yamamura Chemical Laboratory) for preparative purposes. [ 3 H]PDBu (17.0 Ci/mmol) was purchased from PerkinElmer Life Science. COS-7 cells were obtained from the Riken Cell Bank (Tsukuba, Japan). Unless otherwise noted, reagents were purchased from Sigma, Allexis, Wako Pure Chemical Industries, or Nacalai Tesque.
Synthesis of the C1 Peptides of DGK Isozymes-The 49 -52-mer peptides corresponding to the cysteine-rich sequences of DGK isozyme C1 domains (Table I and Fig. 2) were synthesized in a stepwise fashion on 0.1 mmol of preloaded Fmoc-Gly-PEG-PS resin (Applied Biosystems) by Pioneer TM using the Fmoc method as reported previously (23,24). The coupling reaction was carried out using each Fmoc amino acid (0.4 (29), and N,N-diisopropylethylamine (0.8 mmol) in N,N-dimethylformamide for 30 min (flow rate, 30 ml/min). After completion of the chain assembly, each peptide resin was cleaved, and the resultant crude peptide was precipitated by diethyl ether. The crude peptide was purified by gel filtration, followed by HPLC as reported previously (23,24). Lyophilization gave a corresponding pure C1 peptide, the purity of which was confirmed by HPLC (Ͼ98%). Each purified peptide exhibited satisfactory mass spectrometric data. The yields and mass data of the C1 peptides synthesized in this study are summarized in Table I.
Plasmid Construction for FLAG-tagged Rat-DGK␥-The plasmid for FLAG-tagged DGK␥-(1-789) (FLAG-DGK␥) was generated by PCR using BS412 (rat-DGK␥) as a template as reported previously (28). The cDNA of DGK␥-(1-789) was digested with MunI and BamHI and then  l The A12T mutant of human-DGK␤-C1A. To prevent racemization and oxidation during the synthesis, glycine was added to the C-terminal cysteine in each peptide. For example, rat-DGK␤-C1A means rat-DGK␤-(244 -292) ϩ Gly (50-mer peptide). All these sequences derive from Refs. 3-15. subcloned into the EcoRI and BglII sites of pTB701-FL, a mammalian expression vector, to express fusion protein with the N-terminal FLAG epitope.
Preparation of FLAG-DGK␥-Transient transfection into COS-7 cells was performed by electroporation as reported previously (28). After transfection, the cells were cultured at 37°C for 48 h and were harvested with phosphate-buffered saline(Ϫ), followed by centrifugation at 600 ϫ g for 5 min at 4°C. The cells were resuspended in 300 l of homogenization buffer (250 mM sucrose, 10 mM EGTA, 2 mM EDTA, 50 mM Tris/HCl, 20 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 0.5% Triton X-100, pH 7.4) and sonicated on ice. The homogenate was centrifuged at 10,000 ϫ g for 30 min at 4°C, and the resultant supernatant was saved for the PDBu binding assay described below.
[ 3 H]PDBu Binding Assay of DGK Isozyme C1 Peptides and FLAG-DGK␥-The PDBu binding assay was carried out using the procedure of Sharkey and Blumberg (30) with slight modification (23,24). The standard assay mixture (250 l) in 1.5-ml Eppendorf tubes contained 50 mM Tris maleate, pH 7.4, 50 g/ml 1,2-di-(cis-9-octadecenoyl)-sn-glycero-3phospho-L-serine (phosphatidylserine dioleoyl), 3 mg/ml bovine ␥-glob- For chemically synthesized DGK C1 peptides, metal coordination was carried out in a helium-purged distilled water solution (pH 5.5-6.0) of each C1 peptide. Five mol eq of 10 mM ZnCl 2 in helium-purged distilled water was added to the peptide solution, and the solution (174 M) was allowed to stand at 4°C for 10 min. After 10 l of the peptide solution was diluted with 990 l of helium-purged distilled water, the resultant solution (1.5-2.9 l) was added to the standard assay mixture described above (247.1 l), and the solution was incubated at 4°C for 10 min. For FLAG-DGK␥, 5 l of the enzyme solution after homogenization was similarly added and incubated at 30°C for 20 min. To the tubes was added 187 l of 35% (w/w) poly(ethylene glycol) (average molecular weight, 8000), and the mixture was vigorously stirred. The tubes were allowed to stand at 4°C for 10 min and then centrifuged for 10 min at 12,000 rpm in an Eppendorf microcentrifuge at 4°C. A 50-l aliquot of the supernatant of each tube was removed, and its radioactivity was measured to determine the free [ 3 H]PDBu concentration. The remainder of the supernatant of each tube was removed by aspiration. The tips of the tubes were cut off, and the radioactivity in the pellets was measured to determine the bound [ 3 H]PDBu. Specific binding represents the difference between the total and nonspecific binding. The nonspecific binding for each tube was calculated from its measured free [ 3 H]PDBu concentration and its partition coefficient to the pellet (about 3%).
In competition experiments using rat-DGK␥-C1A, various concentrations of an inhibitor in ethanol solution were added to the reaction mixture mentioned above. The effective concentration of [ 3 H]PDBu and rat-DGK␥-C1A was 20 and 5 nM, respectively. The final ethanol concentration of the mixture was less than 2%. Binding affinity was evaluated by the concentration required to cause 50% inhibition of the specific [ 3 H]PDBu binding, IC 50 , that was calculated by a computer program (Statistical Analysis System) with a probability unit procedure (31). The binding constant (K i ) was calculated from the IC 50 values and K d for PDBu of rat-DGK␥-C1A by the method of Sharkey and Blumberg (30).
Construct of Plasmids Encoding Domain Deleted Mutants of DGK␥-Domain deleted mutants of DGK␥ were produced using an ExSite PCR-based Site-directed Mutagenesis kit (Stratagene, La Jolla, CA). The plasmid (BS465) encoding rat-DGK␥ (27) with an XhoI site in the 5Ј terminus and a SmaI site in the 3Ј terminus was used as a template. The sense and antisense primers for producing a C1A-deleted mutant (⌬C1A-DGK␥) were 5Ј-GTGAAAACATACTCCAAAGCCAAAAGG-3Ј and 5Ј-GCGTCCATCCCCCTTGGAG-3Ј, respectively. The primers for a mutant DGK␥ lacking the C1B region (⌬C1B-DGK␥) were 5Ј-GATG-GTGGGGAGCTCAAAGAC-3Ј and 5Ј-CTGCATCACCTCACCGCTC-3Ј. The sequence was confirmed by verifying sequences. Each C1A-or C1B-deleted mutant of DGK␥ cDNA was subcloned into SalI and SmaI sites in pEGFPC1.
Observation of Translocation of the GFP Fusion Proteins-Plasmids (ϳ5.5 g) were transfected into 1.0 ϫ 10 6 cells by lipofection using FuGENE 6 transfection reagent (Roche Applied Science), 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 (Mat-Tek 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, and 10 mM glucose at pH 7.3 (Ringer's solution). Translocation of the GFP fusion protein was triggered by a direct application of a 10 times higher concentration of TPA 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 bandpass barrier filter. All experiments were performed at 37°C.
To prevent racemization and oxidation during the synthesis, the C terminus of each peptide was extended from the final cysteine to a glycine. These peptides exhibited satisfactory mass spectrometric data (Table I), and their purity was confirmed by HPLC analysis (Ͼ98%).
These peptides were folded using zinc chloride by a method reported previously (23,24) and subjected to a PDBu binding assay. Because our recent investigation found that some C1 domain fragments of PKC isozymes suffered from temperaturedependent inactivation (24,33), the incubation temperature of the binding assay was set at 4°C for the DGK isozyme C1 peptides. The specific binding to PDBu of the DGK C1 peptides is summarized in Fig. 3. Only the DGK␤-C1A and DGK␥-C1A peptides showed significant binding at 20 nM, whereas other DGK C1 peptides with over 50% sequence homology to PKC C1 domains were completely inactive even at 100 and 500 nM (data not shown). Rat-DGK␥-C1A and human-DGK␥-C1A showed quite similar PDBu binding although their sequences were different. In a control experiment, rat-DGK␥-C1B did not show any binding.
The weak PDBu binding affinity of human-DGK␤-C1A (K d ϭ 14.6 nM) compared with human-DGK␥-C1A (K d ϭ 2.8 nM) suggests that the Ala-12 residue is responsible for the decrease in binding affinity because the Thr-12 residue is strictly conserved in the conventional and novel PKC isozymes that show strong PDBu binding (24). The Thr-12 residue along with the Leu-21 and Gly-23 residues of PKC␦-C1B is shown to be involved in the phorbol ester binding in the x-ray structure (20). To prove this, the A12T mutant of human-DGK␤-C1A was synthesized, and its PDBu binding affinity was tested. The K d value for PDBu of the A12T mutant was 4.9 Ϯ 1.0 nM with a B max value of 42.4 Ϯ 2.8% (Fig. 4e), suggesting that the Thr-12 residue plays a significant role in the PDBu binding of DGK␤-C1A.

FIG. 2. Cysteine-rich sequences of all DGK isozymes (r, rat; h, human; ha, hamster) (3-15) along with rat-PKC␥ and rat-PKC⑀. a,
aromatic; h, large hydrophobic; ϩ, basic. After the completion of this study, we found that rat-DGK␤-C1A as well as human-DGK␤-C1A also had the valine residue at position 44. The cysteine-rich sequence of rat-DGK␤-C1A is, therefore, identical with that of human-DGK␤-C1A (N. Saito, unpublished results).  vestigate whether whole rat-DGK␥ as well as rat-DGK␥-C1A peptide binds PDBu. After homogenization and centrifugation of the COS-7 cells, the supernatant was saved for PDBu binding assay. Scatchard analysis of FLAG-DGK␥ in the PDBu binding was performed with the incubation temperature set at 30°C as for whole PKC isozymes (24,34). The FLAG-DGK␥ solution (5 l) showed remarkable specific PDBu binding (24,600 Ϯ 790 dpm) when 20 nM [ 3 H]PDBu and 50 g/ml phosphatidylserine dioleoyl were employed. The corresponding background binding because of endogenous PKC isozymes was 3320 Ϯ 560 dpm. These data indicate that the background binding does not significantly affect the K d value of FLAG-DGK␥. The K d value for PDBu binding of FLAG-DGK␥ was 4.4 Ϯ 0.3 nM as shown in Fig. 4f. This is in good agreement with that of rat-DGK␥-C1A (3.6 nM), indicating that DGK␥ is a new receptor of tumor-promoting phorbol esters.
Affinity of PKC Activators Other than PDBu to Bind DGK␥-C1A-The binding affinity of PKC activators other than PDBu for rat-DGK␥-C1A was evaluated by inhibition of the specific binding of [ 3 H]PDBu. Dose-response curves were plotted for each compound, and the concentration at which 50% of the specific [ 3 H]PDBu binding was inhibited (IC 50 ) was calculated. The binding constant (K i ) was calculated from the K d for PDBu of rat-DGK␥-C1A by the method of Sharkey and Blumberg (30). We determined the K i values of OAG, bryostatin 1, IL-V, and Ing-3-Bz. OAG is a membrane-permeable analogue of natural diacylglycerols without tumor promoting activity (35). Bryostatin 1 is a macrocyclic lactone with potent antileukemic activity (36). IL-V (37-39) and Ing-3-Bz (40) are naturally occurring tumor promoters. Table III summarizes the binding data for these compounds. The K i values of PKC␥ and PKC⑀ C1 peptides were also determined as references. Rat-DGK␥-C1A as well as PKC C1 peptides bound significantly OAG, bryostatin 1, IL-V, and Ing-3-Bz. The ability of each ligand to bind rat-DGK␥-C1A was similar to that to bind PKC C1 peptides. It is noteworthy that DG can bind to rat-DGK␥-C1A. This is the first evidence that DG specifically binds to the C1 domain of DGK isozymes.
Translocation of Rat-DGK␥ to the Plasma Membrane with TPA Treatment-Because TPA induces an irreversible translocation of conventional and novel PKC isozymes by binding to the C1 domains (41,42), we examined the TPA-induced translocation of rat-DGK␥ with green fluorescent protein (GFP-DGK␥) in CHO-K1 cells. To clarify the role of each C1 domain in the translocation, C1A or C1B deletion mutants of DGK␥ fused with GFP (GFP-⌬C1A-DGK␥ and GFP-⌬C1B-DGK␥) were also analyzed. As shown in Fig. 5, the fluorescence of GFP-DGK␥, GFP-⌬C1A-DGK␥, and GFP-⌬C1B-DGK␥ was observed throughout the cytoplasm and in the nucleus. Treatment with 1 M TPA induced an obvious translocation of GFP-DGK␥ and GFP-⌬C1B-DGK␥ from the cytoplasm to the plasma membrane (Fig. 5, top and bottom panels). This translocation began within 30 s and remained for at least 60 min after the TPA treatment. In contrast, GFP-⌬C1A-DGK␥ did not respond to TPA (Fig. 5, middle panel). DISCUSSION It has long been believed that the cysteine-rich C1 domains in DGK isozymes do not bind phorbol esters (1, 2). However, there was circumstantial evidence that DGK␥ binds phorbol esters; TPA induced the translocation of DGK␥ from the cytoplasm to the plasma membrane of CHO-K1 cells (27). We have recently synthesized the cysteine-rich sequences of all conventional and novel PKC isozyme C1 domains by the solid-phase Fmoc strategy in high purity (23). These peptides folded by zinc chloride showed strong PDBu binding in the presence of phosphatidylserine with K d values in the nanomolar range comparable with the respective whole PKC isozymes (24). We adopted this method to prove the phorbol ester binding ability of DGK isozymes. Among the DGK C1 peptides with good sequence homology to the PKC C1 domains, the DGK␥-C1A peptides bound most strongly PDBu. There was no significant difference in PDBu binding affinity between rat-DGK␥-C1A and human-DGK␥-C1A (Fig. 4) although 6 amino acid residues differ between them (Fig. 2). This remarkable binding was also proved by using whole rat-DGK␥; the K d value for PDBu binding of rat-DGK␥-C1A (3.6 nM) was almost equal to that of whole rat-DGK␥ (4.4 nM). This is the first unambiguous evidence that DGK␥ is one of the specific receptors of phorbol esters. The lack of PDBu binding by rat-DGK␥-C1B clearly showed that the major binding site in whole rat-DGK␥ is the C1A domain. The TPA-induced translocation experiment using the C1A or C1B deletion mutants of rat-DGK␥ also indicated that TPA binds directly to the C1A domain of DGK␥ to cause its translocation from the cytoplasm to the plasma membrane of CHO-K1 cells as observed for PKC isozymes (41,42). These results suggest that DGK␥ as well as PKC isozymes should be taken into account to interpret the biological phenomena evoked by phorbol esters. Furthermore, popular PKC activators other than phorbol esters (bryostatin 1, IL-V, and Ing-3-Bz) bound to rat-DGK␥-C1A like to PKC isozymes (Table III), suggesting that DGK␥ also serves as the receptor of these PKC activators.
Contrary to our results, Sakane et al. (26) did not detect specific PDBu binding in whole human-DGK␥. In experiments using a whole enzyme, there are several factors to attenuate the binding potency as follows: instability or low concentration of the enzyme, problem with the folding, and so on. The contribution of the molecular chaperon would be large because rat-DGK␥ expressed in the E. coli system did not show any detectable PDBu binding (data not shown). Moreover, the quality of the commercially available phosphatidylserine might be one of the reasons for the discrepancy between Sakane et al. (26) and our group. We have found recently that a synthetic phosphatidylserine such as phosphatidylserine dioleoyl (Sigma) gives more reproducible PDBu binding data than natural phosphatidylserine derived from the bovine brain. 2 We experienced that some lots of natural phosphatidylserine did not work at all as a cofactor in the PDBu binding assay using whole PKC isozymes. Long storage or long purification process of natural phosphatidylserine would result in oxidation to produce reactive oxygen species, which might have oxidized the cysteine-rich sequences of the DGK isozyme C1 domains to abolish their PDBu binding ability.
According to the sequence analysis, the DGK␤-C1A peptides as well as the DGK␥-C1A peptides have the highest sequence homology to the PKC isozyme C1 peptides (Table II): 95% homology for rat-DGK␤-C1A and 100% homology for human-DGK␤-C1A. Both DGK␤-C1A peptides showed significant PDBu binding as expected (Fig. 3). However, the K d values for rat-DGK␤-C1A and human-DGK␤-C1A were considerably different. Human-DGK␤-C1A showed potent PDBu binding affinity (K d ϭ 14.6 nM), whereas the binding affinity of rat-DGK␤-C1A was very weak (K d ϭ 202 nM). The Val-44 deletion in the rat-DGK␤-C1A sequence is deduced to be responsible for this weak binding affinity because the C-terminal cysteine residue is one of the zinc coordination sites (20 -22). After the completion of this study, we reexamined the sequence of rat-DGK␤-C1A and found that rat-DGK␤-C1A as well as human-DGK␤-C1A have a valine residue at position 44. 3 These results indicate that both rat-DGK␤ and human-DGK␤ are targets of tumor-promoting phorbol esters. The Val-44 deletion in rat-DGK␤ proved to be reported accidentally. 4 The K d value of human-DGK␤-C1A was ϳ5 times larger than that of human-and rat-DGK␥-C1A peptides (Fig. 4), suggesting that there are additional important residues other than the 20 residues (Fig. 2) that play a significant role in the PDBu binding. After careful comparison of the sequence of DGK␤-C1A with DGK␥-C1A, we focused on the Ala-12 residue which might decrease the PDBu binding affinity of the DGK␤-C1A peptides. The Thr-12 residue of PKC␦-C1B was deduced to contribute to the hydrogen bonding with phorbol 13-acetate in an x-ray analysis (20). Molecular modeling and site-directed mutagenesis studies also suggest the involvement of the ␤-hydroxyl group of the Thr-12 residue in phorbol ester binding (43). Moreover, the Thr-12 residue is strictly conserved in the conventional and novel PKC isozymes that have potent PDBu binding ability (24). These considerations led us to synthesize the A12T mutant of human-DGK␤-C1A. This mutant peptide showed a K d value for PDBu quite similar to that of the DGK␥-C1A peptides (Fig. 4), indicating that the ␤-hydroxyl group of Thr-12 is involved in the hydrogen bonding between DGK␤-C1A and PDBu.
The potent PDBu binding ability of the DGK␤-C1A peptide suggests that DGK␤ is also a new receptor of tumor-promoting phorbol esters. In fact, Caricasole et al. (15) have reported recently that human-DGK␤ was redistributed from the cytoplasm to the plasma membrane of HEK-293 cells upon TPA treatment. However, we could not confirm this translocation in our assay system using CHO-K1 cells because almost all GFP-DGK␤ exists in the plasma membrane and not in the cytoplasm (data not shown   not show any significant PDBu binding (Fig. 3). What then is the intrinsic role of the C1 domains of DGK isozymes? They might bind DG by analogy to PKC isozymes. However, DG binding cannot be demonstrated directly using radioisotopelabeled DG because DG is highly hydrophobic like TPA whose specific binding to PKC isozymes cannot be proved by the usual method (44). The sequence homology of the DGK C1B and C1C domains to the PKC C1A and C1B domains is considerably lower than that of the DGK C1A domains, but the DGK-C1B and C1C domains have the conserved sequence of 15 amino acid residues at their C terminus which is characteristic to DGK isozymes. Because of this sequence, DGK-C1B and C1C domains are called "extended cysteine-rich domains" (1, 2), and must be closely related to the function of DGK isozymes. It has been shown recently (2,45) that the extended cysteine-rich sequence of DGK-C1C is necessary for DGK activity. If DG binds to the C1B or C1C domains of DGK isozymes, these C1 domains might assist DG to bind to the catalytic domain for phosphorylation. To verify this hypothesis, it is indispensable to develop a new ligand with moderate hydrophobicity that shows strong affinity to bind these extended cysteine-rich domains. Although OAG bound to the C1A domain of DGK␥-C1A (Table III), this binding would be related to the translocation of DGK␥ rather than to the phosphorylation of DG.
Overall, we have identified DGK␥ and DGK␤ as novel phorbol ester-binding sites by the chemical synthesis of the cysteine-rich sequences of DGK isozyme C1 domains. These C1 peptide receptors obtained in high yield and high purity gave more reproducible binding results than the whole enzymes prepared by DNA recombination techniques. The synthetic PKC C1 homology domain peptides are now established as an effective screen for identifying new phorbol ester receptors. They are also useful for identifying selective binding agents for individual PKC and DGK isozymes.