Diacylglycerol Kinase δ Phosphorylates Phosphatidylcholine-specific Phospholipase C-dependent, Palmitic Acid-containing Diacylglycerol Species in Response to High Glucose Levels*

Background: Diacylglycerol (DG) kinase (DGK) δ is activated by acute high glucose stimulation. Results: DGKδ high glucose-dependently phosphorylates 30:0-, 32:0-, and 34:0-DG and interacts with phosphatidylcholine-specific phospholipase C (PC-PLC). Conclusion: DGKδ utilizes palmitic acid-containing DG species and metabolically connects with PC-PLC. Significance: The newly identified PC-PLC/DGKδ pathway could play an important role in insulin signaling and glucose uptake. Decreased expression of diacylglycerol (DG) kinase (DGK) δ in skeletal muscles is closely related to the pathogenesis of type 2 diabetes. To identify DG species that are phosphorylated by DGKδ in response to high glucose stimulation, we investigated high glucose-dependent changes in phosphatidic acid (PA) molecular species in mouse C2C12 myoblasts using a newly established liquid chromatography/MS method. We found that the suppression of DGKδ2 expression by DGKδ-specific siRNAs significantly inhibited glucose-dependent increases in 30:0-, 32:0-, and 34:0-PA and moderately attenuated 30:1-, 32:1-, and 34:1-PA. Moreover, overexpression of DGKδ2 also enhanced the production of these PA species. MS/MS analysis revealed that these PA species commonly contain palmitic acid (16:0). D609, an inhibitor of phosphatidylcholine-specific phospholipase C (PC-PLC), significantly inhibited the glucose-stimulated production of the palmitic acid-containing PA species. Moreover, PC-PLC was co-immunoprecipitated with DGKδ2. These results strongly suggest that DGKδ preferably metabolizes palmitic acid-containing DG species supplied from the PC-PLC pathway, but not arachidonic acid (20:4)-containing DG species derived from the phosphatidylinositol turnover, in response to high glucose levels.

Type 2 diabetes is expected to afflict over 300 million people worldwide by 2015 (1). The characteristic features of type 2 diabetes include insulin resistance, glucose intolerance, hyper-glycemia, and often, hyperinsulinemia (2). Glucose-induced insulin resistance is associated with a temporal increase in the intracellular diacylglycerol (DG) 2 mass in skeletal muscle (3).
DGK␦ is highly expressed in skeletal muscle (13), which is a major insulin-target organ for glucose disposal (14). Chibalin et al. (15) demonstrated that DGK␦ regulates glucose uptake and that a decrease in DGK␦ expression resulted in the aggravation of type 2 diabetes. Long term exposure (96 h) to high glucose medium decreased DGK␦ protein levels in primary cultured skeletal muscle cells, and the transcription of DGK␦ and the levels of DGK␦ protein were also reduced in skeletal muscles from type 2 diabetes patients (15). Moreover, DGK␦ haploinsufficient mice (DGK␦ ϩ/Ϫ ) exhibited decreased total DGK activity, reduced DGK␦ protein levels, and the accumulation of DG in skeletal muscle. The increase in the amount of DG caused the increase in the phosphorylation of protein kinase C (PKC) ␦ and a reduction in the expression of the insulin receptor and insulin receptor substrate-1 proteins involved in insulin signaling (15). Furthermore, Miele et al. (16) reported that acute high glucose exposure (within 5 min) increased DGK␦ activity in skeletal muscle cells followed by a reduction of PKC␣ activity and transactivation of the insulin receptor signal.
Hence, these studies indicate that DG consumed by DGK␦ in response to high glucose exposure is a key regulator of glucose uptake in skeletal muscle cells. DGK␦1 translocated from the cytoplasm to the plasma membrane in mouse myoblast C2C12 cells within 5 min of short term exposure to a high glucose concentration, whereas DGK␦2 was located in punctate vesicles irrespective of the glucose concentration (17).
Mammalian cells contain at least 50 structurally distinct molecular DG species because DG contains a variety of fatty acyl moieties at positions 1 and 2 (18). In general, DGs containing arachidonic acid (20:4), especially 18:0/20:4-DG (38:4-DG), in the phosphatidylinositol (PI) turnover are important molecules that serve as second messengers for PKC activation (18). Moreover, previous studies have demonstrated that DGK⑀ preferably phosphorylates arachidonic acid-containing DGs derived from PI turnover (19,20). Therefore, it is generally believed that all DGKs preferentially metabolize 38:4-DG for the regulation of signal transduction. However, the DG molecular species phosphorylated by DGK␦ in response to glucose stimulation remain unknown.
Establishment of a Stable Cell Line Overexpressing DGK␦-To establish C2C12 cells stably expressing human DGK␦2, the cells were transfected with pAcGFP-DGK␦2 (11,17) using PolyFect (Qiagen) according to the instruction manual and were selected with 800 g/ml G418 for 2 weeks. Single colonies were isolated and then were then grown in DMEM containing 10% FBS.
Glucose Stimulation and Treatment with Lipid Metabolism Enzyme Inhibitors-Glucose stimulation was performed as reported previously (16). Briefly, untransfected C2C12 myoblasts and C2C12 myoblasts transfected with Stealth RNAi duplexes were grown on poly-L-lysine (Sigma-Aldrich)-coated culture dishes. The cells were rinsed and incubated in glucosefree medium (16)  Lipid Extraction and Western Blot Analysis-The cells grown under each culture condition were harvested and lysed in icecold lysis buffer (50 mM HEPES, pH 7.2, 150 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, cOmplete TM EDTA-free protease inhibitor (Roche Diagnostics)) followed by centrifugation at 1,000 ϫ g for 5 min at 4°C. Total lipids were extracted from the cell lysates (1.0 mg of protein), in which DGK␦ expression was confirmed by Western blot analysis using an anti-DGK␦ antibody (13), according to the method of Bligh and Dyer (26). The extracted lipids were used for subsequent MS analyses.
Analysis of PA Molecular Species-PAs in extracted cellular lipids (5 l) containing 40 pmol of the 14:0/14:0-PA internal standard (Sigma-Aldrich) were analyzed separately by LC/ESI-MS using an Accela LC system (Thermo Fisher Scientific) coupled online to an Exactive Orbitrap MS (Thermo Fisher Scientific) equipped with an ESI source as described previously (21). The MS peaks are presented in the form of X:Y, where X is the total number of carbon atoms and Y is the total number of double bonds in both acyl chains of the PA.
For the identification of fatty acid residues in PA molecular species by ESI-MS/MS, PA molecular species (28:0 -40:0-PA) were fractionated using the above LC/ESI-MS system equipped with an FC 203B fraction collector (Gilson). The mixture of these isolated PA molecular species was infused into an Exactive Orbitrap MS (Thermo Fisher Scientific) equipped with a syringe pump (an infusion rate of 5 l/min) and an ESI source. A collision energy of 40 eV was used to obtain fragment ions.
Analysis of DG Molecular Species-The isolation of DG was performed according to previously reported procedures (27). The extracted cellular lipids (per 1 mg of protein) were developed on Silica Gel 60 high performance thin layer chromatography plates (Merck, 10 ϫ 20 cm) using hexane/diethyl ether/ acetic acid (75:25:1, v/v). After development, DG was extracted from silica gel and redissolved in 200 l of methanol:chloroform (9:1, v/v) containing 1 g/ml 12:0/12:0-DG (Avanti Polar Lipids), and 10 l of 100 mM sodium acetate were added to each sample (28). MS analysis was performed on an Exactive Orbitrap MS (Thermo Fisher Scientific) equipped with a Fusion 100T syringe pump (an infusion rate of 5 l/min, Thermo Fisher Scientific) and an ESI source. The ion spray voltage was set to 5 kV in the positive ion mode. The capillary temperature was set to 300°C.
Immunoprecipitation and Measurement of PC-PLC Activity-The glucose-stimulated cells stably expressing human DGK␦2 were harvested and lysed in ice-cold lysis buffer (50 mM HEPES, pH7.2, 150 mM NaCl, 5 mM MgCl 2 , 1% Nonidet P-40, 1 mM dithiothreitol, cOmplete TM EDTA-free protease inhibitor (Roche Diagnostics)) for immunoprecipitation. The mixtures were centrifuged at 12,000 ϫ g for 5 min at 4°C to yield the cell lysates. 500 g of the cell lysates were incubated with normal rabbit IgG (2 g, Santa Cruz Biotechnology) or rabbit anti-DGK␦ antibody (2 g) (13,29) at 4°C overnight and incubated with protein A/G PLUS-agarose (Santa Cruz Biotechnology) for an additional 1 h. The bead-bound proteins were washed with ice-cold wash buffer (50 mM HEPES, pH 7.2, 100 mM NaCl, 5 mM MgCl 2 , 0.1% Triton X-100, 10% glycerol, 20 mM NaF) four times and resolved in 70 l of 1ϫ reaction buffer (50 mM Tris-HCl, pH 7.4, 140 mM NaCl, 10 mM dimethylglutarate, 2 mM CaCl 2 ) in the Amplex Red PC-PLC assay kit (Molecular Probes-Life Technologies). In this enzyme-coupled assay, PC-PLC activity is monitored indirectly using 10-acetyl-3,7-dihydroxyphenoxazine (Amplex Red reagent), a sensitive fluorogenic probe for H 2 O 2 . First, PC-PLC converts PC to form phosphocholine and DG. After the action of alkaline phosphatase, which hydrolyzes phosphocholine, choline is oxidized by choline oxidase to betaine and H 2 O 2 . Finally, H 2 O 2 , in the presence of horseradish peroxidase, reacts with Amplex Red reagent in a 1:1 stoichiometry to generate the highly fluorescent product, resorufin. Resorufin has absorption and fluorescence emission maxima of ϳ571 nm and 585 nm, respectively. 50-l aliquots of the mixtures were used for the measurement of PC-PLC activities, and 10 l of the mixtures were used for Western blot analysis.
Statistics-All LC/ESI-MS data were normalized based on the protein content and the intensity of the internal standard. The data were represented as the mean Ϯ S.D. Statistical analysis was performed using the two-tailed t test or analysis of variance followed by Tukey's post hoc test.

Increase in the Amount of PA by Acute Stimulation with High
Glucose-We first examined whether the amount of total PA was increased in C2C12 myoblasts stimulated with 25 mM glucose. As shown in Fig. 1A, LC/ESI-MS analysis indicated that exposure to high glucose levels (for 5 min) statistically increased the total PA amounts (1.23-fold, p Ͻ 0.005). In addition, the stimulation significantly increased the amounts of C30 to C36 PA molecular species, with the exception of 36:1-PA (Fig. 1B). However, the stimulation did not substantially affect the production of C38 to C40 PA molecular species, including 38:4-PA, with the exception of 38:6-PA.
We investigated the high glucose-dependent increases of total PA amount and PA molecular species in C2C12 myoblasts at different time points. After 5 min of glucose stimulation, the levels of total PA and PA molecular species were significantly increased ( Fig. 1, C and D). However, total PA and PA molecular species levels returned close to basal levels by prolonging the incubation with high glucose concentrations for up to 15 and 30 min. We confirmed that DGK activity in vitro was increased by glucose stimulation for 5 min (data not shown). These results strongly suggest that C2C12 myoblasts and L6 myotubes (16) have essentially the same lipid metabolism pathway to produce PA in response to acute glucose stimulation.
We confirmed the changes in the amounts of PA molecular species in C2C12 myotubes in response to acute high glucose stimulation (5 min). The glucose-stimulated C2C12 myotubes showed essentially the same results ( Fig. 1, E and F) as those obtained with C2C12 myoblasts (Fig. 1, A and B). The results support that C2C12 myoblasts and myotubes possess essentially the same lipid metabolism pathway to produce PA in response to high glucose stimulation. Because C2C12 myoblasts were more efficiently transfected with siRNAs than C2C12 myotubes, C2C12 myoblasts were used for identification of PA molecular species produced by DGK␦ in response to high glucose stimulation.
Glucose stimulation substantially increased the amounts of various DG species (Fig. 3). However, DGK␦-siRNA-1 failed to significantly affect the amounts of 30:0-, 32:0-, 34:1-, and 34:0-DG molecular species both in the absence and in the presence of high glucose levels. Therefore, it is likely that the decreases in the amounts of 30:0-, 32:0-, 34:1-, and 34:0-PA were not caused by decreased amounts of the corresponding DG species.
Effect of Overexpression of DGK␦ on the Production of PA Molecular Species-To confirm the results of the siRNA experiments, we evaluated the result of DGK␦2 overexpression on high glucose-dependent production of PA species in C2C12 cells. In response to high glucose, the levels of 30:0-, 32:0-, and 34:0-PA statistically increased in C2C12 cells stably expressing DGK␦2 when compared with control cells (Fig. 4B). Moreover   (Table 1).
In Vitro DGK␦ Activity-We examined whether the preference of DGK␦2 for palmitic acid (  substrates. As shown in Fig. 5, the levels of 32:0-and 34:1-PA generated by DGK␦2 were similar to or slightly lower than that of 38:4-PA. These results indicate that DGK␦2 does not exhibit intrinsic substrate selectivity for particular DG molecular species, 32:0-DG, in vitro. Therefore, we hypothesized that DGK␦ accomplishes apparent substrate selectivity in C2C12 cells by accessing a DG pool containing only 30:0-, 32:0-, and 34:0-DG, and not based on the intrinsic properties of the enzyme. Effects of Inhibitors of Lipid Metabolism Enzymes on High Glucose Level-induced PA Production-To test this hypothesis, we next searched for the lipid metabolic pathway that supplies 30:0-, 32:0-, and 34:0-DG species as a substrate for DGK␦2. There are three pathways that produce DG, 1) the de novo pathway (30,31), 2) the PLD/PA phosphatase pathway (32), and 3) the PC-specific PLC pathway (33). The treatment with 20 M TOFA, which inhibits acetyl-CoA carboxylase involved in the de novo synthesis of DG (23, 24), did not decrease the glucosestimulated production of PA molecular species (Fig. 6A). Moreover, 100 nM FIPI, which inhibits PLD involved in DG generation from PC through the action of PA phosphatase (25), reduced the amounts of most of the PAs in the absence of high glucose stimulation (data not shown). However, this compound failed to attenuate the glucose-stimulated production of PA molecular species (Fig. 6B). These results strongly suggest that these pathways are not involved in the DG supply to DGK␦2.   D609 is an inhibitor of PC-PLC (22), which generates DG via PC hydrolysis (34). Treatment with 100 M D609 strongly inhibited the high glucose stimulation-responsive production of 30:0-, 32:0-, and 34:0-PA to their basal levels (Fig. 7A), suggesting that DGK␦2 utilizes DG species supplied from the PC-PLC pathway.
We next confirmed that D609 inhibited the production of DG molecular species, including 30:0-, 32:0-, and 34:0-DG. This inhibitor statistically attenuated the amounts of 30:0-, 32:0-, and 34:0-DG in the absence of high glucose stimulation (Fig. 7B). However, D609 inhibited high glucose-dependent increases for all of the C30-C34 DG species (Fig. 7C). These results suggest that, in response to acute high glucose stimulation (5 min), DGK␦2 can utilize DG species that are supplied from the PC-PLC pathway, in both high glucose-independent and high glucose-dependent manners.
Linkage between PC-PLC and DGK␦-To further examine the linkage between the PC-PLC pathway and DGK␦2, we determined whether D609 and DGK␦-siRNA-1 additively affected the high glucose-dependent increases of 30:0-, 32:0-, and 34:0-PA. If DGK␦2 utilizes DG species supplied from the PC-PLC pathway, it would be expected that reduced expression of DGK␦2 via DGK␦-siRNA-1 would not enhance the effect of the PC-PLC inhibitor. It was confirmed that the expression of DGK␦2 was substantially reduced by DGK␦-siRNA-1, even in the presence of D609 (Fig. 8A). As shown in Fig. 8B, DGK␦-siRNA-1 failed to further inhibit the glucose-dependent increases of 30:0-, 32:0-, and 34:0-PA in the presence of D609. These results strongly suggest that 30:0-, 32:0-, and 34:0-DG phosphorylated by DGK␦2 in response to acute high glucose exposure are generated, at least in part, by PC hydrolysis catalyzed by PC-PLC.
We next examined whether DGK␦ directly or indirectly interacted with PC-PLC. To this end, we used C2C12 cells stably overexpressing DGK␦2 (Fig. 4) and stimulated the cells with high glucose. We confirmed that DGK␦2 was immunoprecipitated with the anti-DGK␦ antibody (Fig. 8C). Because the molecular identity of mammalian PC-PLC remains unclear (35), its antibody is unavailable. Therefore, we determined PC-PLC activity in the immunoprecipitates using the Amplex Red PC-PLC assay kit, which detects phosphocholine generated by PC-PLC. As demonstrated in Fig. 8D, PC-PLC activity was clearly co-immunoprecipitated with DGK␦2. The assay does not detect the activity of sphingomyelin synthase, which produces DG and sphingomyelin, but not phosphocholine. The contribution of PLD, which hydrolyzes PC to PA and choline, can be accounted for by elimination of alkaline phosphatase from the assay (see "Experimental Procedures"). However, when the assay was performed in the absence of alkaline phosphatase, the activity was not detectable. Taken together, these results strongly suggest that DGK␦2 utilizes DG species supplied from PC-PLC-dependent PC hydrolysis in response to high glucose (Fig. 9).

DISCUSSION
The increase in PA molecular species by stimulation with high glucose levels has not been identified until now. Moreover, it has not been reported that high glucose induces total PA production. The main reasons for this are that PA species are minor components and it is difficult to quantify the amounts of PA molecular species using conventional LC/ESI-MS methods. To overcome this difficulty, we recently established an LC/ESI-MS method specialized for PA species (21). In this study, we revealed for the first time that acute high glucose stimulation statistically increased the PA mass and number of molecular species using the newly developed method (Fig. 1).
The role of sphingomyelin synthase as a potential PC-PLC was indicated (36). We cannot rule out the possibility that DGK␦2 partly utilizes sphingomyelin synthase-dependent DG. However, it is likely that DGK␦2 phosphorylates DG species generated, at least in part, by PC-PLC because the co-immunoprecipitates with DGK␦2 contained PC-PLC activity.
The molecular identity of PC-PLC remains unclear (35). In this study, DGK␦ was revealed to directly or indirectly associate with PC-PLC. With the pulldown of PC-PLC activity with DGK␦2, there may be an opportunity to identify the unidentified PC-PLC enzyme by proteomics approaches. Therefore, DGK␦2 may serve as a good tool to search for the PC-PLC molecule.
Recently, Shulga et al. (37) and Lowe et al. (38) reported that DGK␦ positively regulated lipid synthesis, including DG and PA, during adipocyte differentiation. However, unlike for acute high glucose stimulation, a significant preference against DG and PA was not found. The increases were, at least in part, a result of promoting the de novo synthesis of fatty acids. However, in this study, an inhibitor of acetyl-CoA carboxylase TOFA did not decrease glucose-stimulated PA production (Fig.  6A). Because differentiation is a long term event, the difference between acute high glucose stimulation in C2C12 myoblasts and adipocyte differentiation may be due to distinct supply pathways and/or fatty acid conversion during long term culture through the remodeling pathway (Lands' cycle) (39).