Protein Kinase Cδ Plays a Non-redundant Role in Insulin Secretion in Pancreatic β Cells*

Protein kinase C (PKC) is considered to modulate glucose-stimulated insulin secretion. Pancreatic β cells express multiple isoforms of PKCs; however, the role of each isoform in glucose-stimulated insulin secretion remains controversial. In this study we investigated the role of PKCδ, a major isoform expressed in pancreatic β cells on β cell function. Here, we showed that PKCδ null mice manifested glucose intolerance with impaired insulin secretion. Insulin tolerance test showed no decrease in insulin sensitivity in PKCδ null mice. Studies using islets isolated from these mice demonstrated decreased glucose- and KCl-stimulated insulin secretion. Perifusion studies indicated that mainly the second phase of insulin secretion was decreased. On the other hand, glucose-induced influx of Ca2+ into β cells was not altered. Immunohistochemistry using total internal reflection fluorescence microscopy and electron microscopic analysis showed an increased number of insulin granules close to the plasma membrane in β cells of PKCδ null mice. Although PKC is thought to phosphorylate Munc18-1 and facilitate soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors complex formation, the phosphorylation of Munc18-1 by glucose stimulation was decreased in islets of PKCδ null mice. We conclude that PKCδ plays a non-redundant role in glucose-stimulated insulin secretion. The impaired insulin secretion in PKCδ null mice is associated with reduced phosphorylation of Munc18-1.

Altered regulation of insulin secretion is a common feature of type 2 diabetes mellitus, although the underlying mechanism is not fully understood. The mechanism of glucose-stimulated insulin secretion involves closure of ATP-sensitive potassium channels by increased levels of glucose metabolites followed by depolarization of the plasma membrane and increased influx of Ca 2ϩ via voltage-dependent gating of Ca 2ϩ channels. The resultant elevation of intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) triggers transport of insulin into the plasma membrane and exocytosis (1). Although the events involved in initiating glucose-stimulated insulin secretion are well studied, those of regulated exo-cytosis are considerably less clear, although it is reported that several molecules expressed in ␤ cells play a crucial role in exocytosis (2).
Various protein kinases are activated by downstream signals of glucose metabolism; however, their precise contribution to insulin secretion is not clear yet (3). Protein kinase C (PKC) 2 is one such protein kinase. Previous studies revealed that phorbol 12-myristate 13-acetate, which activates PKC, could trigger insulin secretion (4 -10). PKC is a serine/threonine kinase characterized by molecular structure and activation requirement. The isoforms consist of the conventional PKCs (␣, ␤, and ␥), which are sensitive to Ca 2ϩ and diacylglycerol (DAG), novel PKCs (␦, ⑀, , and ), which are sensitive to DAG only, and atypical PKCs (, ), which do not respond to either Ca 2ϩ or DAG (11,12). The ␤ cells express PKC␣ and -␦ as major isoforms of PKC. However, other isoforms of PKC are also expressed in ␤ cells (7,(13)(14)(15)(16). The importance of each PKC isoform in ␤ cell function is not clear yet, mainly due to the scarcity of true isoform-specific inhibitors.
Pancreatic ␤ cells express the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins such as vesicular-associated membrane protein (VAMP), synaptosomal-associated protein of 25 kDa (SNAP25), and syntaxin-1A, which form the core machinery for vesicular docking on target membrane and membrane fusion (17,18). In the machinery of regulated exocytosis, additional families of other proteins function as SNARE regulators, such as Sec/Munc18 proteins (19). Munc18-1, a family member of Munc18, binds with high affinity to syntaxin-1A. In the Munc18-1-syntaxin-1A complex, syntaxin is in a closed conformation, whereby its N-terminal helices are folded back onto the C-terminal SNARE motif. Release of syntaxin into an open conformation seems to be required for the association of syntaxin with the SNARE complex (19). Thus, it seems that Munc18-1 must be released from syntaxin-1 for the fusion event. Previous data showed that PKC could phosphorylate Munc18-1 and allow it to be released for binding with syntaxin-1A (20). However, several studies have cast doubt on the role of Munc18-1 and its modulators (19). In addition, the involvement of these events in ␤ cell is not fully elucidated yet.
In the present study, as a first step to elucidate the involvement of PKC␦ in ␤ cell function, we investigated the function of PKC␦ in ␤ cells using mice with disruption of the PKC␦ gene (PKC␦ Ϫ/Ϫ ). The results showed that disruption of the PKC␦ gene impaired exocytosis of insulin containing granules, concomitant with decreased glucose-stimulated phosphorylation of Munc18-1.

MATERIALS AND METHODS
Animals-The study protocol was reviewed and approved by the Animal Care and Use Committee of Juntendo University. All mice were housed in specific pathogen-free barrier facilities, maintained 12-h light/dark cycle, and fed a standard rodent food (Oriental Yeast Co., Osaka, Japan) and provided with water ad libitum. In this study, we used 8 -20-week-old male mice with C57BL/6J background. We mated PKC␦ ϩ/Ϫ mice to generate PKC␦ Ϫ/Ϫ mice, PKC␦ ϩ/Ϫ mice, and wild type mice. Then these mice were used for further analyses. PCR analysis for determination of genotype was performed as described previously (21).
Measurement of Blood Glucose and Insulin Levels for Phenotypic Characterization-After an overnight fast (16 h), agematched 8-, 12-, 16-, and 20-week-old male mice were injected intraperitoneally with glucose at 1.0 g/kg of body weight. For the insulin tolerance test, age-matched 12-week-old mice were injected intraperitoneally with insulin (0.75 units/kg). A few microliters of blood samples were taken from the tail vein in awake mice, and glucose levels were measured using whole blood with a compact glucose analyzer (Free Style, KISSEI Pharmaceutical Co., Nagano, Japan) at the indicated time points. To measure insulin levels, whole blood samples were collected and centrifuged with EDTA and aprotinin. After centrifugation, plasma was stored at Ϫ80°C until analysis. Insulin levels were measured by enzyme-linked immunosorbent assay kit (Morinaga Co., Kanagawa, Japan).
Assay of Insulin Secretion from Isolated Islets and Islet Insulin Content-Pancreatic islets were isolated from 12-week-old mice by collagenase digestion, as described previously (17). The islets were preincubated for 30 min in Krebs-Ringer solution (KRS) at 37°C under 5% CO 2 and saturated humidity. Sizematched 5 islets were incubated in 400 l of KRS containing various insulinotrophic agents for 30 min at 37°C. For measurement of islet insulin content, islets were solubilized in acidethanol solution overnight at 4°C. All samples were stored at Ϫ80°C until assay. Insulin concentration using isolated islets was analyzed by radioimmunoassay (LINCO insulin kit, LINCO Research Inc., St. Charles, MO). Islet insulin content was calculated after correction by DNA contents.
Islet Perifusion-Pancreatic islets were isolated as mentioned above and preincubated for 30 min in KRS. Size-matched 50 islets were placed in each column. Then, the kinetics of insulin release was studied using the perfusion system described previously (22).
Immunohistochemical Analysis and Quantification of ␤ Cell Area-After anesthetization, 12-week-old mice were perfused intracardially with 4% paraformaldehyde, and the pancreas was removed and fixed overnight in a solution of 4% paraformaldehyde at 4°C. The fixed tissue was embedded in paraffin and then cut into 5-m-thick sections and mounted on slides. The sections were subjected to immunohistochemical staining for insulin using guinea pig anti-human insulin antibody (Linco Inc.) as described previously (23).
The ␤ cell area was determined in 10 insulin-stained sections separated by 200 m by evaluating the mean area of the individual cell. Images were captured with a microscope (E800; Nikon, Tokyo) connected to a digital camera (Sony, Tokyo). The area of insulin-stained pancreatic islets was determined automatically using image analysis software Pro4.5J Planetron (Tokyo). ␤ cell area was calculated using the following formula: islet ␤ cell area (%) ϭ (area stained by insulin antibody/whole pancreatic area) ϫ 100. Nonspecific insulin staining was excluded from quantification of ␤ cell area.
Measurement of [Ca 2ϩ ] i in Single Mouse ␤ Cells-Isolated islets were placed in 100 l of Ca 2ϩ -free KRS containing 1 mM EGTA and gently agitated till turbidity due to the floating dispersed islets. After centrifugation at 500 rpm for 1 min, 50 l of the supernatant was discarded, and 150 l fresh KRS was decanted for washing. Finally, 50 l of residual dispersed islets were again gently agitated and incubated on 0.04 mg/ml fibronectin-coated (Koken Co., Tokyo) Petri dishes with a 35-mm cover glass in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal calf serum at 37°C and under 5% CO 2 for 3 days. Cells were loaded with 2 M Fura2/AM (Molecular Probes, Eugene, OR) in 200 l of KRS containing 2.8 mM glucose for 20 min at 37°C. After incubation, Fura2-loaded cells were placed in KRS supplemented with 2.8 or 16.7 mM glucose on a micro-warm plate (Kitazato Co., Shizuoka, Japan) maintained at 37°C and mounted on the stage of TE300 microscope (Nikon, Tokyo). The loaded cells were excited at 340 and 380 nm; the florescence emitted at 510 nm was captured by an intensified charge-couple device camera. The images were analyzed by ARGUS/HiSCA system (Hamamatsu Photonics, Shizuoka).
Electron Microscopy-Each mouse at 12 weeks of age was anesthetized with pentobarbital sodium, and the pancreas was removed after cardiac perfusion and fixed in a solution of 2.5% glutaraldehyde. The pancreas was cut into pieces of ϳ1 mm 3 . The pieces were fixed for 2 h in 2.5% glutaraldehyde buffered to pH 7.4 with phosphate buffer and treated with osmium tetroxide for 2 h at 4°C. The tissues were dehydrated with graded concentrations of ethanol solutions and then embedded with Epon 812. Thin sections were cut with a Leica Ultracut UCT using a diamond knife and then stained with uranyl acetate followed by lead citrate. Electron micrographs were taken with a JEOL JEM-1200EX electron microscope operated at 80 kV.
To measure the number of insulin granules according to their distance from the granule center, isolated islets obtained from 5 mice were incubated in KRBB (2.8 or 16.7 mmol/liter glucose) at 37°C for 60 min. After incubation, islets were fixed with 2% paraformaldehyde, 2% glutaraldehyde, 0.2% picric acid in 0.1 M cacodylate buffer, pH 7.4, for 90 min at room temperature. Fixed islets were further treated by osmium tetroxide and then embedded in epoxy-resin. Micrographs were taken randomly at a ϫ3000 magnification. Using image analysis software (Image-Pro Plus Version 4.0, Planetron), we counted the total number of insulin granules that resided within 500 nm of the plasma membrane. Then, the insulin granules were categorized according to their distance from the granule center to the plasma membrane (0 -100, 100 -200, 200 -300, 300 -500 nm), and the number of insulin granules in each fraction was counted. Data were presented as % distribution below the plasma membrane. In these experiments, 100% corresponds to total counts of insulin granules that resided within 500 nm of the plasma membrane.
Total Internal Reflection Fluorescence (TIRF) Imaging Study-The Olympus total internal reflection system (Olympus, Tokyo) was used for counting the number of insulin granules 80 nm from the plasma membrane in ␤ cells after immunostaining of insulin as described previously (24).
In Vitro Pulldown Assay-Plasmid coding His 6 -fused Munc18-1 protein (pQE-Munc18-1) (20), kindly provided by Dr. Y. Takai (Osaka University), was transformed into BL-21 cells. His 6 fusion protein was prepared as the nickel-nitrilotriacetic acid-agarose bead (Qiagen, Tokyo)-bound form as described previously (25). Mouse full-length PKC␦ cDNA was amplified using RNA from mouse cerebrum by the RT-PCR method and cloned into pBSK vector containing T7 promoter. Then, [ 35 S]methionine-labeled PKC␦ protein was synthesized using the TNT-coupled reticulocyte lysate system (Promega, Madison, WI). Next 25 l of 35 S-labeled protein was mixed with 10 g of His 6 fusion Munc18-1 bound to nickel-nitrilotriacetic acid-agarose beads in a total volume of 600 l of interaction buffer (40 mM, pH 7.5, 50 mM KCl, 5 mM MgCl 2 , 0.2 mM EDTA, 1 mM dithiothreitol, and 0.5 m% Nonidet P-40). Samples were then incubated for 1 h at 4°C with gentle rocking, and the beads were washed 3 times with interaction buffer. The bound protein was eluted with 2ϫ sample buffer, fractionated on 7.5% SDS-PAGE, and visualized by autoradiography.
Immunoprecipitation and Immunoblot Analysis-HIT cells were cultured as described previously (26). Immunoprecipitation and immunoblot analyses were performed as described previously (27). Anti-PKC␣, -␤2, -␦, and -⑀, anti-Munc18-1, anti-Rab3, and anti-SNAP25 antibodies were purchased from BD Transduction Laboratories. Anti-syntaxin-1a antibody was purchased from Sigma. To detect phosphorylated Munc18-1 and SNAP25, we collected phosphorylated protein using Phos-phoProtein purification kit (Qiagen), which is based on affinity chromatography, a technique that delivers complete separation of phosphorylated and unphosphorylated proteins from a cell lysate. Approximately 1000 islets were used to collect phosphorylated proteins as a sample. Then, 20 g of each sample was used for immunoblotting using anti-Munc18-1 or anti-SNAP25 antibody. Immunoreactivity was visualized and quantitated using Fuji LAS 1000 (Fuji Film, Tokyo).
Data Analysis-Results are presented as the mean Ϯ S.E. Differences between groups were examined for statistical significance using the unpaired Student's t test. A p value less than 0.05 denoted the presence of a statistically significant difference.

RESULTS
PKC␦ Ϫ/Ϫ Mice Exhibit Glucose Intolerance and Impaired Insulin Secretion-As reported previously (21), PKC␦ Ϫ/Ϫ mice are born normal and fertile with no apparent abnormalities in general appearance or behavior, but they start to develop B lymphocytic autoimmune disease after the age 24 weeks. Body weight of PKC␦ Ϫ/Ϫ mice increased with age, but the gain was modestly lower than that of PKC␦ ϩ/Ϫ and wild type mice (Fig. 1a). Although fasting glucose level was comparable among the three groups (Fig.  1b), intraperitoneal glucose challenge at the age of 12 weeks identified glucose intolerance with reduced glucose-stimulated insulin secretion in PKC␦ Ϫ/Ϫ mice (Fig. 1, c and d). The glucose intolerance was continuously observed from 8 to 20 weeks of age, but no obvious deterioration was observed with aging (Fig. 1e). An insulin tolerance test did not demonstrate a significant decrease in insulin sensitivity in PKC␦ Ϫ/Ϫ mice (Fig. 1f). These results suggest that the impaired glucose tolerance observed in PKC␦ Ϫ/Ϫ mice is mainly due to impaired glucose-stimulated insulin secretion.
To search for the mechanism of impaired glucose-stimulated insulin release in PKC␦ Ϫ/Ϫ mice, we investigated islet morphology and ␤ cell area by immunohistochemistry. However, we could not find any obvious changes in islet morphology or significant differences in ␤ cell area between PKC␦ Ϫ/Ϫ mice and wild type (Fig. 2, a and b). The insulin content of isolated islets was also comparable (Fig. 2c). Furthermore, electron micro-FIGURE 1. Impaired glucose tolerance and impaired insulin secretion in PKC␦ ؊/؊ mice. a, changes in body weight of wild type , PKC␦ ϩ/Ϫ , and PKC␦ Ϫ/Ϫ mice between 8 and 20 weeks of age. b, changes in fasting glucose levels in mice between 8 and 20 weeks of age. c, blood glucose concentrations measured during intraperitoneal glucose tolerance test in each group at the age of 12 weeks. d, serum insulin levels measured during intraperitoneal glucose tolerance test in each group at the age of 12 weeks. e, changes in glucose tolerance in each group at the age of 8, 16, 20 weeks. f, results of intraperitoneal insulin tolerance test at the age of 12 weeks. Open circles, wild type mice (n ϭ 6); solid triangles, PKC␦ ϩ/Ϫ mice (n ϭ 6); solid circles, PKC␦ Ϫ/Ϫ mice (n ϭ 7). Data are the mean Ϯ S.E. §, p Ͻ 0.05 versus WT. scopic analysis demonstrated no apparent abnormalities of ␤ cell and endothelial cells in PKC␦ Ϫ/Ϫ islets (Fig. 2d) (28). Therefore, the reduced glucose-stimulated insulin secretion in PKC␦ Ϫ/Ϫ mice was not associated with any apparent reduction in insulin-containing granules or abnormal ␤ cell morphology.
Characterization of Impaired Insulin Secretion in Islets of PKC␦ Ϫ/Ϫ Mice-To elucidate whether the deletion of PKC␦ gene in islets resulted in compensatory increases in the expression of other PKC family members, we investigated mRNA and protein expression levels of conventional and novel PKC family members that are known to be expressed in ␤ cells (29). However, we could not find any significant increase in the expression of PKC␣, PKC␤2, and PKC⑀ investigated in the present study (Fig. 3, a and b). In further search for the mechanism of decreased glucose-stimulated insulin secretion in PKC␦ Ϫ/Ϫ mice, we investigated insulin secretion using several secretagogues in isolated PKC␦ Ϫ/Ϫ islets. Insulin secretion stimulated by high glucose was modestly but significantly lower in PKC␦ Ϫ/Ϫ islets than the wild type (Fig. 3c). We further investigated glucose-stimulated insulin secretion in PKC␦ Ϫ/Ϫ islets in a set of perfusion studies (Fig. 3d). The second phases of insulin secretion by stimulation with high glucose (16.7 mM) was significantly decreased in PKC␦ Ϫ/Ϫ islets (Fig. 3e). In addition, insulin secretion in response to depolarization induced by KCl was also reduced in PKC␦ Ϫ/Ϫ islets (Fig. 3c). Arginine, an insulinotrophic agent thought to depolarize ␤ cell membrane to activate voltage-dependent Ca 2ϩ channel by its positively charged form (30,31), also failed to stimulate insulin secretion in PKC␦ Ϫ/Ϫ islets to levels comparable with those in wild type islets. These results suggest impairment of glucose-stimulated insulin secretion in PKC␦ Ϫ/Ϫ islets after membrane depolarization. Insulin secretion was also significantly decreased in PKC␦ Ϫ/Ϫ islets in response to phorbol 12-myristate 13-acetate, an activator of PKC, and to phorbol 12-myristate 13-acetate and forskolin, activators of adenylate cyclase. Thus, the decrease in insulin secretion in PKC␦ Ϫ/Ϫ islets is observed even after application of various secretagogues.
Characterization of Insulin Granules in PKC␦ Ϫ/Ϫ Islets-Next, we determined changes in [Ca 2ϩ ] i in response to glucose stimulation in single ␤ cells by using fura-2, a Ca 2ϩ indicator. As shown in Fig. 4, the amplitude and time course of a glucoseinduced rise in Ca 2ϩ concentration was not different between PKC␦ Ϫ/Ϫ ␤ cells and wild type. These results suggest normal glucose metabolism and ATP production in PKC␦ Ϫ/Ϫ ␤ cells.
This finding together with the experiments of insulin secretion indicates that the decrease in insulin secretion in PKC␦ Ϫ/Ϫ islets might be simply due to a decrease in the releasable pool of insulin granules. Thus, to investigate the number of insulin granules that are located close to the plasma membrane, pancreatic ␤ cells isolated from PKC␦ Ϫ/Ϫ mice were immunostained with insulin and visualized by TIRF microscopy. TIRF imaging can identify single insulin granules within Ͻ100 nm from underneath the plasma membrane (32). Counting the number of insulin granules in the obtained image indicated that the density of insulin granules in PKC␦ null ␤ cells was not decreased but, rather, was increased (Fig. 5, a and b).
To confirm this finding, we counted the relative number of insulin granules according to their distance from the granule center to the plasma membrane in wild type and PKC␦ Ϫ/Ϫ islets after incubation with low and high glucose. As shown in Fig. 5c, the number of granules residing within 100 nm of the plasma membrane was much lower than those located at other distances from the membrane as described previously (33). The number of insulin granules in this fraction was significantly increased in PKC␦ Ϫ/Ϫ islets treated with low glucose. In addition, the number of insulin granules located 100 -200 nm from the membrane was increased in PKC␦ Ϫ/Ϫ islets treated with low glucose. On the other hand, the number of insulin granules in other fractions was not significantly different between wild type and PKC␦ Ϫ/Ϫ islets. Probably the increased number of insulin granules close to the plasma membrane reflects the increase in granular backup in trafficking. Next, we investigated  (Fig. 5d). However, all molecules investigated were not decreased but, rather, were increased, in agreement with the data of increased number of insulin granules that were closely located to plasma membrane in PKC␦ Ϫ/Ϫ ␤ cells. With regard to Munc18-1, syntaxin 1a, and Rab3a, we also investigated the protein expression level in PKC␦ Ϫ/Ϫ islets (Fig. 5e). We confirmed that the protein expression levels of these molecules were not decreased, in agreement with the data of quantitative RT-PCR method. These results suggest that the impaired insulin secretion in PKC␦ Ϫ/Ϫ islets is not simply due to a decrease in the insulin granules located close to plasma membrane or a decreased expression of major molecules involved in regulated exocytosis.
Reduced Phosphorylation of Munc18-1 in PKC␦ Ϫ/Ϫ Islets-PKC is a serine/threonine kinase; thus, a decreased phosphorylation state of target molecules might be involved in impaired insulin secretion in PKC␦ Ϫ/Ϫ islets. Among the molecules involved in exocytosis, VAMP2, synaptotagmin 3 and 7, Rab3a, SNAP25, and Munc18-1 (20,38,(42)(43)(44)(45) have been identified as substrates of PKCs. Previous data demonstrated that Ser residue at position 189 of SNAP-25, a known PKC phosphorylation site, is phosphorylated in islets in response to glucose. However, this phosphorylation does not seem to be related to glucose-sensitive insulin secretion (46). On the other hand, Munc18-1 was reported to bind to syntaxin-1A to inhibit the formation of SNARE complex (19). The PKC-catalyzed phosphorylation of Munc18-1 inhibits its interaction with syntaxin-1A (20); thus, the reduced phosphorylation state might be the underlying mechanism of decreased insulin secretion. However, the association of Munc18-1 and PKC in ␤ cells has not been studied.
Thus, we investigated first the interaction between PKC␦ and Munc18-1. Although there are no data on the exact PKC isoform(s) that can bind to Munc18-1, we demonstrated that PKC␦ directly bound with Munc18-1 by in vitro pulldown assay (Fig. 6a). In addition, co-immunoprecipitation experiments revealed that PKC␦ tightly bound to Munc18-1 (Fig. 6, b  and c) in HIT cells and mouse islets. This interaction was not altered by glucose stimulation and was not observed in PKC␦ Ϫ/Ϫ islets (Fig. 6c).
Next, we investigated the phosphorylation of Munc18-1 by glucose stimulation. In HIT cells, 30 min of stimulation with high glucose concentration increased the phosphorylation of Munc18-1. The increase in Munc18-1 phosphorylation level was suppressed by rottlerin (Fig. 7a). Furthermore, time-series studies showed that the increase in Munc18-1 phosphorylation FIGURE 3. Impaired glucose-stimulated insulin secretion in isolated PKC␦ ؊/؊ islets. a, abundance of mRNAs for the indicated proteins was determined by real-time RT-PCR analysis using total RNA extracted from wild type and PKC␦ Ϫ/Ϫ islets. The amounts of the mRNAs in PKC␦ Ϫ/Ϫ islets are expressed relative to those in wild type islets. Data are the mean Ϯ S.E. of five independent experiments for each group. b, abundance of the indicated proteins was determined by Western blotting analysis. The amounts in PKC␦ Ϫ/Ϫ islets are expressed relative to those in wild type islets. Data are the mean Ϯ S.E. of five independent experiments for each group. KO, knock out. c, insulin secretion induced by various insulinotropic agents in batch-incubated pancreatic islets of wild type (open bars) and PKC␦ Ϫ/Ϫ mice (solid bars) at the age of 12 weeks. Data were derived from at least six independent experiments for each group. Data are the means Ϯ S.E. §, p Ͻ 0.05 versus wild type. PMA, phorbol 12-myristate 13-acetate. IRI, immunoreactive insulin. d, insulin secretory responses to glucose in perifused pancreatic islets of wild type (open circles) and PKC␦ Ϫ/Ϫ (solid circles) islets. Data were obtained from each of five independent experiments for wild type and PKC␦ Ϫ/Ϫ islets. §, p Ͻ 0.05 versus wild type. e, the area under the curve (AUC) of the first phase (0 -2 min after stimulation with high glucose) and second phase of insulin secretion (2-20 min after stimulation by high glucose) for wild type (open bars) and PKC␦ Ϫ/Ϫ (solid bars) islets. Data are the mean Ϯ S.E. of five independent experiments for wild type and PKC␦ Ϫ/Ϫ islets. §, p Ͻ 0.05 versus wild type. occurred within 5 min after glucose stimulation (Fig. 7b). In wild type islets, glucose induced a 2.5-fold increase in Munc18-1 phosphorylation and a 1.8-fold increase in phosphorylation of SNAP25. On the other hand, the phosphorylation of both Munc18-1 and SNAP25 was markedly disturbed in PKC␦ Ϫ/Ϫ islets (Fig. 7c). These results suggest that the increased phosphorylation of Munc18-1 and SNAP25 by high glucose was at least in part regulated by PKC␦.

DISCUSSION
In most cells PKC has multiple regulatory functions in signal transduction, including growth, differentiation, apoptosis, exocytosis, and cell movement (47). Although at least nine PKC isoforms have been identified in pancreatic ␤ cells, much of the evidence implicating a particular PKC isoform is inconclusive because most of these are based on measuring secretory responses to PKC inhibitors of uncertain specificity (7,9,15,16,48). This is the first report to demonstrate the role of PKC␦ on insulin secretion using PKC␦ knock-out mice.
Among the family members of PKCs, ␤ cells contain PKC␣ and PKC␦ as the major isozymes. Similar to other protein kinases, PKCs are known to translocate to their effector sites when activated (49). Previous studies investigated the localization in cells and translocation of PKC␦ after glucose stimulation in ␤ cells (49). PKC␦ was stained as granulated foci with partial association with insulin staining; however, no glucose-dependent translocation was observed (49). Although the translocation assay is only indirect measure of activation, we also found that PKC␦ did not show significant translocation after exposure to high glucose concentration (data not shown). In addition, we also found that isolated PKC␦ Ϫ/Ϫ islets secreted lower amounts of insulin in response to KCl and arginine. Thus, the activation of PKC␦ by the metabolic change is not likely to be involved in the mechanism of glucose-stimulated insulin release. Instead, PKC␦ might be activated by the stimulation of exocytosis by various insulin secretagogues and forms an essential component of the regulated insulin exocytosis. A very recent study reported the unique structural feature of PKC␦ among the members of PKC regarding with the independence of its activity from activation loop phosphorylation (50). Such unique structure might be involved in the manner of the activation.
Recently, Carpenter et al. (7) reported that overexpression of kinase dead PKC␦ in rat islets did not affect glucose-stimulated insulin secretion assessed by batch incubation study. The conflict with our experiments might be due to several reasons. First, species differences might explain this finding. It is possible that PKC␦ could play a nonspecific role in insulin secretion in mice but not in rats. Second, it is also possible that deletion of PKC␦ could have deleterious effects on the proper formation of ␤ cells during the developmental stage. However, we found that rottlerin, a PKC␦ inhibitor, also decreased glucose-stimulated insulin secretion in wild type mice (data not shown). In addition, we did not find obvious morphological changes in PKC␦ knock-out islets using light and electron microscopy. Thus, it is unlikely that PKC␦ plays major roles in islet formation during the differentiation of ␤ cells. Third, the adenoviral infection used in the previous study might change the natural properties of glucose-stimulated insulin secretion through certain factors such as adenovirus-activated cytokines (51). Fourth, although we used a model of complete disruption of PKC␦ gene, infection with a dominant negative adenovirus approach often does not completely suppress the activity of PKC␦. Differences in suppression level might explain the discrepancy.
We observed that Ca ϩ2 influx after glucose stimulation was not altered in PKC␦ null ␤ cells. In addition, the number of insulin granules located close to plasma membrane in PKC␦ null ␤ cells was not decreased but, rather, increased compared with wild type ␤ cells. The increased docked insulin granules might represent the increase in granular backup in trafficking. These results suggest that the decreased insulin release in PKC␦ null islets occurs mainly after the process of docking of insulin granules to the plasma membrane. Thus, we investigated the phosphorylation state of molecules involved in exocytosis. The SNARE proteins play a major role in intracellular membrane fusion events. An additional family of proteins that acts as SNARE regulators, the Sec1/Munc18 proteins, is also an essential component of the fusion machinery (19). Previous data demonstrated that Munc-18-1 protein is present in HIT cells and functions as a negative regulator of the insulin secretory machinery via a mechanism that does not involve syntaxinassociated calcium channels (52). In adrenal chromaffin cells, Munc-18-1 is phosphorylated on Ser-313 by phorbol 12-myristate 13-acetate treatment. The phosphorylated Munc18-1 reduces its affinity for syntaxin (42). Thus, this event allows syntaxin to be released into an open conformation and associate into the SNARE complex. Although the involvement of Munc18-1 in insulin secretion has not been elucidated in islets, it is likely that a similar mechanism is also involved in the process of insulin secretion. Although HIT cells do not always represent the normal machinery of insulin secretion, a strong association between Munc18-1 and PKC␦ was observed in our study. This interaction was not affected by glucose concentration. On the other hand, high glucose stimulation increased the phosphorylation of Munc18-1 within 5 min, which was suppressed by rottlerin. Because the amount of phosphoproteins harvested from mouse islets is limited, we could not investigate the phosphorylation state of Munc18-1 in islets in detail. However, we found 2.5-fold increase in Munc18-1 induced by high glucose. This increase was markedly suppressed in PKC␦ null islets. Thus, it seems that PKC␦ null islets have impaired phosphorylation of Munc18-1. Considering the phenotype of PKC␦ Ϫ/Ϫ islets, our data suggest the importance of Munc18-1 as a substrate of PKC␦ in the fusion event of insulin-containing granules. However, there are no data on the importance of the phosphorylation status of Munc18-1 in insulin secretion in islets. Thus, further assessment will be needed to complete our understanding of this process.
Regulated exocytosis is a process of the fine-tuning of secretion in response to external stimuli. Other than molecules investigated in this study, recent data demonstrate the implication of several molecules such as Noc2 (53), collectrin (54), and granuphiline (55) in regulated exocytosis in ␤ cells. Whereas in this study we suggested the implication of Munc18-1 as a target molecule of PKC␦, it is also possible that PKC␦ modifies such molecules and, thus, is involved in insulin secretion.  (n ϭ 12 cells, each). §, p Ͻ 0.05 versus wild type. c, relative number of insulin granules categorized according to the distance from the granule center to the plasma membrane. Data are expressed as percent distribution below the plasma membrane (100% corresponds to total counts of insulin granules that resided within 500 nm of the plasma membrane) (n ϭ 9 cells each). §, p Ͻ 0.05 versus wild type. d, abundance of mRNAs for the indicated proteins was determined by real-time RT-PCR analysis using total RNA extracted from wild type and PKC␦ Ϫ/Ϫ islets. The amounts of the mRNAs in PKC␦ Ϫ/Ϫ islets are expressed relative to those in wild type islets. Data were obtained from five independent experiments for each group. Data are the mean Ϯ S.E. e, abundance of the indicated proteins was investigated by Western blotting analysis (n ϭ 3). In conclusion, our data clearly demonstrated that PKC␦ is an essential player in glucose-stimulated insulin secretion. Impaired glucose-stimulated insulin release is considered one of the possible genetic mechanisms of type 2 diabetes mellitus. Thus, it is important to identify each player involved in the mechanism of glucose-stimulated insulin release. Accumulation of such knowledge will finally contribute to isolation of candidate genes of type 2 diabetes and drug targets for its treatment.