Dynamics of Glucose-induced Membrane Recruitment of Protein Kinase C βII in Living Pancreatic Islet β-Cells* 210

The mechanisms by which glucose may affect protein kinase C (PKC) activity in the pancreatic islet β-cell are presently unclear. By developing adenovirally expressed chimeras encoding fusion proteins between green fluorescent protein and conventional (βII), novel (δ), or atypical (ζ) PKCs, we show that glucose selectively alters the subcellular localization of these enzymes dynamically in primary islet and MIN6 β-cells. Examined by laser scanning confocal or total internal reflection fluorescence microscopy, elevated glucose concentrations induced oscillatory translocations of PKCβII to spatially confined regions of the plasma membrane. Suggesting that increases in free cytosolic Ca2+ concentrations ([Ca2+]c) were primarily responsible, prevention of [Ca2+]c increases with EGTA or diazoxide completely eliminated membrane recruitment, whereas elevation of cytosolic [Ca2+]c with KCl or tolbutamide was highly effective in redistributing PKCβII both to the plasma membrane and to the surface of dense core secretory vesicles. By contrast, the distribution of PKCδ·EGFP, which binds diacylglycerol but not Ca2+, was unaffected by glucose. Measurement of [Ca2+]c immediately beneath the plasma membrane with a ratiometric “pericam,” fused to synaptic vesicle-associated protein-25, revealed that depolarization induced significantly larger increases in [Ca2+]c in this domain. These data demonstrate that nutrient stimulation of β-cells causes spatially and temporally complex changes in the subcellular localization of PKCβII, possibly resulting from the generation of Ca2+ microdomains. Localized changes in PKCβII activity may thus have a role in the spatial control of insulin exocytosis.

forms can be divided into three subfamilies. Conventional PKCs are activated via recruitment to membranes, mediated by the Ca 2ϩ -dependent binding of a C2 domain to phospholipids, and this effect is further potentiated by the binding of diacylglycerol (DAG) to C1 domains (1). By contrast, novel PKCs bind DAG, but not Ca 2ϩ and phospholipids, while atypical PKCs are not affected by any of the above activators (1).
Biochemical studies of the activation of PKC are complicated by the need for cell disruption and isolation of membrane and cytosol fractions (2) or for cell fixation and immunocytochemistry (2)(3)(4). Each of these approaches is limited by the difficulty of detecting any changes in subcellular localization, which are spatially or temporally complex. To overcome this limitation, fusion constructs between enhanced green fluorescent protein (EGFP) (5) and PKC␥ (6), PKC␣ (7), PKC␦ (8), and PKC␤II (9) have recently been used to monitor the dynamics of membrane translocation of PKCs in a number of non-excitable cell types and appear faithfully to reflect the behavior of the endogenous PKC isoforms. However, while PKC may play an important role in agonist stimulation of exocytosis from neurosecretory cells (10), no data are presently available on the dynamics of conventional PKCs in any excitable cell type.
Elevated glucose concentrations stimulate insulin secretion from ␤-cells via metabolism of the sugar (11,12) and increases in cytosolic free ATP concentration (13). Closure of ATP-sensitive K ϩ channels (14) then leads to depolarization of the plasma membrane, influx of Ca 2ϩ through voltage-gated Ca 2ϩ channels (15), and secretory vesicle fusion (16). PKC activity is present in both primary pancreatic islets (17) and derived ␤-cell lines (18,19). Furthermore, conventional (␣, ␤I, ␤II; sensitive to Ca 2ϩ and DAG), novel (␦; sensitive to DAG but not Ca 2ϩ ), and atypical (, ; insensitive to Ca 2ϩ and DAG) PKC isoforms (20 -23) have all been reported in islet cells. However, the role of PKC in the stimulation of insulin secretion is controversial. Acute activation of conventional and novel PKCs with the phorbol ester 12-O-tetradecanoyl-phorbol-13-acetate strongly stimulates insulin secretion (19,24) without affecting ␤-cell electrical activity or cytosolic free Ca 2ϩ ([Ca 2ϩ ] c ) (25,26). On the other hand, inhibition of PKC activity with the broad specificity inhibitor staurosporine (27), or an inhibitor specific for classical PKC isoforms (Go6976), slightly enhances the first phase of glucose-stimulated insulin release from rat islets (28) while diminishing the sustained phase. Down-regulation of conventional PKC isoforms with phorbol esters has little effect on glucose-stimulated insulin release (29).
To determine whether active PKCs may play a role in the spatial coordination of exocytosis in individual ␤-cells without necessarily affecting total insulin release, we have therefore generated fusion constructs between EGFP and PKC␤II, PKC␦, and PKC. PKC␤II and PKC␣ represent the major conventional PKC isoforms in ␤-cells (20), and PKC␤II activity has recently been shown to be important for the regulation of the preproinsulin gene (23). Expression of these constructs has allowed the dynamics of each isoform to be studied in real time in both primary islet and clonal ␤-cells. Using confocal and total internal reflection fluorescence (TIRF)/evanescent wave (30 -34) imaging, we show that elevated glucose concentrations cause complex, oscillatory translocations to the plasma and other membranes of PKC␤II in primary ␤-cells and clonal MIN6 cells. These changes appear to be produced largely by transient depolarizations of the plasma membrane and stimulated Ca 2ϩ influx. The formation of microdomains of [Ca 2ϩ ] c immediately beneath the plasma membrane, demonstrated directly by targeting a green fluorescent protein-based Ca 2ϩ probe ("pericam") (35) exclusively to this domain, may be critical for the generation of complex movements of PKC.

Materials and Methods
Cell culture reagents were obtained from Invitrogen or Sigma, and molecular biologicals from Roche Molecular Biochemicals.
Adenoviral Generation-Adenoviruses were constructed and amplified using the pAdEasy system (36) as previously described (37). The PKC␤II⅐EGFP, PKC␦⅐EGFP, and PKC⅐EGF (8) cDNAs were transferred into plasmid pShuttleCMV as KpnI/XhoI fragments. Adenoviral generation from the recombinant shuttle vectors was performed, and infection of cells and islets was performed as previously described (37).
Recombination with pAdEasy-1, transfection into HEK 293 cells, and viral amplification of the pShuttle-CMV based plasmids encoding each recombinant PKC isoform-GFP fusion proteins was performed essentially as previously described (37). Determination of viral concentration was by comparison of the absorbance at 260 nm with a viral stock of known titer (37). MIN6 cells were infected with a multiplicity of infection of 30, ϳ16 h prior to imaging.
Cell Culture and Adenoviral Infection-Primary isolated islet ␤-cells and MIN6 cells (passages nos. 20 to 30) were cultured and infected with adenoviruses as previously described (38). In each case, the concentration of glucose was lowered to 3 mM for 16 h before experiments.
Confocal Imaging Analysis-Coverslips (24 mm in diameter) were placed in a thermostatted Leyden chamber, (model TC-202A, Medical Systems Corp.) on the stage of an inverted Leica SP2 confocal imaging system using a 63X (numerical aperture ϭ 1.45) oil immersion objective. All experiments were carried out in Krebs-Ringer bicarbonate buffer (KRB): 125 mM NaCl, 3.5 mM KCl, 1.5 mM CaCl 2, 0.5 mM NaH 2 PO 4 , 0.5 mM MgSO 4 , 3 mM glucose, 10 mM Hepes, 2 mM NaHCO 3 , pH 7.4, containing, initially, 3 mM glucose and equilibrated with O 2 /CO 2 (19:1). Images were acquired at a rate of 0.5 s Ϫ1 and processed off line. Green and cyan fluorescent protein (ECFP) fluorescence were imaged simultaneously through alternate excitation (0.2 s Ϫ1 ) at 430 and 488 nm with emitted fluorescence filtered between 450 and 490 nm and between 520 and 560 nm, respectively. Under these conditions, crosscontamination of the two signals was negligible.
Calcium Crimson Imaging-MIN6 cells were infected with adenoviral PKC␤II⅐EGFP as described above and cultured overnight in medium containing 3 mM glucose. 1 h prior to imaging, cells were micro-injected using an Eppendorf 5171/5242 micromanipulator/pressure microinjector with a solution of 0.5 M Calcium Crimson conjugated to 10,000 Da dextran (Molecular Probes, Eugene, OR), to give an approximate final concentration of 20 nM in the cytosol. The cells were washed once and incubated at 37°C until use.
During imaging, cells were incubated in KRB medium and maintained at 37°C on a heated stage. Cells that had been successfully injected with Calcium Crimson dye and were expressing the PKC␤II⅐EGFP were identified by epifluorescence and imaged on a Leica confocal imaging spectrophotometer system (TCS-SP) running on a DM/IRBE inverted microscope (ϫ40 objective). Fluorescence of Calcium Crimson (568-nm excitation, Kr laser; 580 -640-nm emission) and EGFP (488-nm excitation, Ar laser; 520 -560-nm emission) were monitored simultaneously and analyzed using Leica TCS software. Additions were made via a small volume of a stock solution (3.5 M KCl) followed by rapid mixing with a pipette.
Changes in Calcium Crimson fluorescence were determined throughout the whole cell and presented as an increase relative to basal fluorescence. EGFP fluorescence was determined in the vicinity (ϳ1 m) of the plasma membrane and in the bulk cytosol. The ratio of the average fluorescence of these regions was used as a measure of PKC␤II⅐EGFP translocation. Relative changes in this ratio, normalized to basal conditions, are given.
TIRF Microscopy-To assess translocation of PKC␤II⅐EGFP, we employed a TIRF (also known as evanescent wave microscopy) microscope similar to that described previously by Tsuboi et al. (32)(33)(34). The incident light for total internal reflection illumination was introduced from the objective lens (Olympus, numerical aperture ϭ 1.65, 100X magnification) through a single mode optical fiber and two illumination lenses. To observe the EGFP fluorescence image, we used a 488-nm laser (argon ion laser, 30 mW, Spectra-Physics) for total internal fluorescence illumination and a long pass filter (515 nm) for barrier. The laser beam was passed through an electromagnetically driven shutter (Till Photonics). The shutter was opened synchronously with camera exposure under control by TillvisilON software (Till Photonics). Images were acquired every 2 s. To analyze the data, translocation events were manually selected and the average fluorescence intensity of individual plasma membrane regions was calculated.

Statistical Analysis
Data are given as means Ϯ S.E. of at least three individual experiments. Comparisons between means were performed using one-tailed Student's t test for paired data with Microsoft Excel TM or Origin 7 TM (OriginLab, Northampton, MA) software. Fig. 1A shows the responses to a stepped increase in glucose concentration from 3 to 25 mM of adenovirally expressed PKC␤II⅐EGFP, imaged by laser-scanning confocal microscopy in primary ␤-cells. An increase in fluorescence ratio (plasma membrane: cytosol) was observed in 7 of 16 cells examined (from two separate preparations; mean increase 13.5 Ϯ 4.2%), with partial oscillations (i.e. retranslocation to the cytosol) observed in three of seven cells. In some cases, recruitment was "patchy" with evidence of localization on membrane-associated organelles (Fig. 1A, arrow). Implicating [Ca 2ϩ ] c increases in these effects of glucose, cell depolarization with 35 mM KCl ( Fig.  1B) or stimulation of muscarinic receptors with carbachol, 100 M (Fig. 1C), also caused a clear increase in the proportion of plasma membrane-bound PKC␤II and in each case the appearance of focal points of high fluorescence (arrows).

Responses of PKC␤II to Glucose and Other Agonists in Primary ␤-Cells-
In contrast to PKC␤II⅐EGFP, neither PKC␦⅐EGFP nor PKC⅐EGFP displayed any detectable change in subcellular distribution in primary ␤-cells in response to the above stimuli, while phorbol 12-myristate 13-acetate (PMA) caused translocation of PKC␦ from the cytosol to the nuclear periphery (not shown; see also Fig. 7 for response in MIN6 cells).

Responses of PKC␤II⅐EGFP Distribution to Elevated [Glucose] and Other Stimuli in MIN6
␤-Cells-To explore the mechanisms involved in the glucose-stimulated translocation of PKC␤II⅐EGFP in more detail we next used clonal MIN6 ␤-cells.
In contrast to primary ␤-cells, these well differentiated and glucose-responsive cells (43) can be easily microinjected with both plasmid cDNAs and with Ca 2ϩ indicator dyes (37) without marked deterioration of cell function.
Examined first by laser-scanning confocal microscopy, PKC␤II⅐EGFP translocated to the plasma membrane in response to 25 mM glucose in 7 of 22 MIN6 cells examined ( Fig.  2A). Retranslocation to the cytosol was clearly evident in almost half (three of seven) of the cells examined. To provide greater temporal and spatial resolution we next employed TIRF microscopy (31)(32)(33)(34). This technique involves the generation of a thin (Ͻ100 nm) field of fluorescence at the surface of the coverslip and thus at the surface of an attached cell. Hence a fluorophore such as PKC␤II⅐EGFP will only fluoresce as it approaches very close to (within ϳ50 nm) the plasma membrane while molecules in the cytosol remain in darkness. Since MIN6 cells display a flattened morphology, this technique was anticipated to permit a more precise quantification of plasma membrane-associated PKC␤II.
PKC␤II EGFP was translocated to the plasma membrane with a half-time of ϳ60 s and a peak increase in membrane: cytosolic PKC␤II⅐EGFP of ϳ1.5 (Fig. 2B). Translocation was not provoked by a non-metabolizable sugar (galactose, not shown) and was completely suppressed by chelation of extracellular Ca 2ϩ with EGTA or by cell hyperpolarization with the ATP-sensitive K ϩ channel opener, diazoxide (Fig. 2B, trace 4). In some cells, "hot spots" and waves of PKC were clearly detectable (see movie " Fig. 2B" at http://www.jbc.org). These effects were not observed in cells expressing a membrane-targeted GFP chimera (not shown) and are thus unlikely to result simply from changes in the shape of the cell. Moreover, the effects of glucose upon translocation were only marginally reduced by inhibition of phospholipase C activity with U73122 (  Fig. 2C), with a half-time similar to that for the increases in PKC␤II associated with the plasma membrane (Fig. 2, A and B). In some cells (4 of 10 examined) the glucoseinduced increases were more oscillatory, consisting of spikes on a steadily increasing baseline (Fig. 2D).
We next tested the possibility that highly localized changes in [Ca 2ϩ ] c immediately beneath the plasma membrane may be involved in PKC␤II membrane recruitment. The formation of such a Ca 2ϩ microdomain would be expected to permit phospholipid-dependent interaction of the PKC␤II C2 domain with the membrane inner leaflet (45), independently of the DAG binding domain (C1) (46).
To achieve measurements of [Ca 2ϩ ] c close to the inner surface of plasma membrane (Ͻ10 nm), we targeted the Ca 2ϩ sensor, ratiometric pericam (35) to this region of the cell. cDNA encoding the pericam was fused in frame with that encoding the soluble N-ethyl maleimide-sensitive factor receptor (t-SNARE), synaptosome-associated protein of 25 kDa (SNAP25), which binds to membranes after palmitoylation (47). The SNAP25⅐pericam chimera displayed a largely plasma membrane localization with some fluorescence on intracellular structures, possibly corresponding to the Golgi apparatus or mature secretory vesicles (Fig. 2E, monochrome panel) (13). Importantly, the molecular targeting of this construct eliminated the need for spatially selective excitation (i.e. by confocal or TIRF microscopy), permitting ratiometric measurement of  D) or plasma membrane-targeted (E) pericams prior to imaging. Cells were maintained initially in KRB containing 3 mM glucose and imaged (A) on the confocal microscope or (B) by total internal reflection fluorescence microscopy during the increases in glucose concentration indicated. In A, the traces show the increases in total plasma membrane fluorescence in the single cell shown relative to cytosolic fluorescence (see Fig. 1). The increase in plasma membrane-associated fluorescence (calculated as a ratio of average intracellular fluorescence by quantification of regions of interest with 1 m, Ͼ2 m, respectively, from the cell surface) or the increase in fluorescence normalized to the prestimulatory level in B. In each case, traces represent the mean of more than four cells or are from a single typical cell. Cells in C, D, and E were transfected with constructs encoding untargeted or plasma-membrane-targeted pericams, respectively, before ratio metric imaging (pseudocolor) of [Ca 2ϩ ] changes by epifluorescence microscopy as described under "Experimental Procedures." Note the greater heterogeneity in [Ca 2ϩ ] (trace 1 versus 2) and appearance of small transients at the plasma membrane (E). Monochrome images show fluorescence excited at 410 nm; an essentially identical distribution of fluorescence was apparent under excitation at 480 nm and reflects the intracellular distribution of the probe. Scale bars, 5 m.
cation was due to a short-lived increase in DAG generation by phospholipid hydrolysis, glucose-stimulated translocation of PKC␤II⅐EGFP was entirely unaffected by the pharmacological PLC inhibitor, U73122 (Fig. 3B, trace 3).
Imaged by confocal microscopy (Fig. 3C), PKC␤II was found also to translocate to intracellular structures in response to KCl. The identity of the majority of these structures was re-vealed as mature insulin secretory vesicles by simultaneous imaging of a co-expressed dense core vesicle membrane protein, phogrin (39), conjugated to cyan fluorescent protein (47).
Changes in [Ca 2ϩ ] c in the bulk cytosol and beneath the membrane in response to cell depolarization induced by KCl or tolbutamide were explored with targeted pericams. In contrast to untargeted pericam, which reported an increase in intracel-  Fig. 3E).

Simultaneous Imaging of PKC␤II Translocation and Depolorization-induced [Ca 2ϩ ] c Increases in Single
Cells-We next sought evidence that the larger increase in Ca 2ϩ beneath the plasma membrane may be important for the recruitment of PKC␤II⅐EGFP. If the glucose-induced translocation of PKC␤II were due solely to a global increase in intracellular [Ca 2ϩ ] c , it would be predicted that the kinetics of the increases in [Ca 2ϩ ] c and the membrane content of PKC␤II would be very similar. Indeed, glucose induced changes in PKC␤II⅐EGFP distribution, and cytosolic Ca 2ϩ displayed grossly similar kinetics (Fig. 2, A  and B versus C and D). However, when cells were stimulated with KCl, this prediction only held true during the initial recruitment of the chimera (see Fig. 6). At later time points (Ͼ30 s) PKC␤II⅐EGFP dissociated from the membrane while [Ca 2ϩ ] c remained close to maximal. These data suggest that PKC␤II association with the plasma membrane may be controlled by locally high Ca 2ϩ concentrations.
Impact of Intracellular Ca 2ϩ Mobilization on PKC␤II Localization-Activation of muscarinic receptors with carbachol and mobilization of intracellular Ca 2ϩ caused a rapid, transient translocation to the plasma membrane (in 25 of 29 cells examined, Fig. 5, A and B). This effect was entirely blocked by the presence of the phospholipase C inhibitor U73122 (Fig. 5B,  trace 3). In contrast to depolarizing stimuli (Figs. 3 and 4) carbachol caused an essentially identical increase in [Ca 2ϩ ] PM (to 2.62 Ϯ 0.42 M, n ϭ 10 cells; Fig. 5D) as [Ca 2ϩ ] c (to 2.48 Ϯ 0.36 M, n ϭ 10 cells; Fig. 5C). Interestingly, the response to carbachol was significantly accelerated at high glucose concentrations (Fig. 5A, images 3 and 4 and lower graph; solid versus dashed trace), presumably reflecting glucose-induced Ca 2ϩ influx and/or DAG production (see "Discussion").
Effects of Glucose on the Subcellular Distribution of Novel and Atypical PKC Isoforms-The ineffectiveness of the phospholipase C inhibitor to prevent glucose or KCl-induced recruitment of PKC␤II to the plasma membrane (Fig. 3B) suggested that DAG production and binding to C1 domains may have played a relatively small part in translocation. In line with this view, the distribution of neither the novel isoform PKC␦ (no C2 domain) (48) nor PKC (lacking both C1 and C2 domains) were affected by glucose (Fig. 7) or depolarizing stimuli (not shown). By contrast, PKC␦ was rapidly translocated (half-time ϳ20 s in each case) to both the nuclear membrane and cell surface in response to addition of the phorbol ester, PMA (Fig. 7, A and B).

DISCUSSION
Dynamics of PKC␤II⅐EGFP Translocation-We show here, for the first time in single living ␤-cells, that elevated glucose concentrations cause complex and dynamic changes in the localization of a conventional PKC isoform PKC␤II. This behavior was observed in both primary islet ␤-cells (Fig. 1A) and, more dramatically, in clonal MIN6 ␤-cells (Fig. 2, A and B). In the latter case, TIRF microscopy revealed the creation by elevated glucose concentrations of hot spots and waves of PKC␤II at the plasma membrane (see also movie " Fig. 2B" at http:// www.jbc.org). In this respect, the behavior of PKC␤II (9) as well as the conventional PKC isoforms PKC␥ (6) and PKC␣ (7) is reminiscent of that previously described in non-excitable cells using GFP chimeras and confocal microscopy. However, by the use of TIRF microscopy, we also reveal the creation by elevated glucose concentrations of hot spots and waves of PKC␤II at the plasma membrane, phenomena recently described for PKC in astrocytes (49). Arguing against the possibility that this behavior reflects a nonspecific coagulation of GFP molecules on the membrane, such hot spots are rarely observed using phospholipid-dependent membrane-targeted EGFP chimeras that incorporate pleckstrin homology domains using either confocal (50) or TIRF microscopy. 2 The present data are also consistent with the findings of Yedovitzky et al. Interestingly, we failed to find any change in the localization of either PKC␦ or PKC in response to glucose or other secretagogue stimuli (Fig. 7). These results contrast with reports of an important role of PKC in the regulation of the preproinsulin gene by glucose (51), although it should be emphasized that we did not explore the localization of this isoform beyond relatively short (ϳ30 min) time points after glucose stimulation.
Mechanisms Involved in PKC␤II Translocation, Role of Ca 2ϩ Microdomains-We provide evidence that the changes in PKC␤II distribution are likely to result from localized changes in cytosolic Ca 2ϩ concentration generated beneath the plasma membrane during the depolarization-induced opening of L-type Ca 2ϩ changes (15). Thus, depolarizing concentrations of KCl (Fig. 3, D and E) or tolbutamide (Fig. 4, C and D) increased [Ca 2ϩ ] c in this domain ([Ca 2ϩ ] PM ) to concentrations 1.3-1.5fold higher than those in the bulk cytosol and caused robust translocation of PKC␤II⅐EGFP to the membrane. However, the partial inhibition of glucose-induced PKC␤II translocation by blockade of phospholipase C activity (Fig. 2B, trace 5) suggests that the local generation of DAG, caused by phospholipid hydrolysis, may contribute to the recruitment of conventional PKCs to the membrane in response to glucose. In this regard it should be mentioned that total islet DAG content is reported to increase only slightly (52) if at all (53) at elevated glucose concentrations, largely through de novo synthesis of DAG from glucose-derived palmitate (52). Importantly, such changes are not expected to be blocked by inhibitors of phospholipase C (Fig. 2B). However, arguing that glucose-induced increases in DAG content are small in the MIN6 cell system studied here, we failed to observed any translocation of PKC␦ to the cell surface in response to elevated glucose concentrations (Fig. 7, A  and B). On the other hand, because PKC␣ activity is regulated by several long chain acyl-CoA esters (54), a possible role for glucose-induced increase in the concentrations of these latter species (55) in the observed recruitment of PKC␤II to the plasma membrane cannot be ruled out.
Our observations (Fig. 6) that cytosolic [Ca 2ϩ ] and PKC membrane localization could be dissociated in the same single cell are perhaps most simply explained by the fact that in the absence of generation of DAG a "threshold" concentration of Ca 2ϩ , probably Ն 1 M, is required to ensure the binding of the C2 domain of PKC␤II to membrane phospholipids (44) as previously proposed for PKC␣ (56). Interestingly, the concentra-tions of Ca 2ϩ measured here immediately beneath the membrane of stimulated MIN6 ␤-cells (2-3 M) are similar to, if somewhat lower than, those previously reported at greater distances from the plasma membrane (0.5-1.0 m) of ␤-cells using diffusible dyes (6 -10 M) (57). Thus, the present data, which were obtained using a molecularly targeted probe, would seem to rule out the notion of a generalized large gradient of Ca 2ϩ concentration stretching across the whole interior surface of the cell membrane. However, more localized [Ca 2ϩ ] c domains (for example at the mouth of individual Ca 2ϩ channels) (58,59) cannot be excluded. In contrast to the impact of stimulated Ca 2ϩ influx, the stimulation of intracellular Ca 2ϩ release and DAG production with a muscarinic agonist elicited efficient membrane localization of PKC␤II (Fig. 5), presumably reflecting a slightly larger increase in plasma membrane [Ca 2ϩ ] c as well as the cooperation of C1 and C2 domains in membrane association (44). Interestingly, this effect of carbachol was significantly accelerated by elevated glucose concentrations (see legend to Fig. 5), possibly reflecting the de novo synthesis of DAG from glucose (52).
Potential Roles of PKC␤II Translocation in Regulated Insulin Secretion and Gene Expression-What may be the consequences of the translocation of PKC␤II (and other conventional PKCs) to the plasma membrane? Arguing that the enzyme is at least partly activated upon membrane translocation in ␤-cells, only kinase-active PKC␤II, but not an active site (K371R) mu- tant, was found to retranslocate into the cytosol after antigen stimulation of HEK 293 cells (9). Although targets of PKC are not well characterized in the ␤-cell, possibilities include both the pore-forming subunit of K ATP channels (60) and proteins of the secretory machinery (e.g. SNAP25) (61,62).
By demonstrating that activated PKC␤II can migrate to the surface of secretory vesicles (Fig. 3C) the current studies provide evidence for a new mechanism whereby vesicle fusion may be controlled locally. Thus, efflux of stored Ca 2ϩ from vesicles (40,63), possibly by gating of vesicle-associated receptors for ryanodine (63) or nicotinic acid adenine dinucleotide phosphate (64), 3 may lead to the recruitment of the kinase to a highly localized domain of [Ca 2ϩ ] at the vesicle surface. A similar mechanism has recently been proposed for the translocation of PKC␣ to internal ryanodine receptor-gated Ca 2ϩ release sites in vascular smooth muscle cells (65).