Translocation inhibitors define specificity of protein kinase C isoenzymes in pancreatic beta-cells.

The protein kinase C (PKC) family consists of 11 isoenzymes. Following activation, each isoenzyme translocates and binds to a specific eceptor for ctivated inase (RACK) (Mochly-Rosen, D. (1995) Science 268, 247-251) that provides an anchoring site in close proximity to the isoenzyme's specific substrate. Pancreatic islet cells contain at least six PKC isoenzymes (Knutson, K. L., and Hoenig, M. (1994) Endocrinology 135, 881-886). Although PKC activation enhances insulin release, the specific function of each isoenzyme is unknown. Here we show that following stimulation with glucose, αPKC and εPKC translocate to the cell's periphery, while δPKC and ζPKC translocate to perinuclear sites. βC2-4, a peptide derived from the RACK1-binding site in the C2 domain of βPKC, inhibits translocation of αPKC and reduces insulin response to glucose. Likewise, εV1-2, an εPKC-derived peptide containing the site for its specific RACK, inhibits translocation of εPKC and reduces insulin response to glucose. Inhibition of islet-glucose metabolism with mannoheptulose blocks translocation of both αPKC and εPKC and diminishes insulin response to glucose while calcium-free buffer inhibits translocation of αPKC but not εPKC and lowers insulin response by 50%. These findings illustrate the unique ability of specific translocation inhibitors to elucidate the isoenzyme-specific functions of PKC in complex signal transduction pathways.

cytoskeletal elements (11,12), and the nucleus (13,14), as well as other subcellular compartments (15). Many cells are known to contain several isoenzymes (16,17), each localizing to a different cellular site upon stimulation (18). The multiplicity of isoforms of a single enzyme renders the analysis of enzymefunction relationship difficult. Recent work revealed that activated PKC isoenzymes bind anchoring proteins termed RACKs (1)(2)(3), believed to be positioned in close proximity to the isoenzyme's substrate. It was further shown that the functional specificity of the PKC isoenzyme is determined, in part, by the differential localization of the isoenzyme-specific RACKs (19). The RACK for ␤PKC, RACK1, has been cloned, and at least part of its binding site on ␤PKC has been mapped to a short sequence within the C2 domain (1). ␤C2-4, a nonopeptide derived from this region, inhibits phorbol ester-induced translocation of the C2-containing isoenzymes but not the translocation of C2-less isoenzymes such as ␦and ⑀PKC when tested in intact cells (1). A short peptide derived from the V1 region of ⑀PKC, ⑀V1-2, was similarly shown to inhibit the translocation of ⑀PKC, but not ␣-, ␤-, and ␦PKC (20). Furthermore, these isozyme-specific inhibitors blocked the specific function of individual isoenzymes; for example, ⑀V1-2, but not ␤C2-4, inhibited phorbol 12-myristate 13-acetate-induced regulation of the contraction rate in intact cardiomyocytes. Here we use these novel PKC isozyme-specific inhibitors to determine that PKC activation is part of the signals involved in the regulation of glucose-induced insulin secretion and to identify the specific isoenzymes that mediate this glucose effect.

MATERIALS AND METHODS
Islets obtained from 200-g male Sprague-Dawley rats were cultured for 3-5 days in glass chamber slides coated with extracellular matrix of bovine corneal endothelial origin (21). When more than 75% of the islet cells spread out to form a monolayer, the media were replaced with modified Krebs Ringer solution (KRB) (22) containing either 2.5 or 20 mM glucose. Following a brief wash, the islets were fixed in cold acetone, blocked with 1% normal goat serum for 1 h, and treated overnight with the specific anti-PKC isoenzyme antibody (Research and Diagnostic Antibodies, Berkeley, CA). Fluorescein isothiocyanate-linked goat-antirabbit IgG (Sigma Israel Chemicals, Rehovot, Israel) was applied for 2 h, and the slides were mounted in 90% glycerol, 10% phosphatebuffered saline, 0.1% sodium azide, 3% diazabicyclo[2.2.2]octane, pH 9.0, for microscopic imaging. Histochemical imaging was conducted on a PhioBos 1000 confocal microscope (Sarastro Inc., Ypsilanti, MI), equipped with Zeiss Universal Optics and argon laser illumination. The anti-PKC antisera each exhibited a single band of the appropriate size on Western blot (10 g of rat brain or rat islet homogenate/lane). No bands were observed when the antibodies were preincubated with excessive amounts of their corresponding antigen derived from the V5 region of the isoenzyme (PKC-epsilon-(728 -737) or PKC-␣-(Tyr 663 -(664 -672)), Research and Diagnostic Antibodies, Berkeley, CA). Preincubation of anti-␣PKC with excess PKC-␣-(Tyr 663 -(664 -672)) and anti ⑀PKC with excess PKC-epsilon-(728 -737) completely abolished the fluorescence images of the glucose-dependent isoenzyme translocation. The cells were identified as ␤-cells by counterstaining with tetramethylrhodamine B isothiocyanate-conjugated rabbit anti-guinea pig IgG and guinea pig anti-insulin serum (Dako Corp., Carpinteria, CA).
Freshly isolated islets were used for insulin release studies. Islets were preincubated for 1 h in KRB-BSA buffer containing 2.5 mM glucose, then transferred, 1-2 islets per tube, for an additional 60-min incubation in KRB-BSA buffer containing either 2.5 mM or 20 mM glucose as described previously (22). When calcium-free buffer was used, 5 mM EGTA was added to the calcium-free KRB-BSA buffer during the test period.
Peptides were synthesized at the Hebrew University School of Medicine Intradepartmental Equipment Services. Peptide introduction into * This work was supported in part by a grant from the JDF International (to R. N.) and from the Piccioto Foundation (to E. C.) and National Institutes of Health Grant HL43380 (to D. M.-R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Insulin release was measured by radioimmunoassay using specific guinea pig anti-rat insulin antiserum (Linco Research, St Charles, MO) and rat insulin standard (Novo Research Institute, Bagsvaerd, Denmark) (22). Data presented are mean net insulin values after subtraction of non-stimulated level (2.5 mM glucose). Statistical significance was determined by paired, non-parametric comparison to control or to control skinned islet, using the Wilcoxon test.

RESULTS AND DISCUSSION
The role of PKC as an amplifier of the glucose-generated signal to release insulin has been well established (5,(23)(24)(25). Six PKC isoenzymes, ␣, ␤, ␦, ⑀, , and , were found thus far in rat pancreatic islets of Langerhans (5,6,25). The specific function of the individual isoenzymes remains unknown. We used adult rat islet cultures (21) to assess the direct effect of glucose on PKC isoenzymes in intact islet cells. Site-specific translocation could be demonstrated for ␣-, ␦-, ⑀-, and PKC (Fig. 1). Confocal microscopy imaging revealed that ␣PKC and ⑀PKC redistributed following glucose stimulation to the cell's periphery, ␦PKC concentrated in an asymmetric structure in perinuclear region (possibly the Golgi apparatus), and PKC concentrated as a ring around the cell nucleus (Fig. 1).
Moreover, administration of ␤C2-4 into freshly isolated islets resulted in 35% reduction in the insulin response to glucose stimulation (Fig. 2e); introduction of the scrambled analog, ␤C2-4-s, had no inhibitory effect on insulin secretion. To rule out the possibility that the partial inhibition of the insulin response was the result of unequal penetration of the peptide, freshly isolated islets in suspension were skinned in the absence or presence of 10 M ␤C2-4 and subsequently incubated with 20 mM glucose. At the end of a 60-min stimulation, the islets were fixed in paraformaldehyde, dehydrated in ethanol, and further treated in propylene oxide Surre mixture. Following resin polymerization, 5-and 10-m slices were prepared and stained with anti-␣PKC antibodies. Confocal imaging revealed uniform inhibition of the glucose-induced isoenzyme translocation throughout the sections (not shown), indicating homogeneous penetration of the octapeptide throughout the islet. ␤C2-4 has been shown to be equally effective against all members of the classical PKC subfamily (19). While ␣PKC often has been reported to be the predominant PKC isoenzyme in islet ␤-cells, ␤PKC exhibits only scant expression in these cells (5); we, however, assume that all cPKC isoenzymes were equally inhibited by ␤C2-4. The fact that maximally effective concentrations of ␤C2-4 had only a partial inhibitory effect on glucose-induced insulin release may therefore indicate that in addition to the cPKC subfamily, other PKC isoenzymes or non-PKC stimulus amplifiers are involved in the glucose-generated stimulus-secretion coupling.
5 M was found to be a maximally effective concentration for both ␤C2-4 and ⑀PKC in 1-h incubations, never exceeding 43% inhibition of glucose-induced insulin release. The effect of the two peptides was additive at that concentration in fresh islets incubated for 60 min at 20 mM glucose (Fig. 4b). The fact that both peptides together inhibited only 67% of the glucose-mediated insulin release suggests that ␣PKC and ⑀PKC are each independently involved in one of several distal coupling systems (see below). Stimulation of islet monolayer in the absence of Ca 2ϩ (and in the presence of EGTA) abolished glucose-dependent translocation of ␣PKC but not of ⑀PKC (Fig. 4a). It also diminished the glucose-induced insulin response in fresh islets by 85% (Fig. 4b), providing further evidence that ␣PKC but not ⑀PKC is involved in the calcium-mediated regulatory signal to release insulin. However, addition of mannoheptulose, an inhibitor of glucose metabolism, completely abolished the glucose-induced translocation of both ␣and ⑀PKC isoenzymes (Fig. 4a) as well as the islet insulin response to the sugar (Fig.  4b), indicating that the glucose-stimulatory coupling signal for the activation of PKC is linked to glycolysis, the primary signal for ␤-cell insulin response (26).
The islet ␤-cell coupling mechanism regulating the insulin secretory response to glucose is a complex sequence of metabolic events (for review, see, for example, Ref. 26), leading to a unique multiphasic dynamic of hormonal secretion (for review, see Ref. 30). While the primary coupling signal originates from glucose metabolism, leading to calcium mobilization and activation of poorly defined calcium-dependent pathways, the same primary signal also activates numerous distal potentiating pathways, some calcium-dependent, others calcium-independent (27)(28)(29). Since insulin secretion is strongly modulated by  4. a, islet monolayer subjected to calcium-free (ϩ5 mM EGTA) medium exhibits no glucose-dependent translocation of ␣PKC, as compared with glucose-stimulated control culture. Glucose-dependent translocation of ⑀PKC is not affected by calcium omission. In fresh islet, omission of calcium diminishes the insulin response to 1-h stimulation with glucose by 85% (b). Mannoheptulose (30 mM), a glycolytic inhibitor, impedes glucose-stimulated translocation of both ␣PKC and ⑀PKC (a). Addition of 30 mM mannoheptulose fully inhibits glucose-stimulated insulin response in fresh islets (b). In fresh islets, the inhibitory effect of ␤C2-4 and ⑀V1-2 (5 M each) on glucose-mediated insulin response is additive (b).* denotes statistical significance of p Ͻ 0.02 compared with glucose-stimulated control; ** denotes statistical significance p Ͻ 0.05 compared with either ␤C2-4 or ⑀V1-2 alone; *** denotes statistical significance of p Ͻ 0.005 compared with glucose-stimulated control. the duration of glucose stimulation (30), it is possible that different PKC isoenzymes have distinct roles in the different phases of release. This subject is presently under investigation. Islets coupling signals originating from activation of adenylate cyclase, phospholipase C, and phospholipase A 2 are among the more thoroughly investigated potentiating signals involved in the modulation of the insulin response to glucose (31)(32)(33). Messengers generated from these key metabolic pathways are known to activate multiple PKA, PKC, and CaM kinases, each controlling a specific amplifying branch of the insulin stimulussecretion coupling pathway, resulting in the multiphasic dynamics of hormonal secretion in response to glucose stimulus (30,31). The identification of PKC isoenzyme-specific binding proteins offers novel tools to resolve the specific contribution of each isoform to this complex of interrelated signals. Furthermore, RACK-binding translocation inhibitors should prove to be valuable tools in resolving the specific function of the individual PKC isoenzyme in cells expressing multiple forms of the enzyme as well as in identifying their specific substrates.