C2 region-derived peptides inhibit translocation and function of beta protein kinase C in vivo.

RACK1 is a protein kinase C (PKC)-binding protein that fulfills the criteria previously established for a receptor for activated C-kinase (RACK). If binding of PKC to RACK anchors the activated enzyme near its protein substrates, then inhibition of this binding should inhibit translocation and function of the enzyme in vivo. Here, we have identified such inhibitors that mimic the RACK1-binding site on βPKC. We first found that a C2-containing fragment, but not a C1-containing fragment of βPKC, bound to RACK1 and inhibited subsequent βPKC binding. The RACK1-binding site was further mapped; peptides βC2-1 (βPKC(209-216)), βC2-2 (βPKC(186-198)), and βC2-4 (βPKC(218-226), but not a number of control peptides, bound to RACK1 and inhibited the C2 fragment binding to RACK1. Peptides βC2-1, βC2-2, and βC2-4 specifically inhibited phorbol ester-induced translocation of the C2-containing isozymes in cardiac myocytes and insulin-induced βPKC translocation and function in Xenopus oocytes. Therefore, peptides corresponding to amino acids 186-198, and 209-226 within the C2 region of the βPKC are specific inhibitors for functions mediated by βPKC.

RACK1 is a protein kinase C (PKC)-binding protein that fulfills the criteria previously established for a receptor for activated C-kinase (RACK). If binding of PKC to RACK anchors the activated enzyme near its protein substrates, then inhibition of this binding should inhibit translocation and function of the enzyme in vivo. Here, we have identified such inhibitors that mimic the RACK1-binding site on ␤PKC. We first found that a C2containing fragment, but not a C1-containing fragment of ␤PKC, bound to RACK1 and inhibited subsequent ␤PKC binding. The RACK1-binding site was further mapped; peptides ␤C2-1 (␤PKC(209 -216)), ␤C2-2 (␤PKC(186 -198)), and ␤C2-4 (␤PKC(218 -226), but not a number of control peptides, bound to RACK1 and inhibited the C2 fragment binding to RACK1. Peptides ␤C2-1, ␤C2-2, and ␤C2-4 specifically inhibited phorbol esterinduced translocation of the C2-containing isozymes in cardiac myocytes and insulin-induced ␤PKC translocation and function in Xenopus oocytes. Therefore, peptides corresponding to amino acids 186 -198, and 209 -226 within the C2 region of the ␤PKC are specific inhibitors for functions mediated by ␤PKC.
Protein kinase C (PKC) 1 isozymes are phosphatidylserine (PS)-and diacylglycerol (DG)-dependent kinases (1) that translocate from the soluble fraction to the cell particulate fraction following activation (2,3). Several PKC isozymes are present in each cell type. The isozymes are localized to different subcellular compartments (4 -6), and following stimulation, each translocates to distinct intracellular structures (7). The translocation and binding of PKC isozymes to different intracellular structures suggests distinct physiological roles for individual isozymes. We previously demonstrated that inhibition of PKC translocation inhibits its function (8,9). Therefore, inhibitors that prevent the translocation of specific isozymes, can provide pharmacological tools to determine the function of each isozyme.
The PKC isozymes can be divided into three subfamilies: conventional cPKC, novel nPKC, and atypical aPKC (10). Each PKC isozyme contains unique (V) regions. In addition, there are regions common to all the isozymes. The subfamilies differ from each other in the common (C) regions within their regu-latory domains. The regulatory domain of the cPKC isozymes contains two common regions, C1 and C2. The C1 region consists of two cysteine-rich loops that mediate DG and phorbol ester binding (11)(12)(13). C1 is also found in the nPKC subfamily and one of the cysteine loops is found in the aPKC subfamily (14 -16). The C2 region is present only in the cPKC subfamily and mediates calcium binding; all the C2-containing isozymes require calcium for their activity (10). This region may also serve as a low affinity calcium sensor (17). In addition, the C2 region mediates PS binding (18). However, recent studies indicate that the V1 and C1 regions also mediate calcium and PS binding (19,20). Finally, the C2 regions of other C2-containing proteins were proposed to mediate direct binding of these translocating proteins to lipids at the plasma membrane (21).
Translocation of PKC to the cell particulate fraction was thought to reflect direct binding of the enzyme to lipids at the plasma membrane. However, data from several laboratories including our own indicate that translocated PKC interacts with proteins at the site of translocation (2,(22)(23)(24)(25)(26)(27)(28). We have identified several proteins from the cell particulate fraction that bind PKC only in the presence of its activators (24). Binding of PKC to these proteins was concentration dependent, saturable, and specific, suggesting that these binding proteins are receptors for activated C-kinase, or RACKs (24). Recently, we cloned RACK1, a gene encoding for a 36-kDa homolog of the ␤ subunit of G proteins (accession number U03390) that fulfills the criteria for RACKs (25). RACK1 is neither a PKC substrate nor an inhibitor (25). Rather, it increases PKC phosphorylation of substrates presumably by stabilizing the active form of PKC (25).
The RACK1-binding site on PKC is unknown. If this site is identified, peptides that mimic the binding site could serve as specific inhibitors of PKC translocation and function. Our previous studies suggested that the C2 region of cPKC contains at least part of the RACK-binding site on the enzyme; other C2containing proteins such as synaptotagmin (29) and phospholipase C␥ (30) also bind to a mixture of RACKs prepared from the cell particulate fraction. In addition, recombinant fragments of synaptotagmin containing the C2 homologous region bind to RACKs and inhibit PKC binding to RACKs (29). However, these studies were carried out with heterologous C2containing fragments from synaptotagmin and a preparation containing a mixture of RACKs. Here, we used recombinant RACK1, recombinant fragments of ␤PKC containing the C1 or the C2 regions, and short synthetic peptides derived from the C2 region, to identify the RACK1-binding site on ␤PKC. We found that the C2 region contains at least part of the RACK1binding site on PKC and that some C2-derived peptides act as specific inhibitors of hormone-induced translocation and functions of the C2-containing ␤PKC isozymes.
Preparation of RACK1-Recombinant RACK1 was overexpressed in Escherichia coli as a fusion protein with the maltose-binding protein and tagged with a FLAG sequence (DTKDDDDK; Kodak) down-stream from the Xa proteolysis site. The overexpressed protein was then purified on amylose affinity column (Biolabs Inc.). The 36-kDa RACK1 containing the FLAG sequence was recovered after removal of the maltose-binding protein by incubating with factor Xa (5 g/ml; Biolabs Inc.) for 48 h at 4°C.
Recombinant Fragments of ␤PKC-The plasmids L9 and L10 expressing the C1 (L9) and C2 (L10) regions as fusion proteins with gluthatione S-transferase (GST) were a generous gift from Dr. Bernard Weinstein, Columbia University, New York, and the GST-fusion proteins expressed in E. coli were produced as described (20).
Overlay Assay-RACK1 was blotted onto nitrocellulose as described elsewhere (24). Strips of the nitrocellulose sheet (ϳ0.5-1 g RACK1/ strip) were incubated in overlay buffer (50 mM Tris-HCl, pH 7.5, containing 0.1% bovine serum albumin, 5 g/ml leupeptin, 10 g/ml soybean trypsin inhibitor, 0.1% polyethylene glycol, 0.2 M NaCl, 0.1 mM CaCl 2 , and 12 mM ␤-mercaptoethanol) or preincubated for 30 min at room temperature, in overlay buffer with the indicated peptide (10 M). L9 or L10 (ϳ10 M) were added in the presence or absence of 50 g/ml PS, and 1 mM calcium, and the mixture was further incubated for 30 min at room temperature. Where indicated, immobilized RACK1 was first incubated with C2-derived peptides in the absence of PKC activators. PKC fragments in the presence or absence of PKC activators were then added, unbound material was removed, and the strips were washed three times for 5 min with overlay wash buffer (0.1% polyethylene glycol, 0.2 M NaCl, 0.1 mM CaCl 2 , 12 mM ␤-mercaptoethanol, and 50 mM Tris-HCl, pH 7.5). Binding of fragments L9 and L10 to RACK1 was detected with anti-GST polyclonal antibodies (1:5000, a gift from Dr. Richard Scheller, Stanford University). The strips were then incubated with anti-rabbit horseradish peroxidase-linked antibodies diluted 1:1000 (Amersham Life Science) followed by chemiluminescent reaction. Quantitation of L9 and L10 binding was carried out by analyzing autoradiograms using a Microscan 1000 gel analyzer (Galai Inc., Israel) at linear range of detection.
Column Assay-An amylose suspension (0.5 ml; Biolabs Inc.) was loaded onto a poly-prep chromatography column (Bio-Rad) and washed with column wash buffer (20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA, 1 mM sodium azide, and 10 mM ␤-mercaptoethanol). Crude extracts of RACK1-maltose-binding protein fusion protein (1 ml, ϳ10 g/ml) were loaded and the column washed extensively (20 column volumes). L9, L10 (ϳ10 M), or partially purified rat brain PKC (containing a mixture of PKC isozymes; ϳ10 nM; 200 units/mg) were then added, in the presence or absence of 50 g/ml PS and 1 mM calcium or 50 g/ml PS, 0.8 g/ml DG, and 1 mM calcium in overlay buffer. After incubation while shaking for 30 min at room temperature, unbound material was removed, the column was washed with column wash buffer (20 column volumes), the proteins were eluted with 10 mM maltose in column wash buffer, and the amount of bound ␤PKC and L10 determined by Western blot analysis. To this end SDS-PAGE sample buffer (0.3 M Tris-HCl, 5% SDS, 50% glycerol, 0.01% bromphenol blue, and 5% ␤-mercaptoethanol) was added, the samples were boiled for 10 min, and loaded on 10% SDS-PAGE gel. Western blot analysis with anti-FLAG antibodies (Kodak; 1:10,000) demonstrated that equal amounts of RACK1 were eluted from all columns by amylose. PKC and L10 levels were then determined using anti-␤PKC monoclonal antibodies (Seikaguku America, Inc; 1:1000). L9 and L10 binding were also detected with anti-GST antibodies, followed by a chemiluminescent reaction as described above.
RACK1 Binding to the C2-derived Peptides-Peptides (ϳ1 nmol/slot) were blotted onto nitrocellulose paper using a slot-blot apparatus (Schleicher & Schuell). Unbound material was removed, and the nitrocellulose was incubated for 2 h in overlay block (20 mM Tris-HCl, pH 7.5, 3% bovine serum albumin, 0.1% PEG, 0.2 M NaCl). RACK1 (ϳ20 nM) was added in overlay buffer, and the blot was incubated for 30 min at room temperature. Binding of RACK1 to these peptides was determined as described above with anti-FLAG monoclonal antibodies (1: 10,000). Binding was quantitated using a MicroScan 1000 Gel Analyzer (Galai Inc. Israel) Cell Culture-Primary cultures of ϳ90% pure cardiac myocytes were prepared from hearts of 1-day-old rats by gentle trypsinization at room temperature as described (7). Cells were cultured for 4 days in Lab-Tek chamber slides (Nunc Inc.), pre-coated with laminin (1 g/ml), in the presence of Dulbecco's modified Eagle's medium, with 5% fetal bovine serum. The culture medium was then replaced with serum-free medium containing transferrin (10 g/ml) and insulin (5 g/ml) for 2 days. Bromodeoxyuridine (0.1 mM) was used through the third day of culture to keep the level fibroblasts at 10% or lower (7).
Permeabilization of Cardiac Myocytes and PKC Translocation: Immunofluorescence Studies-The cells were washed once with PBS and treated with cold permeabilization solution (10 mM EGTA, 140 mM KCl, 20 mM HEPES, 50 g/ml saponin, 5 mM sodium azide, and 5 mM potassium dioxalate, pH 7.4) together with the indicated peptide (10 M) for 10 min at room temperature. Permeabilization does not affect the viability of the cells nor does it alter the rate of spontaneous and stimulated contraction, basal or hormone-induced expression of c-fos, and stimulation-induced hypertrophy, indicating that complex cell functions are not altered by the permeabilization. 2 PMA-induced translocation of PKC isozymes was determined by immunofluorescence studies as described before (32). The cells were then washed with PBS, fixed with cold acetone for 3 min, and washed twice with cold PBS. The cells were incubated for 1 h with 1% normal goat serum in PBS containing 0.1% Triton X-100 followed by overnight incubation with anti-␤I, ␤II, or ␦PKC polyclonal antibodies (Research and Diagnostic Antibodies; 1:100), anti-⑀PKC polyclonal antibodies (Santa Cruz Biotechnology; 1:100), or anti-RACK1 (Transduction Laboratories; 1;100) diluted in PBS containing 0.1% Triton X-100 and 2 mg/ml bovine serum albumin. The cells were washed three times with PBS containing 0.1% Triton, incubated for 2 h with fluorescein-conjugated anti-rabbit IgG antibodies (to detect binding of anti-PKC antibodies; Organon Teknika; 1:1000) or anti-mouse IgM antibodies (to detect binding of anti-RACK1 antibodies; Boehringer Mannheim; 1:1000) and washed again three times with PBS containing 0.1% Triton. After mounting with Miowiol 4 -88 (Calbiochem), the slides were viewed with a Zeiss IM35 microscope using a 40X water immersion objective. Multiple fields of cells for each treatment group and each PKC isozyme were monitored, and the number of cells showing the localization of activated isozyme (32) was recorded. Data are presented as the percentage of cells having the tested isozyme at the activated site. Images of RACK1 and PKC localization from the Zeiss microscope were recorded on Kodak TMax 400 film, and the exposure time was 30 s for these micrographs.
Xenopus Oocyte Maturation Assay-Microinjection of Xenopus oocytes was carried out by Wu and associates (Berkeley, CA) as described previously (8). Oocytes were injected with 50 nl of the indicated peptide 1 h before insulin treatment (8.25 g/ml). Insulin-induced oocyte maturation was then determined by monitoring the appearance of a white spot in the animal pole of the oocyte that is indicative of germinal vesicle breakdown and maturation. Ten to 15 oocytes were included in each treatment, and oocytes were scored for up to 35 h after insulin treatment.
Analysis of ␤PKC Distribution in Xenopus Oocytes-One hour after microinjection of the indicated peptide to the oocytes (100 oocytes/ group), the oocytes were incubated for an additional hour with or without insulin and then homogenates were prepared as described (8). Samples were diluted in homogenization buffer (500 l; 20 mM Tris-HCl, pH 7.5, 10 mM EGTA, 20 mM EDTA, 0.25 M sucrose, and 20 g/ml of each soybean trypsin inhibitor, leupeptin, and aprotonin), and centrifuged at 100,000 ϫ g for 30 min at 4°C. The supernatant (cytosolic fraction) was removed, and the pellet (particulate fraction) was dissolved in homogenization buffer (500 l). After addition of sample buffer, these samples were loaded onto 10% SDS-PAGE, and the amounts of ␤PKC in both fractions were determined by Western blot analysis using anti-␤PKC antibodies (Seikaguku America 1:1000) followed by IgG rabbit anti-mouse (Zymed Inc. 1:2000) and 125 I-labeled protein A. RESULTS Previous studies suggested that at least part of the RACKbinding site on ␤PKC lies within the C2-region of the enzyme (29). In addition these studies suggested that the activators of PKC (PS, DG, and calcium) are required to expose the RACKbinding site on ␤PKC rather than for the interaction of the C2 region with RACKs (29). If the RACK1-binding site on ␤PKC is within the C2 region, a C2-containing fragment of ␤PKC should bind to RACK1 in a PS-, DG-, and calcium-independent manner. We used two fragments of ␤PKC expressed as fusion proteins with GST. One of the fragments, L9, includes the V1 region, the pseudosubstrate sequence, and the C1 and V2 regions (amino acids 3-182) (20). The second fragment, L10, includes the V1 region, the pseudosubstrate sequence, and the first cysteine repeat from the C1 region, as well as the entire C2 and V3 regions (amino acids 3-76 and 143-339) (20). In an overlay assay, L10, the C2-containing fragment, but not L9, the C1-containing fragment, bound to RACK1 (Fig. 1). Saturation of binding of L10 to RACK1 was obtained at ϳ1 M. In contrast, the binding of L9 to RACK1 (Fig. 1, lanes 2 and 4) was minimal, not saturable and similar to the nonspecific binding of GST carrier protein alone (not shown). The V1 region, the pseudosubstrate site, and the first cysteine-loop of C1 are present in both L10 and L9. However, L10 also contains the C2 and V3 regions of PKC. Therefore, these results suggest that the C2 and/or the V3 regions bound to RACK1. In addition, PKC activators PS and calcium did not increase the binding of L10 to RACK1 (Fig. 1, lanes 1 versus 3). Because these activators are required for the binding of intact PKC to RACK1 (25), the data are consistent with our previous studies (29) suggesting that the PKC activators are required only to expose the RACKbinding site in the intact PKC and that this site is already exposed in the C2-containing fragment L10.
If the RACK1-binding site is within the C2 and/or V3 regions of ␤PKC, then L10 (which contains these regions) should inhibit the binding of intact ␤PKC to RACK1. RACK1 was immobilized on an amylose column, and ␤PKC binding in the presence of PS and calcium and L10 or L9 was determined (Fig.  2). In the presence of L10 (Fig. 2, lanes 2 versus 1), but not L9 (Fig. 2, lane 3), ␤PKC binding to RACK1 was completely inhibited. Similar results were also obtained when the effect of the fragments on ␤PKC binding to RACK1 was determined using the overlay assay (not shown). Since L10 inhibited ␤PKC binding to RACK1, the C2 region and/or the V3 regions, present in L10 and not in L9, are likely to contain at least part of the RACK1-binding site on ␤PKC.
We found that synaptotagmin fragments that contain the C2 homologous region bind to purified RACKs (29) and to recombinant RACK1 (n ϭ 3, data not shown) and inhibit PKC bind-ing to RACKs (29). We have therefore reasoned that homologous sequences within the C2 region of ␤PKC and synaptotagmin may mediate their binding to RACK1. Three ␤PKC-derived peptides derived from the homologous sequences of ␤PKC and synaptotagmin were prepared (  -207), that shares no homology with synaptotagmin (Fig. 3A), was also synthesized. As seen in Fig. 3B, immobilized peptides ␤C2-1, ␤C2-2, and ␤C2-4, but not ␤C2-3, bound directly to RACK1. Similar binding of these peptides to RACK1 in the absence of PKC activators was observed (not shown). An additional control peptide corresponding to ␤PKC (amino acids 266 -273 in the C2 domain) did not bind to RACK1 (data not shown; n ϭ 3).
Using primary rat neonatal cardiac myocytes in culture, we have previously demonstrated that activation of PKC by PMA or by norepinephrine causes isozyme-specific translocation to distinct subcellular sites (7,32). If these C2-derived peptides mimic the RACK1-binding site on the C2-containing isozymes, they should inhibit stimulation-induced translocation and binding of these isozymes to their RACKs, but not the translocation of the C2-less isozymes. To test this prediction, we first determined whether RACK1 is present in cardiac myocytes. Using anti-RACK1 antibodies, we found RACK1 immunostaining at perinuclear structures and diffusely in the cytosol (Fig. 4A). RACK1 location was not altered by PMA or norepinephrine (3-100 nM and 2 M, respectively; not shown). However, following activation with norepinephrine or PMA, ␤IIPKC immunoreactivity translocated and was co-localized with RACK1 (Fig. 4, A and B, respectively, and not shown). Partial co-localization of activated ␤IPKC with RACK1 was also noted (not shown, see also Ref. 32). In contrast, RACK1 immunoreactivity did not co-localize with inactive or activated C2-less isozymes, ␦ or ⑀PKC (see in the following and Ref. 32), suggesting that RACK1 may be a specific anchoring protein for activated ␤PKC in cardiac myocytes. FIG. 1. Binding of ␤PKC fragments L9 and L10 to RACK1, and effects of calcium and phosphatidylserine. L10 (lanes 1 and 3) and L9 (lanes 2 and 4) (ϳ10 M) were incubated with nitrocellulose blots of SDS-PAGE loaded with RACK1 (ϳ0.5 g-1 g RACK1/strip) in an overlay assay in the presence (lanes 1 and 2) and absence (lanes 3 and  4) of phosphatidylserine and calcium (PS/Ca). After washing, binding of the fragments was detected with anti-GST antibodies (1:5,000). An autoradiograph, representative of three independent experiments is shown. The arrow indicates the position of RACK1, and the intensity of the bands are proportional to the amount of L9 or L10 bound to RACK1. We next determined whether the C2-derived peptides, that inhibit ␤PKC binding to RACK1 in vitro, inhibit activationinduced translocation of these C2-containing isozymes. The peptides were introduced into the cells by transient permeabilization with saponin (50 g/ml), which has been successfully used to introduce various peptides and other compounds into different cell types (33)(34)(35). As was demonstrated in cardiac myocytes 2 and other cells (33, 36 -38), permeabilization with saponin does not affect the viability of the cells nor other cellular functions including contraction rate, gene expression, and cell growth. After permeabilization, the subcellular localization of different PKC isozymes following stimulation with PMA was determined by immunofluoresence as described previously (32).
Transient permeabilization of these cells in the absence of any peptide did not affect the localization of ␤I, ␤II, ␦, or ⑀PKC isozymes before or after stimulation. ␤IPKC in non-stimulated cells was found on cytosolic structures. After exposure to 100 nM PMA for 15 min, antibodies against this isozyme showed localization to perinuclear and intranuclear structures in ϳ80% of the cells (Fig. 5, vehicle). ␤IIPKC was also cytosolic before stimulation, and in ϳ80% of the cells it translocated to perinuclear structures after PMA treatment (Fig. 5, vehicle). In contrast, permeabilization in the presence of peptides ␤C2-1, ␤C2-2, or ␤C2-4 (10 M extracellular concentration) resulted in inhibition of the PMA-stimulated translocation of the ␤I and ␤IIPKC isozymes by 65-95%, with ␤C2-4 causing the largest inhibition (Fig. 5). Other peptides, including scrambled ␤C2-1, a control peptide derived from the C2-region outside the synaptotagmin-C2 homology region (␤PKC(266 -272)), and the C2derived peptide ␤C2-3 that did not inhibit L10 binding to RACK1 (Fig. 3C) did not affect PMA-induced translocation of ␤I and ␤IIPKC in cardiac myocytes (Fig. 5).
Because the C2 region is present in ␤PKC, but not in ␦ or ⑀PKC, for example (1,16), the C2-derived peptides should only affect the translocation of the C2-containing isozymes, but not that of the C2-less isozymes. Similar to non-permeabilized cells, we found that treatment with 100 nM PMA resulted in the translocation of ⑀PKC from the nucleus to cross-striated structures in 80% of the cells, whereas ␦PKC translocated from the nucleus to perinuclear and fibrillar cytosolic structures in 90% of the cells. Moreover, as predicted, the translocation of these C2-less isozymes was not affected by introduction of any of the C2-derived peptides into the cells (Fig. 5). These results indicate that the ␤C2-1, ␤C2-2, and ␤C2-4 peptides are specific inhibitors of translocation for the C2-containing cPKC isozymes such as ␤I and ␤IIPKC, but not for the C2-less nPKC isozymes such as ␦ and ⑀PKC.
If translocation of ␤PKC is required for its function, peptides that inhibit ␤PKC translocation should also inhibit ␤PKC-mediated function. The function of ␤PKC in cardiac myocytes has not yet been determined. Therefore, we used another assay system, insulin-induced maturation of Xenopus oocytes. We previously demonstrated that oocyte maturation is mediated in part by ␤PKC; insulin treatment results in translocation of ␤PKC (but not other PKC isozymes) from the cytosol to the cell particulate fraction (9) and maturation is delayed by the PKC-specific inhibitor pseudosubstrate peptide (8,9). Furthermore, this insulininduced response is also inhibited when PKC translocation is blocked by injection of purified RACKs (8) or a peptide corresponding to the PKC-binding site on RACKs (9). Therefore, inhibition of translocation inhibits PKC-mediated function.
We then determined whether this inhibition of PKC function in oocytes was due to prevention of insulin-induced ␤PKC translocation to the cell particulate fraction. Since immunofluorescence studies in Xenopus oocytes are not possible, we determined ␤PKC translocation by cell fractionation. The distribution of ␤PKC between the soluble and particulate fractions of oocytes (100,000 ϫ g supernatant and pellet, respectively) was determined in oocytes injected with vehicle or ␤C2-1 using anti-␤PKC antibodies (Fig. 7). Microinjection of ␤C2-1 to nonstimulated oocytes did not affect ␤PKC distribution (not shown) and was similar to control non-stimulated oocytes (Fig.  7, lanes 1 and 2). In vehicle-injected oocytes, insulin treatment resulted in a decrease in the level of the 80-kDa ␤PKC from the cytosol and a corresponding increase in the particulate fraction level (Fig. 7, lanes 4 and 3 versus 2 and 1). However, no insulin-induced translocation of ␤PKC was observed following microinjection of ␤C2-1; rather, there was a decrease in the ␤PKC level in the particulate fraction (Fig. 7, lanes 5 versus 1), suggesting degradation of ␤PKC. Similar results were also observed following microinjection of ␤C2-2 (not shown). Therefore, ␤C2-1 and ␤C2-2-inhibition of PKC-mediated function following insulin-induced stimulation appears to be due to inhibition of ␤PKC translocation. DISCUSSION Using the L9 and L10 recombinant fragments of ␤PKC and short peptides derived from the C2 region, we have mapped at least part of the RACK1-binding site on ␤PKC to amino acids 186 -198 and 209 -226 within the C2 region. Furthermore, peptides corresponding to these sequences inhibited the translocation of C2-containing isozymes but not the translocation of C2-less isozymes in neonatal cardiac myocytes. Finally, these peptides inhibited PKC-mediated function in Xenopus oocytes. Since RACK1 immunoreactivity was found in cardiac myocytes (Fig. 4A) and Xenopus oocytes, 3 it appears likely that the C2derived peptides inhibited PKC function by binding to RACK1 and blocking subsequent binding of the intact enzyme.
The ␤PKC fragment L10 also contains the V3 region of the enzyme, and therefore, our study cannot rule out a role for V3 in binding of ␤PKC to RACK1. Since each of the C2-containing isozymes in cardiac myocytes are localized to a different sub-3 D. Ron and D. Mochly-Rosen, manuscript in preparation.  (32)). For each condition, 100 -150 cells were scored. Results for the subcellular localization of ␤I, ␤II, and ⑀PKC are mean Ϯ S.E. of seven independent experiments for ␤C2-1, eight for ␤C2-2, three for ␤C2-3, two for ␤C2-4, five for ␤C2-1 scrambled, and two for the control peptide (␤PKC(266 -273)). Results for the subcellular localization of ␦PKC are representative of two independent experiments. cellular site (39), it is likely that isozyme unique sequences (e.g. V1, V3, and V5) also contain isozyme-specific RACK-binding sites in addition to the site within the common C2 region. Other studies suggested that binding of PKC to proteins different from RACK1 is mediated by the pseudosubstrate sequence (via a phospholipid bridge (40)) or by the catalytic domain of PKC (41,42). However, the role of the interaction of PKC with these PKC-binding proteins in vivo has not yet been determined.
Very recently, the role of the C1 region in localizing ⑀PKC to the Golgi apparatus has been reported (43). Golgi functions were inhibited by overexpression of both intact ⑀PKC and the C1 fragment of ⑀PKC, leading the authors to suggest that the C1 region may mediate subcellular localization. Our studies with the C1 fragment of ␤PKC cannot exclude the possibility that this domain may also participate in localizing the enzyme. However, the combined in vitro and in vivo studies indicate that the C2 domain is required for this interaction; inhibitors of the C2 domain binding to RACK1 prevent ␤PKC translocation and function.
The C2 region of other translocating enzymes also appears to be required for their translocation and function. Recent data indicate that the C2 region of cytosolic phospholipase A2 associates with membranes, whereas a mutant of this lipase lacking the C2 region does not (44). In addition, a fusion protein containing 43 amino acids from the C2 region of Ras GTPase (GAP) confers calcium-dependent interaction with cellular membranes, whereas a GAP mutant lacking this region does not (45). Finally, the binding of synaptotagmin to membranes is abolished by protease treatment of the membrane (46), and peptides derived from the C2 region of synaptotagmin inhibited calcium-induced neurotransmitter release from the giant squid axon (47). Therefore, the C2 region appears to mediate translocation for a number of translocating proteins.
It appears that the inhibitory effects of the C2-derived peptides presented here are due to the inhibition of translocation of PKC rather than of other C2-containing translocating proteins. PMA is not thought to induce translocation of phospholipase C␥ or GTPase activating protein, and synaptotagmin immunoreactivity was not found in cardiac myocytes (not shown). Similarly, insulin treatment does not induce translocation of phospholipase C␥ or GTPase activating protein in Xenopus oocytes, nor is there synaptotagmin immunoreactivity in oocytes (data not shown). In addition, progesterone-induced oocyte maturation that does not involve PKC activation (8,48) was not affected by the C2-derived peptides (n ϭ 3, data not shown). Therefore, the effects of the ␤C2-1, ␤C2-2, and ␤C2-4 peptides are most likely specific for PKC. Because we do not have antibodies that distinguish between ␤I and ␤IIPKC for Western blot analysis, we could not determine whether one or both mediate the insulin-induced effect.
The inhibitory effects of the ␤C2 peptides were sequencespecific. ␤C2-1, ␤C2-2, and ␤C2-4, but not ␤C2-3 or a number of control peptides, inhibited translocation of ␤PKC. Since ␤C2-1 is highly basic (Fig. 3A), it was previously proposed to  1-4) or ␤C2-1 (50 M; lanes 5 and 6) and the distribution of ␤PKC in the particulate (p) and cytosolic (c) fractions was determined 60 min after incubation without (lanes 1 and 2) or with insulin (lanes 3-6). The cell particulate fraction (lanes 1, 3, and 5) and cytosolic fractions (lanes 2, 4, and 6) of the oocytes were prepared as described under "Experimental Procedures," using 100 oocytes for each treatment and PKC was detected using anti-␤PKC antibodies (1:1000) in Western blot analysis. The antibodies reacted with an ϳ80-kDa protein that corresponds with ␤PKC. The identity of the two other immunoreactive bands in the particulate fraction of control and insulin-treated oocytes is unknown. (This antibody recognizes both ␤I and ␤IIPKC isozymes. However, only ␤IIPKC appears to translocate on insulin treatment in these cells). The figure is a representative of results obtained in three independent experiments. mediate direct binding of the C2-containing proteins to the negatively charged PS in the membrane (21). However, we found that its inhibitory activity cannot be attributed to charge only, since a scrambled ␤C2-1 peptide was inactive in inhibiting either the translocation of C2-containing ␤PKC isozymes (Fig. 5) or their function (Fig. 6).
Why was there a decrease in the levels of ␤PKC in activated oocytes injected with peptides ␤C2-1 or ␤C2-2? Activated PKC has been previously demonstrated to be more sensitive to proteolysis (49) and subsequent inactivation, and we found that co-incubation with RACKs in vitro partially protects from this inactivation. 4 Since the levels of ␤PKC in the absence of insulin treatment did not decrease following microinjection of ␤C2-1 and ␤C2-2 peptides (not shown), the data are consistent with increased sensitivity of activated ␤PKC to proteolysis in the presence of these translocation inhibitors in vivo. Taken together, the data suggest that ␤C2-1, ␤C2-2, and ␤C2-4 bound to the ␤PKC-specific RACK in Xenopus oocytes and inhibited ␤PKC translocation and function.
Relatively high concentrations (50 M) of the C2-derived peptides were required to produce inhibition of insulin-induced oocyte maturation. This may be due to degradation of the peptides in the oocytes during the course of the experiments (hours), the effect of the yolk on the distribution of the peptide in the oocytes, and/or because of competition of the endogenous intact PKC for binding to RACKs. In cardiac myocytes, the intracellular concentration of the peptides has not been determined directly. However, using a tracer radioactive probe, the final concentration of the peptide was estimated to be ϳ1 M when the applied concentration of the peptide is 10 M. 2 Furthermore, since the peptides are likely to be sensitive to proteolysis, the intracellular concentration in cardiac myocytes is likely to be much lower.
Because the C2-derived peptides specifically inhibited translocation and function of ␤PKC in vivo, they belong to a new class of PKC inhibitors, translocation inhibitors. This term was introduced to describe an inhibitor of 5-lipoxygenase, MK0866. MK0866 inhibits the binding of 5-lipoxygenase to its intracellular receptor, FLAP, a binding that is required for activation of the enzyme (50). We have recently identified additional translocation inhibitors for PKC; a 15 amino acid peptide (peptide I) from other PKC-binding proteins inhibited the interaction between PKC and RACK1 and the function of ␤PKC in vivo (9,51). Therefore, peptides that mimic the interaction sites on either PKC or RACK are inhibitors of PKC-mediated function.
How do short peptides from the C2 region of ␤PKC inhibit PKC translocation and function? Many small ligands are thought to interact with their receptor or enzyme via a "greasy pocket." In contrast, the interaction of two proteins in the cell may reflect multiple binding sites on relatively large surface areas of these proteins. Peptides derived from the interacting areas are likely to have lower affinity for their binding sites as compared to the intact proteins. Yet, if these peptides act as a wedge between the two interacting proteins, they may serve as specific and effective inhibitors.
The structure of the C2 domain of p65 has recently been determined (52). Examination of that structure demonstrates that ␤C2-1, ␤C2-2, and ␤C2-4 but not ␤C2-3 are localized on adjacent exposed regions in the C2 domain. Two of the biologically active peptides (␤C2-1 and ␤C2-2) correspond to sequences that contain both loops and ␤ strand structures. In contrast, ␤C2-4, the peptide with the highest biological activity (Fig. 5), corresponds to sequences from only a surface ␤ strand structure (the fifth ␤ strand (52)). It is possible that the interaction surfaces between PKC and RACK1 are constituted only by ␤ strands. Future experiments utilizing shorter peptides will address this possibility. It is interesting to note that the co-crystal structure of the Ras-binding domain of c-Raf-1 with the small GTP-binding protein Rap1A suggests that the interaction between the two proteins is mediated by single anti-parallel ␤ strands at the edge of the molecules (53). The authors also predicted that such a small interaction site between the two proteins provides an opportunity to design drugs that inhibit this interaction. Indeed, our data on ␤PKC and RACK1 interaction demonstrate that such inhibitors have the predicted biological activity.
Finally, PKC translocation inhibitors can be used to elucidate the cellular role of specific PKC isozymes. Our finding that the ␤C2-1, ␤C2-2, and ␤C2-4 peptides caused a delay in insulin-induced oocyte maturation indicates that C2-containing isozymes, most likely ␤PKC, mediate this function in oocyte maturation. Since in cardiac myocytes, inhibition of translocation of C2-less isozymes was not observed, the inhibitory peptides can be used as tools to identify the PKC isozymes that mediate specific cellular functions in cells in which multiple isozymes are activated by a single stimulus. The role of C2containing isozymes in the PMA-induced regulation of cardiac myocyte function is currently under investigation using these peptides.