A protein kinase C translocation inhibitor as an isozyme-selective antagonist of cardiac function.

Protein kinase C (PKC) isozymes translocate to unique subcellular sites following activation. We previously suggested that translocation of activated isozymes is required for their function and that in addition to binding to lipids, translocation involves binding of the activated isozymes to specific anchoring proteins (receptors for activated protein kinase C. Using cultured cardiomyocytes we identified inhibitors, the V1 fragment of εPKC (εV1), and an 8-amino acid peptide derived from it that selectively inhibited the translocation of εPKC. Inhibition of εPKC translocation but not inhibition of δ or βPKC translocation specifically blocked phorbol ester- or norepinephrine-mediated regulation of contraction. These isozyme-selective translocation inhibitors provide novel tools to determine the function of individual PKC isozymes in intact cells.

Protein kinase C (PKC) isozymes translocate to unique subcellular sites following activation. We previously suggested that translocation of activated isozymes is required for their function and that in addition to binding to lipids, translocation involves binding of the activated isozymes to specific anchoring proteins (receptors for activated protein kinase C. Using cultured cardiomyocytes we identified inhibitors, the V1 fragment of ⑀PKC (⑀V1), and an 8-amino acid peptide derived from it that selectively inhibited the translocation of ⑀PKC. Inhibition of ⑀PKC translocation but not inhibition of ␦ or ␤PKC translocation specifically blocked phorbol ester-or norepinephrine-mediated regulation of contraction. These isozyme-selective translocation inhibitors provide novel tools to determine the function of individual PKC isozymes in intact cells.
Activation of protein kinase C (PKC) 1 isozymes is associated with translocation of the enzymes form the cell soluble to the cell particulate fraction (1). These isozymes are activated by binding to lipid-derived second messengers and negatively charged phospholipids present in the cell particulate fraction (2,3). In addition to binding to lipids, specific anchoring proteins participate in binding the activated PKC isozymes to this fraction (4 -9). We collectively termed these proteins RACKs, for receptors for activated protein kinase C (7,10).
In cultured neonatal cardiomyocytes, immunofluorescence studies demonstrated isozyme-specific subcellular localization following activation with either 4-␤ phorbol 12-myristate-13acetate (PMA) or with norepinephrine (NE) via an ␣ 1 -adrenergic receptor (11,12). Similar isozyme-specific localization was found in other cells following PKC activation (e.g. (13)). This isozyme-specific localization suggests that unique sequences in each isozyme (14) contain at least part of the recognition site for the anchoring molecules, the isozyme-specific RACKs.
Here, we focus on the ⑀PKC unique region, ⑀V1, which is the largest variable region in this isozyme. Some homology between ⑀V1 and the C2 region of the classical PKCs, ␣, ␤, and ␥, was noted (15). Because C2 contains at least part of the RACKbinding site of classical PKCs (16,17), an ⑀PKC specific RACKbinding site may reside within ⑀V1. In that case, an ⑀V1 fragment should bind to the ⑀PKC-specific RACK when introduced into cells and thus inhibit PMA-or hormone-induced ⑀PKC translocation and binding to that RACK. Translocation of other PKC isozymes should not be affected by ⑀V1. The following study confirms these predictions and demonstrates the use of translocation inhibitors to determine the role of specific isozymes in regulating cardiac contraction.
Expression and Purification of the V1 Regions-The V1 regions of ⑀ and ␦PKC (amino acids 2-144) were amplified by polymerase chain reaction from a rat cDNA library (Stratagene) and tagged with a FLAG epitope (DYKDDDDK) at the 5Ј end of the fragments, and the polymerase chain reaction fragments were subcloned into pMAL-c2 vector (New England BioLabs) for overexpression as fusion proteins with maltose binding protein in Escherichia coli. Protein purification and factor Xa proteolysis of the fusion proteins were as before (19), and the resulting V1 fragments were Ͼ90% pure.
Permeabilization of Neonatal Rat Cardiomyocytes Culture and Immunolocalization of PKC Isozymes-Primary neonatal rat cardiomyocytes were cultured on chamber slides as described previously (12) and transiently permeabilized with saponin (50 g/ml) with or without 150 g/ml of rat recombinant PKC fragment, ⑀V1 or ␦V1. 2 PKC isozyme localization was determined by immunofluorescence in cells fixed with methanol and acetone (12). The anti-␦ and ⑀PKC antisera (Research and Diagnostic Antibodies, Inc) do not recognize the corresponding V1 fragments. Multiple randomly selected microscopic fields were monitored for each study to determine the percentage of cells having the tested isozymes at the activated site. There was no bias in scoring; in an additional blind study, essentially identical results were obtained.
The following criteria were used to confirm that the immunostaining is specific for the monitored isozymes. No immunostaining is obtained by antisera preadsorbed with the corresponding immunizing peptide or by the corresponding sf9-expressed isozyme (Ref. 12 and data not shown.) We also showed that there was a complete correlation between translocation from the cell soluble to the cell particulate fraction as measured by Western blot analysis (using antibodies from Life Technologies, Inc.) and translocation measured by immunofluoresence using the same antibodies as in this study (12,21,22). Furthermore, the results obtained with the commercial anti-⑀PKC are indistinguishable from those that we previously obtained with our own monoclonal antibody CK 1.4 (11). Finally, localization of PKC to cross-striated structures in heart was reported by Kuo and collaborators using another antiserum (23).
Monitoring of Cardiomyocyte Contraction Rates-Cells were cultured and permeabilized with saponin as above in the absence or the presence of 150 g/ml ␦V1 fragment, ⑀V1 fragment, or the indicated concentration of various PKC-derived peptides. Basal contraction rates were then monitored for 10 min to ensure stable contraction rates following permeabilization. (The rates of contraction after permeabilization with * This work was supported by Grant HL52141 from the National Institutes of Health. 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. ‡ To whom correspondence should be addressed. Tel.: 415-725-7720; Fax: 415-725-2952. 1 The abbreviations used are: PKC, protein kinase C; RACK, receptor for activated protein kinase C; PMA, phorbol 12-myristate-13-acetate; NE, norepinephrine; AR, adrenergic receptor. obtained by scoring 27 random microscopic fields from a single experiment, with greater than 700 cells scored for each treatment group in each experiment. Similar data on translocation were obtained by several experimenters. An example of the primary data used for A-E is given in F. Cells permeabilized in the presence of ⑀V1 fragment (c, g, and k) or ␦V1 fragment (d, h, and l) or in their absence (a, b, e, f, i, and j) were incubated with either 100 nM 4-␣ (a, e, and i) or 4-␤ PMA (b-d, f-h, and j-l), and the localization of ⑀PKC (a-d), ␦PKC (e-h), and ␣PKC (i-l) was determined as above by immunofluoresence. 4-␤ PMA treatment caused translocation of ⑀PKC from the nucleus and perinucleus to cross-striated structures and cell-cell contacts (intercalated discs; a versus b), ␦PKC translocated from the nucleus to the perinucleus (e versus f), and ␣PKC translocated from diffuse cytosolic sites to the perinucleus, yielding a thick perinuclear outline (i versus j). Also noted is that ⑀V1 fragment inhibited the translocation of ⑀PKC (c versus b) but not that of ␦PKC or ␣PKC (g and k versus f and j), whereas ␦V1 fragment inhibited the translocation of ␦PKC (h versus f) but not that of ⑀PKC or ␣PKC (d and l versus b and j) by 4-␤ PMA. (When PKC isozymes are not at the cell periphery, such as in c and f, it may appear as though the cells are smaller in size. This is not the case, however; phase contrast show that cell size is not altered and that many of the cells in these panels touch each other.) vehicle, ␦V1, or ⑀V1 fragment were 49 Ϯ 5, 50 Ϯ 6, and 40 Ϯ 6 contractions per 15 s (mean Ϯ S.E. of 12, 6, and 12 experiments, respectively)). The cells were then treated with 4-␤ PMA, and the rate of contraction was monitored at the indicated times as described previously (22).

RESULTS
To determine whether the first variable region of ⑀PKC, ⑀V1, contains a specific anchoring site of the enzyme, cardiomyocytes were transiently permeabilized by saponin with or without ⑀V1, and the subcellular localization of different PKC isozymes following PMA stimulation was determined (Fig. 1). Transient permeabilization alone did not affect many cellular functions including cell viability, spontaneous and stimulated contraction rates (see Figs. 2 and 3), gene expression, and hypertrophy. The intracellular concentration of the fragment was ϳ10% of that applied. 2 As in a previous study (12), we found that in cardiomyocytes, activated ␣PKC is localized to the nuclear boundary (Fig. 1, F, j), activated ␤IPKC is localized inside nuclei (not shown), activated ␦PKC is localized to perinuclear structures (Fig. 1, F, f), and activated ⑀PKC is localized to cross-striated structures (Fig. 1, F, b). Introduction of ⑀V1 caused selective inhibition of translocation of ⑀PKC (Fig. 1A) but not of ␣, ␤I, or ␦PKC (not shown) to their corresponding subcellular sites following treatment with 3 nM 4-␤ PMA. ⑀V1 also caused partial inhibition of ⑀PKC (see Fig. 1, F, c versus b) but not ␣, ␦ or ␤IPKC translocation (Fig. 1, B-F) when the isozymes were fully activated (22) with 100 nM 4-␤ PMA. Therefore, the ⑀V1 fragment appears to compete with endogenous activated ⑀PKC and selectively prevents its translocation and binding to cross-striated structures. As a control, we used cells that were transiently permeabilized with the first variable region of ␦PKC, the ␦V1 fragment. We found that ␦V1 selectively inhibited 4-␤ PMA-induced translocation of ␦PKC (Fig. 1, F, h versus f) but not the translocation of ␣, ␤I, or ⑀PKC, to their respective subcellular sites (Fig. 1, B-F). Therefore, ⑀V1 and ␦V1 probably contain a corresponding PKC-specific RACKbinding site and can be used as isozyme-selective inhibitors of activation-induced translocation.
If stimulation-induced translocation of PKC isozymes is required for their unique functions, then introduction of the above isozyme-specific translocation inhibitors into cardiomyocytes should inhibit these functions. We focused on the effect of ⑀V1 on 4-␤ PMA and ␣ 1 -adrenergic regulation of contraction, because activated ⑀PKC and the contractile apparatus in cardiomyocytes are both located in cross-striated structures (10) and because correlation studies using 4-␤ PMA implicated ⑀PKC in this function (22). Cardiomyocytes in culture contract spontaneously at ϳ200 beats/min and incubation with 4-␤ PMA reduces this rate (22). We determined the effect of ⑀V1 on this 4-␤ PMA-induced negative chronotropy. ⑀V1 or ␦V1 fragments were introduced into the cells by transient permeabilization with saponin as above. Following removal of saponin, a stable rate of spontaneous contraction resumed within a few minutes. 2 In cells permeabilized with vehicle alone or with ␦V1, 4-␤ PMA inhibited the basal contraction rate by an average of 70 -85% (Fig. 2, A and B). In contrast, following permeabilization in the presence of ⑀V1, the 4-␤ PMA-induced negative chronotropy was abolished even 60 min after PMA addition (3-10 nM; Fig. 2C). This effect was dependent on PMA dose, with an average of 50% inhibition by ⑀V1 in cells treated with 100 nM 4-␤ PMA for 60 min (n ϭ 3). Of interest, ⑀V1 did not affect the catalytic activity of partially purified ⑀PKC (ϳ100 units/mg) in vitro; phosphorylation of a purified 55-kDa ⑀PKC substrate derived from cardiomyocyte sarcoplasmic reticulum in the presence of ⑀V1 was 118 Ϯ 17% of that in its absence (mean Ϯ S.E., n ϭ 3). 3 Furthermore, because ⑀V1 specifically inhibited 4-␤ PMA-induced translocation of ⑀PKC but not the translocation of ␦, ␤I or ␣PKC (Fig. 1, A and B versus C, D, and E), these data suggest that ⑀V1 antagonizes 4-␤ PMA regulation of contraction by inhibiting translocation of ⑀PKC to its site of action.
We next determined whether ⑀V1 also inhibits hormoneinduced regulation of contraction rate in cardiomyocytes. These cells contain ␣ 1 -and ␤ 1 -adrenergic receptors (ARs), which are coupled to activation of PKC and cAMP-dependent protein kinase, respectively. Stimulation with NE, which activates both ARs, resulted in a transient increase in contraction rate (Fig. 3A). Inhibition of both ARs with prazosin and propranolol followed by stimulation with NE inhibited the effect of NE (Fig.  3B). Inhibition of the ␤ 1 -AR by propranolol allowed the effect of ␣ 1 -AR activity to be manifested and hence caused a 4-␤ PMAlike reduction of contraction rate (albeit to a smaller extent, Figs. 3C versus 2A). In contrast, specific inhibition of the ␣ 1 -AR by prazosin caused a greater increase in the immediate rise in NE-induced contraction rate and a longer period of elevated rate than that observed with NE alone (e.g. see 40 min after drug addition; Fig. 3, D versus A). Therefore, ␣ 1 -AR activation, which results in translocation of all the PKC isozymes in these cells (12), causes negative chronotropic effects and attenuates the positive chronotropic effect of ␤ 1 -AR stimulation.
If these ␣ 1 -AR-stimulated effects on contraction are mediated by ⑀PKC, then cells containing ⑀V1 should respond to NE alone as though prazosin were present. Indeed, NE caused a greater and a more sustained increase in the rate of contraction in cells permeabilized in the presence of ⑀V1 as compared with cells permeabilized in the absence of any fragment or with cells permeabilized in the presence of ␦V1 (Fig. 3E). These data indicate that ␣ 1 -AR effects on contraction rate are me- diated by ⑀PKC and that this ⑀PKC-selective translocation inhibitor also blocks hormone-induced regulation of function in cardiomyocytes.
Finally, short peptides derived from the C2 region of ␤PKC (e.g. ␤C2-4) selectively inhibit translocation and function of C2-containing isozymes (17,24). We synthesized several short peptides derived from the ⑀V1 region and determined their effect on translocation and function of ⑀PKC. The peptides ⑀V1-1, -2, -3, -4, and -5 corresponding to amino acids 5-11, 14 -21, 81-87, 92-100, and 116 -125, respectively, were synthesized based on the sequences conserved between Aplysia and rat ⑀PKC (15). We reasoned that conserved sequences between the evolutionarily remote species are likely to be functionally important. Of the five peptides tested, only ⑀V1-2 inhibited 4-␤ PMA-induced ⑀PKC translocation (Fig. 4, D versus B). Translocation of ␤I or ␦PKC was not inhibited by this peptide (Fig. 4C). Therefore, at least part of the RACK-binding site on ⑀PKC is localized between amino acids 14 -21 in the V1 region. Furthermore, ⑀V1-2 but not ␤C2-4 (a translocation inhibitor specific for C2-containing isozyme (17)) inhibited the 4-␤ PMA-induced negative chronotropic effect in cardiomyocytes (Fig. 4E) at intracellular concentrations as low as 10 nM (not shown). No inhibition of the 4-␤ PMA effect on contraction was observed in the presence of 1-10 M of the other ⑀V1derived peptides, a PKC-derived peptide that is highly homologous to ⑀V1-2 (EAVgLqPT; lower case for amino acid substitution), or a scrambled ⑀V1-2 (not shown). Therefore, an ⑀V1derived octapeptide selectively inhibits ⑀PKC translocation and function. DISCUSSION Although PKC has been a focus of research for many years, understanding of the role of individual members of this family of isozymes was hindered by the lack of isozyme-selective inhibitors. As Wilkinson and Hallam commented in a recent review (25), "A full understanding of the role of the specific in an additional blind study, the percentage of cells with cross-striation after 15 min with 10 nM 4-␤ PMA was 65 and 23% for sham-and V1-2-treated cells, respectively, and 20 and 21%, respectively, after 20 min with 10 nM 4-␣ PMA (an average of 634 cells were scored for each treatment). In E, cells permeabilized in the presence of 10 M ⑀V1-2 (f) or the classical PKC-selective inhibitor ␤C2-4 (Ⅺ) (17) were treated with 10 nM 4-␤ PMA at time 0, and the rate of contraction was monitored at the indicated times, as described in the legend to Fig. 2. The number of experiments, each from a separate myocyte preparation, was six and three for ⑀V1-2 and ␤C2-4, respectively. Similar data were obtained also with cells permeabilized in the presence of 100 nM ⑀V1-2.
PKC isotypes in physiological and pathophysiological processes awaits the development of yet more specific activators and isoenzyme-selective inhibitors." The work described above directly addresses this issue.
Neonatal cardiomyocytes contain at least six different PKC isozymes (12,22,26,27) that translocate following NE or 4-␤ PMA treatment (11,12). Because 4-␤ PMA and NE induce a variety of physiological responses including hypertrophy, modulation of contraction rate, gene expression, and organization of contractile elements (22, 28 -32), it is likely that each isozyme mediates a different function. Our data indicate that translocation of ⑀PKC to the cross-striated structure but not the translocation of other isozymes such as ␣, ␤I, or ␦PKC to their subcellular sites is required for 4-␤ PMA and NE modulation of spontaneous contraction rate. Therefore, as we predicted earlier (10,11), the subcellular localization of the activated isozymes influences the specific function and presumably the substrates of each PKC isozyme.
Recently, the C1 and V3 regions of ⑀PKC were suggested to determine subcellular localization of inactive ⑀PKC. Nonactivated overexpressed ⑀PKC was localized to multiple subcellular sites including the Golgi apparatus, whereas a C1-containing fragment, for example, was localized only to Golgi (33,34). The effect of the C1 fragment on the localization and function of either the overexpressed full-length ⑀PKC or the endogenous enzyme were not determined. In addition, an ⑀PKC-unique C1-derived pentapeptide representing an actin-binding sequence (20) was found to inhibit ⑀PKC binding to actin in vitro. However, the function of this binding in vivo has not yet been determined. In the present work, we showed that the V1 region is required for anchoring of activated ⑀PKC and that inhibition of the anchoring by the V1 region or by amino acids 14 -21 (⑀V1-2) are sufficient to inhibit translocation and function of the activated ⑀PKC in intact cells. Therefore, PKC translocation is required for its function, and translocation inhibitors can be used to obtain isozyme-selective inhibition of PKCmediated functions.