State-specific monoclonal antibodies identify an intermediate state in epsilon protein kinase C activation.

Evaluation of the activation state of protein kinase C (PKC) isozymes relies on analysis of subcellular translocation. A monoclonal antibody, 14E6, specific for the activated conformation of epsilonPKC, was raised using the first variable (V1) domain of epsilonPKC as the immunogen. 14E6 binding is specific for epsilonPKC and is greatly increased in the presence of PKC activators. Immunofluorescence staining by 14E6 of neonatal rat primary cardiac myocytes and the NG108-15 neuroblastoma glioma cell line, NG108-15/D2, increases rapidly following cell activation and is localized to new subcellular sites. However, staining of translocated epsilonPKC with 14E6 is transient, and the epitope disappears 30 min after activation of NG-108/15 cells by a D2 receptor agonist. In contrast, subcellular localization associated with activation, as determined by commercially available polyclonal antibodies, persists for at least 30 min. In vitro, epsilonRACK, the receptor for activated epsilonPKC, inhibits 14E6 binding to epsilonPKC, suggesting that the 14E6 epitope is lost or hidden when active epsilonPKC binds to its RACK. Therefore, the 14E6 antibody appears to identify a transient state of activated but non-anchored epsilonPKC. Moreover, binding of 14E6 to epsilonPKC only after activation suggests that lipid-dependent conformational changes associated with epsilonPKC activation precede binding of the activated isozyme to its specific RACK, epsilonRACK. Further, monoclonal antibody 14E6 should be a powerful tool to study the pathways that control rapid translocation of epsilonPKC from cytosolic to membrane localization on activation.

Several isozymes of protein kinase C (PKC), 1 lipid-dependent protein kinases, are present within a single cell, each mediat-ing unique intracellular functions. Studies using conventional or confocal microscopy reveal a complex and specific localization of PKC isozymes in their inactive as well as their active state (1)(2)(3)(4). Most isozymes are localized to unique sites prior to cell stimulation, and translocate upon activation to new distinct intracellular sites. PKC isozyme localization is determined by binding to specific anchoring molecules termed RACKs (receptors for activated C-kinase) (5). Two RACKs have been identified and characterized to date: the ␤IIPKC-specific RACK (RACK1) (6) and the ⑀PKC-specific RACK (⑀RACK), also known as ␤ЈCOP (7).
Many aspects of PKC activation are not fully understood. Are conformational changes associated with PKC activation? Does lipid binding precede RACK binding? In order to answer these and other questions, it is necessary to develop new "statespecific" reagents to distinguish between active and inactive individual PKC isozymes.
⑀PKC is an isozyme that regulates many cellular functions (8 -14). In cardiac myocytes, ⑀PKC mediates cardioprotection from an ischemic episode (9, 11, 14 -16), cardiac hypertrophy (17,18), regulation of L-type calcium channel (19,20), and regulation of contraction rate (10,21). The V1 domain of ⑀PKC is involved in the binding of activated ⑀PKC to its RACK in these cells (7,10,22). Because binding to RACKs occurs only after activation of ⑀PKC (7,23,24), we expected that a conformational change in this domain must occur upon activation. We therefore predicted that some antigenic determinants on the V1 domain should be exposed only following ⑀PKC activation and that some of the antibodies raised against V1 might recognize the active state of ⑀PKC and be isozyme-selective.
We report here the production and characterization of an isozyme-selective monoclonal antibody (mAb) to ⑀PKC that is specific for the activated form of this isozyme. Activation is required to expose the epitope for this antibody, but 14E6 no longer binds to the activated enzyme when the latter is bound to ⑀RACK. Together, these studies suggest the existence of a previously unidentified state in PKC activation: a transient, activated, but non-anchored state. Immunostaining with 14E6 should help in identifying cells in tissues where ⑀PKC has been activated by a physiological trigger. In addition, 14E6 will be a useful marker to follow the pathway of ⑀PKC translocation and help elucidate the mechanism involved in this pathway.

EXPERIMENTAL PROCEDURES
Materials-Recombinant ␦, ⑀, and ␥ PKCs produced in Sf9 cells were obtained from PanVera (Madison, WI). Phorbol myristate acetate (PMA) was from Alexis (San Diego, CA), and phosphatidylserine (PS) and dioleoylglycerol (DG) were purchased from Avanti (Alabaster, AL). PKC was partially purified from rat brain by DEAE-cellulose chroma-* This work was supported by Grants AA11147 (to D. M.-R.) and AA010030 (to I. D.) from the National Institutes of Health. This research was also supported by funds provided by the State of California for medical research on alcohol and substance abuse through the University of California, San Francisco (to I. D., A. S. G., and 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.
Preparation of Recombinant Proteins-The V1 regions of rat ␦ and ⑀PKC (amino acids 2-144 and 2-145, respectively) were expressed as fusion proteins with MBP (maltose-binding protein; pMAL-c2 vector, New England BioLabs) in Esherichia coli. Vectors were also constructed expressing the C-terminal half of ⑀RACK as an MBP fusion protein (MBP-⑀RACK-C, amino acids 425-905), MBP-LacZ, and MBP-␤IIPKC-V5 (amino acids 622-673). MBP fusion proteins were purified by affinity chromatography on amylose resin according to the New England Bio-Labs protocol. Where indicated, cleavage of the fusion protein and removal of MBP were carried out with 5 g/ml Factor Xa (New England BioLabs) for 24 h at 4°C.
Monoclonal Antibody Production-Purified ⑀PKC-V1 was incubated with 120 g/ml PS, 4 g/ml DG, and ⑀ RACK-C for 15 min at room temperature (PS and DG in ethanol were dried under nitrogen, 20 mM Tris, pH 7.5 was added to bring PS to 1.2 mg/ml and DG to 40 g/ml, the mixture sonicated on ice, with a Branson Sonifier at 20% output at setting 2.5, using 3 cycles of 1 min each, with 30 s cooling periods on ice between cycles). Eight-week-old BALB/c mice (Harlan) were injected subcutaneously three times at 3-4-week intervals with 10 g of ⑀PKC-V1 and 10 g of ⑀RACK-C with PS and DG, emulsified with Freund's adjuvant. Four days after the final boost, splenocytes (10 8 ) were fused with 2 ϫ 10 7 murine myeloma Fox-NY cells (from Thomas A. Stamey) according to Kohler and Milstein (26). Cells were grown in HAT-containing medium as described (25), and positive cultures were subcloned three times. Antibodies were obtained either as ascites in BALB/c mice or as supernatant in serum-free medium in roller bottles, and purified by precipitation with 50% ammonium sulfate, followed by gel filtration chromatography on Sephacryl S-300. SDS-PAGE revealed the expected 75 and 25 kDa heavy and light chains at ϳ95% purity. Immunoglobulin isotypes were determined with a kit from BD Pharmingen.
ELISA Binding Assays-Recombinant MBP fusion proteins and PKCs were incubated at 3-5 g/ml in PBS in 96-well ELISA plates for 2 h at room temperature or overnight at 4°C. Plates were washed three times with PBS plus 0.05% Tween-20 (PBS/Tween), blocked with 1% BSA in PBS for 2 h (this and subsequent incubations were at room temperature), and washed as above. 100 l of antibody was added, and the plate was further incubated for 1 h, followed by three washes with PBS/Tween. 100 l of HRP-goat anti-mouse IgG (Sigma Chemicals and Jackson Laboratories, West Grove, PA; diluted 1:5000 in PBS containing 1% BSA) was added and incubated for 1 h, followed by three washes with PBS/Tween. 100 l of O-phenylenediamine and 0.03% H 2 O 2 in 100 mM citrate buffer, pH 6, was added to assess HRP activity; after 15 min, 50 l of 2 N H 2 SO 4 was added to stop the reaction and absorbance at 490 nM determined. To determine whether binding of ⑀RACK to ⑀PKC affects the subsequent binding of antibodies, 5 g/ml MBP-⑀RACK-C was added to immobilized MBP-⑀PKC-V1 for 1 h at room temperature, followed by three washes with PBS/Tween, prior to antibody addition.
ELISA for Analysis of ⑀PKC Activation-Activation of ⑀PKC was carried out by incubation of 1-5 g of PanVera Sf9 ⑀PKC in a final volume of 100 l containing 20 mM Tris, pH 7.5, 60 g/ml PS, and 2 g/ml DG, at room temperature for 5 min. The mixture was diluted to 10 ml with 20 mM Tris, pH 7.5, and bound to ELISA plates (Costar 3590) at 100 l per well for 1 h. Plates were washed (this and all subsequent washes were carried out three times with PBS/Tween), blocked with 5% nonfat dry milk powder in PBS/Tween for 45 min, washed again, incubated with antibodies in PBS/Tween for 1 h at room temperature, and washed again. Plates were incubated with anti-mouse IgM-horseradish peroxidase conjugate (Pierce Endogen 31440) for 1 h, washed, and developed with 3,3Ј,5,5Ј-tetramethylbenzidine (Sigma T3405; 1 mg per 10 ml of 50 mM sodium phosphate/25 mM citric acid, pH 5.5 with 2 l of 30% hydrogen peroxide). The reaction was stopped with 50 l of 2 M H 2 SO 4 and absorbance determined at 450 nm.
Cardiac Myocyte Preparation-Primary cardiac myocytes were prepared from the hearts of 1-day-old Sprague-Dawley rats by gentle trypsinization as described (27). Non-myocytes were removed by preplating for 30 min. Myocytes were plated at 800 cell/mm 2 in M-199 medium (Invitrogen, Life Technologies, Inc.) with 0.1 mM bromodeoxyuridine, 2 g/ml vitamin B 12 , 100 units/ml penicillin, 100 g/ml streptomycin sulfate, and 10% fetal bovine serum (Hyclone, Logan, UT) in 1% CO 2 . After 24 h, the culture medium was changed to the above medium containing 80 M ascorbic acid, and cells were cultured an additional 72 h. Myocytes were then incubated in a defined medium (M-199 with penicillin and streptomycin, 2 g/ml vitamin B 12 , and 10 g/ml each of insulin and transferrin) for 24 h, after which experiments were initiated.
NG108-15 Cell Culture-NG108 -15 cells were cultured as described (28), and maintained in a serum-free defined medium for 6 days prior to immunofluorescence or immunoprecipitation experiments. Experiments with the dopamine receptor agonist NPA were carried out with NG108 cells stably transfected with the dopamine D2 receptor, NG108-15/D2 (29).
Immunofluorescence-Cardiac myocytes were plated on glass cover slips or in chamber slides coated with 1.2 g/ml laminin from mouse sarcoma cells (Invitrogen, Life Technologies, Inc.). Following the indicated treatments, cells were fixed for 3 min on ice in methanol/acetone (1:1, Ϫ20°C). Slides were washed three times with cold PBS, blocked with PBS plus 0.1% Triton X-100 (PBS/Triton) containing 1% normal goat serum, and incubated with primary antibodies overnight at 4°C in PBS/Triton containing 1% normal goat serum. After three washes with PBS/Triton, cells were incubated with fluorescein labeled-goat antimouse IgM, anti-mouse IgG, or anti-rabbit IgG (Organon Teknika Corp., Durham, NC) in PBS/Triton containing 1% normal goat serum for 2 h at room temperature, washed three times with PBS/Triton, and mounted in Vecta Shield (Vector Labs, Burlingame, CA). Fluorescence microscopy was carried out with a ϫ40 water immersion objective or by confocal microscopy. NG108 -15/D2 cells were fixed with methanol and confocal microscopy carried out as described (28). These fixation conditions were found optimal to conserve antibody epitopes and cell structure.
Immunoprecipitation of NG108 -15 Cell Lysates-For immunoprecipitation with 6E4 antibody, 2 g of 6E4 was incubated with 30 l of packed volume of protein G beads (Invitrogen) overnight at 4°C. For immunoprecipitation with 14E6 antibody, 2 g of 14E6 in PBS was added to 30 l of packed volume of anti-mouse IgM agarose (Sigma) overnight at 4°C. Antibody-bound beads were then washed twice with PBS and blocked with 3% BSA for 2 h at 4°C. Triton-soluble material was diluted 5 times in homogenization buffer to dilute the Triton X-100 to 0.2%, then precleared with protein G or anti-IgM beads for 30 min at 4°C, incubated with antibody-bound beads overnight at 4°C, and subsequently washed four times with PBS containing 0.1% Triton X-100. Bound material was eluted with SDS sample buffer, run on an 8% SDS-PAGE, transferred and probed for ⑀PKC (anti-V5 Santa Cruz) and ⑀RACK (monoclonal antibody 4B12).
Immunoprecipitation of Rat Brain PKC-For immunoprecipitation, 5 g of each monoclonal antibody was added to 50 l of goat anti-mouse IgM (Jackson Laboratories) conjugated to Sepharose beads using cyanogen-bromide-activated Sepharose 4B (Amersham Biosciences), or to 50 l of protein G-Agarose beads (Invitrogen, Life Technologies, Inc.), and rocked for 2 h at 4°C. The beads were blocked with 3% BSA in PBS for 1 h at 4°C, and washed three times with PBS/0.05% Tween. Approximately 30 g of rat brain PKC (partially purified by DEAE cellulose and gel filtration chromatography, Ref. 17) was diluted in 20 mM Tris, pH 7.4, with 20 mg/ml soybean trypsin inhibitor, 20 mg/ml aprotinin, and 10 mg/ml phenylmethylsulfonyl fluoride (protease inhibitors) without EDTA or EGTA, and incubated for the indicated times at room temperature with 120 g/ml PS and 4 g/ml DG. Brain PKC was then added to the beads and the incubation continued for 2 h at 4°C, followed by washing the beads three times with PBS/0.1% Triton X-100.

RESULTS
Our purpose was to generate mAbs that distinguish between active and inactive ⑀PKC. The V1 domain of this isozyme was chosen as the immunogen because it contains a RACK binding site (7, 10) that is exposed only after ⑀PKC activation. We therefore expected that certain epitopes in this domain may be exposed only following activation of ⑀PKC and therefore should be recognized by antibodies that would only recognize activated ⑀PKC. Mice were injected with recombinant protein comprising the V1 region of rat ⑀PKC (amino acids 2-145). Hybridoma culture supernatants were screened by ELISA for binding to MBP-⑀PKC-V1.
14E6 Competes with ⑀RACK for Binding to ⑀PKC-V1-Because the 14E6 binding site on ⑀PKC is within the V1 domain ( Fig. 1), the same domain that contributes significantly to the binding of ⑀PKC to full-length ⑀RACK (7), we next determined whether the 14E6 epitope overlaps with the ⑀RACK binding site. We used recombinant MBP-fusion proteins containing either the ⑀ V1 domain, MBP-⑀PKC-V1 (⑀PKC 2-145) or the C-terminal half of ⑀RACK, MBP-⑀RACK-C (␤Ј-COP 425-905), which contains the ⑀PKC binding site. Both 14E6 and the control anti-⑀PKC-V1 mAb, 6E10, bound to MBP-⑀PKC-V1 (Fig. 2, A and B, solid bar and Fig. 1), and did not bind to MBP-⑀RACK-C or to MBP (Fig. 2, A and B, striped bars). However, when ⑀PKC-V1 and ⑀RACK-C were preincubated before addition of 14E6, the binding of 14E6 to this complex ( Fig. 2A, open bar) was 70% lower than to ⑀PKC-V1 alone ( Fig.  2A, filled bar), suggesting that binding of ⑀RACK to ⑀PKC prevents the binding of 14E6 to the enzyme. In contrast, binding of 6E10 to ⑀PKC-V1 was not significantly decreased by concurrent ⑀RACK-C binding (Fig. 2B, open bar), making it unlikely that the inhibition of 14E6 binding to ⑀PKC by ⑀RACK was due to a nonspecific steric effect. Our data indicate that the epitope for 14E6 overlaps with or is close to an ⑀RACK binding site on ⑀PKC-V1 and predict that 14E6 will not bind to complexes of active ⑀PKC with ⑀RACK in vivo.
In Vitro Activation of ⑀PKC with PS and DG Increases Binding by 14E6 -PKC is activated in vivo through binding of lipid activators such as PS and DG. To determine whether 14E6 or 6E10 could distinguish between the active and inactive forms of ⑀PKC, recombinant ⑀PKC was activated by addition of PS and DG, and 14E6 or 6E10 binding was assessed by ELISA. Antibody binding to ⑀PKC is expressed as the ratio of binding in the presence of lipid activators to binding in the absence of lipid activators (Fig. 3). Binding of 14E6 was increased almost 3-fold by inclusion of lipid activators to 10 ng of ⑀PKC (Fig. 3). The concentration of enzyme appeared to be critical for optimal sensitivity in this assay. When 25 or 50 ng of ⑀PKC was used, binding of 14E6 was already high (data not shown) and was minimally affected by lipid activation of the enzyme (Fig. 3). This explains the binding of 14E6 to ⑀PKC in the absence of activators shown in Fig. 1, since the standard ELISA protocol used over 300 ng of protein per well. The increased binding to ⑀PKC in the presence of lipid activators when limiting amounts of ⑀PKC are used suggests that the 14E6 epitope on the enzyme becomes exposed upon activation of ⑀PKC. We also found that the binding of mAb 6E10 to activated ⑀PKC is slightly but significantly lower than binding to inactive ⑀PKC (Fig. 3). Taken together, our data suggest that mAb 14E6, but not 6E10, specifically recognizes the active form of PKC in an ELISA assay.
We next determined whether 14E6 recognizes the active form of ⑀PKC in solution using an immunoprecipitation assay. When a partially purified preparation of PKC from rat brain was incubated with 14E6, little ⑀PKC was immunoprecipitated (Fig. 4A). However, incubation of brain PKC with the activating lipids PS and DG for 3 or 6 min at room temperature resulted in significant immunoprecipitation of ⑀PKC by 14E6 (Fig. 4A). These activators increase the catalytic activity of calcium-independent PKC in the brain PKC preparation by over 10-fold as determined by phosphorylation of myelin basic protein (data not shown). In contrast, 6E10 immunoprecipitated more ⑀PKC in the absence of PKC activators (Fig. 4B) and commercial antiserum to the C terminus of ⑀PKC-V5 immunoprecipitated equal amounts of enzyme regardless of the presence of PKC activators (data not shown). The level of ␦PKC in the 14E6 immunoprecipitate was very low, indicating the specificity of 14E6 for ⑀PKC (Fig. 4A). In addition, SDS electrophoresis of the supernatants of the immunoprecipitates demonstrated that degradation of ⑀PKC did not occur during the experiment (Fig 4A, upper blot), indicating that the differences in immunoprecipitation observed in Fig. 4 reflect differences in epitope exposure in the various samples. Therefore, 14E6 is a unique antibody; it specifically recognizes an antigenic determinant on ⑀PKC that becomes exposed following activation by PS and DG.
The 14E6 Epitope Is Exposed Following Stimulation of NG108-15/D2 Cells-To determine whether 14E6 identifies activated ⑀PKC in cells, we used a rat/mouse neuroblastoma X glioma cell line stably expressing the dopamine D2 receptor, NG108-15/D2. These cells respond to the dopaminergic agonist NPA with a robust translocation of ⑀PKC (29). We compared staining with 14E6 and with commercial ⑀PKC monoclonal antibodies (raised against the last 15 amino acids in the V5 domain of ⑀PKC) as a function of time following cellular activation. Immunofluorescence confocal microscopy with the commercial anti-⑀PKC mAbs revealed an NPA-induced translocation from the nucleus and a narrow perinuclear region to a broader perinuclear/Golgi distribution as well as to cytosolic sites (Fig. 5A). In contrast, very little staining with 14E6 was observed in resting cells (Fig. 5B, left panel), but was induced by NPA treatment, yielding staining in broad perinuclear and Golgi regions (Fig. 5B); there was little cytosolic staining, however.
With the commercial anti-⑀PKC mAb, translocation was not obvious until 5 min after exposure to NPA, and persisted at 30 min. In contrast, 14E6 staining of NG108-15/D2 cells increased within 1 min after addition of NPA, was maximal by 10 min, and was low again by 30 min. The level of exposure of the 14E6 epitope in these cells is very low, as the signal from 14E6 (Fig.  5B) was amplified 10 times compared with the signal from the commercial antibody (Fig. 5A). Notably, 14E6 staining is not observed in the nucleus either before or after NPA is added, in contrast to staining with the commercial anti-⑀PKC antibody. These results suggest that 14E6 does not recognize inactive ⑀PKC in the nucleus of NG108-15/D2 cells. Translocation of ⑀PKC, as determined by the commercial antibodies, persisted at 30 min (Fig. 5A), when activated enzyme was no longer detected by 14E6 (Fig. 5B), suggesting that exposure of the epitope for 14E6 is more transient than is localization to sites usually associated with activated enzyme. Further, 30 min after treatment of cells, the commercial anti-⑀PKC antibody (Fig. 5A), but not 14E6 (Fig. 5B), stained the cell cytosol peripherally.
We also carried out immunoprecipitation experiments on the solubilized particulate fraction of lysates from NG108-15/D2 cells that were incubated with NPA over the same time course as the immunofluorescence studies. Western blots of immunoprecipitation studies with monoclonal anti-⑀PKC indicate that the total amount of ⑀PKC in the particulate fraction increases with time of incubation with NPA, as expected (Fig. 5C). In contrast, 14E6 only immunoprecipitates a protein recognized by antibodies to ⑀PKC at the 1 min time point (Fig. 5D), supporting the transient nature of the 14E6 epitope indicated in the immunofluorescence studies (Fig. 5B). The exact time of appearance of the 14E6 epitope in the immunofluorescence and immunoprecipitation studies cannot be compared since the cells are fixed immediately for the former and must be lysed and fractionated after the indicated time of incubation with NPA before immunoprecipitation can be carried out. Nevertheless, the immunoprecipitation studies support the conclusion of the immunofluorescence studies that the 14E6 epitope only occurs transiently after activation of the D2 receptor.  4. 14E6 binding to ⑀PKC in vitro increases following activation of the enzyme with PS and DG. A, partially purified PKC from rat brain (ϳ100 units/mg) was incubated at room temperature for 3 or 6 min with or without PS and DG (upper blot) and immunoprecipitation by 14E6 was carried out as described under "Experimental Procedures." The immunoprecipitated material was analyzed for ⑀PKC and ␦PKC by Western blot (lower blots). The molecular weight of ␦PKC was confirmed using the brain extract (data not shown). B, partially purified PKC was incubated at room temperature for 5 min with or without PS and DG, immunoprecipitated by either 14E6 or 6E10, and analyzed for ⑀PKC by Western blot.
Both the immunofluorescence and immunoprecipitation data indicate that the anti-⑀PKC antibodies recognize activated ⑀PKC at 30 min. However, there is no 14E6 staining at this time. One possible interpretation of this data is that ⑀PKC is active but bound to ⑀RACK and therefore cannot bind to 14E6 (Fig. 2). Therefore we incubated the blots from the immunoprecipitation experiments with antibodies to ⑀RACK. We found that ⑀RACK is co-immunoprecipitated by the anti-⑀PKC antibody (Fig. 5C), but not by 14E6 (Fig. 5D). Taken together, our data suggest that the 14E6 epitope becomes inaccessible when ⑀PKC is bound to ⑀RACK in cells. To further investigate this possibility, we used neonatal cardiac myocytes, where the localization of activated ⑀PKC is very characteristic and more easily discernable (3,4).
Activation of Cardiac Myocytes with Phorbol Ester Results in Induction of the Epitope for 14E6 -Our previous studies in resting cardiac myocytes using commercially available anti-⑀PKC antibodies raised against the V5 region of ⑀PKC localized ⑀PKC to the nucleus, with only some cells showing ⑀PKC at the perinucleus and in cross-striated structures in the cell body (3,4). After activation, ⑀PKC is localized to cross-striated structures and the perinucleus in most cells. This suggests that inactive ⑀PKC is found in the nucleus and activated ⑀PKC in the perinucleus and cross-striated structures (4). The localization of ⑀PKC in resting and activated cardiac myocytes was compared using commercial polyclonal anti-⑀PKC antiserum and the ⑀PKC-V1 antibody, 14E6. In agreement with our published observations, when cells were stained with the commercial anti-⑀PKC antibodies, the ratio of perinuclear and cell body staining to nuclear staining increased upon activation of the cells with PMA (Fig. 6A, top panels). In contrast, staining with 14E6 was always non-nuclear, consisting of perinuclear, punctate, and cross-striated patterns in the cell body (Fig. 6A, middle panels). In addition, there was an increase in the intensity of 14E6 immunofluorescence staining in PMA-treated cells compared with that in unstimulated cells (left versus right middle panels, Fig. 6A). 14E6 staining was not seen when the FIG. 6. 14E6 stains cardiac myocytes in an activation-specific pattern. A, cardiac myocytes were untreated (control) or treated with 100 nM PMA for 10 min, fixed, and immunofluorescence localization of ⑀PKC determined with either a polyclonal anti-⑀PKC V5 antibody or with 14E6. In lower panel, 1 mg/ml MBP-⑀PKC-V1 was added with 14E6 to block specific staining. B, cardiac myocytes were treated with 1 nM PMA or 1 nM PMA plus the ⑀PKC agonist ⑀RACK for 1 min.
FIG. 5. The epitope for 14E6 is induced by NPA treatment of NG-108/D2 cells. NG108 -15/D2 cells were treated with 50 nM NPA for the indicated times. Cells were fixed and stained with a commercial monoclonal anti-⑀PKC-V5 antibody (A) or 14E6 (B) and analyzed by confocal microscopy. Laser power was 10fold higher for B. The false color images represent staining intensity (black, green, yellow, orange, in order of increasing intensity). C, Triton-soluble lysates were prepared from NG108 -15/D2 cells treated with 50 nM NPA for the indicated times and immunoprecipitated using monoclonal antibody for ⑀PKC. D, Triton-soluble lysates were prepared as in C and the 14E6 monoclonal antibody was used for immunoprecipitation. Both C and D were probed for ⑀PKC using a commercial rabbit polyclonal antibody against ⑀PKC-V5 and a monoclonal antibody against ⑀RACK.
antibody was pre-incubated with 1 mg/ml MBP-⑀PKC-V1 (Fig.  6A, lower panels), indicating again that 14E6 is specific for the V1 domain of ⑀PKC. Both the extranuclear localization of the 14E6 epitope and the increase in staining following activation of PKC suggest that 14E6 is specific for activated ⑀PKC. Moreover, as was observed in NG108-15/D2 cells stimulated with a dopaminergic agonist, translocation of ⑀PKC observed with 14E6 could be detected in myocytes before that seen with the polyclonal antibody and it was more transient (not shown). Taken together, these data suggest that 14E6 is selective for activated ⑀PKC, recognizing a transient form of the activated enzyme.
We recently identified a peptide activator of ⑀PKC. This peptide, ⑀RACK, is thought to disrupt intramolecular interaction within PKC and thus expose both the RACK binding site and the catalytic site, rendering the enzyme active (9,30). Addition of this peptide to cardiac myocytes causes selective ⑀PKC translocation and function (9). In addition, it increases the function of ⑀PKC in the presence of suboptimal levels of PMA. We predicted that if 14E6 recognizes activated ⑀PKC, we should see an increase in staining when cells are treated with ⑀RACK. As seen in Fig. 6B, cardiac myocytes treated for 1 min with 1 nM PMA had immunostaining levels with 14E6 that were not different from those in control-treated cells (compare with Fig. 6A, middle left panel). However, in the presence of ⑀RACK and 1 nM PMA, immunostaining with 14E6 was similar to that seen with fully activated ⑀PKC (compare Fig. 6A, middle right panel and 6B, right panel.) These data support our hypothesis that 14E6 specifically recognizes the active state of ⑀PKC.
14E6 Recognizes Active but Not RACK-associated ⑀PKC in Cells-The ELISA data shown in Fig. 2 and the co-immunoprecipitation experiments shown in Fig. 5D support our hypothesis that following cellular activation, binding of ⑀PKC to ⑀RACK in cells prevents binding of 14E6 to ⑀PKC, thus leading to the transient appearance of the 14E6 epitope observed in NG108 -15/D2 cells (Fig. 5). To determine whether ⑀RACK binding to ⑀PKC precludes 14E6 binding to ⑀PKC in cardiac myocytes, we used confocal microscopy to assess co-localization of ⑀PKC and ⑀RACK in cardiac myocytes. Because both 14E6 and the only anti-⑀RACK antibodies available are mouse IgM mAbs, we first examined simultaneous staining of ⑀PKC with a rabbit polyclonal antibody (Fig. 7A) and anti-⑀RACK mAbs (Fig. 7B). In resting cells, there is very little overlap between ⑀RACK and ⑀PKC as stained with the polyclonal ⑀PKC antibodies (data not shown; see also Refs. 5,7,10). However, after activation most of the ⑀PKC co-localizes with ⑀RACK (Fig. 7C,  yellow). After even brief and mild activation, only small areas of unique staining for polyclonal anti-⑀PKC (red staining) remain. This implies that most of the ⑀PKC stained by the polyclonal antibody is also bound to ⑀RACK. If binding of activated ⑀PKC to ⑀ RACK precludes 14E6 binding, then the 14E6 epitope should not co-localize with polyclonal anti-⑀PKC staining in activated cells. Therefore, we next examined simultaneous staining of ⑀PKC using the polyclonal anti-⑀PKC (Fig. 8, left, red) and 14E6 (Fig. 8, right, green). After activation, although both antibodies indicated translocation of ⑀PKC to cross-striated structures (see also Fig. 5), we observed very few areas where the cross-striated staining by the two antibodies merged (Fig. 8, merged). In most areas, there was alternate green-red staining. These data are consistent with the existence of at least two populations of activated ⑀PKC in the cross striations of cardiac myocytes. One is ⑀PKC that is stained by the polyclonal anti-⑀PKC antibody, which is co-localized with ⑀RACK and that makes up the overwhelming majority of ⑀PKC. The second, is a transient and small population of ⑀PKC that is stained by 14E6. We propose therefore that the commercial polyclonal antibodies stain the cross-striated structures by binding to activated ⑀PKC bound to its RACK; 14E6 recognizes a transient, activated, but non-anchored state of ⑀PKC that has not yet reached the site of ⑀RACK.

DISCUSSION
Monoclonal antibody 14E6, raised against the V1 domain of ⑀PKC, appears to recognize an epitope exposed only after activation of the enzyme. First, binding of 14E6 to ⑀PKC in ELISA or immunoprecipitation experiments is increased in the presence of lipid PKC activators when using limiting amounts of the enzyme (Figs. 3 and 4). Second, immunofluorescence localization of ⑀PKC indicates that the 14E6 epitope is induced upon activation of cardiac myocytes by PMA, by the selective ⑀PKC activator peptide, ⑀RACK (Fig. 6), or by stimulation with norepinephrine (data not shown); in NG108 -15/D2 cells, staining with 14E6 is only observed after activation of the D2 receptor. Third, in cardiac myocytes (Figs. 6 and 8) and in NG108 -15 cells (Fig. 5), 14E6 does not stain subcellular compartments where inactivate ⑀PKC is localized. In cardiac myocytes, for example, activation with PMA leads to an increased Cardiac myocytes were treated with 3 nM PMA for 1 min. Immunostaining was carried out with anti-⑀PKC-V5 polyclonal antibodies (red, left) and with 14E6 (green, right). Merge of the images (middle panel) suggests lack of co-localization of staining with the two antibodies as indicated by the distinct alternate red and green staining (arrows) and the almost complete absence of yellow in these panels. (Note that the intensity of staining and the ability to merge two images to give a yellow image if they are co-localized depends on the intensity of each staining; it is therefore limited to detect overlap in localization of the majority of each staining). ratio of extranuclear to nuclear ⑀PKC (Fig. 6), supporting our earlier observation (4) that nuclei contain a pool of inactive ⑀PKC that translocates to extranuclear structures upon activation. 14E6 staining of PMA-treated cardiac myocytes is exclusively non-nuclear, localized to sites near those where activated ⑀PKC is found (4), including a perinuclear ring, a punctate pattern in the cell body, and in cross-striated structures. Importantly, 14E6 only stains the activated but nonanchored enzyme, since staining by 14E6 of activated ⑀PKC does not overlap with staining of ⑀RACK where activated ⑀PKC is anchored (Fig. 8).
The V1 domain of ⑀PKC was used as the immunogen for raising activation-specific antibodies because it contains an activation -specific binding site for ⑀ RACK (7,22). The activation-specific epitope for 14E6 on ⑀PKC appears to be within or near the binding site for ⑀ RACK (Fig. 2), suggesting that 14E6 only recognizes activated but non-anchored ⑀PKC. Indeed, the data in Figs. 7 and 8 show that 14E6 stains cross-striations distinct from those stained by both an antibody to ⑀ RACK and a polyclonal anti-⑀PKC antibody, supporting this conclusion. Activation by DG, ⑀RACK, or PMA causes ⑀PKC to unfold into an active conformation (30) that is recognized by 14E6. We propose that this conformational change unmasks the ⑀ RACK binding site in the V1 domain, leading to subsequent binding of the enzyme to ⑀ RACK, which in turn prevents the binding of 14E6.
Our data suggest a model of activation of ⑀PKC where the enzyme has at least three states of activation: State I, inactive ⑀PKC; State II, a lipid-activated transient state of ⑀PKC to which 14E6 can bind; and State III, active, RACK-associated ⑀PKC to which 14E6 cannot bind (Fig. 9). This model agrees with the time course of immunofluorescence and immunoprecipitation studies in NG108-15/D2 cells (Fig. 5). In the immunofluorescence experiments, the 14E6 epitope appeared within 1 min following treatment with NPA, before translocation of the majority of the ⑀PKC from distinct perinuclear localization to a broader perinuclear/Golgi and cytosolic localization (Fig. 5,  A and B; step 1, Fig. 9). Thirty minutes after treatment with NPA, the 14E6 epitope disappeared (Fig. 5B; step II, Fig. 9), whereas commercial anti-⑀PKC antibodies showed staining at sites where ⑀PKC is bound to ⑀RACK. 2 The absence of 14E6 staining appears to be due to masking of the 14E6 epitope by interaction with ⑀ RACK (Fig. 2). The transient nature of the 14E6 epitope was confirmed by immunoprecipitation studies of the solubilized particulate fraction of NG108-15 cells; 14E6 staining was only observed 1 min after activation of the D2 receptor. Moreover, at no time point was ⑀RACK co-immunoprecipitated with 14E6 antibodies (Fig. 5D). In cardiac myocytes, where the localization of ⑀PKC is more distinct, 14E6 and ⑀RACK do not appear to be co-localized, since 14E6 yields cross-striated staining similar to but distinct from the localization of ⑀RACK or the ⑀PKC-⑀RACK complex (Figs. 7 and 8). An elegant study of Walker and co-workers (22) using kinetic analysis of ⑀PKC and ⑀PKC-V1 translocation in cardiac myocytes demonstrated that ⑀PKC translocation is controlled by the rate of intramolecular conformational changes within the V1 domain. They also suggest that in vivo anchoring of PKC may involve sequential activation events due to binding via the C1 domain, the phorbol ester-and DG binding domain and binding via the V1 domain, the RACK binding domain (22). A recent study from our laboratory using real-time imaging of GFP-tagged ⑀PKC supports this conclusion (31).
In conclusion, 14E6, raised against the V1 domain of ⑀PKC, appears to recognize a transient epitope formed upon activation of purified ⑀PKC by lipid cofactors; this site is masked once the enzyme binds to its RACK. Therefore, this highly specific antibody has allowed us to identify an intermediate stage of ⑀PKC activation, which results from lipid activation, and which is translocated from its site in non-activated cells, but has not yet reached its final localization site bound to ⑀RACK. These data suggest that lipid binding precedes binding of the activated enzyme to ⑀RACK, since 14E6 recognizes lipid-activated ⑀PKC but not activated enzyme anchored to ⑀RACK. Because there is little staining of 14E6 in non-stimulated cells, this antibody should be a useful diagnostic antibody to study acute activation of ⑀PKC in tissues and in vivo. FIG. 9. A model of ⑀PKC activation. Shown are three stages of PKC activation: cytosolic inactive PKC (State I, red) anchors to membranes on elevation of DG (State II, yellow). This transient state (II) is selectively detected by 14E6 and is induced by PKC binding to DG in the presence of lipids (bold in lipid scheme). The activated enzyme then binds to ⑀RACK, but is not detected by 14E6 (green, state III). This last state represents the active stable form of PKC. On activation (step 1), PKC translocates from the cytosol to the membrane. This activation state (detected by 14E6) is transient and then this lipid-bound PKC binds to its RACK (step 2). Finally, by an as yet unknown mechanism, PKC detaches from its RACK and returns to the inactive state in the cytosol (step 3).