Binding Specificity for RACK1 Resides in the V5 Region of βII Protein Kinase C

Identification of selective anchoring proteins responsible for specialized localization of specific signaling proteins has led to the identification of new inhibitors of signal transduction, inhibitors of anchoring protein-ligand interactions. RACK1, the firstreceptor for activated C kinase identified in our lab, is a selective anchoring protein for βII protein kinase C (βIIPKC). We previously found that at least part of the RACK1-binding site resides in the C2 domain of βIIPKC (Ron, D., Luo, J., and Mochly-Rosen, D. (1995) J. Biol. Chem.270, 24180–24187). Here we show that the V5 domain also contains part of the RACK1-binding site in βIIPKC. In neonatal rat cardiac myocytes, the βIIV5-3 peptide (amino acids 645–650 in βIIPKC) selectively inhibited phorbol 12-myristate 13-acetate (PMA)-induced translocation of βIIPKC and not βIPKC. In addition, the βIIV5-3 peptide inhibited cardiac myocyte hypertrophy in PMA-treated cells. Interestingly, βIV5-3 (646–651 in βIPKC), a selective translocation inhibitor of βIPKC, also inhibited PMA-induced cardiac myocyte hypertrophy, demonstrating that both βI- and βIIPKC are essential for this cardiac function. Therefore, the βIIV5 domain contains part of the RACK1-binding site in βIIPKC; a peptide corresponding to this site is a selective inhibitor of βIIPKC and, hence, enables the identification of βIIPKC-selective functions.

The localization of signaling enzymes within cells is highly specific and often regulated by selective anchoring proteins (1,2). A number of these proteins have recently been identified; some anchor and coordinate multiple enzymes in the same signaling cascade (3,4) and can bind to their selective proteins or enzymes depending on their activation state (2). Selective localization of signaling enzymes in cells results in tethering them in the proper subcellular location for their function. Disruption of the selective protein-protein interactions between the signaling enzymes and their anchoring proteins alters the specialized localization of the signaling enzymes and thus disrupts their function (5).
We have studied the mechanism leading to selective localization of protein kinase C (PKC). 1 PKC isozymes are a family of serine/threonine, phospholipid-dependent protein kinases (6) that translocate after stimulation to select subcellular sites where they bind their corresponding selective anchoring proteins, RACKs (receptor for activated C kinase) (2). RACKs bind only the active form of their respective PKCs. Our lab has identified some of the RACK-binding sites on ␤, ⑀, and ␦PKC and demonstrated that RACK binding is essential for both proper localization and function of these PKC isozymes (5). So far we have cloned and characterized two RACKs and demonstrated that RACK1 is selective for ␤IIPKC (7,8), whereas RACK2, also known as ␤ЈCOP (a coatomer protein involved in vesicle transport) is selective for ⑀PKC (9).
␤Iand ␤IIPKC, members of the classical family of PKCs, are differentially spliced products of the same gene and therefore differ only in their C-terminal variable domain, the V5 domain (10,11). Immunofluorescence studies demonstrate that ␤Iand ␤IIPKC are differentially localized in both their inactive and active states (12,13) in a number of cell types. We have demonstrated that the second conserved domain in ␤PKC, the C2 domain, contains part of the RACK-binding site in ␤PKC (7,14). The ␤C2 domain binds RACK1 in vitro and peptides derived from this domain inhibit this interaction (7). In addition, the C2-derived peptides block translocation of both ␤Iand ␤IIPKC in cells (7). However, the C2 domains of ␤Iand ␤IIPKC are identical and, therefore, cannot account for the differential localization of ␤Iand ␤IIPKC. We hypothesize that the distinct sequences in the ␤IV5 and ␤IIV5 domains should confer the RACK-binding specificity and differential localization of these isozymes. A selective RACK for ␤IPKC has yet to be identified. We therefore used RACK1, the selective anchoring protein for ␤IIPKC, to test our hypothesis.
Using short peptides derived from the ␤IIV5 domain, we show here that unique sequences within ␤IIPKC contain part of the RACK1-binding site. In addition, we show that one of the V5-derived peptides functions as an isozyme-selective translocation inhibitor of ␤IIPKC in neonatal rat cardiac myocytes. This peptide was used to demonstrate that ␤IIPKC mediates phorbol 12-myristate 13-acetate (PMA)-induced cardiac myocyte hypertrophy. Of interest, a peptide-selective translocation inhibitor of ␤IPKC identified in this study also inhibited PMAinduced myocyte hypertrophy, suggesting that the two isozymes are required for this function.

EXPERIMENTAL PROCEDURES
Materials-PMA was purchased from LC Laboratories. Diacylglycerol and phosphatidylserine were purchased from Avanti. Luminol, p-coumaric acid, IGEPAL detergent, saponin, and Triton X-100 were purchased from Sigma. Polyclonal anti-␤I-PKC and anti-␤IIPKC antibodies were purchased from Santa Cruz Biotechnologies, and R&D Antibodies. Amylose resin was purchased from New England Biolabs. Monoclonal anti-␤PKC antibodies were purchased from Seikagaku, Inc. * This work was supported in part by National Institutes of Health Grant HL52141 (to D. M.-R.). In addition, this work was performed during the tenure of a research fellowship from the American Heart Association, Western States Affiliate (to E. G. S.). 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.: 650-725-7720; Fax: 650-723-2253; mochly@stanford.edu. 1 The abbreviations used are: PKC, protein kinase C; PMA, phorbol and Transduction Laboratories. Anti-RACK1 antibodies were purchased from Transduction Laboratories. The secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit and HRP-conjugated goat anti-mouse antibodies, glutathione Sepharose 4B beads, and [ 14 C]phenylalanine were purchased from Amersham Pharmacia Biotech. Recombinant ␤Iand ␤IIPKC were purchased from PanVera. Protein Expression and Purification-Recombinant ␤IIPKC was purchased from PanVera, or a clone (received from Alexandra Newton) was expressed in Tn5 insect cells and partially purified to homogeneity as previously described (15). Fragments of the ␤PKC C2 (amino acids 175-289), ␤IV5 (amino acids 622-671), and ␤IIV5 (amino acids 622-673) domains were expressed in Escherichia coli as fusion proteins with maltose-binding protein (MBP) using the pMAL-c2 expression vector (New England Biolabs). Bacterial pellets were resuspended in amylose column buffer (10 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, 10 mM ␤-mercaptoethanol) and lysed by sonication. Fusion proteins were purified by immobilization on an amylose resin column. The column was washed with 8 column volumes of column buffer, and bound protein was eluted in 10 mM maltose in column buffer. Protein concentration was determined by Bradford assay.
RACK1 was expressed in bacteria as a fusion protein with glutathione S-transferase (GST) using the pGEX-4T-1 GST gene fusion expression vector (Amersham Pharmacia Biotech). Bacterial pellets were resuspended in STE buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA) and lysed by sonication. Triton X-100 was added to the lysate to a final concentration of 1%, and the mixture was incubated on ice for 30 min with occasional mixing. The lysate was then centrifuged at 12,000 ϫ g for 15 min, and the supernatant was stored in 50% glycerol at Ϫ20°C.
Binding Assay-Recombinant GST or GST-RACK1 bacterial lysate was incubated with 25 l (ϳ20 l packed bead volume) of pre-equilibrated glutathione-Sepharose 4B beads, and the beads were washed with overlay wash (200 mM NaCl, 50 mM Tris-HCl, pH 7.5, 0.1% polyethylene glycol, 12 mM ␤-mercaptoethanol). For competition experiments the complex was preincubated with the PKC-derived peptides (10 M) or protein fragments (1-2.5 M) in overlay buffer (200 mM NaCl, 50 mM Tris-HCl, pH 7.5, 0.1% polyethylene glycol, 12 mM ␤-mercaptoethanol, 0.1% bovine serum albumin, 20 g/ml leupeptin, 20 g/ml soybean trypsin inhibitor, 20 g/ml aprotinin, and 10 g/ml phenylmethylsulfonyl fluoride) for 15 min before the addition of PKC and PKC activators. The complex was then incubated with or without PKC or PKC fragments in the presence or absence of PKC activators (2 g/ml diacylglycerol and 60 g/ml phosphatidylserine and 1 mM CaCl 2 ) for 15 min at room temperature in overlay buffer. The beads were washed three times with overlay wash containing 1% IGEPAL detergent, and the third wash was used to transfer the beads to fresh Eppendorf tubes to help decrease background. Bound proteins were eluted in sample buffer, followed by SDS-polyacrylamide gel electrophoresis and Western analysis. ␤PKC-selective antibodies were used to detect ␤PKC holoenzyme and C2 and V5 fragments.
Isolation and Permeabilization of Neonatal Rat Cardiac Myocytes-Cardiac myocytes were isolated from 4 litters of 1-day-old Sprague-Dawley rats (each with 8 -10 animals) as described previously (16). Cells were plated in 12-well plates for [ 14 C]phenylalanine incorporation experiments or laminin-coated 8-well chamber slides for immunofluorescence studies. Cells were maintained in media M199 (Life Technologies, Inc.) with 10% serum after plating. For [ 14 C]phenylalanine incorporation experiments, cells were transferred to serum-free media on day 3, and experiments were initiated on day 4. Immunofluorescence experiments were performed on days after a change to serum-free media on day 4. All peptides were delivered into cells via transient permeabilization with the detergent saponin (50 g/ml) as described previously (16).
Immunofluorescence-Cardiac myocytes were permeabilized in the presence or absence of peptide and then treated with 4␣-or 4␤-PMA. Cells were washed with phosphate-buffered saline and fixed with icecold acetone:methanol (1:1). Cells were then washed with phosphatebuffered saline followed by two brief washes with water. Vectashield (Vector Laboratories) mounting media was used for mounting of a coverslip. Cells were scored using a Zeiss fluorescence microscope. In immunofluorescence studies, distinct changes in ␤Iand ␤IIPKC subcellular localization are apparent after activation (13). Active ␤IPKC translocates from the cytosol into the nucleus, whereas active ␤IIPKC is found in both the perinuclear region and the cell periphery but not on fibrillar structures, where it is found before activation. We confirmed that scoring translocation in this manner is not subjective; when the experimentor was blinded to the identity of the treatment, the same results were obtained. Moreover, in several previous studies, we com-pared the quantitation method described above using immunofluorescence to others (i.e. Western blot analysis of cell fractions), and the same results were obtained (e.g. Ref. 17).
[ 14 C]Phenylalanine Incorporation-Isolated cardiac myocytes plated in 12-well plates were permeabilized in the presence or absence of peptide. Cells were transferred to serum-free media M199 containing 0.15 Ci/ml [ 14 C]phenylalanine (Amersham Pharmacia Biotech), treated with 10 nM 4␤-PMA or the inactive analog 4␣-PMA, and incubated for 48 h at 37°C. Cells were then harvested as described (16). Briefly, media was aspirated, cells were washed three times with phosphate-buffered saline, and protein was precipitated with ice-cold 10% trichloroacetic acid for at least 1 h at 4°C. Trichloroacetic acid was aspirated, and wells were washed three times with ice-cold 10% trichloroacetic acid. Precipitated protein was solubilized in 1% SDS for 2 h at 37°C. Solubilized protein was mixed with Universol TM scintillation fluid (ICN) and counted with a scintillation counter (Beckman Instruments). Phase contrast pictures were obtained using a Zeiss light microscope.

FIG. 1. Dose-dependent binding of ␤IIPKC to RACK1 in vitro.
A, RACK1, bacterially expressed as a fusion protein with GST (GST-RACK1), was immobilized on glutathione-Sepharose-4B beads. The complex was incubated with purified ␤IIPKC in the presence (Active PKC) or absence (Inactive PKC) of PKC activators. Bound protein was eluted followed by SDS-polyacrylamide gel electrophoresis and Western blot analysis. ␤IIPKC was detected with a ␤IIPKC isozyme-selective antibody. A representative Western blot is shown. B, quantitative results from four independent experiments. C, selective binding of ␤IIPKC to RACK1 in vitro. Recombinant ␤IPKC or ␤IIPKC was incubated with immobilized GST (lane 3) or GST-RACK1 (lane 4) in the presence of PKC activators in vitro. Bound protein was detected by Western analysis using an antibody against the regulatory domain of ␤PKC that recognizes ␤Iand ␤IIPKC equally well (lanes 3 and 4). Lanes 1 and 2 contain 5 and 10 ng of ␤Iand ␤IIPKC used as standards. A representative of three Western blots is shown.
precipitate. 2 Here, we determined the binding affinity of ␤IIPKC for RACK1 in vitro. RACK1, expressed as a fusion protein with GST, was immobilized on glutathione-conjugated Sepharose beads and incubated with increasing concentrations of recombinant ␤IIPKC in the presence of activators. Binding of activated ␤IIPKC to RACK1 is both dose-dependent and saturable with a half-maximal binding of 3 Ϯ 2 nM (n ϭ 4; Fig. 1, A and B). Furthermore, at all concentrations, binding of ␤IIPKC to RACK1 is at least 2-fold greater than that of ␤IPKC (n ϭ 3; Fig. 1C).
We previously demonstrated that the C2 domain of ␤PKC (␤C2), identical in the two ␤PKC isozymes, contains part of the RACK1-binding site in ␤PKC (7,14). Here, in vitro binding studies show that the half-maximal binding of the ␤C2 domain-MBP fusion protein to RACK1 is ϳ500 nM ( Fig. 2A). Since RACK1 is selective for ␤IIPKC (7, 8) and its subcellular localization overlaps that of ␤IIPKC and not ␤IPKC (7, 13), we reasoned that the unique sequences in the ␤IIV5 domain should confer specificity of ␤IIPKC for RACK1. This suggests that the distinct ␤IIV5 domain may bind RACK1 directly. We incubated the recombinant ␤IIV5 domain, expressed as an MBP fusion protein, with immobilized GST-RACK1 in vitro, as described under "Experimental Procedures" and found that ␤IIV5 binding to RACK1 was dose-dependent and saturable with a half-maximal binding of ϳ400 nM (Fig. 2B). This affinity is similar to that of the ␤C2 domain, having a half-maximal binding of ϳ500 nM ( Fig. 2A). Therefore, the ␤IIV5 domain also contains part of the RACK1-binding site in ␤IIPKC.
␤C2 and ␤V5 Domains Compete with ␤IIPKC for RACK1 Binding-If part of the RACK1-binding site in ␤IIPKC is within the ␤IIV5 domain, then ␤IIV5 should inhibit ␤IIPKC binding to RACK1. Furthermore, if both C2 and V5 domains are required for ␤IIPKC binding to RACK1, an additive inhibitory effect may be seen when combining the ␤IIV5 domain along with the ␤C2 domain. To this end, recombinant ␤IIV5, ␤IV5, and/or ␤C2 fusion proteins were preincubated with immobilized GST-RACK1, and then full-length ␤IIPKC was added in the presence of PKC activators. We found that the ␤IIV5 domain competed with ␤IIPKC for RACK1 binding. Furthermore, an additive effect in competition for ␤IIPKC binding to RACK1 was observed in the presence of both ␤C2 and the ␤IIV5 domains (Fig. 3). However, similar results were obtained when using the ␤IV5 domain, both alone or in combination with the ␤C2 domain.
␤V5and ␤C2-derived Peptides Compete with ␤IIPKC-RACK1 Binding-The non-selective inhibition of ␤IIPKC binding to RACK1 with both ␤IV5 and ␤IIV5 fragments (Fig. 3) was somewhat surprising, as ␤IPKC and ␤IIPKC display differences in binding to RACK1 (Fig. 1C), and active ␤IPKC and RACK1 is not co-localized in cells (7,13). We hypothesized that the inhibitory effect of the ␤IV5 fragment on ␤IIPKC binding to RACK1 is due to some interactions with RACK1 or ␤IIPKC holoenzyme via conserved sequences within the ␤Iand ␤IIV5 domains.
Although ␤Iand ␤IIPKC differ in the V5 domain, they display high homology (ϳ60%) within that domain (Fig. 4A). Therefore, we expect that the least similar sequences within the ␤IIV5 domain should confer ␤IIPKC RACK1-binding specificity. We synthesized short peptides corresponding to the least similar sequences in the ␤V5 domains, since we expected that they would contain the selective RACK1-binding sequences in ␤IIPKC. Three peptides corresponding to unique regions were selected from each of the ␤I and ␤II V5 domains:

V5 of ␤IIPKC Contains the RACK1-binding Site
␤IV5-1 and ␤IIV5-1 comprise part of the antigenic peptides used for production of many of the commercially available anti-␤I and -␤II PKC isozyme-specific antibodies).
Isozyme-selective Translocation Inhibitors of ␤I and ␤II PKC in Cardiac Myocytes-In cells it is thought that RACKs act as isozyme-selective anchoring proteins, functioning to tether specific PKC isozymes nearby their respective substrates (2,5). We previously showed that disruption of intracellular PKC-RACK interactions inhibits PKC translocation and proper subcellular localization, therefore preventing PKC substrate phosphorylation and blocking downstream function (2,5,18).
We show here that the combination of ␤C2and ␤IIV5derived peptides are necessary to inhibit ␤IIPKC-RACK1 interactions in vitro (Fig. 4). However, we previously demonstrated that each of the peptides derived from the RACK1-binding site in the ␤C2 domain (␤C2-1, ␤C2-2, and ␤C2-4) is sufficient alone to inhibit translocation of both ␤Iand ␤IIPKC in neonatal rat cardiac myocytes (7). To determine if individual peptides derived from the ␤IIV5 domain can also act alone as translocation inhibitors, we first determined their effects on ␤IIV5 fragment binding to RACK1 in vitro. We found that each of the ␤IIV5-1, ␤IIV5-2, and ␤IIV5-3 peptides (10 M) alone inhibits binding of MBP-␤IIV5 fusion protein (500 nM) to RACK1 in vitro by 37, 30, and 34%, respectively (average of 3 measurements), suggesting that each peptide contains part of the RACK1-binding site in the ␤IIV5 domain and, therefore, may be an effective inhibitor of translocation in cells. Consequently, we set out to test the effects of the individual V5-derived peptides on ␤IIPKC translocation in cells. Since none of the peptides stood out as the overall strongest inhibitor of ␤IIV5-RACK1 binding in vitro, we chose to start our in-cell studies with the ␤IIV5-3 peptide. We propose that the ␤IIV5-3 peptide, containing part of the RACK1-selective binding sequence in ␤IIPKC, may function as isozyme-selective translocation inhibitor by binding to the isozyme-selective RACK and inhibiting translocation of ␤IIPKC isozyme. The ␤IV5-3 peptide was used both as a control for ␤IIV5-3 and a possible selective inhibitor of ␤IPKC translocation.
Neonatal rat cardiac myocytes were permeabilized in the presence or absence of 10 M peptide and then treated with or without 10 nM PMA for 5 min (we showed previously that ϳ10% of the applied peptide is internalized by the cells (16)). Cells were then fixed and stained with isozyme-selective anti-␤I-and anti-␤IIPKC-selective antibodies followed by a fluorescein isothiocyanate-conjugated secondary antibody, as described under "Experimental Procedures." In norepinephrine-or PMAtreated neonatal rat cardiac myocytes, active ␤IPKC localizes in the nucleus of the cell, whereas active ␤IIPKC translocates to both the perinuclear region and the cell periphery upon activation (13). In these experiments, cells were scored for the number of cells staining for active ␤IPKC or ␤IIPKC at their respective sites in the cell, as previously described (7). PMAtreated cells permeabilized in the absence of peptide showed 89% Ϯ 4 of the cells staining ␤IIPKC at the perinuclear region and cell periphery (Fig. 5). Whereas there was no change in ␤IIPKC translocation in cells treated with the ␤IV5-3 peptide followed by PMA (77% Ϯ 6 versus 89% Ϯ 4 of ␤IV5-3 and control, respectively, Fig. 5), treatment with ␤IIV5-3 reduced ␤IIPKC translocation to 17% Ϯ 5 of the cells (Fig. 5). Therefore, ␤IIV5-3 selectively inhibited ␤IIPKC translocation, whereas ␤IV5-3 had no effect. Additionally, ␤IIV5-3 peptide had no effect on ␤IPKC translocation with 85% Ϯ 3 of the cells dis- FIG. 4. In vitro selectivity of peptides derived from the ␤Iand ␤IIV5 domains. A, the sequence of ␤Iand ␤IIV5 domains are shown (single letter amino acid code). Lines between the sequences denote similarity. Black bars above and below the sequences mark the sequences of the peptides synthesized from the ␤IV5 and ␤IIV5 domains. B, a combination of ␤C2and ␤IIV5-derived peptides inhibit ␤IIPKC binding to RACK1. GST-RACK1 was immobilized on glutathione-Sepharose-4B beads and preincubated with or without a mixture of three peptides from the ␤C2 domain (␤C2-1, ␤C2-2, and ␤C2-4, 10 M of each) in the presence or absence of peptides from the ␤IV5 domain (␤IV5-1, ␤IV5-2, and ␤IV5zϪ3, 10 M of each) or the ␤IIV5 domain (␤IIV5-1, ␤IIV5-2, and ␤IIV5-3, 10 M of each). The complex was then incubated with PKC activators and recombinant ␤IIPKC (5 nM). Bound protein was detected by Western analysis using anti-␤IIPKC antibodies. C, average results are presented as the percent of ␤IIPKC bound (n ϭ 4; *, S.E., p Ͻ 0.0001).
We proposed that the ␤IIV5-derived translocation inhibitor, ␤IIV5-3, determines ␤IIPKC-selective functions in primary cultures of neonatal rat cardiac myocytes. ␤PKC has recently been reported to mediate cardiac hypertrophy (22), a normal process occurring during development as well as a compensatory mechanism after an insult to the adult heart (23). We therefore set out to determine if phorbol ester-induced hypertrophy requires ␤IIPKC using the translocation inhibitor, ␤IIV5-3. Cardiac hypertrophy involves an overall increase in the size of the cardiac myocyte, due primarily to increased expression of specific contractile proteins. Simpson et al. (24) demonstrate that this increased protein expression directly correlates with the size of the cell, enabling the use of total protein synthesis as a quantitative measure for increased cell size. In our experiments, hypertrophy of isolated primary neonatal rat cardiac myocytes was induced using limiting amounts of PMA (10 nM), and hypertrophy after 48 h was measured via 14 C-labeled phenylalanine ([ 14 C]Phe) incorporation into protein as a measure of protein synthesis (16). Cardiac myocytes were transiently permeabilized in the presence or absence of peptide, incubated with or without 10 nM PMA in media containing 14 C-labeled phenylalanine for 48 h, and total protein was harvested as described under "Experimental Procedures." Fluorescence-activated cell sorter analysis, used to compare vehicle and PMAtreated cells by size, confirmed that [ 14 C]phenylalanine incorporation correlates with increased cardiac myocyte cell size. 3 Cells treated with 10 nM PMA in the absence of peptide displayed an increase in cell size (Fig. 6B) as well as a 2-fold increase in protein synthesis (Fig. 6A). Pretreatment with the ␤IIV5-3 or ␤IV5-3 translocation inhibitor peptides resulted in a 77% Ϯ 20 and 82% Ϯ 14 decrease in PMA-induced protein synthesis, respectively (Fig. 6A). Additionally, the ␤IIV5-3 peptide inhibited basal hypertrophy by 26% Ϯ 3, with a 21% Ϯ 8 decrease observed with the ␤IV5-3 peptide (Fig. 6A). Fig. 6B shows phase contrast pictures of cells after pretreatment with peptides and after PMA-induced hypertrophy. Unlike the control cells, the cells treated with either the ␤IIV5-3 peptide or the ␤IV5-3 peptide did not increase in size in response to PMA but, instead, were much closer in size to the non-PMA-treated control cells (Fig. 6B). Therefore, the ␤IV5-3 and ␤IIV5-3 peptides inhibited the PMA-induced increase in cell size. Additionally, cells treated with a C2-derived peptide, ␤C2-4, previously shown to inhibit ␤PKC-mediated cellular functions (7,21), did not increase in size in response to 10 nM PMA, whereas a 3 R. R. Begley and D. Mochly-Rosen, unpublished observation. V5 of ␤IIPKC Contains the RACK1-binding Site control peptide (with non-relevant sequence) had no effect on PMA-induced cell size (data not shown). Taken together, these data demonstrate that both ␤Iand ␤IIPKC are essential for PMA-induced cardiac myocyte hypertrophy. DISCUSSION This study demonstrates that a domain other than the C2 domain of ␤IIPKC (7) is required for binding of ␤IIPKC to RACK1. Using fragments and peptides derived from the V5 domain of the ␤I and ␤II isozymes of PKC, we have shown that the ␤IIV5 domain bound RACK1 directly (Fig. 2B) and partially inhibited ␤IIPKC binding to RACK1 (Fig. 3). Furthermore, a combination of the ␤C2 domain and the ␤IIV5 or the ␤IV5 domain nearly abolished ␤IIPKC binding to RACK1 (Fig.  3). The RACK1 selectivity was mapped to the unique sequences in the ␤IIV5 domain. When combined with ␤C2-derived peptides, known to contain part of the RACK1-binding site in ␤IIPKC, peptides derived from the unique sequences in the ␤IIV5 domain selectively competed with ␤IIPKC binding to RACK1 in vitro (Fig. 4). Importantly, when introduced into cardiac myocytes, the ␤IV5and ␤IIV5-derived peptides ␤IV5-3 and ␤IIV5-3, selectively inhibited translocation of their respective PKC isozymes (Fig. 5). Therefore, ␤IV5-3 and ␤IIV5-3 function as isozyme-selective translocation inhibitors of ␤IPKC and of ␤IIPKC, respectively. We conclude that the ␤IIV5 domain contains part of the RACK1-binding site in ␤IIPKC. Our data on the role of the V5 domain of the ␤PKCs in binding to their respective RACKs suggest that the V5 domains of other PKC isozymes may be important for binding to their RACKs. Supporting this conclusion is the observation that the V5 domains of ␣-, ␤I-, ␤II-, ␥-, ␦-, and ⑀PKC are greater than 88% conserved between species. Therefore, a combination of the C2 and V5 domains may be useful as bait for identifying unknown RACKs for other classical and novel PKCs.
The importance of the V5 domain of ␤PKC in enzyme localization and function has been previously demonstrated. Overexpression of chimeras of the regulatory and catalytic domains of ␣and ␤IIPKC in K562 erythroleukemia cells demonstrated that the C-terminal 13 amino acids of ␤IIPKC were sufficient to confer proper localization and function (lamin B phosphorylation) of an ␣/␤II chimera (25,26). Fields and co-workers (26,27) attributed the selectivity of this region to a direct interaction with the nuclear membrane lipid, phosphatidylglycerol. Furthermore, these authors reported that ␤IIPKC-phosphatidylglycerol interaction was inhibited with a peptide derived from the C-terminal 13 amino acids (26). This peptide corresponds to our peptide labeled ␤IIV5-1, which we found to partially inhibit ␤IIV5 binding to RACK1 in vitro (see "Results"). Therefore, the ␤IIV5-1 sequence may be important for both ␤IIPKC binding to lipids as well as for RACK1 binding. Using a number of ␤IIPKC C-terminal deletion mutants lacking parts of the ␤IIV5 domain, Cooper and co-workers (28) further demonstrate the requirement of an intact ␤IIV5 domain for proper enzyme function. The work described here adds to the above studies and shows that a short peptide, ␤IIV5-3, corresponding to part of the RACK1-binding site in the V5 domain of ␤IIPKC is sufficient to inhibit ␤IIPKC function in cells.
Our data cannot exclude the possibility that at least part of the effects exerted by the V5-derived peptides is due to their interaction with the C2 domain. Newton and co-workers (29,30) show that the ␤V5 region regulates calcium binding, a function of the C2 domain (31). This suggests a direct interaction between the ␤C2 and ␤V5 domains. The data presented here demonstrating direct interactions of both the ␤IIV5 and ␤C2 fragments with RACK1 further support this hypothesis. It is interesting to note that the half-maximal binding of the ␤C2 and ␤IIV5 domains for RACK1 (400 and 500 nM, respectively, Fig. 2, A and B) is 2 orders of magnitude higher than that of ␤IIPKC holoenzyme RACK1 binding (3 nM Ϯ 2, Fig. 1A). This may reflect cooperativity between the ␤C2 and ␤IIV5 domains upon ␤IIPKC holoenzyme binding to RACK1.
Although only a combination of the ␤C2 and ␤IIV5 fragments or peptides was sufficient to completely inhibit ␤IIPKC binding to PKC in vitro (Figs. 3 and 4), we found previously that each of the ␤C2-derived peptides alone, ␤C2-1, ␤C2-2, and ␤C2-4, inhibit translocation of both ␤Iand ␤IIPKC in neonatal rat cardiac myocytes (7). Furthermore, we show here that a single peptide derived from the ␤IV5 or ␤IIV5 domains (␤IV5-3 and ␤IIV5-3, respectively) is each sufficient to inhibit translocation of its corresponding isozyme in an isozyme-selective manner (Fig. 5). Why is a peptide from either the ␤C2 or the ␤IIV5 domain sufficient to inhibit translocation of ␤IIPKC in cells when in vitro interference with both ␤C2and ␤IIV5-RACK1 interactions is required to inhibit ␤IIPKC-RACK1 binding? This may be due to the local relative concentrations of ␤IIPKC and RACK1 in cells. Using known amounts of recombinant proteins as standard, we estimate the intracellular concentration of ␤IIPKC to be ϳ1 nM and that of RACK1 to be ϳ10 nM. Moreover, the local concentration of each may vary from one cell compartment to the next. The intracellular concentration of the peptide is estimated to be ϳ1 M (10% of the extracellular concentration (16)), ϳ2 orders of magnitude above that of RACK1. Therefore, a single peptide containing part of the RACK1-binding site in ␤IIPKC may be sufficient to selectively bind RACK1 and inhibit ␤IIPKC from binding in cells. Furthermore, additional intracellular components not present in an in vitro assay, such as other binding proteins and modulators of ␤IIPKC and RACK1, may affect ␤IIPKC-RACK1 interactions in cells. In this case, a perturbation of the ␤IIPKC-RACK1 interaction induced by the selective translocation inhibitor peptide may be sufficient to prevent the translocation of ␤IIPKC to its RACK in cells and, hence, prevent its cellular function.
We also show here that complete inhibition of ␤IIPKC binding to PKC in vitro (Figs. 3 and 4) occurs only when the ␤C2 and ␤IIV5 fragments (500 nM each) or peptides (10 M each) are added together; the concentration of inhibitors used in vitro here is in great excess of that of the ␤IIPKC holoenzyme (5 nM). Such a difference in relative affinities of signaling enzymes to their anchoring proteins and a peptide inhibitor derived from one of them for the same binding protein was previously reported for several other inhibitors of protein-protein interactions. For example, a concentration of 25 M or more of a peptide derived from the cAMP-dependent protein kinase (protein kinase A)-anchoring protein, AKAP 79, is necessary to inhibit in vitro binding of the regulatory RII subunit of protein kinase A to AKAP 79 (32). Yet the half-maximal binding of the protein kinase A RII subunit to AKAP79 is obtained at a concentration of ϳ1 nM (33). Therefore, to inhibit this proteinprotein interaction, a 25,000-fold excess of the inhibitory peptide over the holoenzyme was required. Additionally, Iyengar and co-workers used a 2000-fold excess of a peptide (100 M versus 50 nM) to inhibit G protein ␤␥ subunit (G␤␥) binding to adenylyl cyclase in vitro (34) and ϳ1000-fold more peptide to stimulate phospholipase C␤2 to a similar extent to that seen with full-length G␤␥ (35). Therefore, when examined in in vitro studies, it is common to see a relatively low affinity of inhibitory peptides relative to the affinity of the proteins from which they are derived.
Active ␤IIPKC, and not active ␤IPKC, localizes to sites in cardiac myocytes where RACK1 is located (7). These data suggest that RACK1 may bind to ␤IIPKC, and not ␤IPKC, in cells. However, although the in vitro binding of ␤IIPKC to RACK1 is greater than that of ␤IPKC, some binding of ␤IPKC to RACK1 is observed (Fig. 1C). Why does ␤IPKC bind to RACK1 in vitro? The lack of absolute selectivity of interaction between a kinase and its anchoring protein in vitro is a common phenomenon. For example, Baltimore and co-workers (36), investigating the differential binding specificity of the SH3 domains of Src, NCK, Grb2, and Abl to the Abl SH3 binding protein, 3BP2, found little specificity in vitro. Binding affinities were not determined in that study; however, binding of each SH3 domain to the 3BP2 SH3 binding domain appeared to be within an order of magnitude of the others when examined by an in vitro binding assay (36). Yet subsequent studies show exquisite specificity of the SH3 domains in cells. Work from the laboratories of Bar-Sagi et al. (37) demonstrates that the SH3 domains of phospholipase C␥ and GRB2 determine the differential localization of the two proteins in cells (37). This specificity indicates the high selectivity of protein-protein interactions in cells in contrast to the findings in vitro. Similarly, we showed that the RACK for ⑀PKC, ␤ЈCOP, also binds ␤IPKC in vitro, albeit ϳ10 times less well than the binding to ⑀PKC (9). Yet, only ⑀PKC co-localizes with and binds to this RACK in cells (9), and inhibition of this interaction using an ⑀PKC-derived fragment, but not inhibition of interaction of ␤PKC with its RACK, causes fatal cardiomyopathy (38).
Finally, using the ␤V5-derived PKC translocation inhibitors, we showed that both ␤Iand ␤IIPKC are required for phorbol ester-induced cardiac myocyte hypertrophy, as measured by increased protein synthesis (Fig. 6). ␤PKC was previously reported to mediate cardiac hypertrophy (22), as shown by overexpression of ␤PKC in neonatal cardiac myocytes (39,40) and in transgenic mice (41). Furthermore, ␤Iand ␤IIPKC protein levels were elevated in hearts from a pressure overload-induced cardiac hypertrophy rat model (42) and in humans with congestive heart failure (43). Together, these data demonstrate that, similar to our findings here, both ␤Iand ␤IIPKC have been implicated in cardiac hypertrophy and heart failure in animal models and in humans. This suggests that the two PKC isozymes carry out redundant cellular functions; however, their differential subcellular localization (13) suggests that they are performing these cellular functions by phosphorylating different protein substrates. Both opposing and parallel roles of individual PKC isozymes in a single cell function have been previously observed (19,44,45). Therefore, multiple PKC isozymes can play distinct roles in common cellular functions, working together or in opposition and providing multiple levels of regulation for the same cellular process. Selective inhibitors, like the ␤Iand ␤IIPKC isozyme-selective translocation inhibitors described here, should be useful in determining the roles of individual PKC isozymes in various cellular functions.