Modulation of the Protein Kinase Cδ Interaction with the “d” Subunit of F1F0-ATP Synthase in Neonatal Cardiac Myocytes

The F1F0-ATP synthase provides ∼90% of cardiac ATP, yet little is known regarding its regulation under normal or pathological conditions. Previously, we demonstrated that protein kinase Cδ (PKCδ) inhibits F1F0 activity via an interaction with the “d” subunit of F1F0-ATP synthase (dF1F0) in neonatal cardiac myocytes (NCMs) (Nguyen, T., Ogbi, M., and Johnson, J. A. (2008) J. Biol. Chem. 283, 29831–29840). We have now identified a dF1F0-derived peptide (NH2-2AGRKLALKTIDWVSF16-COOH) that inhibits PKCδ binding to dF1F0 in overlay assays. We have also identified a second dF1F0-derived peptide (NH2-111RVREYEKQLEKIKNMI126-COOH) that facilitates PKCδ binding to dF1F0. Incubation of NCMs with versions of these peptides containing HIV-Tat protein transduction and mammalian mitochondrial targeting sequences resulted in their delivery into mitochondria. Preincubation of NCMs, with 10 nm extracellular concentrations of the mitochondrially targeted PKCδ-dF1F0 interaction inhibitor, decreased 100 nm 4β-phorbol 12-myristate 13-acetate (4β-PMA)-induced co-immunoprecipitation of PKCδ with dF1F0 by 50 ± 15% and abolished the 30 nm 4β-PMA-induced inhibition of F1F0-ATPase activity. A scrambled sequence (inactive) peptide, which contained HIV-Tat and mitochondrial targeting sequences, was without effect. In contrast, the cell-permeable, mitochondrially targeted PKCδ-dF1F0 facilitator peptide by itself induced the PKCδ-dF1F0 co-immunoprecipitation and inhibited F1F0-ATPase activity. In in vitro PKC add-back experiments, the PKCδ-F1F0 inhibitor blocked PKCδ-mediated inhibition of F1F0-ATPase activity, whereas the facilitator induced inhibition. We have developed the first cell-permeable, mitochondrially targeted modulators of the PKCδ-dF1F0 interaction in NCMs. These novel peptides will improve our understanding of cardiac F1F0 regulation and may have potential as therapeutics to attenuate cardiac injury.

increase its activity in response to cellular ATP demands (3). In addition, in chronic cardiomyopathy, ATP levels have been estimated to decline by ϳ20% along with an ϳ80% drop in phosphocreatinine levels (11), suggesting that chronic energy starvation in the heart may be a major contributor to congestive heart failure (12).
Following severe cardiac ischemia/reperfusion (IR) injury, the type of damage sustained in a heart attack, ATP levels also decline substantially (9,10). A major component of this drop involves the loss of the mitochondrial IM proton gradient that supplies the energy for F 1 F 0 -ATP synthase. Therefore, in the early phases of myocardial ischemia, the F 1 F 0 -ATP synthase becomes inhibited. It then makes a futile attempt to re-establish mitochondrial IM potential by pumping protons out of the mitochondrial matrix. This process requires energy, which is supplied by the F 1 F 0 complex, then operating as an ATPase (7,10). If ischemia is not interrupted, F 1 F 0 -ATPase activity contributes heavily to the loss of myocardial ATP (7,10). In addition, the return of aerobic ATP synthesis is impaired following IR injury, but the mechanisms for this delay are not completely known. Under ischemic conditions, the heart attempts to compensate by using glucose as a preferred substrate (instead of predominantly fatty acids) in anaerobic glycolysis. However, anaerobic ATP production is not sufficient to supply the extensive cardiac demands for energy, and it also generates considerable lactic acid, which damages the heart further. The consequent decreases in pH inhibit glycolytic enzymes (7,10). An accelerated return of aerobic ATP production, following IR injury, would therefore improve the survival of cardiac cells, making the F 1 F 0 -ATP synthase/ATPase a logical focus of cardiac IR research.
Previous studies have proposed that cardiac pre-conditioning (PC), a cardioprotective response against IR injury, inhibits F 1 F 0 -ATPase mode activity (13,14). This could contribute protection by reducing ATP hydrolysis after ischemic insults. There have also been reports that PC induces an improved recovery of cardiac ATP levels following prolonged IR injury (10,15,16). This is consistent with the hypothesis that cardiac PC may improve F 1 F 0 -ATP synthase-mode activity as well. However, to date, the precise roles of the F 1 F 0 complex in cardiac IR injury and PC are not fully understood.
Mitochondrial mechanisms involving protein kinase C (PKC) isozymes in cardiac PC and IR injury have received considerable attention (17)(18)(19)(20)(21). Identification of the particular isozyme(s) that is involved in these responses is an essential prerequisite for the development of clinically useful pharmacological mimetic agents. In support of this, previous studies have reported a PKC⑀-selective facilitator peptide, and a PKC␦ inhibitor peptide can induce cardio-protection from IR injury in cardiomyocytes (22), isolated hearts, and, in vivo, transgenic mice (23)(24)(25)(26). The above-mentioned PKC␦ inhibitor has also shown great promise in human clinical trials targeting damage from acute myocardial infarction (27). Furthermore, a PKC␦selective translocation inhibitor, administered immediately following a prolonged ischemic insult, improved the recovery of mouse heart ATP levels during oxygenated reperfusion (25).
In this study, we have identified the first cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 interaction inhibitor and facilitator peptides. We have determined that these peptides modulate the interaction between PKC␦ and dF 1 F 0 and, consequently, F 1 F 0 activity in NCMs. We propose that excessive inhibition of the F 1 F 0 -ATP synthase by PKC␦ may play a significant role in cardiac pathology and that our novel peptides could provide a mechanism for investigating these effects.

EXPERIMENTAL PROCEDURES
Primary NCMs-NCMs were isolated from the hearts of 1-day-old Sprague-Dawley rats as described previously (28,29). This study was conducted in accordance with Institutional, State, and Federal guidelines for the humane care and use of laboratory animals. Briefly, cells were obtained from hearts by gentle trypsinization at room temperature, and dissociated cells were preplated for 40 min onto 100-mm dishes in medium 199 (M-199) (Invitrogen) containing 10% fetal bovine serum (Hyclone Laboratories). The nonattached cells were plated onto 35-or 100-mm Corning Petri dishes at a density of 800 cells/mm 2 and incubated at 37°C in humidified air with 1% CO 2 . Myocytes were cultured in M-199 supplemented with vitamin B 12 (1.5 mmol/liter), penicillin G (50 units/ml), 0.1 mmol/liter bromodeoxyuridine, and 10% fetal bovine serum through day 4. After the 4th day, cells were then placed in defined M-199 containing 50 units/ml penicillin G, 1.5 mmol/ liter vitamin B 12 , and 10 g/ml each of transferrin and insulin.
Peptide Synthesis-The 14 sequential dF 1 F 0 -derived peptides (1 15-mer, 12 16-mers, and 1 18-mer) covering the entire 161amino acid sequence of dF 1 F 0 were synthesized by 21st Century Biochemicals (Marlborough, MA) using Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry. Each peptide had a 5-amino acid overlap with the previous peptide. Final sequences of these peptides are shown in Table 1. Several modifications were made to some of the peptides to more closely mimic the physiological forms of the amino acids found in the native dF 1 F 0 protein. For example, we did not include the first methionine in dF 1 F 0 sequence, as the National Center for Biotechnology Education database indicates that this methionine is not present in the final mature form of dF 1 F 0 . In addition, we have added N-acetyl and C-carboxyamide groups to the N and C terminus of each peptide, respectively, to properly mimic the peptide bonds normally present in the native protein. Because peptide 14 represents the actual C terminus of the protein, this peptide was made as a free C-terminal carboxyl. Finally, there are several lysines in the native protein (e.g. amino acids 63, 78, 85, 99, 117, and 149) that were incorporated into peptides as N-acetyl-lysines. All peptides were high pressure liquid chromatographypurified (purity Ն90%), and their masses and sequences were verified by collision-induced fragmentation using a Qo-TOF electrospray ionization-mass spectrometer (QSTAR XL Pro, MDX Sciex). All of the peptides were finished as acetate salts, and endotoxin-free water was used for all peptide purification and formulations.
Syntheses for cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 interaction inhibitor and facilitator peptides were as follows. Each peptide included either the PKC␦-dF 1 F 0 interaction antagonist NH 2 -2 AGRKLALKTIDWVSF 16 -COOH or agonist NH 2 -111 RVREYEKQLEKIKNMI 126 -COOH sequence. An HIV-Tat protein-transducing sequence YGRKKRRQRRR was also placed at the N terminus of each peptide using a cysteine-cysteine linkage as described previously (30). Following the Tat sequence, we included a mitochondrial targeting sequence (MLATRALSLIGKRAISTSVC) derived from the number IV subunit of cytochrome oxidase (31). Finally, a FLAG epitope (DYKDDDDK) was also attached to the C-terminal end of each peptide to monitor their uptake into mitochondria. As a control, a scrambled (inactive) (ADKIGWAVLRTKSLF) version of the PKC␦-dF 1 F 0 inhibitor peptide was synthesized with all of the above HIV-Tat, mitochondrial targeting and FLAG domains. The final overall sequence of each of these peptides is shown in Fig. 1A.
F 1 F 0 -ATP Synthase Purification-Isolation of the F 1 F 0 -ATP synthase holoenzyme by chromatography was performed as described previously in detail by Buchanan and Walker (32). Briefly, mitochondria were isolated from five adult Sprague-Dawley rats and were then washed in phosphate buffer and solubilized in buffer containing 20 mM ATP, 20 mM MgSO 4 , 0.001% phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 1% n-dodecyl-␤,D-maltoside. The solubilized mitochondria were precipitated with ammonium sulfate. The ammonium sulfate pellet was resuspended in buffer and dialyzed overnight to remove lipids. The sample was then loaded onto a Q-Sepharose anion exchange column. The F 1 F 0 -ATP synthase holoenzyme was then eluted using a linear NaCl gradient (from 0 to 1 M), and peak F 1 F 0 -ATPase fractions were initially identified by the F 1 F 0 -ATPase activity assay. Subsequently, the presence of the ␣, ␤, and d subunits of the F 1 F 0 -ATPase was confirmed in peak F 1 F 0 -ATPase activity fractions by Western blots using antisera against each of these subunits. Peak fractions were then pooled and used for overlay binding and other assays. Antisera for the ␣, ␤, and d subunits of the F 1 F 0 -ATPase were obtained from Molecular Probes and Mitosciences and were used in Western blots.
PKC Binding Overlay Assays-The overlay method to detect PKC-binding partners was described previously by Mochly-Rosen and co-workers (33)(34)(35)(36). Required activators and co-factors of PKC such as calcium (Ca 2ϩ ), phosphatidylserine (PS, Avanti, Alabaster, AL), and diacylglycerol (DG, Avanti) were included in the assay. Purified F 1 F 0 holoenzyme (50 g) was separated by 13.5% SDS-polyacrylamide gel and transferred onto nitrocellulose paper (NCP). The blots were then cut into strips (0.3 ϫ 5.5 cm) and washed briefly with distilled water. The membrane was then blocked with overlay blocking buffer (50 mM Tris/HCl, pH 7.5, 200 mM NaCl, 3% bovine serum albumin, and 0.1% polyethylene glycol) for 1 h at room temperature. Next, the strips were incubated with overlay buffer containing purified PKC (50 g/ml) and co-factors for 1 h at room temperature. The strips were washed four times for 5 min with overlay wash buffer. Bound PKC was detected with PKC-isozyme antisera and ECL detection as described below for Western blot analyses.
Cell Lysis and Isolation of Mitochondria from NCMs-100-mm dishes of NCMs were used for mitochondria isolation. The culture medium was removed, and cells were washed twice with Ca 2ϩ /Mg 2ϩ -free phosphate-buffered saline containing 1.4 nM calyculin A and 1.44 mM tetrasodium pyrophosphate (NaPP i ). Cells were scraped from the dishes in 400 l of isotonic MSE buffer (10 mM Tris/HCl, pH 7.5, 220 mM mannitol, 70 mM sucrose, 1 mM EGTA, 0.025% bovine serum albumin, 2 mM taurine, 1.6 mM carnitine, 1.5 mM NaPP i , 1.5 nM calyculin A, and 10 g/ml protease inhibitors) containing 100 units/ml trypsin (Sigma) for 20 min on ice. After the incubation period, cells were then subjected to Dounce homogenization. The resulting homogenate was centrifuged twice at 600 ϫ g for 5 min to pellet nuclei and cell debris. Protease inhibitors were added to the final supernatant at 10 g/ml each, and it was then loaded onto a Percoll/Optiprep gradient as described below (37,38).
Preparations of Percoll/Optiprep Gradient-The postnuclear supernatant (prepared as described above) was layered over a combination Percoll/Optiprep (Accurate Chemical and Scientific, Westbury, NY) gradient prepared on the same day as follows. Each gradient was prepared in Beckman-Ultraclear 14 ϫ 89-mm centrifuge tubes. The first step in gradient formation involved overlaying 1.74 ml of a 17% (v/v) Optiprep solution on a 1.74-ml cushion of 35% Optiprep solution. Next, 4.35 ml of a 6% (v/v) Percoll solution was layered on top of the 17% Optiprep solution. All Optiprep and Percoll solutions were prepared using MSE buffer as the diluent. Gradients were stored on ice until use. Next, 4.2 ml of postnuclear supernatant was gently layered on top of the 6% Percoll portion of the gradient. All tubes were centrifuged in a Beckman SW.41 swinging bucket rotor at 50,000 ϫ g for 30 min using the lowest acceleration and deceleration speeds. Mitochondria were collected at the 17/35% Optiprep interface from each gradient using a Pasteur pipette and placed on ice until use.
Immunoprecipitation (IP) Experiments-The antisera against the "d" subunit of the F 1 F 0 -ATPase (dF 1 F 0 ) from Molecular Probes (50 g) was coupled with Bio-Rad Affi-Gel (1 ml) according to the manufacturer's instructions. The IP was performed as described previously (37,38) with the addition of 150 mM NaCl to the IP buffer in the three final wash steps. F 1 F 0 -ATPase Activity Measurement-The oligomycin-sensitive activity of F 1 F 0 -ATPase activity was measured using a spectrophotometric assay as described previously (32). Briefly, cells were scraped and lysed by sonication in 400 l of buffer containing 20 mM Tris/HCl, pH 7.5, 1 mM MgCl 2 , 5 mM KCl, and 1 mM EGTA. The F 1 F 0 -ATPase activity then was measured spectrophotometrically by monitoring the disappearance of NADH, which manifests as a decline in absorbance at 340 nm as NADH is oxidized NAD ϩ . The final F 1 F 0 -ATPase assay buffer contained 25 mM Tris/HCl, pH 7.5, 83 mM sucrose, 4 mM MgCl 2 , 25 mM KCl, 1 mM KCN, 1 mM EDTA, 1 mM EGTA, 2 mM ATP, 1.5 mM phosphoenolpyruvate, 5 units of pyruvate kinase, 5 units of lactate dehydrogenase, and 75 M NADH in a 1-ml cuvette. All assay reagents were obtained from Sigma. The assay was monitored continuously for 1-3 min at 25°C in the absence and presence of oligomycin (8 g/ml).
Western Blot Analyses-Western blotting was carried out as described previously (28,29) using ECL detection (GE Healthcare). Samples were subjected to SDS-PAGE on 12-13.5% acrylamide gels and then transferred onto NCP. The resulting blots were probed for PKC isozymes and F 1 F 0 subunits. PKC antisera were obtained from BD Transduction Laboratories or Santa Cruz Biotechnology; F 1 F 0 antisera were obtained from Invitrogen, and the FLAG antibody was obtained from Sigma.
Statistical Analyses-Differences between two groups were assessed using unpaired Student's t test, and comparisons among multiple groups were made using one-way analysis of variance with Bonferroni's post hoc test. A p value Յ 0.05 was considered significant.

Design of dF 1 F 0 -derived Peptides-Our previous studies in
NCMs demonstrated an interaction between PKC␦ and the d subunit of F 1 F 0 -ATP synthase (dF 1 F 0 ), which correlated with 4␤-PMA-and hypoxia-induced inhibition of F 1 F 0 -ATPase activity (39). We next surveyed which amino acid sequence(s) within the dF 1 F 0 protein was responsible for its binding to PKC␦. Currently, there are no commercially available recombinant sources of mammalian dF 1 F 0 protein. For this reason, we synthesized 14 sequential peptides (Table 1) covering the entire amino acid sequence of dF 1 F 0 as described under "Experimental Procedures." PKC Overlay Assays Reveal an Inhibitor and Facilitator of the PKC␦ dF 1 F Interaction-We have previously demonstrated that PKC␦, but not PKC␣, -⑀, or -, selectively binds to the dF 1 F 0 protein in overlay assays (39). In Fig. 1, B and C, purified F 1 F 0 -ATPase holoenzyme was isolated from adult rat heart mitochondria by chromatography (32). Next, individual F 1 F 0 -ATPase subunits were resolved by SDS-PAGE and then transferred to NCP. The resulting NCP blots were incubated with a mixture of purified rat brain PKC isozymes (40) and PKC activators, in the presence of increasing concentrations of each of the 14 dF 1 F 0 -derived peptides shown in Table 1. For brevity, data for only the two peptides with the most optimal effects on the PKC␦-F 1 F 0 binding interaction are shown (Fig. 1, B and C). PKC␦ binding was detected by anti-PKC␦ antisera and ECL techniques as described previously (39). It should be noted that four different F 1 F 0 -ATPase preparations, three different PKC preparations, and two different lots of peptides were used to obtain the data shown in Fig. 1, B-D. A major finding of our study was that one of the peptides (NH 2 -2 AGRKLALKTID-WVSF 16 -COOH) demonstrated a dose-dependent inhibition of the PKC␦-dF 1 F 0 binding interaction (Fig. 1B). When 0.3 M concentrations of this peptide were included in the overlay assay, we observed a 37 Ϯ 9% (n ϭ 9) inhibition of PKC␦ binding to dF 1 F 0 . The calculated IC 50 for this peptide was 0.216 M. NCBI database searches revealed that the amino acid sequence of this inhibitor is not found in any known proteins except the mammalian F 1 F 0 d subunit.
A second major finding of our study was that one of the remaining peptides (NH 2 -111 RVREYEKQLEKIKNMI 126 -COOH) facilitated the binding of PKC␦ to dF 1 F 0 in the overlay assay (Fig. 1C). In the presence of PKC activators, 0.3 M concentrations of this peptide potentiated PKC␦ binding to dF 1 F 0 by 229 Ϯ 76% (n ϭ 6). The EC 50 for this peptide was 0.71 M. These results suggested that we had identified two peptides that modulate PKC␦ binding to dF 1 F 0 in opposite directions; one peptide was an inhibitor and the second peptide acted as a facilitator of this interaction.
Addition of HIV-Tat, Mitochondrial Targeting, and FLAG Sequences Do Not Reduce the Efficacy of the PKC␦-dF 1 F 0 Inhibitor in Overlay Assays-In vitro binding of PKC␦ to dF 1 F 0 in the presence of increasing concentrations of the HIV-Tat/mitochondrial targeting domain-containing PKC␦-dF 1 F 0 inhibitor ( Fig. 1A) was monitored using the approach outlined in Fig. 1, B and C. When this inhibitor peptide (100 nM) was included in overlay assays, we observed a 42 Ϯ 14% (n ϭ 6) reduction in PKC␦ binding to dF 1 F 0 (Fig. 1D). The inactive version of this peptide had no effect on the binding of PKC␦ to dF 1 F 0 (Fig. 1D). Therefore, the modified antagonist/inhibitor exhibited a dosedependent inhibition of the PKC␦-dF 1 F 0 binding interaction (Fig. 1D), which appeared to be modestly more potent (IC 50 ϭ 0.144 M) than the inhibitor sequence alone (Fig. 1B, IC 50 ϭ 0.216 M).
Incubation of NCMs with the Cell-permeable, Mitochondrially Targeted PKC␦-dF 1 F 0 Inhibitor and Facilitator Peptides Results in Their Delivery into Mitochondria-We next incubated NCMs in the presence or absence of the mitochondrially targeted PKC␦-dF 1 F 0 facilitator ( Fig. 2A) or inhibitor (Fig. 2B) peptides (100 nM) for 2 h at 37°C. Extracellular media were then removed, and cells were washed twice with phosphate-buffered saline. Next, each group of cells was homogenized, and mitochondria were isolated using differential centrifugation and Percoll/Optiprep density gradients as described previously (38,39). In Fig. 2, lane 1, 0.5 g of the FLAG epitope-tagged facilitator ( Fig. 2A) or inhibitor (Fig. 2B) peptide alone was subjected to SDS-PAGE to confirm that we could use anti-FLAG antisera in Western blots to detect our peptides. As shown, each peptide was readily detected. In Fig. 2, A and B, lane 2 is similar to lane 1 except each peptide was first incubated with trypsin (100 units/ml) for 20 min on ice prior to conducting SDS-PAGE and Western blots with anti-FLAG antisera. Note the substantial loss in FLAG immunoreactivity confirming that each peptide was sensitive to trypsin digestion. In Fig. 2, A and B, lanes 3 and 4, we subjected mitochondria isolated from NCMs that were not incubated with any peptide to SDS-PAGE and Western blot analyses with anti-FLAG antisera. As predicted, no FLAG Shown are the amino acid sequences of the 14 peptides synthesized as described under "Experimental Procedures" used to determine their effects on PKC␦ binding to dF 1 F 0 . Peptides are listed in order beginning with the N terminus of dF 1 F 0 in peptide 1 and ending with the COOH-terminal dF 1 F 0 sequence in peptide 14. Each peptide contains a 5-amino acid overlap with the previous sequential peptide. Also note that the N-terminal methionine has been deleted from peptide 1 because it is not thought to be present in the mature dF 1 F 0 protein. Other modifications to certain amino acid side chains of our peptides were made to better mimic their in vivo chemistry and are described in detail under "Experimental Procedures." Shown in boldface superscripts are the actual amino acid numbers corresponding to where each peptide is located in the dF 1 F 0 sequence. Amino acids are indicated by universally accepted single letter abbreviations.

Peptide
Amino acid sequence immunoreactivity was observed in these samples. In contrast, Fig. 2, A and B, lanes 5 and 6 represent samples prepared by isolating density gradient-purified mitochondria from NCMs that were preincubated for 2 h with 100 nM concentrations of either the PKC␦-dF 1 F 0 facilitator ( Fig. 2A) or inhibitor (Fig. 2B) peptides. Note the strong anti-FLAG immunoreactivity observed in Fig. 2, A and B, lane 5. These results indicate that extracellular treatment of NCMs with our peptides resulted in substantial mitochondrial uptake of each peptide. To confirm that the peptides were not simply adhering to extramitochondrial surfaces, we treated the mitochondria with 100 units/ml trypsin for 20 min on ice (Fig. 2, A and B, lane 6). Note that this treatment did not diminish the anti-FLAG immunoreactivity (Fig. 2A, lane 5 versus 6), despite the extreme sensitivity of our peptides to digestion by trypsin (Fig. 2, A and B, lane  1 versus 2). These results suggested that our FLAG-tagged peptides were not likely to be simply adhering to extramitochondrial surfaces and that they do enter NCM mitochondria.
The effects of the mitochondrially targeted PKC␦-dF 1 F 0 facilitator peptide was somewhat predictable in that it promotes PKC␦ binding to dF 1 F 0 by itself in overlay assays (Fig. 1C). We hypothesized that because the interaction of PKC␦ with dF 1 F 0 induces PKC␦-mediated inhibition of F 1 F 0 -ATPase activity that the PKC␦-dF 1 F 0 facilitator peptide, administered by itself, should inhibit F 1 F 0 -ATPase activity. In fact, this turned out to be the case as is shown in Fig. 3, last set of bars. Interestingly, the combination of the PKC␦-dF 1 F 0 facilitator with 4␤-PMA did not induce synergistic inhibition of F 1 F 0 -ATPase activity. We believe that this is because at the concentrations of 4␤-PMA and PKC␦-dF 1 F 0 facilitator peptide used, we observe maximal co-IP of PKC␦ with dF 1 F 0 (Fig.  4B), which is to say that treatment of NCMs with either FIGURE 1. Modulation of PKC␦ binding to the d subunit of F 1 F 0 -ATP synthase (dF 1 F 0 ) by dF 1 F 0 -derived peptides. A, peptides contain the putative PKC␦-dF 1 F 0 inhibitor, facilitator, or scrambled (inactive) amino acid sequences, an HIV-Tat protein-transducing (PTD) sequence (YGRKKRRQRRR) for cell uptake, a mitochondrial targeting sequence (MLATRALSLIGKRAISTSVC), and a FLAG epitope (DYKDDDDK). B and C, PKC overlay assays reveal an inhibitor (B) and a facilitator (C) of the PKC␦-dF 1 F 0 -binding interaction. Purified F 1 F 0 -ATPase holoenzyme (32) was subjected to SDS-PAGE to resolve individual subunits and transferred onto NCP. NCP strips were "overlaid" with a mixture of purified PKC isozymes (40) in the presence of PKC activators (diacylglycerol and phosphatidylserine) and 0 -10 M concentrations of the putative PKC␦-dF 1 F 0 inhibitor (B) or facilitator (C), followed by Western blot analyses using PKC␦-selective antisera (39). Amino acid sequences of each peptide are shown at the top center of B and C. It is important to note that B and C are the only experiments in this study that tested the PKC␦-dF 1 F 0 inhibitor and facilitator peptides in the absence of HIV-Tat PTD, mitochondrial targeting, and FLAG domains. Representative autoradiographs are shown, and the histogram values represent mean Ϯ S.E. % of "no peptide control" densitometry values from nine independent experiments for B and six experiments for C. D, cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor attenuates PKC␦ binding to dF 1 F 0 . PKC overlay assays were conducted as in B and C in the presence of 0 -10 M concentrations of the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor or scrambled (inactive) peptide. The complete sequences of these peptides are as in A. A representative autoradiograph is shown, and the histogram values represent mean Ϯ S.E. % of densitometry values from six independent experiments. B-D, * indicates statistically significant differences from the no peptide control.
Treatment of NCMs with the Cell-permeable, Mitochondrially Targeted PKC␦-dF 1 F 0 Inhibitor Attenuates the 4␤-PMAinduced Co-IP of PKC␦ with dF 1 F 0 -In the experiments represented in Fig. 4A, NCMs were incubated for 2 h at 37°C in the absence or presence of our mitochondrially targeted peptides ( Fig. 1A) followed by a 30-min exposure to 100 nM 4␤-PMA. Cells were next homogenized, and mitochondria were isolated by differential centrifugation and Percoll/Optiprep density gradients (38). Mitochondria were then lysed and subjected to IP experiments using dF 1 F 0 antisera chemically coupled to Affi-Gel. IPs were subjected to SDS-PAGE and Western blot analyses using anti-PKC␦ antisera. Consistent with our previous study (39), 4␤-PMA treatment of NCMs induced the co-IP of PKC␦ with dF 1 F 0 (Fig. 4A, lane 2). This co-IP is preferential for the PKC␦ isozyme as we have previously shown that phorbol ester treatment of NCMs does not induce co-IP of PKC␣,-⑀, or -isozymes with dF 1 F 0 (39). When cells were incubated in the presence of the mitochondrially targeted versions of the inactive scrambled sequence or the PKC␦-dF 1 F 0 facilitator peptides (100 nM), there was no effect on the 4␤-PMA-induced PKC␦-dF 1 F 0 co-IP (Fig. 4A, lane 2 versus lanes 3 and 5). However, the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor attenuated the 4␤-PMA-induced co-IP of PKC␦ with dF 1 F 0 antisera by 50 Ϯ 15% (n ϭ 4) (Fig. 4A, lane 4). These results are consistent with the hypothesis that the PKC␦-dF 1 F 0 inhibitor diminishes the 4␤-PMA-induced inhibition of F 1 F 0 -ATPase activity (Fig. 3) by attenuating the PKC␦-dF 1 F 0 interaction in NCMs.
Treatment of NCMs with the Cell-permeable, Mitochondrially Targeted PKC␦-dF 1 F 0 Facilitator Peptide Induces the PKC␦-dF 1 F 0 Co-IP-To determine the effects of the PKC␦-dF 1 F 0 facilitator peptide (10 nM) (Fig. 1A) on the PKC␦ co-IP with dF 1 F 0 , we conducted the experiments represented in Fig. 4B. We initially hypothesized that this peptide would enhance this 4␤-PMA-induced effect. However, we did not observe any such enhancement (Fig. 4A). It is possible that at these concentrations, each stimulus could induce the PKC␦-dF 1 F 0 co-IP maximally, and therefore we did not observe synergy when the two compounds were applied together.   tide were dose-dependent as 10 nM concentrations of the PKC␦-dF 1 F 0 facilitator increased the PKC␦-dF 1 F 0 co-IP by 2.1 Ϯ 0.1-fold (n ϭ 3) greater than 1 nM concentrations of the peptide (Fig. 4B). These results are consistent with the PKC␦-dF 1 F 0 facilitator inhibiting F 1 F 0 -ATPase activity (Fig.   3) by induction of the PKC␦-dF 1 F 0 interaction in NCMs (Fig. 4B).
PKC␦-dF 1 F 0 Inhibitor Attenuates and the PKC␦-dF 1 F 0 Facilitator Induces Inhibition of F 1 F 0 -ATPase Activity in Vitro-We have previously demonstrated that co-incubation of recombinant PKC␦ with chromatographically purified F 1 F 0 -ATPase from rat heart mitochondria inhibits F 1 F 0 -ATPase activity (39). This inhibition occurs in the absence of PKC activators and is amplified in their presence (39). We next determined if our HIV-Tat-coupled, mitochondrial targeted PKC␦-dF 1 F 0 interaction modulators (Fig. 1A) could inhibit or facilitate PKC␦mediated inhibition of F 1 F 0 -ATPase activity in vitro (Fig. 5). Addition of recombinant PKC␦ to purified F 1 F 0 -ATPase in the absence of PKC activators leads to a 1.7 Ϯ 0.1-fold inhibition of F 1 F 0 -ATPase activity. In the presence of the PKC activators DG and phosphatidylserine (PS), PKC␦ inhibited F 1 F 0 activity by 92 Ϯ 5% (Fig. 5). The scrambled sequence, inactive control peptide had no impact on these PKC␦-mediated responses. In contrast, 10 nM concentrations of the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor attenuated the PKC␦-mediated inhibition of F 1 F 0 -ATPase in the absence of PKC activators by 28 Ϯ 3% (n ϭ 3) (Fig. 5, lanes 2 versus 3). A much more dramatic effect of the PKC␦-dF 1 F 0 inhibitor peptide was observed in the presence of PKC activators as it reduced recombinant PKC␦-induced inhibition of F 1 F 0 -  Fig. 1A. Next, all groups, except lane 1, were exposed to 100 nM 4␤-PMA for 30 min. Mitochondria were then isolated as in Fig. 2. Mitochondrial proteins were subjected to immunoprecipitation reactions using dF 1 F 0 antisera coupled to Affi-Gel. The resulting IPs were probed for PKC␦ in Western blot analyses. As is shown, only the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor peptide attenuated 4␤-PMA-induced co-IP of PKC␦ with dF 1 F 0 antisera (lane 4). A typical autoradiograph is shown at the top of the figure. Bars in the histogram represent mean Ϯ S.E. densitometry values taken from five independent experiments each conducted on mitochondria isolated from a separate myocyte preparation. * indicates statistically different from co-IP values obtained with 4␤-PMA alone in the absence of any peptide (lane 2). Con, control. B, extracellular treatment of NCMs with the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 facilitator peptide induces the PKC␦-dF 1 F 0 co-IP. NCMs were administered no treatments (lane 1) or treatments with 30 nM 4␤-PMA for 30 min (lane 2) and with 1 (lane 3) or 10 (lane 4) nM extracellular concentrations of the PKC␦-dF 1 F 0 interaction facilitator peptide (Fig. 1A) for 2 h (lanes 3 and 4). Mitochondria were isolated; co-IP reactions were carried out, and Western blots were probed for PKC␦ as in Fig. 4A. Under basal conditions there is little co-IP between dF 1 F 0 and PKC␦ (lane 1). In contrast, 4␤-PMA treatment causes a dramatic increase in the co-IP of PKC␦ with dF 1 F 0 (lane 2). Lanes 3 and 4 are co-IPs taken from NCM mitochondria isolated from cells that were treated with 1-10 nM concentrations of the PKC␦-dF 1 F 0 facilitator in the complete absence of 4␤-PMA. Note the dose-dependent induction of the PKC␦-dF 1 F 0 co-IP by the facilitator peptide alone. The autoradiograph shown is from a typical experiment. Histogram bars represent mean Ϯ S.E. densitometry values expressed as % of maximal from three independent experiments each conducted on a separate myocyte preparation. * indicates statistically different from control co-IP values. # indicates statistically different from co-IP values obtained for 4␤-PMA alone in the absence of any peptide. FIGURE 5. Cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 inhibitor and facilitator peptides have opposing effects on PKC␦-mediated inhibition of F 1 F 0 -ATPase activity in vitro. Rat heart F 1 F 0 -ATPase was purified from isolated mitochondria using the methods of Buchanan and Walker (32) and recombinant PKC␦, purified from an Sf9 cell expression system, was obtained from BIOSOURCE. F 1 F 0 -ATPase (50 g) activity was monitored spectrophotometrically in the absence (lane 1) or presence (lanes 2-7) of 1.5 g of recombinant PKC␦. Assays were conducted in the absence (lanes 1-4) or presence (lanes 5-7) of the PKC activators DG (3.2 g/ml) and PS (240 g/ml). The effects of 10 nM concentrations of the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 interaction inhibitor (lanes 3 and 5) or facilitator (lanes 4 and 7) on PKC␦-mediated inhibition of F 1 F 0 -ATPase activity were then evaluated. Note that the PKC␦-dF 1 F 0 inhibitor attenuated PKC␦-mediated inhibition of F 1 F 0 -ATPase activity whether or not DG/PS was present in the assay. In contrast, the facilitator modestly enhanced PKC␦-mediated inhibition of F 1 F 0 -ATPase in the absence of DG/PS. Results shown are mean Ϯ S.E. results from four independent experiments. * indicates statistically different from control (F 1 F 0 only). # indicates statistically different from F 1 F 0 ϩ recombinant PKC␦ group.
We next determined whether inclusion of the cell-permeable, mitochondrially targeted PKC␦-dF 1 F 0 facilitator in the PKC␦/F 1 F 0 -ATPase add-back assay amplified PKC␦ inhibitory actions. Consistent with this, in the absence of DG/PS, we found 55 Ϯ 13% (n ϭ 3) greater PKC␦-mediated inhibition of F 1 F 0 -ATPase when 10 nM concentrations of the PKC␦-dF 1 F 0 facilitator peptide were included in the assay (Fig. 5, compare  lanes 2 and 4 to lane 1). Furthermore, the PKC␦-dF 1 F 0 facilitator peptide modestly enhanced PKC␦-mediated inhibition of F 1 F 0 activity when the PKC activators DG/PS were present in the assay (Fig. 5, lanes 5 versus 7). Collectively, these data support the hypothesis that both DG/PS and the PKC␦-dF 1 F 0 facilitator peptide promote the PKC␦ interaction with dF 1 F 0 to inhibit F 1 F 0 activity.

DISCUSSION
The human heart consumes up to 20 times its mass in ATP each day, and F 1 F 0 -ATP synthase may be the most crucial enzyme for meeting these intense energy requirements. Therefore, a better understanding of the regulation of the F 1 F 0 complex under normal and pathological states is an extremely important goal. Unfortunately, very little has been learned over the past 2 decades regarding cardiac F 1 F 0 function. Recently, our laboratory identified a novel and potent inhibitory regulation of this enzyme complex by the PKC␦ isozyme. The response involves PKC␦ binding to the d subunit of the F 1 F 0 -ATP synthase (dF 1 F 0 ) and is induced by phorbol ester and hypoxia in NCMs (39).
The role of mitochondrial PKC␦ in cardiac PC (41)(42)(43) and IR injury (44 -47) is controversial. Our studies support the notion that whether PKC␦ mediates cardiac protection or injury may be context-dependent. For example, it would be beneficial if PKC␦ inhibited ATP hydrolysis during prolonged ischemia. In support of this, numerous studies have reported that cardiac PC inhibits F 1 F 0 -ATPase activity (13,14). Alternatively, PKC␦ could exacerbate injury if it inhibits F 1 F 0 -ATP synthase-mode activity during early reperfusion after a prolonged ischemic insult. It is well known that one characteristic of cardiac IR injury involves a delayed recovery of aerobic ATP synthesis (10,44). We hypothesize that such a delay in the recovery of aerobic ATP production could contribute heavily to cardiac myocyte cell death. Consequently, disrupting the PKC␦-dF 1 F 0 interaction at the time of cardiac reperfusion could facilitate a more rapid return of aerobic ATP production following IR injury. We therefore developed novel tools to pursue the regulation of the F 1 F 0 complex by PKC␦ interaction with dF 1 F 0 in live heart cells.
There are no published reports for the expression of mammalian F 1 F 0 -ATP synthase subunits in bacteria, baculovirus/ Sf9 systems. The first phase of this study was therefore to prepare a sequential series of peptides derived from the 161-amino acid dF 1 F 0 protein (Table 1). We then used these peptides in PKC overlay assays to determine which ones could antagonize the in vitro binding of PKC␦ to dF 1 F 0 . We reasoned that such a peptide would act as a competitive antagonist of PKC␦ binding to dF 1 F 0 . The peptide shown in Fig. 1B fulfilled these criteria and was selected for the development of a compound that could disrupt the PKC␦-dF 1 F 0 interaction in NCMs.
An unanticipated finding was that the dF 1 F 0 -derived peptide NH 2 -111 RVREYEKQLEKIKNMI 126 -COOH caused a dose-dependent potentiation of PKC␦ binding to dF 1 F 0 in overlay assays (Fig. 1C). The mechanism by which this occurs is at present unknown, but there have been reports of peptides derived from individual PKC isozymes that can enhance their binding to target proteins and facilitate their intracellular trafficking (47,48). For example, peptide sequences derived from individual PKC isozymes called pseudo-RACK sequences have been reported to accomplish this (34). In fact, expression of pseudo-RACK sequences for the PKC⑀ and -␦ isozymes into mouse hearts in vivo produces cardioprotection and exacerbates damage (50), respectively, following a severe IR challenge. Complete characterization of the mechanisms of the PKC␦-dF 1 F 0 facilitator peptide will require further study and is outside of the scope of the current work. However, we are very intrigued by this molecule because it not only has effects in in vitro PKC overlay assays (Fig. 1C), it also inhibits F 1 F 0 -ATPase activity (Fig. 3) and induces the PKC␦-dF 1 F 0 co-IP in NCMs (Fig. 4B).
The next phase of our study involved application of a technology that would allow high efficiency delivery of our PKC␦-dF 1 F 0 inhibitor and facilitator peptides into NCM mitochondria to determine whether they could modulate F 1 F 0 -ATPase activity. It has been previously reported by Dowdy and co-workers (51) that intraperitoneal injection of rodents with the ␤-galactosidase protein chemically coupled to the HIV-Tat protein transduction domain sequence, used in this study, allowed high efficiency delivery of the protein to virtually all tissues in the body of the animal. Our peptides presented an additional challenge in that the F 1 F 0 -ATP synthase exists in the mitochondrial IM. Therefore, to enhance or disrupt PKC␦ interaction with dF 1 F 0 , our modulatory peptides would not only need to enter cells, they would also need to be delivered to the IM where F 1 F 0 -ATPase exists. Of interest, the HIV-Tat protein transduction domain has previously been reported to carry the apoptosis repressor with caspase recruitment domain (ARC) protein into NCM mitochondria when the fusion protein was applied extracellularly (52). The ARC protein is ϳ40 kDa in size, which is considerably larger than any of the peptides we tested in this study. However, we used a cysteinecysteine chemical linkage (30,50) to attach the HIV-Tat protein transduction domain to our peptides, which supposedly allows the Tat protein transduction domain sequence to be cleaved off once the peptide enters the cell. We therefore felt it was necessary to include a mitochondrial targeting sequence (31) to our peptides.
It is well known that ϳ99% of mitochondrial proteins are encoded by nuclear genes and are translated in the cytosol before being transported into mitochondria. All proteins taken up by mitochondria are thought to have the mitochondrial targeting amino acid sequences located, for the most part, on their N terminus (53). This sequence binds to various chaperones and an elaborate array of mitochondrial protein carriers located in the outer and inner mitochondrial membranes (53). The mitochondrial targeting sequence we selected is one of the best studied and is derived from the number IV subunit of cyto-chrome oxidase (31). Our results indicate high efficiency mitochondrial uptake of our inhibitor, facilitator, and scrambled sequence peptides when NCMs are treated extracellularly with them (Fig. 4). In addition to their detection in mitochondria using FLAG immunoreactivity in Western blots, we found our peptides able to achieve sustainable levels in mitochondria suggesting they were not being completely proteolyzed or otherwise lost. In support of this, FLAG immunoreactivity in Western blots co-migrated with pure peptide alone and did not change in molecular weight (Fig. 2, A and B, lanes 1 versus 5). In addition, the cell-permeable inhibitor and facilitator peptides altered F 1 F 0 -ATPase activity and the PKC␦-dF 1 F 0 co-IP (Figs. [3][4][5] in opposing directions, whereas the scrambled sequence control peptide had no effect. There have been reports of nonspecific effects of HIV-Tat protein transduction sequences when introduced into cells (49). In general, toxicities were observed using much higher concentrations (20 -500 M) of the HIV-Tat protein transduction domain peptide than we used. In addition, our peptides were prepared as acetate salts (rather than the generally more toxic trifluoroacetate counter-ions) and using endotoxin-free water, which may also reduce nonspecific toxicities. Therefore, we do not believe that our peptides caused nonspecific actions in these studies. In support of this, our cell-permeable, mitochondrial targeted PKC␦-dF 1 F 0 inhibitor had similar effects to attenuate PKC␦ binding to dF 1 F 0 in in vitro overlay assays (Fig.  1D). We also developed a control peptide that contains all of the HIV-Tat, mitochondrially targeting, and FLAG domains as well as a scrambled amino acid sequence, derived from the inhibitor peptide (Fig. 1A). This scrambled sequence peptide was taken up into mitochondria, but it failed to modify basal or 4␤-PMAinduced changes in F 1 F 0 -ATPase activity (Fig. 3) or co-IPs (Fig.  4A). Finally, using Live/Dead vital staining (Invitrogen) and cardiac troponin I release, we observed no cytotoxicity following a 2-h incubation with up to 1 M extracellular concentrations of the peptides (data not shown).
Collectively, these results provide the first proof-of-principle evidence in support of the opposing mechanisms of the PKC␦-dF 1 F 0 inhibitor and facilitator peptides. We propose that these novel molecules will allow us to either attenuate or induce PKC␦-mediated inhibitory effects on the F 1 F 0 complex to assess the role(s) of this interaction in different cardiac responses. We are currently focusing on the utility of these novel peptides for reducing injury and improving energetics in adult rat models of IR injury in our laboratory.