Protein Kinase C Contributes to the Maintenance of Contractile Force in Human Ventricular Cardiomyocytes*

Prolonged Ca2+ stimulations often result in a decrease in contractile force of isolated, demembranated human ventricular cardiomyocytes, whereas intact cells are likely to be protected from this deterioration. We hypothesized that cytosolic protein kinase C (PKC) contributes to this protection. Prolonged contracture (10 min) of demembranated human cardiomyocytes at half-maximal Ca2+ resulted in a 37 ± 5% reduction of active force (p < 0.01), whereas no decrease (2 ± 3% increase) was observed in the presence of the cytosol (reconstituted myocytes). The PKC inhibitors GF 109203X and Gö 6976 (10μmol/liter) partially antagonized the cytosol-mediated protection (15 ± 5 and 9 ± 2% decrease in active force, p < 0.05). Quantitation of PKC isoform expression revealed the dominance of the Ca2+-dependent PKCα over PKCδ and PKCϵ (189 ± 31, 7 ± 3, and 7 ± 2 ng/mg protein, respectively). Ca2+ stimulations of reconstituted human cardiomyocytes resulted in the translocation of endogenous PKCα, but not PKCβ1, δ, and ϵ from the cytosol to the contractile system (PKCα association: control, 5 ± 3 arbitrary units; +Ca2+, 39 ± 8 arbitrary units; p < 0.01, EC50,Ca = 645 nmol/liter). One of the PKCα-binding proteins were identified as the thin filament regulatory protein cardiac troponin I (TnI). Finally, the Ca2+-dependent interaction between PKCα and TnI was confirmed using purified recombinant proteins (binding without Ca2+ was only 28 ± 18% of that with Ca2+). Our data suggest that PKCα translocates to the contractile system and anchors to TnI in a Ca2+-dependent manner in the human heart, contributing to the maintenance of contractile force.

Protein kinase C (PKC) 3 is a family of serine/threonine kinases (1). Multiple PKC isozymes are often expressed in the same cell, mediating specific functions. Conventional and novel PKCs can be activated by lipids, like the endogenous diacylglycerol (DAG) or the exogenous phorbol ester PMA. It was reported decades ago that PMA activation of PKC leads to the translocation of PKC from the soluble to the particulate fraction (2). This observation has been confirmed by later works, and some of the binding proteins for activated PKC isozymes were identified (receptors for activated C kinases) (3,4). Binding to its respective receptors for activated C kinase localizes each PKC isozyme in the vicinity of a subset of substrates and away from others, and hence this spatial organization may well explain the specificity of PKC isozymes in their intracellular signaling.
It is of interest that from the many PKC isozymes expressed in the heart (5), PKC␣ is the single isozyme that translocates to the contractile system upon Ca 2ϩ stimulation in the rat heart (6), suggesting a unique physiological role for PKC␣ in the Ca 2ϩ -dependent regulation of myofibrillar contractility. As a matter of fact, PKC␣ has been implicated in models of ischemic heart failure, myocardial hypertrophy, hypertension, and atherosclerosis (7). In addition, PKC-dependent phosphorylation of myofibrillar proteins such as desmin (8), myosin light chain (9), troponin I (TnI), and troponin T (TnT) (10) has been documented with suggested functional consequences ranging from changes in mechanical integrity of the cardiac sarcomere to decreased actin-myosin ATPase activity and force generation.
Long term activation of PKC is an essential step in ischemic preconditioning (11), although it is also well established that PKC overexpression contributes to heart failure (12)(13)(14)(15)(16)(17)(18) at least in part by enhanced phosphorylation of troponin proteins resulting in decreased cardiac contractility (12, 19 -25) in various animal models. In contrast, limited information is available on the possible role of PKC under physiological conditions in human heart and the possible PKC-binding proteins responsible for myofibrillar targeting of PKC.
Here we show that endogenous PKC␣ participates in the maintenance of contractile force upon long term Ca 2ϩ stimulations (a condition associated with ischemia-reperfusion) and identify the sarcomeric regulatory protein TnI as a PKC␣ targeting protein. Our data suggest a new physiological role for PKC␣ and pinpoints its intermolecular interaction with TnI as a possible pharmacological target to regulate in vivo PKC␣ activity in a sarcomer-specific manner.

EXPERIMENTAL PROCEDURES
Tissue Source-Healthy human hearts were obtained from five general organ donor patients (two men and three women with a mean age of 39.2 years) whose hearts were explanted to obtain pulmonary and aortic valves for valve replacement surgery (donor hearts). The donors did not show any sign of cardiac abnormalities and did not receive any medication except plasma volume expanders, dobutamine, and furosemide. The causes of death included cerebral contusion caused by accidents and cerebral hemorrhage and subarachnoidal hemorrhage caused by stroke. These experiments complied with the Helsinki Declaration of the World Medical Association and were approved by the Hungarian Ministry of Health (number 323-8/2005-1018EKU). Left ventricular wall samples were frozen in liquid nitrogen and stored at Ϫ80°C.
Myocyte Preparation and Measurement of the Mechanical Properties-Left ventricular tissue samples (about 0.2 g of wet weight) were treated and isolated as described earlier (26). In short, the tissue was mechanically disrupted in isolating solution (containing 4 mmol/liter Na 2 ATP, 1 mmol/liter Mg 2ϩ , 145 mmol/liter KCl, 2 mmol/liter EGTA, 10 mmol/liter imidazole, pH 7.0), and the particulate fraction was separated from the cytosol by centrifugation (1,000 ϫ g, 5 min; supernatant is referred as cytosol). The mechanically isolated cardiac myocytes in the particulate fraction were demembranated by Triton X-100 (0.3%, stirring on ice for 5 min) containing isolation solution. Then the demembranated myocytes were separated by centrifugation (1,000 ϫ g, 5 min; pellet: demembranated cardiomyocytes). The cytosol fraction was kept on ice until further experiments, and the myocytes were washed three times in cell isolation solution without Triton X-100 before measurements. Protein concentrations were determined by the BCA method (Sigma-Aldrich) using bovine serum albumin as a standard.
Measurement of Physiological Properties-Demembranated single cardiomyocytes were mounted between two thin needles with silicone adhesive (Dow Corning, Midland) while being viewed under an inverted microscope (Axiovert 135, Zeiss, Germany). One needle was attached to a force transducer element (SensoNor, Horten, Norway), and the other was attached to an electromagnetic motor (Aurora Scientific Inc., Aurora, Canada). The measurements were performed at 15°C, and the average sarcomere length was adjusted to 2.2 m as described previously (26). A mounted cardiomyocyte is shown on Fig. 1A from both vertical and horizontal directions. The pCa (Ϫlog[Ca 2ϩ ]) values of the relaxing and activating solutions (pH 7.2) were 9 and 4.75, respectively. Solutions with intermediate free Ca 2ϩ levels were obtained by mixing activating and relaxing solutions. All of the solutions for force measurements contained 1 mmol/liter Mg 2ϩ , 5 mmol/liter MgATP, 15 mmol/ liter phosphocreatine, and 100 mmol/liter BES. The ionic equivalent was adjusted to 150 mmol/liter with KCl. Isometric force was measured after the preparation had been transferred from the relaxing solution to a Ca 2ϩ -containing solution. When a steady force level was reached, the length of the myocyte was reduced by 20% within 2 ms and then quickly restretched (release-restretch maneuver). As a result, the force first dropped from the peak isometric level to zero (differ-ence ϭ total peak isometric force) and then started to redevelop. The passive force component was determined in relaxing solution following the Ca 2ϩ contractures. The Ca 2ϩ -activated isometric force was calculated as the difference of the passive and maximal active isometric force. Representative recordings are shown on Fig. 1 (B and C).
After the first maximal activation at pCa 4.75, resting sarcomere length was readjusted to 2.2 m, if necessary. Then cells were subsequently exposed to a series of solutions with intermediate pCa to construct a force-pCa relationship. After obtaining the first (control) force-pCa, the myocytes were incubated in the presence of a Ca 2ϩ concentration, which evoked ϳ50% of maximal force for 10 min and then the functional parameters (force-pCa relationship and passive tension) were measured again to reveal the effects of the Ca 2ϩ stimulation (representative experiments are shown on Fig. 1).
Ca 2ϩ -force relations were fitted to a modified Hill equation as described earlier (26), where F is the steady-state force, F 0 is the steady isometric force at saturating Ca 2ϩ concentration, the Hill coefficient nHill is a measure of the steepness of the relationship, and pCa 50 is the mid-point of the relation. The values are given as the means Ϯ S.E. for n myocytes. The differences were tested by means of Student's paired t test comparing the values before and after treatments. The level of significance was p Ͻ 0.05. Reconstitution of Human Cardiomyocytes-Human cardiomyocytes were reconstituted by adding the cytosol to the demembranated myocytes at a dilution of ϳ1 mg of myocyte protein/ml. In case of stimulation, reconstituted cardiomyocytes were incubated in the presence or absence of PMA (10 mol/liter), Ca 2ϩ and PKC inhibitors GF 109207X and Gö 6976 (10 mol/liter) for 10 or 30 min (physiological and biochemical assays, respectively). For the biochemical assays, the myocytes were pelleted and subsequently washed three times (centrifuged at 1,000 ϫ g for 2 min) in 450 l of cell isolation solution, containing the same concentration of Ca 2ϩ and PMA as during the incubations mentioned above. The pellets were then resuspended in 60 l of SDS-PAGE sample buffer (Sigma-Aldrich), and 30 l was subjected to SDS-PAGE and Western immunoblot analysis.
Western Immunoblot-Human left ventricular tissue samples were homogenized in radioimmune precipitation assay buffer containing 50 mmol/liter Tris-Cl, pH 7.4, 150 mmol/liter NaCl, 1% Triton X-100, 0.1% SDS, and 1% sodium deoxycholate. Protein concentrations were determined by BCA assay (Sigma) using bovine serum albumin as standard, and the concentration of the homogenates were adjusted to 4 mg/ml. Then the homogenates were mixed with equal volumes of 2ϫ SDS sample buffer (Sigma) and were boiled for 10 min before electrophoresis. 50 g of proteins was loaded onto 10% SDS-polyacrylamide gels and after separation transferred to nitrocellulose membranes. The membranes were probed with antibodies against PKC␣ (Sigma-Aldrich; dilution, 1:5 000) PKC␦ and PKC⑀ (both from Santa Cruz Biotechnology; dilution, 1:1 000), and the signal was detected by a peroxidase-conjugated antirabbit IgG-specific antibody (Sigma-Aldrich; dilution, 1:50 000). For the assessment of PKC isozyme expression in the human heart, recombinant human PKC isozymes were loaded on the same gels/membranes as control proteins (Sigma-Aldrich). Then the bands were visualized by ECL and evaluated by Image J software. Only bands on the same membrane were used to avoid alterations because of different transfer efficiency or other technical reasons.
Investigation of the Translocation of PKC Isozymes-Demembranated myocytes were incubated with cytosol (1 mg/ml concentration, 450 l/tube) in the presence or absence of Ca 2ϩ (5 mmol/liter) and PMA (10 mol/liter) for 30 min. Then the myocytes were washed three times with isolation solution (centrifugation by 1,000 ϫ g, 2 min; supernatants were discarded) containing the same concentration of Ca 2ϩ and PMA, as it was used during the incubations to avoid dissociation of the translocated proteins. After the last washing, the supernatants were carefully removed, and the myocytes (pellet) were boiled with 60 l of 2ϫ SDS sample buffer (Sigma) for 5 min. 30 l of this solution was subjected to Western blot analysis with antibodies against PKC␣, ␤1, ␦, and ⑀. The bands were visualized by peroxidase-conjugated secondary antibodies and ECL reaction. The intensities of the bands were quantitated by ImageJ, and the results are expressed as optical densities (arbitrary units).
Determination of Free Ca 2ϩ Concentrations in the Presence of Cytosol Fraction-In the case of measurement of Ca 2ϩ dependence of PKC␣ association (Fig. 4B), demembranated cardiomyocytes were incubated with the cytosol, and 0, 0.001, 0.01, 0.1, 1, 5, or 10 mmol/liter Ca 2ϩ was added to the supernatant. The free Ca 2ϩ concentrations were determined using a Fluo-3 (Sigma-Aldrich) calibration curve, measured by a BMG NOVOstar fluorescent plate reader (BMG Labtech, Offenburg, Germany).
The calculated values are shown on Fig. 6B.
Gel Overlay Assay for PKC␣-binding Proteins in Demembranated Myocytes-Proteins from isolated and demembranated cardiomyocytes were resolved by SDS-PAGE and transferred to nitrocellulose membranes. After blocking, the membranes were incubated with 2 g/ml purified, recombinant PKC␣ (Sigma-Aldrich) in the presence of 10 mol/liter PMA and 5 mmol/liter Ca 2ϩ in 1% milk powder containing TBS for 2 h, followed by three rinses with TBS, supplemented with 5 mmol/liter Ca 2ϩ . The membranes were then processed for Western blotting with an anti-PKC␣ antibody (Sigma-Aldrich; 1:20,000) as described above, except that all of the solutions were supplemented with 5 mmol/liter Ca 2ϩ . The controls were incubated with 1% milk powder containing TBS and 5 mmol/ liter Ca 2ϩ without the recombinant PKC␣.
In Vitro Binding of Recombinant Human PKC␣ to Recombinant Cardiac TnI or Troponin Complex-Recombinant, purified PKC␣ (0.1 g; Sigma-Aldrich) was incubated with recombinant, purified TnI (0.1 g/ml TnI) or reconstituted troponin complex (0.3 g/ml; 1:1:1 stoichiometry of TnI, TnC, and TnT) in the presence or absence of Ca 2ϩ (5 mmol/liter) for 2 h at room temperature. TnI was precipitated with an anti-TnI antibody (clone 16A11; Research Diagnostics Inc., Flanders, NJ) or with mouse IgG (control, Zymed Laboratories Inc., San Francisco, CA) and then pulled down with a protein A-Sepharose CL-4B resin (Amersham Biosciences). The precipitates were washed five times with TBS containing 0.1% Triton X-100 and the respective concentration of Ca 2ϩ (Fig. 7B). The precipitates were analyzed by Western immunoblot for PKC␣ as detailed above.

Effects of Endogenous Protein Kinase C on the Ca 2ϩ -activated
Contractile Force of Human Ventricular Cardiomyocytes-Incubation of demembranated human cardiac myocytes with a Ca 2ϩ concentration evoking about half-maximal force production (pCa 5.8, active force during the incubation: 62 Ϯ 1% of the maximum) for 10 min resulted in a 37 Ϯ 5% reduction in the maximal Ca 2ϩ -activated (active) force determined at pCa 4.75 (n ϭ 6, p Ͻ 0.05, representative experiment; Figs. 1B and 2). In contrast, demembranated cardiac myocytes stimulated with Ca 2ϩ (active force during the incubations: 56 Ϯ 8% of the maximum) in the presence of the cytosol were completely protected from the decrease in the contractile force (2 Ϯ 3% increase in the maximal active force, n ϭ 5, representative experiment; Figs. 1C and 2).
In parallel, the structural effects of the maintained contractions were also recorded as a means of light microscopy. Apparently, the maintained contractions evoked some deterioration of the cross-striation pattern of the myocytes, both in the presence and in the absence of cytosol (Fig. 1, B and C).
Next, the possible contribution of cytosolic proteins (probably with Ca 2ϩ -regulated activity) was investigated. From the many candidates, our attention has been attracted to PKC because of its ability (i) to regulate contractility, (ii) to translocate from the soluble fractions to the particulate fractions, and (iii) to be activated by Ca 2ϩ (classical PKC isoforms).
The administration of the general PKC activator PMA (10-min application) apparently affected neither the contractility (maximal active force: 96 Ϯ 1% of the maximal active force obtained before the treatment; Fig. 3, n ϭ 5), nor the cytosol-mediated protection upon prolonged Ca 2ϩ contractions (maximal active force: 98 Ϯ 1% of the maximal active force obtained before the treatment; Fig. 3, n ϭ 10).
In contrast, inhibition of PKC by GF 109203X and by Gö 6976 (10 mol/liter) partially antagonized the cytosol mediated protection (a decrease of active force by 15 Ϯ 5%, p Ͻ 0.05, n ϭ 6 and by 9 Ϯ 2%, p Ͻ 0.05, n ϭ 8, respectively, as shown on Fig. 4). In addition, control experiments with GF 109203X showed no effects on the active force in the absence of Ca 2ϩ incubations (2 Ϯ 3% increase in the maximal active force, n ϭ 5; Table 1).
Effects of Endogenous Protein Kinase C on the Passive Tension of Human Ventricular Cardiomyocytes-Isolated cardiac myocytes possess a Ca 2ϩ -independent passive tension (force) at a sarcomere length of 2.2 m. In contrast to the Ca 2ϩ -dependent (active) force, the passive force of the myocytes was elevated (passive force: 302 Ϯ 46%, p Ͻ 0.05, n ϭ 6; Fig. 5) after prolonged Ca 2ϩ contractions. The addition of cytosol significantly antagonized this elevation (passive force: 146 Ϯ 9%, p Ͻ 0.05, n ϭ 5; Fig. 5). As the matter of the contribution of PKC to these changes, PKC activation by PMA or inhibition by GF 109203X and by Gö 6976 was without effect on this parameter both in the absence and in the presence of Ca 2ϩ , as summarized in Table 1.

Effects of Endogenous Protein Kinase C on the Ca 2ϩ Sensitivity of Force Development in Human Ventricular Cardiomyocytes-
There was no apparent relationship between PKC activation (PMA) or inhibition (GF 109203X and Gö 6976) and the pCa 50 or Hill coefficient (summarized in Table 1). In general, pCa 50  Experiments similar to those in Fig. 1 were performed (n ϭ 6 for Ca 2ϩ alone (graph on the left) and n ϭ 5 for Ca 2ϩ ϩ cytosol (graph on the right) treatments). The means are represented by the symbols, and the S.E. is shown by the error bars. Note that in many cases the error bars are smaller than the size of the symbols; therefore they are overlapped (not visible). Significant differences between values obtained before (control) and after treatments are labeled by asterisks (p Ͻ 0.05).

Functional Targeting of PKC␣ to the Contractile System
values tended to be higher after the treatments, whereas the Hill coefficients were not affected.

Expression of Protein Kinase C Isozymes in Human
Ventricular Cardiomyocytes-The expression levels of three PKC isoforms (␣, ␦ and ⑀) were quantified in human left ventricular tissue samples. These assays suggested that PKC␣ is an abundant isoform in the human heart with an expression level about 20 times higher than those for PKC␦ or PKC⑀ (expression was 189 Ϯ 31, 7 Ϯ 3, and 7 Ϯ 2 ng/mg protein for PKC␣, PKC␦, and PKC⑀, respectively; n ϭ 4 for PKC␣ and n ϭ 3 for PKC␦ and PKC⑀; data not shown).
Phosphorylation of Human Ventricular Proteins by Protein Kinase C Isozymes-The ability of PKC isoforms to phosphorylate myofibrillar regulatory proteins was checked by in vitro phosphorylation experiments. The Ca 2ϩ -dependent PKC␣ and ␥ showed similar activity (4.3 Ϯ 0.9 and 4.4 Ϯ 2.1 pmol/min, respectively) and substrate specificity, whereas the Ca 2ϩ -independent isoforms produced a rather unique pattern. PKC␦ possessed the highest activity (6.6 Ϯ 3.3 pmol/min) and selectively phosphorylated a protein with a molecular mass of 26 kDa. PKC⑀ showed the lowest overall activity (1.6 Ϯ 0.3 pmol/min) and seemed to be specific to phosphorylate a protein with a molecular mass of 60 kDa. PKC (activity, 2.7 Ϯ 0.8 pmol/min) selectively phosphorylated two proteins with molecular masses of Ͼ200 and 48 kDa (data not shown). Overall, the expression and in vitro phosphorylation data suggest the dominance of PKC␣ over PKC␦ and ⑀ in the human heart (calculated relative activities (expression multiplied by the activity) are: 813, 46, and 11, respectively).
Intracellular Targeting of Protein Kinase C␣ in the Human Ventricular Cardiomyocytes-As a matter of in vivo regulation, experiments were performed to investigate the possible translocation of endogenous PKC␣, ␤1, ␦, and ⑀ from the cytosol to the contractile protein machinery. In accordance with a predominantly cytosolic localization under unstimulated conditions, only a low level of association of PKC isozymes were found in the absence of Ca 2ϩ (Fig. 6A, control). Importantly, in the presence of Ca 2ϩ , the level of PKC␣ bound to the contractile system was selectively increased (Fig. 6A). Interestingly, the widely employed PKC activator PMA alone was without significant effects on the interaction between PKC isozymes and myofilaments, although it evoked significant translocation when applied together with Ca 2ϩ (PKC␣ and ⑀), suggesting a decisive role for Ca 2ϩ in the regulation of translocation. When the Ca 2ϩ dependence of PKC␣ translocation was assayed, an EC 50 ϭ 645 nmol/liter for the free [Ca 2ϩ ] was obtained (Fig. 6B).
Ca 2ϩ -regulated Interaction of Human Cardiac TnI and Protein Kinase C␣-The interactions of PKC␣ and its putative anchoring proteins were further explored by overlay assays (Fig.  7A). Five potential PKC␣-binding proteins were found in the myofibrillar system of human cardiomyocytes. Moreover, one of the most prominent bands migrated with the thin filament regulatory protein TnI. In vitro binding assays, using purified, recombinant human TnI and purified recombinant PKC␣ verified the existence of the Ca 2ϩ -dependent interaction between these two molecules (Fig. 7, B and C). In the cardiac myocytes, TnI functions as a member of the troponin complex, consisting   Fig. 2 were performed in the presence of the PKC inhibitor Gö -6976 (n ϭ 8; left panel) and GF109203X (n ϭ 6; right panel). Note that in cases where error bars (S.E.) are not seen they are overlapped by the symbols (mean). Significant differences between values obtained before (control) and after treatments are labeled by asterisks (p Ͻ 0.05).

TABLE 1 Catalogue of values determined in the mechanical measurements
There are four parameters determined in the mechanical measurements: active force (force evoked by increased Ca 2ϩ concentrations), passive force (passive tension of the cardiomyocytes at 2.2-m sarcomere length), pCa 50 value (Ca 2ϩ sensitivity of the contraction), and Hill value (representing the cooperativity in the development of force upon Ca 2ϩ stimuli). These values are shown in respect to the applied treatments. The number of individual experiments are also shown. Absolute force values stand for the measured contractile force (in millinewtons) divided by the cross-sectional area (in mm 2 ) of the cardiomyocytes, whereas relative force values are the percentage of the response after treatments compared with the measured values before treatments.  0.05, paired t test) between the values obtained before and after treatments.

Functional Targeting of PKC␣ to the Contractile System
of TnT, TnC, and TnI in equimolar concentrations. Hence, the association of PKC␣ to TnI may be influenced by other constituents of the troponin complex. In vitro reconstitution of the troponin complex, however, did not affect the Ca 2ϩ -dependent interaction between TnI and PKC␣ (Fig. 7B).
Colocalization of Human Cardiac TnI and Protein Kinase C␣ in the Human Ventricle-In accordance with the biochemical data, we found significant colocalization of PKC␣ with TnI in human ventricular tissue samples, although the majority of PKC␣ was expressed in the cytosol (Fig. 8), consistent with its cytosolic location under low Ca 2ϩ conditions.
Our data add to the range of these effects, suggesting that PKC is also involved in the maintenance of contractile force in case of prolonged increased intracellular Ca 2ϩ concentrations, a condition that generally occurs upon ischemia-reperfusion (38). This beneficial effect of PKC was totally unexpected in light of the wealth of evidence for decreased contractility upon PKC activation. In particular, Belin et al. recently reported (12) that heart failure is accompanied by increased PKC-dependent myofibrillar protein phosphorylation and decreased contractility in the rat. Moreover, this effect of PKC was probably mediated by increased phosphorylation of troponin proteins (19). The differences in the experimental set-ups provide a plausible explanation for the controversial findings (increase versus decrease in contractility). In the earlier reports the effects of PKC were tested under conditions, where PKC activity was several times higher than the control, like heart failure or transgenic models (16,19,21,34), or phosphorylation of myofibrillar proteins by in vitro kinase treatments (12,20,37), or target proteins altered by site directed mutagenesis (20,22,24,25,35,36). In contrast we used reconstituted cardiomyocytes containing physiological levels of PKC isoforms, in addition to the endogenous mixture of myofibrillar substrates and the respective targeting proteins. It is therefore possible that the development of heart failure is characterized by a dysregulation of PKC pathway, leading to the disruption of its physiological targeting and pathological phospho-  rylation of myofibrillar proteins, which may conceal the physiological effects, revealed in this report (Fig. 9). It needs to be also noted, that PKC␣ not only regulates myofibrillar response to Ca 2ϩ but has a pivotal role in the regulation of intracellular Ca 2ϩ concentrations (14). Although this latter feature is important determinant of cardiac contractility, it was not investigated here.
In addition to the differences in the experimental conditions, a striking difference was found in the PKC isoform expression between rodent and human hearts. For example, whereas PKC⑀ seems to be the predominant isoform in rodent hearts (6, 7, 13, 14, 16, 39 -43), PKC␣ expression is the highest in human (17). In particular, we found about 20-fold higher PKC␣ expression than that of PKC␦ and PKC⑀ in donor human hearts. Moreover, although an increase in PKC expression is a hallmark of heart failure independently of the species, significant differences were found in the isoform expression pattern in human ventricular samples in end stage dilated cardiomyopathy and in severe aortic stenosis (44), again highlighting the complexity of PKC pathway in the heart.
In this context, PKC isoform selectivity can be achieved by at least three ways, namely (i) by selective expression, (ii) by different substrate specificities (iii), or by different targeting of the enzyme. Here an effort was made to investigate all of these possibilities.
As a matter of substrate selectivity, in vitro phosphorylation assays were performed on human ventricular myocytes and confirmed earlier data (4,31), suggesting a pivotal role for intracellular targeting in the determination of the apparent PKC isoform selectivity, in vivo. In particular, although differences in kinase activities and specificities were clearly recognized, all of the studied PKC isozymes were able to phosphorylate the main myofibrillar targets of PKC, such as TnI and TnT, in vitro.
In accordance, the intracellular targeting of PKC isozymes and in particular PKC␣ was investigated in detail. First of all, increase in Ca 2ϩ concentration selectively induced the translocation of PKC␣ to the particulate fraction in the rat heart (6), which was confirmed in this study using human ventricular tissue samples. This pinpointed PKC␣ as the most possible mediator of the sarcomeric effects, resulting in depression of contractility (12,19) or in protection of contractile force (this report), depending on the conditions.
Moreover, PKC␣ appears to be a promising therapeutic target to improve cardiac contractility (12,14,17). To exploit this potential, the ubiquitous expression and the importance of PKC␣ in various physiological processes should be taken into account. The optimal drug candidate would act selectively on the heart muscle located PKC␣ and specifically modulate its effects on Ca 2ϩ handling (14,17) or on thin filament-mediated regulation of sarcomeric sensitivity to Ca 2ϩ (12). Demembranated cardiomyocytes were tested for PKC␣-binding proteins using an overlay assay (A). Proteins of demembranated cardiomyocytes (20 g of protein/well) were separated and transferred to nitrocellulose membranes. The membranes were incubated with TBS (control) or with recombinant PKC␣ (2 g/ml in TBS; Sigma) in the presence of 5 mmol/liter Ca and 10 M PMA (as indicated). The binding of PKC␣ was detected by an anti-PKC␣ antibody. The relative position of troponin I and alfa-actinin were also determined by Western blot on parallel strips as indicated. The interaction between PKC␣ and the troponin I was confirmed in in vitro binding experiments (B and C). Human recombinant proteins (PKC␣ (0.1 g), troponin I (0.1 g), or reconstituted troponin complex (0.3 g/ml, 1:1:1 stoichiometry of troponin I, C, and T) were incubated in the presence or absence of Ca 2ϩ . The formed complexes were pulled down by a mouse IgG (1 g, background binding) or by a troponin I antibody (1 g; R & D, 16A11) and precipitated using protein A-Sepharose resin. The bound complexes were separated, and PKC␣ was visualized by anti-PKC␣ antibody. PKC␣-specific signal and the nonspecific crossreaction between the heavy and light chains of the antibodies used for precipitation are shown on B. The PKC␣-specific band intensities were analyzed by densitometry in case of PKC␣ and troponin I interactions (left three lanes on B). The graph represents the mean intensity of PKC␣ bands as the percentage of the maximal binding Ϯ S.E. from three separate experiments. FIGURE 8. PKC␣ and troponin I colocalization in human hearts. Human ventricular heart samples were sectioned in a cryostate at Ϫ20°C to obtain 10-m-thick slices. The slides were fixed with ice-cold acetone and blocked in normal goat serum. Rabbit anti-PKC-␣ (1:100 dilution) and mouse anti-troponin I (clone 22B11, dilution 1:100) were used as primary antibodies. Localizations of PKC␣ and troponin I were visualized by a fluorescence microscope and represented by green (PKC␣) and red (troponin I) colors on the figure. Overlap images were created by ImageJ software. No significant background staining was found in the same experiments performed by the omission of the primary antibody (data not shown).
We made an effort to reveal the molecular mechanism of PKC␣ targeting to sarcomeric protein machinery. The Ca 2ϩ concentration required for half-maximal translocation of PKC␣ to the contractile system (645 nmol/liter) was in the physiological range. Next, the sarcomeric PKC␣-anchoring proteins were investigated. The thin filamental regulatory TnI was identified as a potential binding protein. Importantly, the interaction between TnI and PKC␣ was modulated by Ca 2ϩ and was not affected by interaction of TnI with the other members (TnT and TnC) of the troponin complex. Moreover, immunohistochemical analysis revealed PKC␣ and TnI colocalization in the human ventricle. These data suggest that pharmacological modulation of TnI-PKC␣ interaction may be a strategy to regulate PKC␣ effects selectively on the sarcomeric proteins.
In the biochemical point of view, PKC-binding proteins are classified as substrates that interact with protein kinase C, receptors for inactive protein kinase C, and receptors for activated protein kinase C (4). In general, PKC activation involves the binding of DAG (or experimentally its stable analogue PMA) to the soluble PKC, which then anchors to membranous structures. Although TnI is apparently one of the substrates that interact with protein kinase C in the human heart, the interaction between TnI and PKC␣ seems to be regulated by Ca 2ϩ alone, independently of lipids. One of the proteins with similar properties is the sdr protein, which targets PKC␣ to the caveolae in a Ca 2ϩ -dependent manner, in the absence of DAG or its analogues (45). This finding suggested that Ca 2ϩ evokes a conformation change in PKC, revealing new interaction sites for nearby proteins. Analysis of the binding of PKC␣ to sdr suggested that although Ca 2ϩ facilitates the sdr-PKC interaction, it is stabilized by phosphatidyl serine. In contrast, we found an apparently stable interaction between TnI and PKC␣ in the absence of lipids, although it may be further stabilized by phosphatidyl serine, in vivo.
It was suggested that temporal and spatial changes in intracellular free Ca 2ϩ concentrations regulate the localization of PKC␣ in vascular smooth muscle cells (46). The same concept may be applied for ventricular cardiomyocytes. They have similar localized changes in the free intracellular Ca 2ϩ concentrations; moreover, they spontaneously produce a rise in intracellular Ca 2ϩ concentrations during the contractions. It is an exciting possibility that PKC␣ moves between the cytosol and the thin filament, where TnI temporarily anchors it to the thin filaments, during the contraction-relaxation cycle of cardiac myocytes and probably contributes to the maintenance of contractile force upon ischemia-reperfusion. On the other hand, dysregulation of PKC␣ expression/targeting apparently leads to pathological phosphorylation of sarcomeric proteins and to a decrease in contractility (12,19).
In summary, our data suggest that PKC plays a role in the maintenance of contractile force in human ventricular cardiomyocytes. The proposed mechanism of the PKC-mediated protection is that PKC␣ translocates to the contractile protein machinery in a Ca 2ϩ -dependent manner, where it is anchored to the TnI. A practical application of these findings may be the pharmacological modulation of PKC␣ targeting in ischemiareperfusion to improve human cardiac contractility.