Epac and Phospholipase Cϵ Regulate Ca2+ Release in the Heart by Activation of Protein Kinase Cϵ and Calcium-Calmodulin Kinase II*

Recently, we identified a novel signaling pathway involving Epac, Rap, and phospholipase C (PLC)ϵ that plays a critical role in maximal β-adrenergic receptor (βAR) stimulation of Ca2+-induced Ca2+ release (CICR) in cardiac myocytes. Here we demonstrate that PLCϵ phosphatidylinositol 4,5-bisphosphate hydrolytic activity and PLCϵ-stimulated Rap1 GEF activity are both required for PLCϵ-mediated enhancement of sarcoplasmic reticulum Ca2+ release and that PLCϵ significantly enhances Rap activation in response to βAR stimulation in the heart. Downstream of PLCϵ hydrolytic activity, pharmacological inhibition of PKC significantly inhibited both βAR- and Epac-stimulated increases in CICR in PLCϵ+/+ myocytes but had no effect in PLCϵ–/– myocytes. βAR and Epac activation caused membrane translocation of PKCϵ in PLCϵ+/+ but not PLCϵ–/– myocytes and small interfering RNA-mediated PKCϵ knockdown significantly inhibited both βAR and Epac-mediated CICR enhancement. Further downstream, the Ca2+/calmodulin-dependent protein kinase II (CamKII) inhibitor, KN93, inhibited βAR- and Epac-mediated CICR in PLCϵ+/+ but not PLCϵ–/– myocytes. Epac activation increased CamKII Thr286 phosphorylation and enhanced phosphorylation at CamKII phosphorylation sites on the ryanodine receptor (RyR2) (Ser2815) and phospholamban (Thr17) in a PKC-dependent manner. Perforated patch clamp experiments revealed that basal and βAR-stimulated peak L-type current density are similar in PLCϵ+/+ and PLCϵ–/– myocytes suggesting that control of sarcoplasmic reticulum Ca2+ release, rather than Ca2+ influx through L-type Ca2+ channels, is the target of regulation of a novel signal transduction pathway involving sequential activation of Epac, PLCϵ, PKCϵ, and CamKII downstream of βAR activation.

Phospholipase C (PLC) 3 -mediated hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP 2 ) results in inositol triphosphate (IP 3 )-mediated Ca 2ϩ release from intracellular stores and diacylglycerol-mediated activation of protein kinase C. This ubiquitous signaling pathway plays an integral role in regulating many physiological processes, including those of the cardiovascular system. PLC⑀ is a recently identified bifunctional PLC isoform that possesses both PIP 2 hydrolytic and Rap guanine nucleotide exchange factor (GEF) activity (1)(2)(3)(4). The activity of PLC⑀ is uniquely regulated by direct binding of small G-proteins including Ras, Rap, and Rho (5,6). PLC⑀ activity is also stimulated by the heterotrimeric G-protein subunits G␣ s , G␤␥, and G␣ 12/13 (5,7,8) but direct binding of these subunits to PLC⑀ has not been demonstrated. In primary astrocytes isolated from PLC⑀ ϩ/ϩ and PLC⑀ Ϫ/Ϫ mice, multiple G protein-dependent upstream signals rely critically on PLC⑀-dependent generation of IP 3 and diacylglycerol (9).
We recently discovered a surprising role for PLC⑀ regulation downstream of the ␤-adrenergic receptor (␤AR) in cardiac myocytes (10). Compared with normal mice, PLC⑀ Ϫ/Ϫ mice exhibit reduced left ventricular developed pressure in response to strong ␤AR stimulation (10). This deficit results from a decrease in isoproterenol (Iso)-dependent stimulation of electrically evoked Ca 2ϩ release from the sarcoplasmic reticulum (SR) in single ventricular cardiac myocytes. ␤AR stimulation increases cardiac Ca 2ϩ release in a cAMP/ protein kinase A (PKA)-dependent mechanism through phosphorylation of multiple targets of the cardiac excitability and Ca 2ϩ handling machinery (11). Recently, we identified a PKA-independent, PLC⑀-mediated pathway that contributes to maximal Iso-dependent enhancement of Ca 2ϩ -induced Ca 2ϩ release (CICR) in cardiac myocytes (12) * This work was supported, in whole or in part, by National Institutes of and explains the decreased ␤AR function in PLC⑀ Ϫ/Ϫ mice. This novel pathway requires cAMP-dependent activation of the RapGEF, Epac (13), which subsequently stimulates Rapdependent activation of PLC⑀.
Here, we establish a mechanistic link between PLC⑀ activity and CICR by showing that Epac/Rap/PLC⑀-mediated enhancement of CICR in the heart requires both PLC⑀-PIP 2 hydrolytic and PLC⑀-RapGEF activities and that downstream of PLC⑀, both PKC⑀ and CamKII are required for Epac-dependent enhancement of Ca 2ϩ release. In addition, voltage clamp experiments reveal that Iso-dependent activation of the Epac/PLC⑀ pathway in the heart does not significantly alter Ca 2ϩ influx through L-type calcium channels indicating that Ca 2ϩ release from the sarcoplasmic reticulum is the ultimate target of this pathway.
Transduction of AVM with Adenovirus-Adenoviruses were prepared using the AdEasy system (Stratagene) with the murine cytomegalovirus promoter used to drive expression of YFP, PLC⑀ wild type, PLC⑀H1460L, PLC⑀⌬CDC25, or PLC⑀K2150E. For wild type and domain mutant PLC⑀ adenoviruses, a second murine cytomegalovirus promoter was used to drive the expression of YFP. Adult AVM were isolated and adhered to laminin-coated coverslips for 2-h pre-infection. Plating media were removed and replaced with fresh media containing 300 multiplicity of infection of YFP control, wild type PLC⑀, or PLC⑀ domain mutant adenovirus. After 2 h, the virus was removed and fresh media were added to the cells. The appearance of YFP fluorescence was used to determine the percentage of cells transduced at 24 h post-infection. PLC message was measured by semiquantitative PCR and protein was detected by Western blotting.
PCR-For detection of PLC⑀ mRNA in PLC⑀ Ϫ/Ϫ AVMs transduced with either wild type or domain mutant PLC⑀ adenovirus constructs, total RNA was isolated using the RNAeasy mini kit (Qiagen, Inc., Valencia, CA) following the manufacturers recommendations. The Superscript III RT-PCR kit (Invitrogen) was used with 100 ng of total RNA template for reverse transcriptase-PCR with mouse PLC⑀ primers 5Ј-ACCCTGCGGTAAATGTTCTG-3Ј and 5Ј-ATG-TGAATTCCGTGCTACCC-3Ј to yield a 300-bp product. Glyceraldehyde-3-phosphate dehydrogenase primers 5Ј-CAA-CGGGAAGCCCATCACCAT-3Ј and 5Ј-CCTTGGCAGCAC-CAGTGGATGC-3Ј yielding a 350-bp product was used as con-trol. Reverse transcriptase was performed for 30 min at 42°C followed by incubation at 94°C for 2 min . The PCR parameters  were denaturation at 94°C for 30 s, annealing at 45°C for 45 s,  and extension at 72°C for 30 s. The number of PCR cycles was  30 for glyceraldehyde-3-phosphate dehydrogenase and 35 for  PLC⑀. Electrically Evoked Ca 2ϩ Transients-Electrically evoked Ca 2ϩ transients were measured as previously described (10). For each experiment, data were collected in the absence of agonist for 5-15 cells to determine naïve Ca 2ϩ transient amplitude. 1 M Isoproterenol and 10 M cpTOME were prepared in control Ringer solution (145 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 and 10 mM Hepes, pH 7.4) and locally perfused for 20 s followed by 60 s of electrical stimulation (20 ms, 8 V, 0.5-1 Hz, 60 s) in the continued presence of agonist.
Pharmacological Inhibition of PKC, IP 3 Receptors, and CamKII-To determine the effect of PKC inhibition on electrically evoked Ca 2ϩ transient amplitude, cells were pre-treated for 5 min with 1 M bisindolylmaleimide-1 (BIM) (Calbiochem), a broad specificity PKC inhibitor, followed by constant perfusion of BIM in the presence of 1 M Iso or 10 M 8-4-(chlorophenylthio)-2Ј-O-methyladenosine-3Ј,5Ј-monophosphate (cpTOME). For IP 3 receptor inhibition, cells were pretreated with 20 M 2-aminoethoxydiphenyl borate (2-APB) for 5 min followed by constant perfusion of 2-APB in the presence of 1 M Iso. For CamKII inhibition, cells were pretreated with 1 M KN93 (or inactive KN92) for 30 min followed by constant perfusion of KN93 in the presence of 1 M Iso or 10 M cpTOME.
PKC Translocation-AVM were isolated as described and plated at a density of 50,000 cells/60-mm tissue culture dish in serum-free minimal essential medium culture supplemented with 2 mM L-glutamine, 1% penicillin/streptomycin, and 2.5 M blebbistatin. After 2 h, the media were changed to remove dead cells and debris. Cells were treated with 1 M Iso for 30 s or 10 M cpTOME for 3 min in a 37°C incubator. Following treatment, dishes were placed immediately on ice and media and agonist were removed. Cells were washed two times with icecold phosphate-buffered saline supplemented with protease inhibitors. Cells were scraped into a lysis buffer (50 mM Hepes, pH 8.0, 3 mM MgCl 2 , 100 M EDTA, 100 mM NaCl, 50 M NaVO 4 , and protease inhibitors) and probe sonicated. Samples were then centrifuged at 100,000 ϫ g at 4°C for 15 min to pellet the membrane fraction. The membrane pellet was washed two times with ice-cold lysis buffer, and then re-suspended in sample buffer (125 mM Tris-HCl, pH 6.8, 20% glycerol, 4% SDS, 10% ␤-mercaptoethanol (1.42 M), and 0.25% bromphenol blue). Samples were resolved by 10% SDS-PAGE and Western blotted for specific PKC isoforms. PKC⑀ and PKC␣ antibodies (Santa Cruz) were used at a 1:1000 dilution. Horseradish peroxidaseconjugated anti-rabbit IgG (Bio-Rad) was used at 1:1000.
PKC siRNA-PKC⑀-specific and CY3-labeled negative control siRNAs (Ambion) were reconstituted at 100 pM. Wild type AVM were isolated as before and media were changed prior to transfection. For each siRNA, 600 pmol was added to 600 l of Opti-MEM. In a separate tube, 6 l of Lipofectamine 2000 was added to 600 l of Opti-MEM. After 5 min, siRNA and Lipofectamine tubes were mixed and incubated at room tempera-ture for 20 min. The 200 pmol of the siRNA mixture was then added to each 35 mM dish of AVM. Efficiency of transfection was determined by fluorescence microscopy of AVM transfected with CY3-labeled negative control siRNA. Cells transfected with PKC⑀-specific siRNA or negative control siRNA were harvested in sample buffer at 24, 36, or 48 h post-transfection. Knockdown of PKC⑀ protein was determined by quantitative Western blotting.
Perforated Patch Clamp-Briefly, individual AVM adhered to laminin-coated dishes were preloaded with 5 M fluo-4 AM for 30 min at 37°C in a Ringer solution. Myocytes were then washed 2 times with an external Ca 2ϩ current recording solution containing: 140 mM tetraethylammonium-Cl, 2 mM MgCl 2 , 1.8 mM CaCl 2 , 0.005 mM blebbistatin, 10 mM glucose, and 10 mM Hepes, pH 7.4. Patch clamp electrodes had a pipette resistance of 1-2 megohms when backfilled with internal solution containing: 135 mM CsCl, 1 mM MgCl 2 , and 10 mM Hepes, pH 7.2. Ca 2ϩ currents and transients were elicited using 200-ms test pulses from Ϫ50 to ϩ70 mV in 10-mV increments delivered at 10-s intervals (0.1 Hz). Peak L-type Ca 2ϩ current magnitude was normalized to total cell capacitance (pA/pF), plotted as a function of membrane potential (V m ), and fitted according to Equation 1, where G max is the maximal L-channel conductance, V m is the test potential, V G1/2 is the voltage of half-maximal activation of G max , V rev is the extrapolated reversal potential, and k G is a slope factor. The kinetics of Ca 2ϩ current inactivation was described by fitting the inactivation phase to the following single exponential function using Equation 2, where I(t) is the current at time t after the depolarization, A is the amplitude of the inactivating component of current, inact is the time constant of inactivation, and C represents the steadystate non-inactivating component of current. Ca 2ϩ transients recorded during each test pulse were expressed as ⌬F/F, where F represents baseline fluorescence and ⌬F represents the fluorescence change from baseline.
Rap and Ras Activation-Hearts were excised from 4 -6month-old PLC⑀ ϩ/ϩ or PLC⑀ Ϫ/Ϫ mice, cannulated through the aorta, perfused in the presence or absence of 1 M isoproterenol for 10 min, and snap frozen in liquid nitrogen. Heart lysates were prepared by Polytron in a buffer containing 50 mM Tris-HCl, pH 7.4, 500 mM NaCl, 2.5 mM MgCl 2 , 1% Nonidet P-40 in 10% glycerol, and protease inhibitors. 1 mg of heart lysate was incubated with GST-tagged fusion protein corresponding to amino acids 788 -884 of human RalGDS-Rap-GTP binding domain or 1-149 of human Raf-1-Ras GTP binding domain bound to glutathione-agarose in heart lysis buffer. Following incubation, beads were harvested by centrifugation, the supernatant removed, and beads were extensively washed with lysis buffer. After washing, beads were pelleted by centrifugation, resuspended in 4ϫ SDS-sample buffer, resolved on a 15% polyacrylamide gel, and transferred to nitrocellulose for Western blotting.
Statistics-Data are given as mean Ϯ S.E. Statistical significance was determined using unpaired Student's t test and a one-way analysis of variance (ANOVA) for multiple comparisons. Differences were considered statistically significant at p Ͻ 0.05.

RESULTS
To determine the relative roles of PLC⑀-RapGEF and PLC activities in the regulation of cardiac Ca 2ϩ handling, we transduced freshly isolated PLC⑀ Ϫ/Ϫ myocytes with adenoviruses directing expression of either wild type PLC⑀ or mutants of PLC⑀ previously shown to eliminate PLC hydrolytic activity (PLC⑀H1460L) (3), RapGEF activity (PLC⑀⌬CDC25), or Rap (and other small GTPases) binding to the RA2 domain (PLC⑀K2150E) (1) (Fig. 1A). Based on PCR analysis of PLC⑀ transcripts, all constructs were expressed to similar levels 24 h after transduction (Fig. 1B). Western blots of extracts from AVM infected with the PLC⑀ mutant viruses indicate that the mutations do not affect expression of PLC⑀ (Fig. 1C). Electrically evoked Ca 2ϩ transients in transduced myocytes were then assessed in the presence or absence of either the ␤AR agonist isoproterenol (1 M) or the direct Epac activator cpTOME (10 M). As previously observed (12), isoproterenol-dependent enhancement of electrically evoked Ca 2ϩ release was significantly increased 24 h after transduction of wild type PLC⑀ in PLC⑀ Ϫ/Ϫ AVM (Fig. 1D). Wild type PLC⑀ expression also restored the 2-fold increase in evoked Ca 2ϩ transient amplitude in response to Epac activation with cpTOME ( Fig. 1E). In contrast, PLC⑀ Ϫ/Ϫ AVM transduced with either PLC⑀H1460L or PLC⑀K2150E failed to respond to cpTOME and showed no increase in isoproterenol responsiveness compared with YFP control (Fig. 1, C-E) (supplemental Fig. S1). These results confirm that direct stimulation of PLC⑀ hydrolytic activity by binding of a Ras family GTPase (likely Rap1) to the PLC⑀ RA2 domain is required for maximal ␤AR-mediated increases in electrically evoked SR Ca 2ϩ release. PLC⑀ Ϫ/Ϫ AVM transduced with PLC⑀⌬CDC25, the RapGEF-deficient PLC⑀ mutant, failed to exhibit increased responsiveness to either isoproterenol or cpTOME (Fig. 1, D and E) (supplemental Fig. S1), indicating that PLC⑀-RapGEF activity is also required for the proper execution of the Epac/PLC⑀ pathway during ␤AR-mediated regulation of SR Ca 2ϩ release. We have previously shown that the Rap GEF deletion, PLC⑀⌬CDC25, does not significantly affect intrinsic PIP 2 hydrolysis activity (9).
The inability of a RapGEF-deficient mutant of PLC⑀ to rescue cpTOME and maximal isoproterenol-stimulated enhancement of CICR in PLC⑀ Ϫ/Ϫ AVM suggests that Rap activation downstream of ␤AR stimulation is at least partially dependent on PLC⑀. Lysates prepared from control and Iso-perfused hearts of PLC⑀ Ϫ/Ϫ or PLC⑀ ϩ/ϩ mice were analyzed for acti-vated Rap (RapGTP) by pulldown with GST-RalGDS and activated Ras (RasGTP) by pulldown with GST-Raf-1-RBD. A significant increase in active Rap over basal levels was observed in lysates from Iso-treated PLC⑀ ϩ/ϩ myocytes (Fig. 2, left). In lysates prepared from PLC⑀ Ϫ/Ϫ mice, detectable Rap activation was not observed under either basal or Iso-treated conditions (Fig. 2, right). On the other hand, similar levels of Ras activation were observed under basal and Iso-treated conditions in hearts from both PLC⑀ ϩ/ϩ and PLC⑀ Ϫ/Ϫ mice. Total Rap and Ras levels were identical between PLC⑀ Ϫ/Ϫ and wild type heart lysates. Together, these results are consistent with previous in vitro findings that PLC⑀ acts specifically as a GEF for Rap, but not Ras (2).
PLC-mediated hydrolysis of PIP 2 results in generation of IP 3 and diacylglycerol and the subsequent activation of Ca 2ϩ release through IP 3 receptors in the sarcoplasmic reticulum and/or PKC activation, respectively. A definitive role for IP 3 -mediated Ca 2ϩ release in EC coupling in cardiac myocytes has not been identified despite intensive investigation (14). Type 2 IP 3 receptors are the predominant IP 3 R isoform present in the heart. Analysis of type 2 IP 3 R knock-out mice indicates that the type 2 IP 3 receptor is not required for ␤AR enhancement of Ca 2ϩ release in atrial cardiac myocytes, but is important for endothelin-dependent regulation of Ca 2ϩ release (15). To assess the potential role of IP 3 in our cells, we compared electrically evoked Ca 2ϩ release in control, Iso-, and cpTOME-treated wild type AVM after treatment with the IP 3 receptor inhibitor, 2-APB (20 M). 2-APB treatment did not alter either the Iso (Fig. 3A) or cpTOME (data not shown) responsiveness, supporting conclusions of previous studies that IP 3 receptors do not directly contribute to Iso-dependent enhancement of CICR.
To test for PKC involvement in the ␤AR and Epac responses, electrically evoked Ca 2ϩ transients in the presence or absence of isoproterenol were compared in AVM pretreated with 1 M BIM, a broad specificity PKC inhibitor. BIM treatment significantly inhibited isoproterenol-stimulated enhancement of SR-Ca 2ϩ release in wild type AVM (Fig. 3B, left). In addition, BIM treatment abolished the increase in electrically evoked Ca 2ϩ transient amplitude observed following direct activation of Epac with cpTOME (Fig. 3C). However, BIM treatment did not alter Iso stimulation of electrically evoked Ca 2ϩ transients in PLC⑀ Ϫ/Ϫ myocytes (Fig. 3B, right) or baseline evoked transients (data not shown). These results indicate that the effects of BIM are specific to the PLC⑀-dependent pathway downstream of ␤AR stimulation and implicate PKC activation in this pathway.
There are 11 distinct isoforms of PKC, four of which are consistently detected in cardiac myocytes: ␣, ␤ II , ␦, and ⑀ (16). To determine whether a specific isoform of PKC is activated downstream of PLC⑀, we monitored translocation of specific PKC isoforms to the particulate fraction following treatment of freshly isolated PLC⑀ ϩ/ϩ and PLC⑀ Ϫ/Ϫ cardiac myocytes with either isoproterenol or cpTOME. Western blot analysis of the particulate fraction of wild type cardiac myocytes revealed a specific increase in PKC⑀ in the membrane fraction in response to both isoproterenol and cpTOME treatment relative to non-treated control (Fig. 4A). In contrast, PKC␣ did not translocate to the membrane fraction in response to isoproterenol. In PLC⑀ Ϫ/Ϫ AVM, neither isoproterenol nor cpTOME triggered translocation of PKC⑀ to the membrane (if anything a small decrease in PKC⑀ at the membrane was observed), placing PKC⑀ downstream of the ␤AR/Epac/Rap/PLC⑀ pathway. To further test the role of PKC⑀ downstream of the Epac/ Rap/PLC⑀ pathway, wild type AVM were transfected with either PKC⑀-specific siRNA or a CY3-labeled negative control siRNA. Transfection efficiency was nearly 100% (supplemental Fig. S2) and Western blot analysis revealed that PKC⑀ (but not PKC␣) protein levels were knocked down by at least 95% at 36 h post-transfection (Fig. 4B). PKC⑀ protein levels were not substantially decreased in cells transfected with the negative control siRNA at all time points monitored (24, 36, and 48 h). Electrically evoked Ca 2ϩ transients were monitored 36 h after transfection of wild type AVM with either PKC⑀ or negative control siRNAs. Baseline electrically evoked Ca 2ϩ transients were not different between control and PKC⑀ siRNA-treated AVM. On the other hand, peak Ca 2ϩ transient amplitude in the presence of isoproterenol was significantly inhibited and cpTOME responses abolished (Fig. 4C) in AVM treated with PKC⑀ siRNA. These data are consistent with results obtained following pharmacological PKC inhibition with BIM (Fig. 3, B and C) and Iso/cpTOME-dependent PKC⑀ membrane translocation (Fig. 4A), indicating that PKC acts downstream of PLC⑀ and specifically implicates PKC⑀ as the relevant PKC isoform involved.
A recent report demonstrated that CamKII is activated following Epac stimulation with cpTOME in rat cardiac myocytes, however, the mechanism for CamKII activation by Epac was not determined (17). PKC has also been shown to activate CamKII in rat ventricular myocytes (18) and CamKII is directly phosphorylated at Thr 286 by PKC in vitro (19). Therefore, we determined if CamKII activation is required for Iso-and Epacdependent enhancement of CICR, and if it is downstream of PKC. Iso-and cpTOME-induced enhancement of electrically evoked Ca 2ϩ transients in wild type AVM were determined in the absence and presence of KN93, a specific CamKII inhibitor. KN93, but not the control compound KN92, attenuated Isoinduced enhancement of evoked release (Fig. 5A) and completely blocked cpTOME-induced enhancement (Fig. 5B). Cotreatment with BIM and KN93 did not further diminish the response to isoproterenol relative to treatment with either compound alone (data not shown). Additionally, KN93 had no effect on the Iso response in PLC⑀ Ϫ/Ϫ AVM (Fig. 5A) or on baseline evoked transients (data not shown), supporting specific involvement of CamKII in the PLC⑀-dependent pathway.
To determine whether CamKII activation was dependent on PKC, wild type cardiac myocytes were treated with either isoproterenol or cpTOME alone or in the presence of BIM and CamKII phosphorylation at Thr 286 was measured by Western blotting (Fig. 5C). Both isoproterenol and cpTOME treatment increased CamKII phosphorylation at Thr 286 relative to nontreated control. The cpTOME-dependent increase was blocked in the presence of BIM and the Iso-dependent increase was partially blocked by BIM. These data support the conclusion that the Epac/PLC⑀ pathway can control CamKII activation in a PKC⑀-dependent, Ca 2ϩ -independent manner. That the Iso-dependent increase in phosphorylation was only partially blocked suggests that there are multiple mechanisms for CamKII activation downstream of Iso, one of which includes the Ca 2ϩindependent Epac and PKC pathway, but may also result from changes in Ca 2ϩ that occur with Iso-dependent regulation of PKA.
CamKII phosphorylates numerous Ca 2ϩ handling proteins, including the L-type Ca 2ϩ channel, RyR2 and PLB, involved in precisely controlling dynamic changes in intracellular calcium levels during the cardiac cycle (20). CamKII-dependent modulation of RyR function by phosphorylation at Ser 2815 has been implicated as a means for positive regulation of SR-Ca 2ϩ release downstream of ␤AR stimulation. CamKII also phosphorylates PLB at Thr 17 to stimulate SR Ca 2ϩ reuptake to increase content available for release. To determine whether Epac/PLC⑀ stimulation results in a PKC-dependent, CamKII-mediated, phosphorylation of RyR2 and/or PLB, AVM were treated with cpTOME in the presence or absence of the PKC inhibitor BIM. Phosphorylation at the CamKII-specific sites was measured by Western blotting (Fig. 5D). cpTOME treatment significantly enhanced phosphorylation of RyR2 at Ser 2815 and PLB at Thr 17 in wild type AVM. These increases were ablated in AVM pretreated with BIM. These data identify at least two effector targets (RyR2 and PLB) of the Epac/ PLC⑀ pathway that could be involved in regulating the magnitude of CICR.
To examine the relationship between Ca 2ϩ influx and release during EC coupling we conducted voltage clamp experiments to determine whether the Epac/PLC⑀ pathway alters the properties of depolarization-induced L-type Ca ϩ channel function and RyR2-mediated Ca 2ϩ release. Perforated patch clamp experiments in fluo-4-loaded AVM were conducted to simultaneously compare the voltage dependence, magnitude, and kinetics of L-type Ca 2ϩ currents and global intracellular Ca 2ϩ transients in wild type and PLC⑀ Ϫ/Ϫ myocytes before and after ␤AR activation ( Fig. 6 and supplemental Fig. S3). L-type Ca 2ϩ currents and intracellular Ca 2ϩ transients were elicited by 200-ms test pulses from Ϫ50 to ϩ70mV at 10-mV increments. No differences in the magnitude or kinetics of L-type Ca 2ϩ currents (Fig. 6A) or global intracellular Ca 2ϩ transients (Fig. 6C) were observed between PLC⑀ Ϫ/Ϫ and PLC⑀ ϩ/ϩ AVM under basal conditions (closed symbols) ( Table 1)  . PLC⑀-dependent enhancement of CICR requires specific activation of PKC⑀. A, left, PKC⑀ translocates to the membrane fraction following treatment with 1 M isoproterenol (30 s) or 10 M cpTOME (3 min) in PLC⑀ ϩ/ϩ but not PLC⑀ Ϫ/Ϫ cardiac myocytes. PKC␣ does not translocate to the membrane in response to ␤AR stimulation. PMA treatment (500 nM, 10 min) was used as a positive control for PKC translocation. 3 g of cardiac myocyte membrane fractions was analyzed for PKC isoform translocation. G␤ subunit was used as a loading control. Right, densitometric quantitation of PKC⑀ membrane translocation from cells isolated from 5 PLC⑀ ϩ/ϩ and 3 PLC⑀ Ϫ/Ϫ mice. Data are represented as a percentage of maximal translocation evoked by PMA treatment. **, p Ͻ 0.01; ***, p Ͻ 0.001; ns, not significant as compared with nontreated PLC⑀ ϩ/ϩ cells. One-way ANOVA, Bonferroni post-test. B, left, PLC⑀ ϩ/ϩ cardiac myocytes were transfected with PKC⑀-specific siRNA or a CY3-labeled negative control siRNA. PKC⑀ protein levels are nearly completely knocked down in cardiac myocytes transfected with PKC⑀-specific siRNA relative to negative control siRNA at 36 h post-transfection. Lower left, PKC␣ protein levels are not significantly affected by PKC⑀ siRNA 36 h post-transfection. Right, densitometric quantitation of PKC⑀ protein expression from myocytes transfected with either PKC⑀ siRNA or Cy3-labeled negative control siRNA pooled from three separate experiments. C, knockdown of PKC⑀ significantly decreases isoproterenol-induced enhancement of CICR and completely eliminates cpTOME responsiveness in PLC⑀ ϩ/ϩ cardiac myocytes. Data are pooled Ca 2ϩ transient amplitudes (⌬404/485) from 20 to 40 cells per condition, n ϭ 3 mice. Results are average (ϮS.E.); ***, p Ͻ 0.001, one-way ANOVA, Bonferroni post-test.
L-type Ca 2ϩ current density (Fig. 6A) and maximum channel conductance ( Fig. 6B and Table 1) in both PLC⑀ Ϫ/Ϫ and PLC⑀ ϩ/ϩ AVM. However, peak Iso-stimulated Ca 2ϩ transient amplitude was significantly attenuated in PLC⑀ Ϫ/Ϫ cardiac myocytes (Fig. 6, C and D, and Table 1), consistent with results observed in intact myocytes (Figs. 3B and 5A). In addition, the kinetics of Ca 2ϩ current inactivation was significantly slower in myocytes from PLC⑀ Ϫ/Ϫ mice (supplemental Fig. S3). These data indicate that alterations in action potential or L-type channel activity are not necessary PLC⑀dependent regulation of CICR.

DISCUSSION
PLC⑀ is unique among PLC enzymes in that it possesses both phospholipase C and RapGEF activities. Physiological roles for both catalytic functions of PLC⑀ are beginning to emerge (9,10,21,22). Here we demonstrate that both PLC⑀ hydrolytic and RapGEF activities are required for maximal ␤AR-mediated (and Epac-dependent) enhancement of electrically evoked Ca 2ϩ release in the heart. We hypothesize that the PLC⑀-RapGEF activity ensures sufficient Rap activation to maintain PLC⑀ hydrolytic activity, and that PLC⑀ is required for sustained Rap activation in the heart. Epac stimulation by cAMP may initiate a low level of Rap activation that (undetectable in the PLC⑀ Ϫ/Ϫ myocytes), in turn, stimulates PLC⑀ to significantly amplify Rap activation that feeds forward to further stimulate PLC⑀ and subsequent regulation of CICR. This model is supported by previous studies in transfected cells demonstrating that PLC⑀ potentiates its own activation by RapGTP (2) and in primary astrocytes where PLC⑀ RapGEF activity is required for sustained Rap activation and downstream ERK signaling (9). In addition, it is important to note that Rap generated from PLC⑀ may also regulate other enzymes in the heart such as ERK5 where Rap-mediated inhibition protects against the development of hypertrophy (23). This would be consistent with our findings that PLC⑀ Ϫ/Ϫ mice exhibit increased susceptibility to stress-induced hypertrophy (10) and that PLC⑀ RapGEF activity modulates ERK signaling in astrocytes (9).
Downstream of PLC⑀ activity, we demonstrate that diacylglycerol-mediated PKC⑀ activation is required for maximal ␤ARdependent enhancement of CICR (Fig. 3). Several reports have suggested that PKC activity modulates CICR in cardiac myocytes. Treatment with norepinephrine or phorbol myristic acid causes PKC⑀ translocation to cross-striated t-tubular regions of cardiac myocytes upon activation, strategically placing this enzyme in position to phosphorylate proteins involved in Ca 2ϩ handling (24). PKC␦ and PKC⑀ have been shown to mediate positive inotropy that is dependent on subcellular localization (25). Our data are consistent with a positive ionotropic effect of  PKC⑀ and indicate that PLC⑀ is an upstream regulator of PKC⑀ in the heart. siRNA knockdown of PKC⑀ strongly suppressed cpTOME-dependent increases in calcium transient amplitudes suggesting that PKC⑀ is the major isoform of PKC involved in Epac/PLC⑀-dependent responses. On the other hand, the inhibition of Iso-dependent responses by PKC⑀ siRNA appeared less than observed with BIM treatment (Figs. 3B versus 4C). One possibility is that another PKC isoform such as PKC␦ is involved in the Iso-dependent response or that a more complete inhibition of PKC⑀ that might be achievable with BIM may be required to fully inhibit the Iso-response.
We also identified PKC-dependent CamKII regulation as an essential downstream component of the Epac/PLC⑀ pathway in AVM. ␤AR and cpTOME enhancement of electrically evoked Ca 2ϩ release was suppressed by PKC inhibition with BIM (Fig.  3, B and C), PKC⑀ knockdown (Fig. 4C), and CamKII inhibition with K93 (Fig. 5) in wild-type PLC⑀ ϩ/ϩ AVM. On the other hand, neither PKC inhibition, CamKII inhibition, nor the combination, had any significant effect on the already reduced Isodependent regulation of evoked release in PLC⑀ Ϫ/Ϫ AVM, indicating that both PKC⑀ and CamKII are downstream from PLC⑀ activation. We also show that phosphorylation of CamKII at Thr 286 is increased by cpTOME in a PKC-dependent manner. These data are consistent with a previous report demonstrating that activation of Epac stimulates CamKII Thr 286 phosphorylation (17) but extend this result to show that Epac-dependent CamKII activation relies on PLC⑀-dependent PKC⑀ activity. Although a mechanism for PKC-dependent activation of CamKII has not been clearly delineated, CamKII Thr 286 has been shown to be directly phosphorylated by PKC in vitro (19). Thr 286 autophosphorylation results in tight binding of calmodulin such that Ca 2ϩ is no longer required for activation. If CamKII Thr 286 is directly phosphorylated by PKC it should also result in Ca 2ϩ -independent regulation. Physiological evidence for PKC-dependent regulation of CamKII is sparse but two previous studies have shown that ␣-adrenergic receptor stimulation facilitates PKC-mediated activation of CamKII (18,26). PKC⑀ may not directly phosphorylate CamKII in cardiac myocytes in response to Epac/PLC⑀ activation but we clearly demonstrate that PKC activation is upstream of CamKII Thr 286 phosphorylation in this pathway. It is also possible that PKC⑀ itself phosphorylates components of the Ca 2ϩ handling machinery, but our data indicate that any such activity alone is insufficient because CamKII inhibition with KN93 completely blocks cpTOME-dependent enhancement of Ca 2ϩ transient amplitudes (Fig. 5B). This indicates that PKC activation by this pathway is not sufficient to cause increases in Ca 2ϩ transients. Nevertheless, it remains formally possible that local Ca 2ϩ release, dependent on PKC activation, could lead to CamKII autophosphorylation. We propose that in cardiac myocytes, linear activation of Epac-PLC⑀-PKC⑀-CamKII mediates a component of Iso-dependent regulation of evoked Ca 2ϩ release in cardiac myocytes.
Potential targets downstream of this pathway include the L-type Ca 2ϩ channel, RyR2, and phospholamban. PLC⑀ ablation did not markedly affect isoproterenol-stimulated increases in L-type Ca 2ϩ channel current density in perforated patch clamp experiments (Fig. 6, A and B) and L-type Ca 2ϩ channel activity was not significantly altered by 20 M cpTOME (data not shown). In the same cells Iso-stimulated enhancement of depolarization-induced Ca 2ϩ transients was significantly attenuated (Fig. 6, C and D). The fact that Ca 2ϩ release elicited by a uniform voltage clamp pulse is reduced in PLC⑀ Ϫ/Ϫ myocytes, suggests that changes in action potential waveform are not likely responsible for the reduction in Iso responsiveness. Together, these data indicate that the Epac/PLC⑀/PKC⑀/ CamKII pathway contributes enhanced release of Ca 2ϩ from the SR during ␤AR stimulation that is not dependent on changes in Ca 2ϩ influx. Two proteins that control release of Ca 2ϩ from the SR, RyR2 and PLB, are phosphorylated at CamKII-specific sites in response to Epac stimulation, supporting this idea. The observed increase in Ca 2ϩ release in response to a uniform Ca 2ϩ influx could arise from a combination of both an increase in RyR2 sensitivity to activation by Ca 2ϩ influx (27,28) and an increase in Ca 2ϩ reuptake and load. The observed increase in Ca 2ϩ release we report here differs from results of Pereira et al. (17) who report that Epac activation decreases evoked Ca 2ϩ release due to CamKII-dependent SR store depletion. Apparent discrepancies between our results and Pereira et al. (17) could be due to differences in protocol or species used, but the overall conclusions that Epac activation results in CamKII activation and RyR(Ser 2815 ) phosphorylation are in agreement.
Roles of PKC and CamKII in both normal and pathological cardiac function have been steadily emerging over the last several years, but our understanding of the physiological mechanisms that control activation of these pathways have lagged behind. Here, we have outlined a novel, PLC⑀/PKC⑀/CamKIIdependent regulatory mechanism for regulating cardiac CICR in adult ventricular cardiac myocytes. Previous studies have implicated CamKII-dependent phosphorylation of RyR2 as a means for regulating SR-Ca 2ϩ release downstream of Epac or ␤AR stimulation (17,28). We extend these findings by identifying key mechanistic links between Epac activation and regulation of CICR such that the majority of the components of the pathway are now defined. This pathway is clearly important for cardiac function because mice lacking PLC⑀ exhibit a significantly impaired ability to respond to ␤AR stimulation. This impairment is manifested by both decreased cardiac function in response to isoproterenol administration in whole animals (10) and decreased ␤AR-dependent enhancement of CICR in AVM isolated from PLC⑀ Ϫ/Ϫ mice (12). PLC⑀ Ϫ/Ϫ mice also show an increased sensitivity to stress-induced cardiac hypertrophy (10) and it remains to be determined how components of the pathway could contribute to this pathology.