S100A1 enhances the L-type Ca2+ current in embryonic mouse and neonatal rat ventricular cardiomyocytes.

S100A1 is an EF-hand type Ca2+-binding protein with a muscle-specific expression pattern. The highest S100A1 protein levels are found in cardiomyocytes, and it is expressed already at day 8 in the heart during embryonic development. Since S100A1 is known to be involved in the regulation of Ca2+ homeostasis, we tested whether extracellular S100A1 plays a role in regulating the L-type Ca2+ current (I(Ca)) in ventricular cardiomyocytes. Murine embryonic (day 16.5 postcoitum) ventricular cardiomyocytes were incubated with S100A1 (0.001-10 microM) for different time periods (20 min to 48 h). I(Ca) density was found to be significantly increased as early as 20 min (from -10.8 +/- 1 pA/pF, n = 18, to -22.9 +/- 1.4 pA/pF; +112.5 +/- 13%, n = 9, p < 0.001) after the addition of S100A1 (1 microM). S100A1 also enhanced I(Ca) current density in neonatal rat cardiomyocytes. Fluorescence and capacitance measurements evidenced a fast translocation of rhodamine-coupled S100A1 from the extracellular space into cardiomyocytes. S100A1 treatment did not affect cAMP levels. However, protein kinase inhibitor, a blocker of cAMP-dependent protein kinase A (PKA), abolished the S100A1-induced enhancement of I(Ca). Accordingly, measurements of PKA activity yielded a significant increase in S100A1-treated cardiomyocytes. In vitro reconstitution assays further demonstrated that S100A1 enhanced PKA activity. We conclude that the Ca2+-binding protein S100A1 augments transsarcolemmal Ca2+ influx via an increase of PKA activity in ventricular cardiomyocytes and hence represents an important regulator of cardiac function.

In 1965, Moore (1) described a novel protein "soluble in 100% saturated ammonium sulfate solution" and therefore denoted it S100 protein. Today 21 different S100 proteins belonging to a multigene family of Ca 2ϩ -binding proteins of the EF-hand type are known to be differentially expressed in a large number of cell types (2). S100A1 is found in the cytosol, displays highest expression levels in the heart (3), and is already expressed at day 8 postcoitum in embryonic mouse hearts (4). Although the physiological relevance of S100 proteins in healthy tissue is still unclear, several findings point to a role of S100A1 in the diseased heart, since it is up-regulated during the course of myocardial hypertrophy (5) and markedly down-regulated in the left ventricle of patients with end stage heart failure (6). Recently, S100A1 was identified as a novel regulator of cardiac function based on the observation that overexpression of S100A1 led to a significant increase of contractility in adult cardiomyocytes (7). This was found to be related to an increased sarcoplasmic reticulum Ca 2ϩ -ATPase (8) and ryanodine receptor 2 activity (9) as well as to a decrease in myofilamental Ca 2ϩ sensitivity (8) and might be explained by enhancement of PKA 3 activity. The physiological relevance of S100A1 is further supported by data of Du and co-workers (10), who identified S100A1 as a regulator of cardiac reserve in a S100A1 mouse knock-out model. Accordingly, transgenic mice overexpressing S100A1 display enhanced myocardial contractility without developing hypertrophy (9).
New evidence suggests a role for S100A1 as a serum marker for myocardial ischemia (11). During heart attacks, S100A1 levels displayed a bell-shaped concentration time course with a fast rise followed by a rapid decline in serum. Being present in the serum compartment, S100A1 may have paracellular effects similar to those of other members of the S100 family (12). In fact, S100A1 translocation into the extracellular space was detected in the intact human heart after reperfusion following prolonged ischemia (13). Since S100 proteins are secreted and/or rapidly taken up by various cell types (12), we have examined the cellular effects of exogeneously applied S100A1 in murine embryonic and neonatal rat cardiomyocytes.
As S100A1 is assumed to be a crucial regulator of myocardial Ca 2ϩ homeostasis, we were particularly interested whether S100A1 has any effect on I Ca , the key trigger mechanism for sarcoplasmatic Ca 2ϩ release and the initiation of contraction. We have employed embryonic murine and neonatal rat cardiomyocytes, since adult cardiomyocytes rapidly dedifferentiate in culture (14). Both murine embryonic and neonatal rat cardiomyocytes are easy to isolate and can be cultivated for several days without changing functional characteristics (15). In particular, neonatal rat cardiomyocytes are known to recapitulate the regulatory components of adult cardiomyocytes (16,17).

MATERIALS AND METHODS
Cell Preparation-Murine embryonic cardiomyocytes were obtained from superovulated mice of the HIM:OF1 strain as described (18,19). Spontaneously beating neonatal Sprague-Dawley rat cardiomyocytes were isolated and purified as reported earlier (20). The purity of the preparation was increased by using the selective adhesion technique (21). All experiments were carried out according to the guidelines provided by the University of Cologne animal welfare committee.
Single Cell Ca 2ϩ Measurements-Ca 2ϩ measurements were performed as previously described (7). In brief, freshly isolated neonatal rat cardiomyocytes were loaded with the Ca 2ϩ indicator Fura 2-AM (5 M) for 30 min. Cytosolic Ca 2ϩ was determined employing a monochromator on the excitation and a CCD camera-based system (TILL Phototon-ics®, Planegg, Germany) on the emission side attached to an Olympus inverted microscope. Cells were paced at 1 Hz using a commercial stimulator (Phywe Systeme GmbH, Göttingen, Germany).
Electrophysiology-I Ca was measured with the whole-cell patch clamp technique as reported in detail earlier (22,23). Briefly, for voltageclamp recordings, the solutions contained the following: internal solution, 120 mM CsCl, 3 mM MgCl 2 , 5 mM MgATP, 10 mM EGTA, 5 mM HEPES, pH 7.4 (CsOH); external solution, 120 mM NaCl, 5 mM KCl, 3.6 mM CaCl 2 , 20 mM tetraethylammonium-Cl, 1 mM MgCl 2 , 10 mM HEPES, pH 7.4 (TEAOH). Only cells with an increase of I Ca density Ͼ5% after drug application were considered responders. As reported previously (22), I Ca did not show significant run down (see also Fig. 6). As a control, heat-inactivated S100A1 (95°C, 1 h) was used, and the effect on I Ca was evaluated. Since no significant changes of I Ca density in embryonic mouse cardiomyocytes were observed 20 min after incubation, we employed, if not otherwise stated, unstimulated cells as control. For the analysis of inactivation, biexponential fits using a simplex optimization algorithm were applied. All measurements were performed 48 h after dissociation. S100A1 Protein-Recombinant human S100A1 protein was purified as described (24). Briefly, after EDTA extraction of recombinant bacteria, S100A1 was purified by octyl-Sepharose hydrophobic interaction and HiTrapQ anion exchange chromatography. The protein concentration was determined using a commercial Bradford protein assay (Bio-Rad). Stock concentrations ranged between ϳ150 and 500 M. Vehicle alone did not affect I Ca (data not shown). Recombinant human S100A1 was labeled custom-based by Eurogentec© (Belgium) with rhodamine dye (rhod-S100A1). Unbound rhodamine dye used for control experiments was inactivated by 50 mM Tris to exclude nonspecific binding. As controls, cardiomyocytes were either not stimulated or exposed to heatinactivated S100A1 protein for capacitance, confocal imaging, and PKA activity measurements in cell homogenates.
Detection of S100A1 Translocation-Cells exposed to rhod-S100A1 were excited (1 s, 0.05 Hz) at 550 nm, employing a monochromator. The emitted fluorescence was recorded with a CCD camera. Inactivated rhodamine served as control. After the extracellular addition of rhod-S100A1 (1 M), intra-and extracellular fluorescence intensities were measured on-line. When fluorescence intensities reached plateau levels, excessive dye was washed out with Tyrode's solution. Intracellular fluorescence intensities were estimated by subtracting background fluorescence and autofluorescence. Furthermore, intracellular accumulation of S100A1 was assessed using confocal laser-scanning microscopy (LSM 410; ϫ63 Plan Neofluar objective; Zeiss) employing the following configuration: wavelength 543 nm, beam splitter FT 488/543 and LP 570.
Capacitance Measurements-Endocytotic uptake of S100A1 was monitored by combining the whole-cell patch clamp technique with on-line capacitance measurements (HEKA-Pulse 8.4 software). Capacitance was determined every 5 s. To avoid interference by sudden changes of capacity, the cellular resistance (R s ) was monitored on-line (25).
Cellular cAMP-Intracellular cAMP concentrations were measured using a quantitative cAMP assay system with a nonacetylation EIA procedure as described elsewhere (26). Controls and S100A1-treated cell homogenates (1 M, 20 min) were analyzed for cAMP concentrations.
To confirm the validity of the assay, cAMP levels were also determined in untreated and S100A1-treated cells after maximal stimulation by isoprenaline (ISO; 1 M) (27) and isobutylmethylxanthine (IBMX; 200 M) (28).
PKA Activity-For the determination of PKA activity in purified cell homogenates of S100A1-treated and control embryonic cardiomyocytes, a nonradioactive assay kit (Promega PepTag-Assay) was used (29). Cell homogenates were prepared using a standard homogenization buffer containing a protease/phosphatase inhibitor mixture. Following incubation of homogenates and a synthetic PKA substrate (Kemptide (Leu-Arg-Arg-Ala-Ser-Leu-Gly)) in reaction buffer and electrophoretic separation of phosphorylated substrate, the intensities of the specific gel bands were quantified using densitometry. For controls unstimulated cell homogenates and cells stimulated with heat-inactivated S100A1 were used.
To gain mechanistic insight into the molecular function of S100A1, highly purified regulatory (R) and/or catalytic (C) subunits of PKA from different suppliers (Promega and Sigma) were incubated with or without cAMP or purified recombinant S100A1 in the presence of 2 Ci of [␥-32 P]ATP/tube as described previously (30). The reaction mixtures comprised 40 mM Tris-HCl (pH 7.4), 10 mM MgCl 2 , 1 mM dithiothreitol and 3 g of histone H1 as the substrate. The reaction was allowed to proceed for 30 min at 30°C before proteins were separated by SDS-PAGE. Western blotted membranes were exposed to Fuji imaging plates, and autoradiographic signals were quantitated with a FLA-5000 Fuji-Imager (Raytest, Straubenhardt, Germany).
Substances-[␥-32 P]ATP (800 Ci/mmol) was purchased from MP Biomedicals (Irvine, CA). All other substances, if not otherwise stated, were obtained from Sigma and were of the highest purity available.
Statistical Analysis-To evaluate statistical significance, we have performed analysis of variance with subsequent Tukey or Dunney tests. Where applicable, paired or unpaired Student's t tests were used. Values below 0.05 were considered significant. Results are given as means Ϯ S.E.
Similar to the effects of ␤-adrenergic stimulation on I Ca in terminally differentiated cardiomyocytes (31), S100A1 (1 M, 20 min) caused a slight left shift of the threshold potential and current voltage (I/V) relationship of I Ca (Fig. 2, b and d) in cardiomyocytes derived from neonatal rats. In addition, we also observed an increase of the fast inactivation component of I Ca (Fig. 2e), probably due to an augmented influx of Ca 2ϩ and ensuing increase of Ca 2ϩ -induced Ca 2ϩ release (31).
To determine the physiological relevance of the S100A1-mediated stimulation of I Ca , intracellular Ca 2ϩ was measured in electrically stimulated neonatal rat cardiomyocytes. S100A1 (1 M, 48 h) led to significantly augmented amplitudes of Ca 2ϩ transients with 195.7 Ϯ 7.2 nmol (n ϭ 51 cells, 204 Ca 2ϩ transients) versus 146.4 Ϯ 6.1 nmol (n ϭ 29 cells, 116 Ca 2ϩ transients) in controls. This is due to a significant increase of peak Ca 2ϩ concentrations (controls: 244 Ϯ 6 nmol versus 278 Ϯ 7 nmol for S100A1-treated cells, p Ͻ 0.02) and lowered resting Ca 2ϩ levels (98 Ϯ 5 nmol versus 82 Ϯ 4 nmol, p Ͻ 0.02). Thus, extracellular application of S100A1 causes a prominent enhancement of both peak I Ca density and cytosolic Ca 2ϩ transients in cardiomyocytes. S100A1 Rapidly Passes the Cell Membrane-To explore the possibility that S100A1 is able to diffuse into the cell, fluorescence imaging experiments were performed using rhod-S100A1 and inactivated rho-FIGURE 1. S100A1 enhances peak I Ca in embryonic (day 16.5 postcoitum) murine cardiomyocytes. a and b, representative current traces and I/V curves of I Ca recorded from a control (a) and a S100A1-treated (b) (1 M, 48 h) cardiomyocyte. I Ca was evoked by applying voltage pulses from Ϫ50 to ϩ50 mV (HP ϭ Ϫ80 mV) at a frequency of 0.2 Hz after an initial prepulse from Ϫ80 to Ϫ40 mV to inactivate I Na . c, significant enhancement of I Ca density was detected at S100A1 concentrations of Ն0.1 M (48 h). d, S100A1 (1 M) enhanced I Ca density from 20 min up to 48 h after application.  (1 M, 2 h). c, the S100A1 effect was observed between 20 min and 48 h of exposure. d, S100A1 (1 M, 20 min) induced a slight left shift of the threshold potential and current voltage (I/V) relationship of I Ca (normalized). e, enhancement of the fast inactivation of I Ca by S100A1 (1 M, 20 min). The insets show fitting of representative normalized current traces ( control (7 ms) versus S100A1 (3 ms)). damine as control. A rapid focal accumulation of rhod-S100A1 was observed in embryonic mouse cardiomyocytes (Fig. 3a), reaching plateau levels within 15 min (Fig. 3b, n ϭ 14). A strong indication for the intracellular location of rhod-S100A1 was given by exclusion of the nucleus and by significantly elevated fluorescence levels after wash-out, whereas inactivated rhodamine was almost instantaneously removed by wash-out (Fig. 3b, n ϭ 5). This finding was further corroborated using confocal laser-scanning microscopy to directly monitor intracellular accumulation of rhod-S100A1 protein (Fig. 3c). Heat-inactivated protein, however, failed to pass the cell membrane (n ϭ 3, data not shown).

S100A1-mediated Modulation of I Ca
To experimentally address the potential endocytotic uptake of S100A1, we performed capacitance measurements using the patch clamp technique. A time-dependent significant change of cell capacitance by Ϫ2.8 Ϯ 0.9 pF (Ϫ6 Ϯ 1.1%) in S100A1-treated (1 M, n ϭ 6) embryonic mouse cardiomyocytes compared with controls (Ϫ0.05 Ϯ 0.2 pF, Ϫ0.4 Ϯ 0.7%, n ϭ 4, p Ͻ 0.03) was observed within 20 s after application of S100A1 (Fig. 3, d and e). By contrast, heat-inactivated S100A1 did not induce changes of membrane capacitance over time (ϩ0.6 Ϯ 0.7 pF, ϩ1.9 Ϯ 1.9%, n ϭ 3, p Ͼ 0.05). Thus, S100A1, but not heat-inactivated control protein, is most likely transported within minutes via an endocytotic process into murine ventricular cardiomyocytes. This rapid intracellular accumulation of S100A1 closely correlates with the early onset of its stimulatory effect on I Ca . S100A1 Does Not Alter Cellular cAMP Levels and Acts Distal of Adenylycyclase/cAMP-The finding that S100A1 increased I Ca density within 20 min after the addition to the extracellular solution suggests a direct action of S100A1 on intracellular signaling mechanisms rather than transcriptional up-regulation of gene expression. As proof, ATP␥S (2 mM) together with forskolin (1 M) were applied to maximally enhance I Ca in embryonic mouse cardiomyocytes (32). Strong stimulation of I Ca reaching similar densities was observed in both S100A1treated (Ϫ29.8 Ϯ 2.8 pA/pF, ϩ76 Ϯ 16.3%, n ϭ 9, p Ͻ 0.001) and control cardiomyocytes (Ϫ27.6 Ϯ 2.4 pA/pF, ϩ115 Ϯ 19%, n ϭ 9, p Ͻ 0.001) (Fig. 4, a-c). These data indicate no increase in the functional expression of voltage-dependent Ca 2ϩ channels.
To elucidate possible mechanisms through which S100A1 enhanced PKA activity, reconstitution experiments using highly purified PKA subunits in a radioactive PKA assay with histone H1 as a substrate were performed. Fig. 7b shows that the regulatory subunit (R) was effectively inhibiting the activity of the catalytic subunit (C) (RC, left lane). Application of C alone resulted in a 7.3 Ϯ 1.5-fold higher amount of 32 Plabeled histone H1 (n ϭ 8) compared with RC. cAMP stimulated RC to similar levels (i.e. 6.5 Ϯ 1-fold increase of basal RC activity) (n ϭ 8). A similar effect (7.3 Ϯ 1.3-fold, n ϭ 8) was seen when S100A1 (1 M) was preincubated with R (right lane). The S100A1-induced stimulation of the enzymatic activity of PKA was concentration-dependent, yielding maximal effects at 10 Ϫ8 M (Fig. 7c). Incubation with purified S100A1 protein alone did not result in phosphorylation of histone H1. In addition, S100A1 and cAMP did not further enhance the enzymatic activity of C (data not shown). DISCUSSION S100A1, a member of the EF-hand Ca 2ϩ -binding protein family, is a novel regulator of cardiac contractility (8), in particular during the course of myocardial hypertrophy (5, 6). This assumption is corrobo- . S100A1 does not alter the functional expression of voltage-dependent Ca 2؉ channels. a and b, time course of I Ca shows that dialysis with ATP␥S (2 mM) and superfusion with forskolin (1 M) led to a significant augmentation of I Ca density in a representative control (a) and S100A1treated (b) embryonic cardiomyocyte. The insets show original current traces. c, statistical analysis proved that coapplication of ATP␥S and forskolin stimulated I Ca to similar densities. S100A1 (1 M) was applied for 48 h, step potential 0 mV. rated by transgenic mouse models where deletion or overexpression of S100A1 results in deterioration or improvement of cardiac contractility, respectively (9,10). Since this protein has been shown to represent an important regulator of Ca 2ϩ homeostasis, we have analyzed whether extracellular S100A1 modulates voltage-dependent Ca 2ϩ channels, key determinants of cardiac contractility.
We show that extracellular application of S100A1 markedly enhanced peak I Ca and Ca 2ϩ transients. S100A1 shifted the I/V relationship of I Ca to negative potentials and accelerated fast inactivation kinetics of the Ca 2ϩ currents. This effect is similar to ␤-adrenergic stimulation of I Ca and ensuing Ca 2ϩ -induced Ca 2ϩ release (31). Interestingly, this was more prominent in neonatal rat cardiomyocytes than in embryonic mouse cardiomyocytes, suggesting differences in L-type Ca 2ϩ channel regulation and the amount of Ca 2ϩ -induced Ca 2ϩ release dur-ing embryonic development. 4 This hypothesis is further validated by the finding that ISO (1 M) shifts the I/V curve in late stage but not early stage embryonic mouse cardiomyocytes (data not shown). The additional enhancement of I Ca after ISO application illustrates a submaximal activation of I Ca by S100A1. Interestingly, the acceleration of I Ca inactivation was clearly visible at 20 min but not at 48 h following S100A1 application, although PKA activity measurements in cell homogenates showed a sustained PKA activation by S100A1 after 48 h. Although at this point, we do not know the underlying mechanism(s), this finding may be of a compensatory nature. In fact, despite unchanged PKA activity, the function of phosphatase type 1 and 2A could be significantly enhanced, leading to an elevated dephosphorylation of L-type Ca 2ϩ channnels as shown earlier in a rat model of long term ␤-adrenergic stimulation (34).
The effect of S100A1 on I Ca occurred at step potentials from Ϫ10 to ϩ20 mV, hence within the physiological voltage range of contracting ventricular cardiomyocytes. The internalization of S100A1 was proven by laser-scanning and fluorescence microscopy. Membrane capacitance measurements clearly suggest, in line with an earlier report (35), that uptake of S100A1 occurs through an endocytotic process.
Electrophysiological and biochemical experiments were performed to identify intracellular target(s) of S100A1 involved in its action on I Ca . The open probability of voltage-dependent Ca 2ϩ channels is critically dependent on PKA activity. Thus, an increase of cAMP-dependent signaling could eventually also lead to stimulation of peak I Ca via activation of PKA. Maximal concentrations of ISO led to a reduced enhancement of I Ca density in S100A1-treated cells, suggesting that S100A1 plays a role in the ␤-adrenergic signaling cascade. A cAMP-dependent effect appears rather unlikely, since the addition of carbachol, IBMX, or submaximal concentrations of ISO had no different effects on I Ca in both S100A1-treated and control cells. This was further strengthened by the finding that S100A1 failed to alter cAMP levels in cell homogenates. Additional experiments indicated that S100A1 acted downstream of cAMP. In fact, the PKA inhibitor PKI accelerated the run down of I Ca in S100A1-treated cardiomyocytes. The critical role of PKA in the S100A1-mediated enhancement of I Ca was strengthened by prestimu-lation of cardiomyocytes with myristoylated PKI which abolished the S100A1 effect. The fact that S100A1 increased PKA-dependent phosphorylation of the Kemptide peptide in cellular homogenates and that S100A1 did not change cAMP concentrations suggested its cAMP-independent activation of PKA as reported by others (36). This assumption was clearly confirmed by reconstitution experiments in vitro using highly purified PKA subunits in order to provide a mechanistic understanding of the S100A1 effect on PKA. Preincubation of S100A1 with the regulatory subunit of PKA disinhibited its function, whereas S100A1 alone had no phosphorylation potential and did not affect the activity of the catalytic subunit of PKA. Our data suggest that S100A1 binds to the regulatory subunit of PKA, thereby preventing its inhibitory action on catalytic subunits, resulting in an augmented PKA activity. Nevertheless, more complex and indirect interactions of S100A1 with the regulatory and/or catalytic subunit cannot be excluded. Since both S100A1 and calmodulin belong to the EF-hand Ca 2ϩ -binding protein family and share a relatively high sequence homology (37), future studies should tackle a potential direct binding of S100A1 to the calmodulin binding site of L-type Ca 2ϩ channels.
Interestingly, a recent report investigating S100A1 action on phospholamban phosphorylation after adenoviral overexpression could not detect S100A1-induced modulation of PKA activity (8). However, enhanced Ca 2ϩ cycling observed in S100A1-overexpressing cardiomyocytes was shown to predominantly involve sarcoplasmic reticulum Ca 2ϩ -fluxes, whereas endocytosed S100A1 was supposed to alter intra- In contrast, rundown was small in control cardiomyocytes without PKI in the pipette (open circles). c, the results were confirmed by statistical analysis. A single asterisk denotes statistical significance of S100A1-treated cells compared with control cells prior to PKI application; a double asterisk reveals statistical significance of treated cells before and after PKI application. d, prestimulation with myristoylated PKI significantly reduced the effect of ISO (see also Fig. 5c) on I Ca and prevented the increase of I Ca by S100A1.
During heart injury cardiac proteins are known to be locally released resulting in an increase in serum concentration. In fact, serum concentrations of troponin I and T, also belonging to the EF-hand protein family, are currently used as sensitive indicators of cardiac injury (38). For S100A1 a correlation between the extent of myocardial damage and the S100A1 serum concentration has also been reported (11). A maximal concentration of 5 ng/ml (0.5 nM) was measured in the peripheral blood. However, the local S100A1 concentration in the healthy or injured heart is still unknown. Troponin T was reported to reach 300fold higher levels in the heart compared with the peripheral blood during myocardial infarction (39). Even if we assume for S100A1 only a 200-fold concentration gradient between the heart and the periphery (see above), a significant enhancement of I Ca would be expected to occur. Thus, we presume that the concentrations of S100A1 used in this study reflect the physiological range and that S100A1 may exert paracrine effects during ischemic myocardial injury.
Taken together, we demonstrate that S100A1 enhances I Ca and Ca 2ϩ transients in ventricular cardiomyocytes via activation of PKA after endocytotic uptake. We currently envision a complex action of S100A1 as follows. First, S100A1 enhances I Ca and triggers Ca 2ϩ via a PKA-dependent mechanism, resulting in elevated systolic Ca 2ϩ concentrations and enhanced Ca 2ϩ -induced Ca 2ϩ release; this is supported by the observed increase of the fast component of I Ca inactivation. Second, since a colocalization of S100A1 with sarcoplasmic reticulum Ca 2ϩ -ATPase and phospholamban (40) and its interaction with the ryanodine receptor 2 have been reported previously (9), S100A1 may also induce PKA-mediated phosphorylation of the ryanodine receptor 2 and phos-pholamban, thereby further increasing sarcoplasmic reticulum Ca 2ϩ release and Ca 2ϩ reuptake. In line with this hypothesis, we and others (35) found besides increased systolic also decreased diastolic Ca 2ϩ levels. Thus, S100A1 is a novel, interesting modulator of PKA signaling and cardiac contractility.