Phospholemman Inhibition of the Cardiac Na+/Ca2+ Exchanger

We have demonstrated previously that phospholemman (PLM), a 15-kDa integral sarcolemmal phosphoprotein, inhibits the cardiac Na+/Ca2+ exchanger (NCX1). In addition, protein kinase A phosphorylates serine 68, whereas protein kinase C phosphorylates both serine 63 and serine 68 of PLM. Using human embryonic kidney 293 cells that are devoid of both endogenous PLM and NCX1, we first demonstrated that the exogenous NCX1 current (INaCa) was increased by phorbol 12-myristate 13-acetate (PMA) but not by forskolin. When co-expressed with NCX1, PLM resulted in: (i) decreases in INaCa, (ii) attenuation of the increase in INaCa by PMA, and (iii) additional reduction in INaCa in cells treated with forskolin. Mutating serine 63 to alanine (S63A) preserved the sensitivity of PLM to forskolin in terms of suppression of INaCa, whereas mutating serine 68 to alanine (S68A) abolished the inhibitory effect of PLM on INaCa. Mutating serine 68 to glutamic acid (phosphomimetic) resulted in additional suppression of INaCa as compared with wild-type PLM. These results suggest that PLM phosphorylated at serine 68 inhibited INaCa. The physiological significance of inhibition of NCX1 by phosphorylated PLM was evaluated in PLM-knock-out (KO) mice. When compared with wild-type myocytes, INaCa was significant larger in PLM-KO myocytes. In addition, the PMA-induced increase in INaCa was significantly higher in PLM-KO myocytes. By contrast, forskolin had no effect on INaCa in wild-type myocytes. We conclude that PLM, when phosphorylated at serine 68, inhibits Na+/Ca2+ exchange in the heart.

Phospholemman (PLM), 2 a 72-amino acid membrane phosphoprotein with a single transmembrane domain (1), belongs to the FXYD gene family of small ion transport regulators (2). With the exception of the ␥-subunit of Na ϩ -K ϩ -ATPase (FXYD2), all other known members of the FXYD gene family have at least one serine or threonine within the cytoplasmic tail (2), indicating potential phosphorylation sites. In particular, PLM (FXYD1) is the only FXYD family member to have a consensus sequence for phosphorylation by PKA (RRXS), PKC (RXXSXR), and NIMA (never in mitosis A) kinase (FRX(S/T)). Indeed PLM has been shown to be phosphorylated by PKA at serine 68 and PKC at both serine 63 and serine 68 (3).
To date, PLM has been demonstrated to modulate ion fluxes through both the Na ϩ -K ϩ -ATPase (4 -8) and the cardiac Na ϩ /Ca 2ϩ exchanger (NCX1) (9 -11). Based on analogy of phospholamban inhibition of sarco(endo)plasmic reticulum Ca 2ϩ -ATPase (SERCA2) (12) and experimental observation on the effects of PLMS (a 15-kDa homologue of PLM isolated from shark rectal glands) on shark Na ϩ -K ϩ -ATPase (13,14), the current working hypothesis is that the Na ϩ pump is inhibited by unphosphorylated PLM. On phosphorylation of PLM, inhibition of Na ϩ -K ϩ -ATPase is relieved. This hypothesis has been given strong support by the observation that the V max of sarcolemmal Na ϩ -K ϩ -ATPase is increased 3-fold after acute cardiac ischemia in association with increased PLM phosphorylation by Ͼ300% (5). In addition, Na ϩ pump current has been demonstrated to directly increase in association with PLM phosphorylation in response to forskolin (6). More recently, comparison of ␤-adrenergic effects on Na ϩ pump function between wildtype and PLM-knock-out (KO) myocytes supports the notion that the inhibitory effects of PLM on Na ϩ -K ϩ -ATPase are relieved by phosphorylation (8). It is at present not clear whether dissociation of the phosphorylated PLM from Na ϩ -K ϩ -ATPase is required to relieve its inhibition on the Na ϩ pump (5,6,8,13,14). With respect to the cardiac Na ϩ /Ca 2ϩ exchanger, previous studies demonstrated that overexpression of PLM inhibits Na ϩ /Ca 2ϩ exchange activity (9,10), whereas down-regulation of PLM enhances NCX1 current (I NaCa ) (11). The importance of PLM phosphorylation in mediating its modulatory effects on NCX1 was not addressed in these early studies except that serine 68 in PLM was found to be important (15).
Here we demonstrated that PKC but not PKA activation enhanced I NaCa when NCX1 was expressed alone in HEK293 cells. Co-expression of PLM with NCX1 resulted in decreased I NaCa in the basal state, an additional decrease in I NaCa when stimulated with forskolin, and attenuation of the magnitude of increase in I NaCa by PKC activation. Mutating serine 68 to glutamic acid (S68E) enhanced whereas substituting serine 68 with alanine (S68A) abolished the inhibitory effect of PLM on I NaCa . Mutating serine 63 to alanine (S63A) preserved the sensitivity of PLM to forskolin in terms of additional inhibition of I NaCa . Using a fundamentally different model system of murine cardiac myocytes, we first showed that endogenous I NaCa was larger in PLM-KO myocytes when compared with wild-type (WT) myocytes despite similar NCX1 protein levels. PKC but not PKA activation increased I NaCa in WT myocytes. PLM-KO myocytes exhibited significantly larger increases in I NaCa when stimulated with phorbol 12-myristate 13-acetate (PMA) as compared with WT myocytes. We conclude that PLM, when phosphorylated at serine 68, inhibits cardiac Na ϩ /Ca 2ϩ exchanger.

EXPERIMENTAL PROCEDURES
Construction of PLM Mutants and NCX1 Clones-PLM serine mutants (S63A, S68A, and S68E) were constructed with PLM in pAlter-1 using Altered Sites II in vitro mutagenesis system (Promega, Madison, WI) as described previously (15). PLM and its serine mutants were authenticated by DNA sequencing and subcloned into the mammalian expression vector pAdTrack-CMV (16). Rat cardiac NCX1 clone in pcDNA3.1(ϩ) was a generous gift from Dr. J. Lytton and subcloned into pAdTrack-CMV as previously described (17). We chose the pAdTrack shuttle vector because it allowed us to identify successfully transfected HEK293 cells through a separate cytomegalovirus (CMV) promoter present on the vector backbone that drives the expression of green fluorescent protein.
For brevity, HEK293 cells expressing NCX1 alone are referred in the text as NCX1 cells, whereas cells co-expressing NCX1 and PLM or its serine mutants are referred as PLM cells or SnnX cells (where nn is either 63 or 68, and X is either Ala or Glu).
Na ϩ /Ca 2ϩ Exchange Current (I NaCa ) Measurements-Whole cell patch-clamp recordings were performed at 30°C as described previously (10,11,18,19). Briefly fire-polished pipettes (tip diameter, 2-3 m) were filled with a buffered Ca 2ϩ solution containing 100 mM Cs ϩglutamate, 7.25 mM Na ϩ -HEPES, 1 mM MgCl 2 , 12.75 mM HEPES, 2.5 mM Na 2 ATP, 10 mM EGTA, and 6 mM CaCl 2 , pH 7.2. Free Ca 2ϩ in the pipette solution was 205 nM, measured fluorometrically with fura 2. Cells were bathed in an external solution containing 130 mM NaCl, 5 mM CsCl, 1.2 mM MgSO 4 , 1.2 mM NaH 2 PO 4 , 5 mM CaCl 2 , 10 mM HEPES, 10 mM Na ϩ -HEPES, and 10 mM glucose, pH 7.4. Verapamil (1 M), ouabain (1 mM), and niflumic acid (30 M) were used to block Ca 2ϩ , Na ϩ -K ϩ -ATPase, and Cl Ϫ currents, respectively. K ϩ currents were minimized by Cs ϩ substitution for K ϩ in both pipette and external solutions. Only cells that fluoresced green (excitation, 380 nm; emission, 510 nm), indicating successful pAdTrack transfection, were selected for current measurements. Membrane potential (E m ) was held at the calculated reversal potential of I NaCa (Ϫ73 mV) for 5 min before stimulation. A descending voltage ramp (from ϩ100 to Ϫ120 mV; 500 mV/s) was immediately followed by an ascending voltage ramp (from Ϫ120 to ϩ100 mV; 500 mV/s) (Fig. 1A). Membrane currents were measured both before and after addition of 1 mM CdCl 2 to the external solution (Fig. 1B). I NaCa was defined as the difference current measured during the descending voltage ramp in the absence and presence of Cd 2ϩ (Fig. 1C). To facilitate comparison of NCX1 currents, I NaCa of each cell was divided by its whole cell membrane capacitance (C m ) to account for variations in cell sizes. Except as otherwise stated, all results were obtained using these standard solutions.
When indicated, PMA (0.1 M) or forskolin (1 M) (both dissolved in Me 2 SO) was added to cells after base-line I NaCa was obtained. Repeat I NaCa was measured ϳ3-5 min after drug addition.
In a second series of experiments, the effects of PMA on I NaCa were measured under Cl Ϫ -free conditions. Pipette solutions consisted of 100 mM Cs ϩ -glutamate, 7.25 mM Na ϩ -HEPES, 1 mM MgSO 4 , 12.75 mM HEPES, 2.5 mM Na 2 ATP, 10 mM EGTA, and 6 mM Ca(OH) 2 , pH 7.2. External solutions contained 130 mM Na ϩ -aspartate, 5 mM Cs ϩ -glutamate, 1.2 mM MgSO 4 , 1.2 mM NaH 2 PO 4 , 5 mM Ca(OH) 2 , 10 mM HEPES, 10 mM Na ϩ -HEPES, and 10 mM glucose, pH 7.4. Verapamil, ouabain, and niflumic acid were added to the bath as before. Holding potential was Ϫ73 mV. I NaCa was defined as the difference current measured during the descending voltage ramp in the absence and presence of Cd 2ϩ (1 mM) or Ni 2ϩ (5 mM).
[Ca 2ϩ ] o was deliberately lowered to 0.2 mM so that the calculated reversal potential of I NaCa (Ϫ103 mV), and thus the holding potential, was closer to the holding potential of Ϫ73 mV used in other experiments. Keeping [Ca 2ϩ ] o at 5 mM would have resulted in a very negative holding potential of Ϫ188 mV. I NaCa was defined as the difference current measured during the descending voltage ramp in the absence and presence of Ni 2ϩ (5 mM).
Generation of PLM-KO Mice-A mouse line deficient in PLM was generated by replacing exons 3-5 of the PLM gene with lacZ and neomycin resistance genes as described in detail previously (20). These mice grow to adulthood and are fertile. Studies were performed using mice backcrossed to a pure congenic C57BL/6 background. Homozygous adult littermates that were 3-6 months old were used in the experiments. Mice were housed in ventilated racks in a barrier facility supervised by the Department of Comparative Medicine at the Pennsylvania State University College of Medicine. Standard care was provided to all mice used for experiments.
PLM, NCX1, and Calsequestrin Immunoblotting-Mouse left ventricles were excised, rinsed in ice-cold phosphate-buffered saline, and cut into small pieces. Approximately 60 mg of tissue were suspended in 700 l of ice-cold lysis buffer containing 50 mM Tris (pH 8.0), 150 mM NaCl, 1 mM Na ϩ -orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 100 mM NaF, 1 mM EGTA, and 0.5% Nonidet P-40. A Complete Mini protease inhibitor mixture tablet (Roche Applied Science) was also added to 10 ml of lysis buffer. The tissue was homogenized with a glass Dounce homogenizer (15-20 strokes) and placed on ice for 15 min before centrifugation at 20,800 ϫ g for 10 min at 4°C. The supernatant was snap frozen with dry ice-ethanol and stored at Ϫ80°C.
Isolation of Murine Myocytes and Measurement of I NaCa -Cardiac myocytes were isolated from the septum and left ventricular free wall of WT and PLM-KO mice (25-37 g) according to the protocol of Zhou et al. (23). Briefly mice were heparinized (1500 units/kg intraperitoneally) and anesthetized (pentobarbital sodium, 50 mg/kg intraperitoneally). The heart was excised, mounted on a steel cannula, and retrograde perfused (100 cm of H 2 O at 37°C) with Ca 2ϩ -free bicarbonate buffer followed by enzymatic digestion (collagenases B and D and protease XIV) as described previously (23). Isolated myocytes were plated on laminin-coated glass coverslips in a Petri dish, and the Ca 2ϩ concentration of the buffer was progressively increased from 0.05 to 0.125 to 0.25 to 0.5 mM in three steps (10-min interval for each). The 0.5 mM Ca 2ϩ buffer was then aspirated and replaced with minimal essential medium (Sigma catalogue number M1018) containing 1.2 mM Ca 2ϩ , 2.5% fetal bovine serum, and antibiotics (1% penicillin/streptomycin). After 1 h (in 5% CO 2 at 37°C), medium was replaced with fetal bovine serum-free minimal essential medium. Myocytes were used within 2-8 h of isolation. The protocol for heart excision for myocyte isolation was approved by the Institutional Animal Care and Usage Committee. I NaCa was measured in isolated murine myocytes with the same protocol and standard solutions used for transfected HEK293 cells except that pipette tip diameter was increased to 4 -6 m, and niflumic acid was decreased to 10 M.
Statistical Analysis-All results are expressed as means Ϯ S.E. For analysis of a parameter (e.g. I NaCa ) as a function of group (e.g. NCX1 versus PLM) and voltage, two-way analysis of variance was used to determine statistical significance. For analysis of C m , Student's t test was used. A commercial software package (JMP, version 4.0.5; SAS Institute, Cary, NC) was used. In all analyses, p Ͻ 0.05 was taken to be statistically significant.

Effects of PMA or Forskolin on I NaCa in HEK293 Cells
Expressing NCX1 Alone-We have shown previously that HEK293 cells do not express NCX1 or demonstrate measurable I NaCa or Na ϩ -dependent Ca 2ϩ uptake (10). When transfected with rat cardiac NCX1, HEK293 cells exhibited characteristic I NaCa demonstrating both forward (inward current, 3 Na ϩ in:1 Ca 2ϩ out) and reverse (outward current, 3 Na ϩ out:1 Ca 2ϩ in) Na ϩ /Ca 2ϩ exchange ( Fig. 2A). In addition, the reversal potential of I NaCa was between Ϫ70 and Ϫ60 mV, close to its theoretical equilibrium potential of Ϫ73 mV under our experimental conditions ( Fig. 2A). There were no significant (p Ͻ 0.76) differences in base-line I NaCa measured with either Cd 2ϩ or Ni 2ϩ (data not shown). Treatment with PMA, which activates PKC, resulted in a large increase in I NaCa in NCX1 cells ( Fig. 2A; p Ͻ 0.0001). For example, at ϩ100 mV, PKC stimulation resulted in an ϳ120% increase in I NaCa . Control experiments performed in Cl Ϫ -free solutions demonstrated that the PMA-induced current increase was not due to an increase in Cl Ϫ currents (Fig. 2B). In addition, PMA induced large increases in currents whether Cd 2ϩ (ϳ122% at ϩ100 mV) (Fig. 2B) or Ni 2ϩ (ϳ81% at ϩ100 mV) (data not shown) was used to define I NaCa under Cl Ϫ -free conditions. To control for the possibility that the observed PMA-induced I NaCa increase was due to small changes in [Na ϩ ] i rather than enhancing intrinsic NCX1 activity, experiments were performed in high [Na ϩ ] i conditions such that I NaCa would not be so sensitive to small changes in [Na ϩ ] i . Fig. 2C shows that base-line I NaCa was significantly (p Ͻ 0.0001) smaller in high [Na ϩ ] i and low [Ca 2ϩ ] o (0.2 mM) when compared with normal [Na ϩ ] i and high [Ca 2ϩ ] o (5 mM) conditions ( Fig. 2A), likely due to the 25-fold reduction of [Ca 2ϩ ] o . However, addition of PMA increased I NaCa (ϳ84% at ϩ100 mV) under high [Na ϩ ] i conditions, similar to the observations obtained under lower but more physiological [Na ϩ ] i conditions. In con- Holding potential was at the calculated reversal potential of I NaCa (Ϫ73 mV) under our experimental conditions. Ca 2ϩ , Na ϩ -K ϩ -ATPase, Cl Ϫ , and K ϩ currents were blocked by appropriate inhibitors. B, membrane currents recorded in a transfected cell during the descending-ascending voltage ramp from ϩ100 to Ϫ120 and back to ϩ100 mV in the absence and presence of 1 mM Cd 2ϩ . C, derived Cd 2ϩsensitive current in the transfected cell shown in B.
trast to results obtained with PMA stimulation, forskolin treatment did not affect I NaCa in NCX1-expressing cells ( Fig. 3B; p Ͻ 0.64).
Effects of PMA or Forskolin on I NaCa in Cells Expressing Both NCX1 and PLM-Co-expression of PLM with NCX1 in HEK293 cells resulted in a significant decrease in I NaCa compared with cells expressing NCX1 alone (Fig. 3, A and B; p Ͻ 0.0005), consistent with our previous observations (10). At ϩ100 mV, PLM inhibited I NaCa by ϳ26%. PMA treatment of PLM cells resulted in a significant increase in I NaCa when compared with unstimulated NCX1 or PLM cells ( Fig. 3A; p Ͻ 0.0001). However, the magnitude of I NaCa increase by PMA was much smaller in PLM cells when compared with NCX1 cells (39 versus 120% at ϩ100 mV).
Despite the absence of an effect of forskolin on I NaCa in cells expressing NCX1 alone, PKA stimulation in PLM cells resulted in a significant decrease in I NaCa compared with unstimulated PLM cells ( Fig. 3B; p Ͻ 0.0001). For example, at ϩ100 mV, forskolin effected an ϳ49% decrease in I NaCa in PLM cells (Fig. 3B).
Effects of PLM Serine 68 Mutants on I NaCa in Transfected HEK293 Cells-Because serine 68 in PLM is the common phosphorylation target for both PKA and PKC, we next investigated the effects of serine 68 mutants on I NaCa in cells co-expressing NCX1 and PLM serine 68 mutants. Mutating serine 68 to alanine (S68A) resulted in abolition of the effect of WT PLM on I NaCa (Fig. 4A; p Ͻ 0.08), consistent with our   MARCH 24, 2006 • VOLUME 281 • NUMBER 12 previous observations (10). Treating S68A cells with PMA, instead of increasing I NaCa as observed in PLM cells (Fig. 3A), resulted in a modest but significant suppression of I NaCa when compared with unstimulated NCX1 cells ( Fig. 4A; p Ͻ 0.0004).

Phosphorylated Phospholemman Inhibits Na ؉ /Ca 2؉ Exchanger
Mutating serine 68 to glutamic acid (S68E) resulted in greater suppression of I NaCa when compared with WT PLM (Fig. 4B; p Ͻ 0.0001). Stimulating S68E cells with PMA, in contrast to increases in I NaCa in PLM cells (Fig. 3A), did not result in appreciable changes in I NaCa when compared with unstimulated S68E cells ( Fig. 4B; p Ͻ 0.70). The lack of I NaCa stimulation by PMA in both PLM serine 68 mutants (Fig. 4, A and  B) as compared with WT PLM (Fig. 3A) suggests altered PLM interaction with NCX1 by serine 68 mutants may somehow interfere with the stimulatory effects of PMA on NCX1.
Effects of PLM Serine 63 Mutant on I NaCa in Transfected HEK293 Cells-Unlike PKA, which phosphorylates serine 68 only, PKC phosphorylates both serine 63 and serine 68 in PLM (3). Co-expressing PLM serine 63 to alanine mutant (S63A) with NCX1 resulted in inhibition of I NaCa compared with cells expressing NCX1 alone ( Fig. 5A; p Ͻ 0.03). The magnitude of inhibition by S63A was quite modest (ϳ8% at ϩ100 mV) when compared with that by WT PLM (ϳ26% at ϩ100 mV; Fig.  3A). Treating S63A cells with forskolin resulted in additional inhibition of I NaCa (ϳ37% at ϩ100 mV) when compared with unstimulated S63A cells or NCX1 cells (Fig. 5A; p Ͻ 0.0001).
In another series of experiments, the effects of PMA on I NaCa in S63A cells were evaluated. Unlike cells co-expressing NCX1 and PLM serine 68 mutants in which I NaCa was not stimulated at all by PMA (Fig. 4, A  and B), S63A cells demonstrated significant PMA-induced enhancement of I NaCa (Fig. 5B; p Ͻ 0.0001). The effects of PMA on I NaCa in cells co-expressing NCX1 and S68A, S68E, or S63A, when considered together, are consistent with the notion that retaining normal serine 68 in PLM is absolutely essential for the stimulatory effect of PMA on I NaCa in cells co-expressing PLM and NCX1.
Effects of PMA on I NaCa in Ventricular Myocytes Isolated from Wildtype and PLM-KO Mice-Results from transfected HEK293 cells strongly suggest that PLM, when phosphorylated at serine 68, inhibits cardiac Na ϩ /Ca 2ϩ exchanger. To put the findings in physiological perspective, we examined the effects of PMA on I NaCa in myocytes isolated from WT and PLM-KO mice. Western blots confirmed the absence of PLM in PLM-KO myocytes (Fig. 6). Importantly NCX1 protein levels (normalized to calsequestrin) were not significantly different (p Ͻ 0.52) between wild-type (99.8 Ϯ 11.7 arbitrary units; n ϭ 8) and PLM-KO myocytes (90.7 Ϯ 7.3 arbitrary units; n ϭ 8) (Fig. 6). Wild-type and PLM-KO myocytes had similar cell sizes as evidenced by no differences (p Ͻ 0.97) in whole cell capacitance (a measure of membrane surface area) between WT (184 Ϯ 7 picofarads, n ϭ 13) and PLM-KO myocytes (184 Ϯ 8 picofarads, n ϭ 13). Unlike HEK293 cells expressing exogenous NCX1 (Fig. 1B), in cardiac myocytes a contaminant inward Na ϩ current was evident during the ascending voltage ramp (Fig. 7B). For this reason, the Cd 2ϩ -sensitive current (I NaCa ) was quantitated during the descending portion of the voltage ramp (Fig. 7C). I NaCa was signifi-  cantly larger in PLM-KO myocytes when compared with WT myocytes ( Fig. 8; p Ͻ 0.0001). In another series of experiments, PKC stimulation resulted in increases in I NaCa in both WT (p Ͻ 0.0001) and PLM-KO (p Ͻ 0.0001) myocytes when compared with their respective unstimulated controls (Fig. 9A). However, the PMA-induced increase in I NaCa was significantly (p Ͻ 0.002) higher in PLM-KO (ϳ132% increase at ϩ100 mV) than wild-type myocytes (ϳ91% at ϩ100 mV).

Effects of Forskolin on I NaCa in Wild-type and PLM-KO Ventricular
Myocytes-In a third series of experiments, we measured the effects of forskolin on I NaCa in murine cardiac myocytes. Base-line I NaCa was again significantly (p Ͻ 0.0001) higher in PLM-KO than WT myocytes (for clarity, KO data not shown in Fig. 9B). PKA stimulation did not result in appreciable changes in I NaCa in WT myocytes ( Fig. 9B; p Ͻ 0.11). In addition, there were no differences in I NaCa between WT and PLM-KO myocytes after forskolin treatment ( Fig. 9B; p Ͻ 0.15).

DISCUSSION
We have demonstrated previously in both rat cardiac myocytes (9,11,15) and transfected HEK293 cells (10) that PLM, in addition to its well known modulatory effects on the Na ϩ pump (4,6,8,24), inhibits cardiac Na ϩ /Ca 2ϩ exchanger. Specifically PLM co-localizes and co-immunoprecipitates with NCX1 and functionally decreases I NaCa and Na ϩ -dependent Ca 2ϩ uptake (9,10). Whereas PLM phosphorylation during ischemia (5) or by ␤-adrenergic stimulation (8) is associated with relief of its inhibition of Na ϩ -K ϩ -ATPase, it is not clear whether inhibition of Na ϩ /Ca 2ϩ exchanger is mediated by phosphorylated or unphosphorylated PLM.
Incorporation of 32 P into PLM in intact guinea pig ventricles is enhanced ϳ2.6-fold with isoproterenol treatment, suggesting WT PLM is partially phosphorylated in the unstimulated state (25). Based on C68PAb and C2Ab, which are antibodies specific for phosphorylated (at serine 68) and unphosphorylated PLM, respectively (24,26), it has been estimated that ϳ41% of PLM in adult rat myocytes (24) and ϳ25% of PLM in guinea pig myocytes (6) are phosphorylated at serine 68 under the basal state. Using another approach of comparing the effects of WT PLM and its serine 68 and serine 63 mutants on I NaCa in adult rat myocytes, ϳ46% of serine 68 and ϳ16% of serine 63 are estimated to be phosphorylated in the resting state (15). The results from these three fundamentally different experimental approaches strongly indicate that PLM is only partially phosphorylated in cardiac myocytes. Overexpression of PLM does not grossly distort the relative level of phosphorylation on serine 68 of PLM in adult rat cardiac myocytes (24). Therefore it is difficult to ascertain which form of PLM (phosphorylated or unphos-phorylated) mediates the inhibition of Na ϩ /Ca 2ϩ exchange in studies using PLM overexpression strategies (9,15).
Because PLM is known to regulate Na ϩ -K ϩ -ATPase (4 -6, 8, 24), it is tempting to explain the effects of PLM on NCX1 as indirect, i.e. changes in [Na ϩ ] i due to alterations in Na ϩ pump activity by PLM would change the driving force of NCX1 and hence I NaCa magnitude. The conditions used in our I NaCa measurements were carefully designed to avoid this ambiguity in that Na ϩ pump activity was eliminated by exclusion of K ϩ in pipette and bathing solutions as well as by the inclusion of ouabain. In addition, the measured and theoretical equilibrium potentials for I NaCa (E NaCa ) were in reasonable agreement, suggesting that under the heavily buffered [Ca 2ϩ ] i conditions used in our I NaCa measurements, the [Na ϩ ] i sensed by NCX1 could be approximated by [Na ϩ ] pipette . Finally the measured E NaCa between NCX1 and PLM cells were in close agreement, indicating that the [Na ϩ ] i sensed by NCX1 were similar in both types of cells. Therefore, the thermodynamic parameters ([Ca 2ϩ ] i, [Ca 2ϩ ] o , [Na ϩ ] i , and [Na ϩ ] o ) that determine E NaCa , and hence the driving force for I NaCa (E m Ϫ E NaCa ), were identical between NCX1 and PLM cells. In . Immunoblots of Na ؉ /Ca 2؉ exchanger (NCX1), calsequestrin, and PLM from murine hearts. Left ventricular homogenates were prepared from wild-type and PLM-KO mice of congenic C57BL/6 background as described under "Experimental Procedures." Proteins were separated by gel electrophoresis under non-reducing conditions for NCX1 (50 g/lane) and calsequestrin (100 g/lane) and reducing conditions for PLM (5 g/lane). After transfer to polyvinylidene difluoride membranes, immunoblotting was performed as described under "Experimental Procedures." Numbers on the left refer to apparent molecular mass. FIGURE 7. Measurement of Na ؉ /Ca 2؉ exchange current (I NaCa ) in murine cardiac myocytes. I NaCa was measured in ventricular myocytes isolated from adult mouse hearts at 5 mM [Ca 2ϩ ] o and 30°C with a descending-ascending voltage ramp protocol (A) as described under "Experimental Procedures." Free Ca 2ϩ in the Ca 2ϩ -buffered pipette solution was 205 nM. Holding potential was at the calculated reversal potential of I NaCa (Ϫ73 mV) under our experimental conditions. Ca 2ϩ , Na ϩ -K ϩ -ATPase, Cl Ϫ , and K ϩ currents were blocked by appropriate inhibitors. B, membrane currents recorded in a wildtype myocyte during the descending-ascending voltage ramp from ϩ100 to Ϫ120 and back to ϩ100 mV in the absence and presence of 1 mM Cd 2ϩ . C, derived Cd 2ϩ -sensitive current in the wild-type myocyte shown in B. addition, we have demonstrated previously that the protein levels of NCX1 in HEK293 cells are similar in the absence or presence of cotransfected PLM (10). The observed differences in I NaCa between NCX1 and PLM cells can thus be unambiguously assigned to the direct inhibitory effects of PLM on NCX1. Similar arguments can be advanced that the observed differences in I NaCa between wild-type and PLM-KO myocytes (with similar NCX1 protein levels) were due to direct inhibition of NCX1 by PLM.
NCX1 is known to be modulated by ␣-adrenergic stimulation (27) presumably mediated via PKC (28). Our finding that, in HEK293 cells expressing NCX1 alone, PKC activation by PMA resulted in a large increase in Na ϩ /Ca 2ϩ exchange activity is similar to that observed in CCL39 fibroblasts expressing NCX1 (28). In our experiments on HEK293 cells expressing NCX1, the increase in current by PMA was not due to activation of Cl Ϫ current because similar current increases were observed under Cl Ϫ -free conditions. Another potential concern is that, although Ca 2ϩ was heavily buffered under our experimental conditions, small changes in [Na ϩ ] i by PMA may have large effects in I NaCa (proportional to third power of [Na ϩ ] i ) with only small effects on E NaCa (proportional to third root of the Na ϩ gradient). Under conditions of high [Na ϩ ] i in which I NaCa would not be expected to be so sensitive to small changes in cytoplasmic Na ϩ , PKC stimulation still effected a large increase in I NaCa . Our control experiments with Cl Ϫ -free solutions and high [Na ϩ ] i conditions indicate that the observed increase in currents by PMA was due to the enhancement of intrinsic NCX1 activity by PKC rather than an artifactual increase in Cl Ϫ currents or changes in driving force for the exchanger.
PKC activation is associated with increased NCX1 phosphorylation at serine 249, serine 250, and serine 357 (29). In normal cardiac myocytes, however, NCX1 is associated with PLM (10,11). Therefore the physiologically more relevant model system is one that co-expresses both NCX1 and PLM. In HEK293 cells co-expressing both NCX1 and PLM, PMA treatment also resulted in enhancement of Na ϩ /Ca 2ϩ exchange activity, similar to that observed in rat sarcolemmal vesicles (27). The magnitude of the I NaCa increase, however, was much smaller in cells co-expressing NCX1 and PLM when compared with cells expressing NCX1 alone. These results suggest that the stimulatory effects of PMA on NCX1 were attenuated by increased PLM phosphorylation. The implication on Na ϩ /Ca 2ϩ exchange in intact myocytes exposed to PKC activators is that the direct stimulatory effects on NCX1 are somewhat opposed by an indirect inhibitory effect by increased phosphorylated PLM.
Because PKC induces phosphorylation at both serine 63 and serine 68 of PLM (3), we next activated PKA to evaluate the effects of PLM phosphorylated only at serine 68 on NCX1. The effects of PKA on the cardiac Na ϩ /Ca 2ϩ exchanger are quite controversial. For example, PKA activation does not enhance phosphorylation of NCX1 expressed in CCL39 fibroblasts (29), but the catalytic subunit of PKA is quite capable of in vitro phosphorylation of NCX1 immunoprecipitated from Xenopus oocytes expressing the Na ϩ /Ca 2ϩ exchanger (30). It is at present equally contentious as to whether the mammalian cardiac Na ϩ /Ca 2ϩ exchange activity is affected by PKA activation. For example, no enhancement of I NaCa by 8-Br-cAMP was observed in HEK cells expressing dog NCX1 (31). Likewise 8-Br-cAMP has no effect on Na ϩ -dependent Ca 2ϩ uptake in CCL39 fibroblasts expressing dog heart NCX1 (29). In giant membrane patches excised from blebs of guinea pig ventricular cells, no stimulatory effect of ␤-adrenergic stimulation or PKA on Na ϩ /Ca 2ϩ exchange activity is observed (32). In isolated rat sarcolemmal vesicles, isoproterenol has no effect on Na ϩ /Ca 2ϩ exchange activity (27). In intact cardiac myocytes, isoproterenol has been reported to increase I NaCa in guinea pig (33) and pig (34) but not in rabbit myocytes (35).  . Effects of PMA and forskolin on I NaCa in wild-type and PLM-KO cardiac myocytes. A, I NaCa was measured in a second group of ventricular myocytes isolated from wild-type (diamonds, n ϭ 7) and PLM-KO (circles, n ϭ 7) mouse hearts at 5 mM [Ca 2ϩ ] o and 30°C as described in Fig. 7. After base-line I NaCa was obtained, PMA (0.1 M) was added to both wild-type (squares, n ϭ 7) and PLM-KO (triangles, n ϭ 7), and I NaCa measurement was repeated. B, I NaCa was measured in a third group of wild-type myocytes both before (diamonds, n ϭ 6) and after (squares, n ϭ 6) addition of forskolin (1 M). Similarly I NaCa was measured in PLM-KO myocytes. For clarity of presentation, only data from PLM-KO myocytes treated with forskolin (triangles, n ϭ 6) are shown. Error bars are not shown if they fall within boundaries of the symbols. pF, picofarad.
Recently an elegant study has shed light on the confusing literature concerning the effects of PKA activation on mammalian cardiac Na ϩ / Ca 2ϩ exchange activity (36). The apparent augmentation of I NaCa by isoproterenol in guinea pig myocytes is due to the activation of a cAMPdependent and Ni 2ϩ -sensitive Cl Ϫ current (36). In rat and mouse ventricular cells in which cAMP does not activate this cAMP-dependent Cl Ϫ current (37), isoproterenol treatment does not increase the amplitude of I NaCa (36). Therefore to date, the weight of current evidence suggests that ␤-adrenergic stimulation with subsequent PKA activation has no discernible effects on mammalian cardiac Na ϩ /Ca 2ϩ exchange activity. Our observations that forskolin had no stimulatory effects on I NaCa in transfected HEK293 cells expressing NCX1 alone and in wildtype mouse myocytes are thus consistent with this view. However, in HEK293 cells expressing both NCX1 and PLM, forskolin resulted in additional suppression of I NaCa . This observation suggests that PLM, when phosphorylated at serine 68, inhibited cardiac Na ϩ /Ca 2ϩ exchange in a heterologous expression system. The importance of phosphorylated serine 68 in mediating the inhibition of I NaCa by PLM is supported by the experimental results with serine 68 mutants. S68A, which cannot be phosphorylated, resulted in loss of function, whereas S68E, which mimicked 100% phosphorylation, resulted in additional suppression of I NaCa when compared with WT PLM both in transfected HEK293 cells (current study) and in adult rat cardiac myocytes overexpressing PLM or its serine 68 mutants (15).
The results of S63A mutant on I NaCa are interesting in three respects. First, leaving serine 68 intact but prohibiting phosphorylation at serine 63 resulted in a much more modest inhibition of I NaCa when compared with wild-type PLM. This suggests that phosphorylation at serine 63 may also contribute to the inhibitory effect of PLM on I NaCa . However, the lack of effects on I NaCa by S68A mutant (with or without PMA stimulation) indicates that serine 68 phosphorylation is of primary importance in the inhibition of NCX1 by PLM. Second, treating S63A cells with forskolin resulted in a more substantial suppression of I NaCa , again indicating the primacy of serine 68 phosphorylation in mediating the inhibitory effect of PLM on I NaCa . Third and perhaps the most intriguing is that although PMA resulted in large I NaCa increases in cells expressing NCX1 alone or NCX1 ϩ PLM, cells that expressed NCX1 and S68A or S68E mutants showed no increases in I NaCa when stimulated with PMA. Cells that expressed NCX1 and S63A mutant (in which serine 68 is intact), on the other hand, were able to increase I NaCa with PKC activation, similar to cells expressing both NCX1 and wild-type PLM. Our results on the serine 63 and serine 68 mutants suggest that changes in conformation in PLM by mutating serine 68 may alter its interaction with NCX1, resulting in NCX1 not being accessible to PKC action perhaps due to steric hindrance.
The relative lack of effects by S68A and S63A mutants on I NaCa in transfected HEK293 cells may be due to loss of interaction between these PLM mutants and NCX1. This is unlikely, however, as we have demonstrated previously that both S68A and S63A mutants, similar to WT PLM, are able to co-immunoprecipitate NCX1 in HEK293 cells co-expressing NCX1 and PLM or its serine mutants (15).
The physiological relevance of serine 68 phosphorylation in PLM on NCX1 function was examined in WT and PLM-KO myocytes. There are many similarities between the results obtained in transfected HEK293 cells and murine myocytes. For example, similar to the observation that I NaCa was higher in HEK293 cells expressing NCX1 alone as compared with cells co-expressing NCX1 and PLM, base-line I NaCa was higher in PLM-KO than WT myocytes. PMA treatment resulted in enhancement of I NaCa in both WT and PLM-KO myocytes, although the increase in I NaCa was much higher in PLM-KO myocytes. This is also similar to our findings in the heterologous expression system. On the other hand, there are some differences between the effects of PKA on I NaCa in HEK293 cells and murine myocytes. For example, forskolin treatment resulted in suppression of I NaCa in HEK293 cells co-expressing NCX1 and PLM. By contrast, PKA stimulation in WT myocytes did not result in any detectable changes in I NaCa in agreement with observations by Ginsburg and Bers (35) and Lin et al. (36). The differences between the results obtained in HEK293 cells and murine myocytes with respect to PKA effects on I NaCa are not intuitively obvious but may relate to association of NCX1 with the catalytic subunit of PKA and protein phosphatase 1 in rat hearts (30). It is known that NCX1 exhibits significant basal phosphorylation in cardiac myocytes (28). In addition, dephosphorylation of NCX1 by protein phosphatase 1 results in reduction of I NaCa (34), whereas increased NCX1 phosphorylation is associated with enhancement of Na ϩ /Ca 2ϩ exchange activity (28). PKA stimulation of intact cardiac myocytes would be expected to simultaneously increase phosphorylation in both NCX1 (stimulatory) (30) and PLM (inhibitory) plus or minus other unknown effects on protein phosphatase 1 such that the net effect would be no measurable changes in I NaCa . In NCX1 expressed heterologously in HEK293 cells, there may not be such close association of PKA with NCX1 in an assembled "macromolecular complex" (38) so that PKA can exert its effects on NCX1. On the other hand, in our simplified heterologous expression system, phosphorylation of PLM by ubiquitous PKA present in these cells 3 or the phosphomimetic S68E mutant would be expected to suppress I NaCa .
In the intact heart, ␤-adrenergic stimulation increases Na ϩ influx into the myocytes because of the chronotropic effect (more frequent depolarizations). In addition, L-type Ca 2ϩ current and SERCA2 activity are also increased in response to ␤-adrenergic stimulation, resulting in increased Ca 2ϩ entry and Ca 2ϩ loading of the sarcoplasmic reticulum. Increased sarcoplasmic reticulum Ca 2ϩ available for release largely accounts for the increased inotropy of ␤-adrenergic agonists. To maintain steady-state Ca 2ϩ balance, the increased myocyte Ca 2ϩ entry must necessitate increased Ca 2ϩ efflux mediated by forward Na ϩ /Ca 2ϩ exchange, thereby bringing more Na ϩ into the cell. Therefore, enhanced Na ϩ -K ϩ -ATPase activity (by PLM phosphorylation) during ␤-adrenergic stimulation is necessary to prevent cellular Na ϩ overload. On the other hand, unchecked stimulation of Na ϩ -K ϩ -ATPase would decrease intracellular Na ϩ concentration, thereby increasing the thermodynamic driving force of forward Na ϩ /Ca 2ϩ exchange, resulting in Ca 2ϩ depletion. The ensuing decreased inotropy is clearly not desirable under the circumstances of fight or flight. Our presented evidence suggests a coordinated paradigm in which PLM, upon phosphorylation at serine 68, enhances Na ϩ -K ϩ -ATPase (5, 8) but inhibits Na ϩ /Ca 2ϩ exchange activities in cardiac myocytes. The consequences of Na ϩ -K ϩ -ATPase stimulation on the one hand and Na ϩ /Ca 2ϩ exchange inhibition on the other on cellular Ca 2ϩ homeostasis and contractility are complex and difficult to predict or model and clearly require further study.
Finally it should be pointed out that the magnitude of inhibition of I NaCa by WT PLM in HEK293 cells was ϳ26% at ϩ100 mV in our current experiments; this is much more modest than our previous results of ϳ80% inhibition at ϩ100 mV (10). This is because we deliberately decreased the amount of plasmid DNA encoding PLM used in the transfection (from 1.5 to 1.0 g/dish) so that we would better be able to detect additional inhibition of I NaCa when PLM was phosphorylated or when a phosphomimetic PLM mutant was used.
In summary, we demonstrated that phospholemman phosphorylated at serine 68 inhibited Na ϩ /Ca 2ϩ exchange in both transfected HEK293 cells and mouse myocytes. We conclude that, in intact cardiac myocytes, phosphorylation of phospholemman results in relief of inhibition of Na ϩ -K ϩ -ATPase and inhibition of Na ϩ /Ca 2ϩ exchange.