Overexpression of phospholemman (PLM) in normal adult rat cardiac myocytes altered contractile function and cytosolic Ca2+ concentration ([Ca2+]i) homeostasis and inhibited Na+/Ca2+ exchanger (NCX1). In addition, PLM coimmunoprecipitated and colocalized with NCX1 in cardiac myocyte lysates. In this study, we evaluated whether the cytoplasmic domain of PLM is crucial in mediating its effects on contractility, [Ca2+]i transients, and NCX1 activity. Canine PLM or its derived mutants were overexpressed in adult rat myocytes by adenovirus-mediated gene transfer. Confocal immunofluorescence images using canine-specific PLM antibodies demonstrated that the exogenous PLM or its mutants were correctly targeted to sarcolemma, t-tubules, and intercalated discs, with little to none detected in intracellular compartments. Overexpression of canine PLM or its mutants did not affect expression of NCX1, sarco(endo)plasmic reticulum Ca2+-ATPase, Na+-K+-ATPase, and calsequestrin in adult rat myocytes. A COOH-terminal deletion mutant in which all four potential phosphorylation sites (Ser62, Ser63, Ser68, and Thr69) were deleted, a partial COOH-terminal deletion mutant in which Ser68 and Thr69 were deleted, and a mutant in which all four potential phosphorylation sites were changed to alanine all lost wild-type PLM's ability to modulate cardiac myocyte contractility. These observations suggest the importance of Ser68 or Thr69 in mediating PLM's effect on cardiac contractility. Focusing on Ser68, the Ser68 to Glu mutant was fully effective, the Ser63 to Ala (leaving Ser68 intact) mutant was partially effective, and the Ser68 to Ala mutant was completely ineffective in modulating cardiac contractility, [Ca2+]i transients, and NCX1 currents. Both the Ser63 to Ala and Ser68 to Ala mutants, as well as PLM, were able to coimmunoprecipitate NCX1. It is known that Ser68 in PLM is phosphorylated by both protein kinases A and C. We conclude that regulation of cardiac contractility, [Ca2+]i transients, and NCX1 activity by PLM is critically dependent on Ser68. We suggest that PLM phosphorylation at Ser68 may be involved in cAMP- and/or protein kinase C-dependent regulation of cardiac contractility.
- adult rat myocyte culture
- edge detection
- excitation-contraction coupling
phospholemman (PLM), a member of the FXYD gene family of small ion transport regulators (25), is synthesized as a 92-amino acid polypeptide. The first 20-amino acid signal peptide is cleaved off during processing so that the mature PLM protein is a 72-amino acid membrane phosphoprotein with a single transmembrane domain (19). With the exception of the γ-subunit of Na+-K+-ATPase (FXYD2), all other known members of the FXYD gene family have at least one Ser or Thr within the cytoplasmic tail (25), indicating potential phosphorylation sites. In particular, PLM is the only FXYD family member to have a consensus sequence for phosphorylation by PKA (RRXS), PKC (RXXSXR), and “never in mitosis” A (NIMA) kinase (FRXS/T). Indeed, PLM has been shown to be phosphorylated by PKA at Ser68 and by PKC at both Ser63 and Ser68 (27). In addition, PLM shares sequence similarity with phospholamban (PLB), a small protein in the sarcoplasmic reticulum (SR) membrane that regulates sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2) transport activity (RSSIRRLST69 in PLM and RSAIRRAST17 in PLB). When phosphorylated by PKA at Ser16, PLB dissociates from SERCA2 and thereby removes its inhibitory effects on Ca2+ transport by SERCA2 (22).
Despite PLM being a major sarcolemmal substrate for PKA (15) and PKC (20), its function in the heart is largely unknown except that PLM phosphorylation in response to β-adrenergic stimulation paralleled the positive inotropic effects (20) and that PLM expression increased twofold in postinfarction rat hearts (21). Using adenovirus-mediated gene transfer, we were the first to show that PLM overexpression in adult rat cardiac myocytes affected cardiac contractility and cytosolic Ca2+ concentration ([Ca2+]i) transients (23). Specifically, at low (0.6 mM) extracellular Ca2+ concentration ([Ca2+]o), a condition that favors Ca2+ efflux, both contraction and [Ca2+]i transient amplitudes were larger in myocytes overexpressing PLM. At high [Ca2+]o (5 mM), a condition that favors Ca2+ influx, cell shortening and [Ca2+]i transient amplitudes were smaller in myocytes overexpressing PLM. This reduced dynamic range in response to increasing [Ca2+]o in PLM-overexpressed myocytes cannot be readily accounted for by the hypothesis that PLM, like other FXYD family members, regulates Na+-K+-ATPase activity (6, 7, 32). Rather, the pattern of contractile and [Ca2+]i transient abnormalities observed in PLM-overexpressed myocytes was similar to that in myocytes in which the Na+/Ca2+ exchanger (NCX1) was downregulated by adenovirus (Adv)-mediated antisense transfer (26), prompting us to suggest that NCX1 might be regulated by PLM (23). Indeed, NCX1 current (INaCa) was depressed in PLM-overexpressed myocytes (31), whereas INaCa was enhanced in PLM-downregulated myocytes (18). In addition, endogenous PLM colocalized with NCX1 to the sarcolemma, t-tubules, and intercalated discs (31) and was physically associated with NCX1 as evidenced by coimmunoprecipitation (18). Our observations strongly suggest that in addition to modulation of Na+-K+-ATPase activity (6, 7, 32), PLM also regulates the activity of cardiac NCX1.
The present study was undertaken to systematically evaluate whether known PKA and PKC phosphorylation sites in PLM (Ser63 and Ser68) are required for PLM's effects on cardiac contractility, [Ca2+]i transients, and INaCa. To accomplish this, a series of COOH-terminal deletion and Ser to Ala mutants of PLM were constructed and tested on adult rat ventricular myocytes.
Myocyte isolation and culture.
Cardiac myocytes were isolated from the septum and left ventricular free wall of adult male Sprague-Dawley rats (∼280 g), as previously described (5). Isolated myocytes were seeded on laminin-coated coverslips and cultured with serum-free medium 199 (Earle's salts without l-glutamine and NaHCO3) supplemented with creatine, carnitine, taurine, and NaHCO3 (23, 33). After 2 h, media were changed to remove nonadherent myocytes. Six hours after isolation, cultured myocytes were electrically paced (1 Hz, [Ca2+]o = 1.8 mM) (23, 33). Culture media were changed daily over the course of the experiments. Under continuous pacing culture conditions, we have previously demonstrated that myocyte contractility did not decline for at least 72 h (23). The protocol for heart excision for myocyte isolation was approved by the Institutional Animal Care and Usage Committee.
Adenoviral infection of cardiac myocytes.
Recombinant, replication-deficient Adv expressing either green fluorescent protein (GFP) alone, GFP and dog PLM, or GFP and dog PLM mutants (see below) were constructed as described previously (23, 33). Two hours after isolation, myocytes were infected with Adv-GFP, Adv-GFP-PLM, or Adv-GFP-PLM mutant at a multiplicity of infection of 2–5 for 3 h. Media were then changed, and myocytes were studied after 72 h in continued pacing culture. We have previously demonstrated that >95% of myocytes were successfully infected (33) and that adenoviral infection of myocytes had no effect on myocyte contractility when examined after 72 h of continuous pacing culture (23). We have also shown that the effects of PLM overexpression on contractility and [Ca2+]i transients were manifested 72 h after Adv-PLM infection (23). In addition, PLM overexpression did not affect the action potential amplitude and morphology, SR Ca2+ uptake, and protein levels of NCX1, calsequestrin, and SERCA2 in adult rat myocytes (23, 31). For the sake of brevity, myocytes infected with Adv-GFP, Adv-GFP-PLM, and Adv-GFP-PLM mutant are referred to as GFP, PLM, and the respective designation for the PLM mutant (e.g., TM43) myocytes, respectively.
Preparation of crude myocyte membranes.
Three days after infection with Adv-GFP or Adv-GFP-TM43 (COOH-terminal deletion mutant of PLM, see below), myocytes were washed three times with ice-cold PBS and scraped into 400 μl of ice-cold buffer I containing (in mM) 10 Tris (pH 7.5), 1 Na+-vanadate, 1 PMSF, 100 NaF, 1 EGTA, and a Complete Protease Inhibitor Cocktail Tablet (Catalog no. 1697498, Boehringer Mannheim; Indianapolis, IN). After sonication (3 × 15 s), 400 μl of ice-cold buffer II containing (in mM) 10 Tris (pH 7.5), 300 KCl, 1 Na+-vanadate, 1 PMSF, 100 NaF, 1 EGTA, 20% sucrose, and a Complete Protease Inhibitor Cocktail Tablet were added. Myocyte sonicates were centrifuged (10,000 g) at 4°C for 10 min. The supernatant was centrifuged (100,000 g) at 4°C for 1 h. After being washed with ice-cold PBS, the pellet (crude membrane fraction) was stored at −80°C until use.
Identification of exogenous dog and endogenous rat PLM by antibodies.
Crude membrane fractions from GFP and TM43 myocytes were resuspended in SDS sample buffer containing 5% 2-mercaptoethanol and subjected to 12% SDS-PAGE (5 μg/lane). After transfer of fractionated proteins to ImmunBlot polyvinylidene difluoride membranes, a monoclonal antibody (B8 Ab, 1:500; a generous gift of Dr. Larry R. Jones) raised against recombinant dog PLM protein expressed in Sf21 insect cells (3) was used to detect exogenous dog PLM. The secondary antibody used was sheep anti-mouse antibody (1:2,000). Rabbit polyclonal antibodies raised against the COOH-terminus of PLM (C2 Ab, 1:10,000; Ref. 23) was used to detect both endogenous rat and exogenous dog PLM, because both dog and rat PLM share identical COOH-terminal sequences (19, 25). Donkey anti-rabbit antibody (1:2,000) was used as the secondary antibody. Immunoreactive proteins were detected with the enhanced chemiluminescence (ECL) Western blotting system.
Immunolocalization of GFP, exogenous dog PLM and its mutants, and endogenous rat PLM.
Freshly isolated rat cardiac myocytes were plated on laminin-coated glass slide chambers (Nunc, Lab-Tek Division; Naperville, IL), allowed to adhere for 2h, and then infected with Adv-GFP, Adv-GFP-PLM, or Adv-GFP-PLM mutants. After 72 h, myocytes were washed three times with PBS containing 2 mM EGTA. Myocytes were fixed for 30 min in 3% paraformaldehyde in PBS with 2 mM EGTA. After two rinses with PBS, myocytes were permeabilized for 2 min in 0.05% Triton X-100. Some myocytes infected with Adv-GFP-TM43 were not permeabilized. Myocytes were rinsed two times with PBS and one time with BLOTTO (5% nonfat dry milk, 0.1 M NaCl, and 50 mM Tris·HCl; pH 7.4). Primary antibodies against PLM (B8 Ab and C2 Ab; both at 1:250) diluted in BLOTTO were added to the cells, incubated in room temperature in the dark for 60 min, and rinsed three times with BLOTTO. Secondary antibodies (see below) diluted in BLOTTO were added to cells, incubated in the dark for 30 min, and followed by three PBS rinses. The secondary antibodies were as follow: for B8 Ab, Alexa fluor 546-labeled goat anti-mouse IgG (1:50, Molecular Probes; Eugene, OR); and for C2 Ab, Alexa fluor 647-labeled goat anti-rabbit IgG (1:50, Molecular Probes). The slide was then removed from the chamber, and a coverslip containing mounting solution (90% glycerol in PBS + p-phenylaminediamine) was applied. Images of myocytes (GFP, excitation 488 nm and emission, 515 nm; B8 Ab, excitation 546 nm and emission 570 nm; and C2 Ab, excitation 633 nm and emission 674 nm) were acquired with a Leica TCS SP2 confocal microscope and processed with LCS software.
Construction of PLM mutants.
PLM COOH-terminal deletion mutants (TM43 and TM65) were generous gifts from Dr. Larry R. Jones (3). PLM Ser substitution mutants were constructed with dog PLM in p-Alter1 (23), using an Altered Sites II In Vitro Mutagenesis System (Promega; Madison, WI). TM43, TM65, and site-directed PLM mutants (Fig. 1) were first authenticated by DNA sequencing. This was followed by subcloning into the shuttle vector pAdTrack-CMV (where CMV indicates the use of a cytomegalovirus promoter), from which recombinant, replication-deficient Adv expressing GFP and each PLM mutant were derived (23). In addition, a control adenovirus expressing GFP and containing the coding sequence of PLM but without the leading 20-amino acid signal peptide (containing the ATG start codon) was made. All studies were performed 72 h after Adv infection.
Myocyte shortening measurements.
Myocytes adherent to the coverslips were bathed in 0.6 ml of air- and temperature-equilibrated (37°C), HEPES-buffered (20 mM, pH 7.4) medium 199 containing either 0.6 or 5.0 mM [Ca2+]o. Measurements of myocyte contraction (1 Hz) were performed as previously described (18, 23, 26, 30, 31, 33).
[Ca2+]i transient measurements.
Myocytes were exposed to 0.67 μM of fura-2 AM for 15 min at 37°C. Fura-2-loaded myocytes mounted in a Dvorak-Stotler chamber situated in a temperature-controlled stage (37°C) of a Zeiss IM 35 inverted microscope were field stimulated to contract at 1 Hz between platinum wire electrodes. [Ca2+]o was either 0.6 or 5.0 mM. [Ca2+]i transient measurements, calibration of fura-2 fluorescent signals, and [Ca2+]i transient analysis were performed as previously described (18, 23, 26, 30, 33).
Whole cell patch-clamp recordings were performed at 30°C as described previously (18, 26, 31). Briefly, fire-polished pipettes (tip diameter of 4–6 μm) with resistances of 0.8–1.4 MΩ when filled with standard internal solution were used. Pipettes were filled with a buffered Ca2+ solution containing (in mM) 100 Cs+-glutamate, 7.25 NaCl, 1 MgCl2, 20 HEPES, 2.5 Na2-ATP, 10 EGTA, and 6 CaCl2; pH 7.2. Free Ca2+ in the pipette solution was 205 nM, measured fluorimetrically with fura-2. Myocyte were bathed in an external solution containing (in mM) 130 NaCl, 5 CsCl, 1.2 MgSO4, 1.2 NaH2PO4, 5 CaCl2, 10 HEPES, 10 Na+-HEPES, and 10 glucose; pH 7.4. Verapamil (1 μM), ouabain (1 mM), and niflumic acid (10 μM) were used to block L-type Ca2+ currents, Na+-K+-ATPase currents, and Cl− currents, respectively. K+ currents were minimized by Cs+ substitution for K+ in both pipette and external solutions. Myocytes were selected for electrophysiological studies on the basis of rod-shaped morphology, clear cross-striations, and the absence of membrane blebs. For current measurements, cell capacitance and series resistance were compensated for as best as possible with the analog circuitry of the patch-clamp amplifier. Membrane potential was held at the calculated reversal potential of INaCa (−73 mV) for 5 min before stimulation. This precaution minimized fluxes through NCX1 before the voltage ramp and thus allowed cytosolic Na+ concentration ([Na+]i) and [Ca2+]i to equilibrate with those present in pipette solution. A descending voltage ramp (from +100 to −120 mV; 500 mV/s) was used to prevent activating the voltage-gated Na+ channel. This was immediately followed by an ascending voltage ramp (from −120 to +100 mV; 500 mV/s). The voltage ramp was repeated after the addition of 1 mM CdCl2 to the external solution. Currents were derived from measurements during the descending voltage ramp. INaCa was defined as the difference current measured in the absence and presence of Cd2+. Currents were filtered at 1 kHz, and data were acquired at 2 kHz. Whole cell capacitance for each myocyte was measured by applying a small hyperpolarizing pulse (−10 mV, 16 ms) and integrating the resulting current change (digitized at 50 kHz, 0.5-kHz filter) over time. To facilitate comparison of NCX1 currents, INaCa of each myocyte was divided by whole cell capacitance to account for variations in cell sizes.
PLM mutants, NCX1, SERCA2, calsequestrin, and Na+-K+-ATPase immunoblotting.
Cultured myocytes were harvested for immunoblotting on day 3 as described previously (18, 23, 26, 31, 32). Briefly, proteins in myocyte lysates were subjected to 7.5% (NCX1, Na+-K+-ATPase, SERCA2, and calsequestrin) or 12% (PLM and its mutants) SDS-PAGE under either nonreducing (10 mM N-ethylmaleimide for NCX1) or reducing (5% β-mercaptoethanol for PLM, PLM mutants, Na+-K+-ATPase, SERCA2, and calsequestrin) conditions. The fractionated proteins were transferred onto ImmunBlot polyvinylidene difluoride membranes. The primary antibodies used were as follows: for PLM and its mutants, monoclonal antibody against dog PLM (B8 Ab) (1:5,000); for NCX1, mouse monoclonal antibody (1:1,000, R3F1, Swant, Bellinzona, Switzerland); for SERCA2, mouse monoclonal antibody (1:2,500, MA3–919, Affinity Bioreagents, Golden, CO); for calsequestrin, rabbit anti-calsequestrin antibody (1:5,000; Swant); and for α-subunits of Na+-K+-ATPase, mouse monoclonal antibody (1:500, α5, Developmental Studies Hybridoma Bank, University of Iowa). The secondary antibodies used were donkey anti-rabbit and sheep anti-mouse IgG (Amersham). Immunoreactive proteins were detected with an ECL Western blotting system. Protein band signal intensities were quantitated by scanning autoradiograms of the blots with a phosphorimager (Molecular Dynamics; Sunnyvale, CA).
Transfection of HEK-293 cells.
HEK-293 cells (American Type Culture Collection; Manassas, VA) were cultured in DMEM-Ham's F-12 (Cellgro; Herndon, VA) containing 10% heat-inactivated fetal bovine serum at a density of 1.2 × 106 cells/100-mm dish. After 24 h, medium was changed, and cells were transfected with 25 μl Lipofectamine (Invitrogen; Carlsbad, CA) and a total of 3 μg plasmid DNA per dish: either pAdTrack-CMV alone (3 μg), pAdTrack-CMV-NCX1 (1.5 μg) + pAdTrack-CMV (1.5 μg), pAdTrack-CMV-NCX1 (1.5 μg) + pAdTrack-CMV-PLM (1.5 μg), pAdTrack-CMV-NCX1 (1.5 μg) + pAdTrack-CMV-S63A (1.5 μg), or pAdTrack-CMV-NCX1 (1.5 μg) + pAdTrack-CMV-S68A (1.5 μg), according to the manufacturer's instructions. Levels of DNA and lipid were optimized in transfection assays to ensure minimal toxicity to cells. Cells were exposed to the lipid-DNA complex for 5 h at 37°C and 5% CO2. Medium was then replaced with DMEM-Ham's F-12 + 10% FBS, and cells were cultured for an additional 48 h before experiments. Transfection efficiency was routinely ∼30–50%.
Coimmunoprecipitation of PLM and NCX1.
Crude membrane fractions from transfected HEK-293 cells were prepared as described for myocyte membranes except that Hanks' balanced salt solution rather than PBS was used for rinsing cells and washing membrane pellets. Crude membrane pellets (adjusted to 2 mg/ml) were resuspended in 800 μl of buffer III containing (in mM) 140 NaCl, 25 imidazole, 1 EDTA, and a combination of protease (catalog no. P-8340, Sigma; St. Louis, MO) and phosphatase inhibitor cocktails (Catalog no. P-2850 and P-5726, Sigma), pH 7.4, and combined with 5 mg/ml C12E8 detergent at room temperature for 20 min. Samples were then subjected to ultracentrifugation at 37,000 g using a Beckman TLA 100.3 rotor. The resulting supernatant was divided equally between two microfuge tubes (∼400 μg each for preimmune control and antibody immunoprecipitation). Samples were precleared before the addition of relevant antibodies by preincubation of supernatants with 50 μl of protein A-agarose for 1 h at 4°C. Precleared supernatants were incubated with either 4 μl of B8 Ab or no Ab (monoclonal Ab control) overnight at 4°C. The next day, 40 μl (50% slurry) of washed suspended protein A-agarose beads were added to each sample and incubated for a further 2 h at 4°C. Beads were pelleted, washed four times with buffer III containing 0.5% C12E8, and then resuspended in 40 μl of Laemmli sample buffer. Beads were boiled for 5 min at 95°C and stored until further use in immunoblotting studies.
Crude membrane input and immunoprecipitated samples were resolved on either 7.5% (NCX1) or 15% (PLM) SDS-PAGE in a Tris-glycine buffer. Proteins were transferred to polyvinylidene difluoride membranes and blocked overnight in 5% nonfat milk-Tris-buffered saline with 0.05% Tween 20 (TBS-T). Primary antibody incubation was performed for 2–3 h at room temperature in 5% nonfat milk-TBS-T containing 1:1,000 R3F1 or 1:5,000 B8 Ab. The econdary antibody used was sheep anti-mouse IgG (1:2,000, Amersham). Immunoreactivity was detected using ECL chemiluminescence.
All results are expressed as means ± SE. For analysis of a parameter (e.g., maximal contraction amplitude) as functions of group (e.g., GFP vs. PLM vs. TM43) and [Ca2+]o, two-way ANOVA was used to determine statistical significance. Likewise, two-way ANOVA was used to analyze INaCa as a function of group and membrane voltage.
For analyses of protein abundance and whole cell capacitance, one-way ANOVA was used. A commercial software package (JMP version 4.0.5, SAS Institute; Cary, NC) was used. In all analysis, P < 0.05 was taken to be statistically significant.
Monoclonal B8 antibody recognizes the NH2-terminus of dog but not rat PLM.
Dog PLM was overexpressed in adult rat myocytes in our previous (23, 31, 32) and present studies. Whereas the COOH-termini of dog and rat PLM are identical, there are three amino acid differences in the extracellular NH2-termini and an additional three amino acid differences in the transmembrane domains between dog and rat PLM (Fig. 1) (19, 25). We exploited these differences to differentiate exogenous dog from endogenous rat PLM using the monoclonal B8 Ab, which was raised against dog PLM (3), and the polyclonal C2 Ab, which was raised against the COOH-terminal peptide of PLM (23). In crude membrane preparations from GFP and TM43 myocytes, B8 Ab did not recognize endogenous rat PLM (GFP) but recognized the dog PLM COOH-terminal truncation mutant TM43 (Fig. 2), suggesting that B8 Ab recognized the NH2-terminus/transmembrane domain of dog but not rat PLM. As expected from COOH-terminus truncation, using C2 Ab, there were no differences in protein band signal intensities between GFP and TM43 myocyte membranes (Fig. 2). This is because only the COOH-terminus from endogenous rat PLM but not from the exogenous dog PLM mutant (TM43) was present for detection. In addition, compared with endogenous rat PLM, TM43 migrated at a lower apparent molecular weight, consistent with COOH-terminal truncation.
To determine whether the NH2-terminus or transmembrane domain of PLM was recognized by B8 Ab, we performed immunofluorescence localization studies on TM43 myocytes. In nonpermeabilized TM43 myocytes (Fig. 3A), distribution of B8 Ab was primarily localized to the sarcolemma and intercalated discs. This observation indicates that B8 Ab recognizes the NH2-terminus of dog PLM, which lies in the extracellular domain. The absence of labeling by B8 Ab in t-tubules of nonpermeabilized myocytes most likely reflects access difficulties by the antibody to the restricted t-tubular space. Indeed, in permeabilized TM43 myocytes, B8 Ab labeling is present in t-tubules (Fig. 3B). As a control, when rat myocytes were infected with Adv containing only the coding sequence of mature dog PLM but without the NH2-terminal signal peptide (no ATG start codon) and therefore no dog PLM could be expressed, no B8 Ab signal was detected in permeabilized myocytes (Fig. 3C).
To further demonstrate the specificities of C2 and B8 antibodies, we obtained wide-field immunofluorescence images. The purpose was to show the different degrees of GFP and/or exogenous dog PLM expression due to variations in Adv infection from myocyte to myocyte. If B8 Ab is specific for dog PLM, then its signal should be proportional to the degree of GFP expression in PLM myocytes and be absent in myocytes infected with control Adv-GFP. On the other hand, the signal intensity of C2 Ab (which recognizes both dog and rat PLM) should bear no relationship to the degree of GFP expression in control GFP myocytes but be proportional to the extent of GFP expression in PLM myocytes. Infection of adult rat myocytes with control Adv-GFP resulted in variable degrees of GFP expression (Fig. 4A), as expected. However, endogenous rat PLM expression as detected by C2 Ab was relatively constant (Fig. 4B) and bears no relationship to the extent of GFP expression (Fig. 4A). Most striking is the absence of B8 Ab signal in GFP myocytes (Fig. 4C), indicating that B8 Ab does not recognize endogenous rat PLM. Infection of adult rat myocytes with Adv-GFP-PLM also resulted in variable GFP expression (Fig. 4D). In contrast to the relatively constant C2 Ab signals observed for control GFP myocytes (Fig. 4B), C2 Ab signal intensities were variable from cell to cell and roughly proportional to the degree of GFP expression (Fig. 4E). In addition, B8 Ab signals were clearly present at the sarcolemma and proportional to the degree of GFP expression (Fig. 4F). These wide-field indirect immunofluorescence images of GFP and PLM myocytes provide another line of evidence that B8 Ab is specific for dog PLM, whereas C2 Ab recognizes both dog and rat PLM.
Our results with Western blots (see Figs. 2 and 6), with myocytes infected with Adv containing the coding sequence of dog PLM but without the ATG start codon (Fig. 3C), and wide-field images of GFP and PLM myocytes (Fig. 4) clearly indicate that B8 Ab recognizes the NH2-terminus of dog but not rat PLM. By contrast, C2 Ab recognizes the COOH-termini of both endogenous rat and the overexpressed dog PLM (23, 31) (see Figs. 4 and 5C).
Overexpressed dog PLM is correctly targeted to the sarcolemma, t-tubules, and intercalated discs.
Figure 5 shows confocal images that demonstrate correct targeting of exogenous dog PLM in adult rat myocytes. GFP expressed under the control of a second CMV promoter of Adv-GFP-PLM was localized in the cytosolic compartment (Fig. 5A). Closer inspection suggests that GFP appears to coassociate with myofibrillar bundles. This is most likely due to fixation and/or permeabilization artifacts. In intact living GFP myocytes, diffuse GFP fluorescence was observed throughout the cytosolic compartment (image not shown). With the use of B8 Ab, which recognizes the NH2-terminus of dog but not rat PLM, correct targeting of exogenous dog PLM to the sarcolemma, t-tubules, and intercalated discs is demonstrated (Fig. 5B). There was little to no detectable accumulation of dog PLM in the cytosol or intracellular compartments (Fig. 5B). Figure 5C shows the localization of both endogenous rat and exogenous dog PLM detected with C2 Ab, which recognizes the COOH-termini of both. Again, PLM was not detected in the cytosol or intracellular compartments. Figure 5D is a merged image of Fig. 5, A–C, and shows colocalization of exogenous dog (blue) with endogenous rat (red) PLM. Note that PLM (blue + red = purple) is distributed in the sarcolemma, t-tubules, and intercalated discs, very different from the distribution of GFP.
We used B8 Ab to detect expression and localization of mutants of dog PLM (TM43, TM65, 3SlT, S63A, S68A, and S68E) in adult rat myocytes. Similar to wild-type dog PLM (Fig. 5B), all the above dog PLM mutants were correctly targeted to the sarcolemma, t-tubules, and intercalated discs, with little to no signal in intracellular organelles. For brevity, only images of TM43 (see Fig. 3B) and S68A (see Fig. 9B) myocytes are shown. Western blots (Fig. 6) demonstrated that overexpression of dog PLM or its mutants did not affect the expression of NCX1, SERCA2, α-subunits of Na+-K+-ATPase, and calsequestrin. Table 1 summarizes the Western blot protein signal band intensities. One-way ANOVA did not detect any significant differences in expression of these transport proteins among myocytes expressing GFP and PLM and its mutants.
Effects of overexpression of PLM mutants on myocyte contraction.
We have previously shown that compared with control GFP myocytes, PLM overexpression resulted in a larger contraction amplitude at 0.6 mM [Ca2+]o but a lower contraction amplitude at 5.0 mM [Ca2+]o (23). This is confirmed in the present study (Table 2). Deletion of the cytoplasmic tail of PLM (TM43) resulted in a loss of wild-type PLM's effect on cardiac contractility (Table 2). Furthermore, deletion of only seven amino acids (amino acids 66–72) from the COOH-terminus of PLM (TM65) also resulted in a complete loss of PLM's effect on contractility (Table 2). Because the last seven amino acids at the COOH-terminus contain Ser68, and because Ser68 is phosphorylated by both PKA and PKC (27), we next evaluated whether Ser68 was critical in PLM's modulation of cardiac contractility. Mutating all the potential phosphorylation sites to Ala (3S1T) resulted in an ineffective PLM mutant (Table 2). Mutating Ser63 to Ala but leaving Ser68 intact (S63A) resulted in a PLM mutant that was partially effective in modulating cardiac contractility (Table 2). A single point mutation of Ser68 to Ala (S68A) abolished PLM's effect on cardiac contractility, whereas substituting Ser68 with a negatively charged molecule (S68E) preserved PLM's influence on contraction amplitudes (Table 2).
Effects of overexpression of Ser63 and Ser68 PLM mutants on [Ca2+]i transients.
Compared with control GFP myocytes, PLM myocytes exhibited similar end-diastolic [Ca2+]i and half-time (t1/2) of [Ca2+]i decline, but higher systolic [Ca2+]i at 0.6 mM [Ca2+]o and lower systolic [Ca2+]i at 5 mM [Ca2+]o in our previous study (23). Because Ser68 was found to be critical in PLM's modulation of cardiac contractility, we next evaluated whether mutations in Ser68 also affected [Ca2+]i transients. Consistent with absence of effects on diastolic [Ca2+]i by wild-type PLM in our previous study (23), S63A, S68A, and S68E PLM mutants also did not affect diastolic [Ca2+]i (Table 3). The PLM mutant with intact Ser68 (S63A) affected systolic [Ca2+]i in a pattern similar to wild-type PLM (Table 3), but mutating Ser68 to Ala (S68A) completely eliminated PLM's effect on systolic [Ca2+]i (Table 3). Interestingly, replacing Ser68 with glutamic acid (S68E) did not affect PLM's ability to modulate systolic [Ca2+]i at low and high [Ca2+]o (Table 3).
The effects of the PLM mutants on the amplitudes of the [Ca2+]i transient, measured as the percent increase in the fura-2 fluorescence intensity ratio, followed the pattern of those on systolic [Ca2+]i (Table 3). Both S63A and S68E but not S68A reproduced the differential effects of low versus high [Ca2+]o on [Ca2+]i transient amplitudes observed in PLM-overexpressed myocytes (23).
The t1/2 of [Ca2+]i decline, an indicator of SR Ca2+ uptake activity (30), was not affected by either wild-type PLM (23) or its COOH-terminal Ser mutants (Table 3). This observation is consistent with the finding that SERCA2 protein levels were not affected by overexpression of PLM or its mutants (Fig. 6 and Table 1). The results of [Ca2+]i transient data suggest that the contraction abnormalities (or lack thereof) observed in myoctyes overexpressing PLM mutants was likely related to alterations in [Ca2+]i homeostasis.
Effects of overexpression of PLM and Ser63 and Ser68 PLM mutants on INaCa.
The pattern of contraction and [Ca2+]i transient amplitude changes observed in PLM-overexpressed myocytes (23) are similar to those observed in NCX1-downregulated rat myocytes (26) but opposite to those in which NCX1 was overexpressed (33). In addition, we have previously demonstrated that PLM colocalized with (31) and coimmunoprecipitated NCX1 (18). Taken together, these data suggest that PLM exerted its effects on cardiac contractility and [Ca2+]i homeostasis by inhibiting NCX1 activity. We therefore measured the effects of wild-type PLM and Ser68 mutants overexpression on INaCa in adult rat myocytes. Overexpression of PLM (157 ± 9 pF, n = 7), S63A (142 ± 6 pF, n = 8), S68A (157 ± 13 pF, n = 7), and S68E (143 ± 5 pF, n = 7) had no significant effect (P < 0.39) on whole cell capacitance (control GFP: 159 ± 7 pF, n = 9), indicating that the myocyte surface area was unchanged by overexpression of PLM or its Ser68 mutants. Figure 7 A shows the descending-ascending voltage-ramp protocol used in measuring INaCa. Figure 7B shows membrane currents in a myocyte overexpressing the S68E mutant of PLM under conditions in which Ca2+, K+, Cl−, and Na+-K+-ATPase currents were blocked and in the absence and presence of Cd2+. Figure 7C shows the difference current, i.e., the Cd2+-sensitive current. Note that with the exception of small contaminating Na+ current during the ascending voltage ramp, Cd2+-sensitive currents measured during the descending and ascending voltage ramps were similar. This suggests that [Ca2+]i and [Na+]i sensed by NCX1 were not appreciably changed by NCX1 fluxes during the voltage ramp. Figure 8A shows current-voltage relationships of INaCa density measured at 30°C and 5 mM [Ca2+]o in control GFP (circles), wild-type PLM (diamonds), S68A (triangles), S68E (squares), and S63A (inverted triangles) myocytes. Note all the currents crossed over at an apparent reversal potential of approximately −65 to −70 mV, close to the calculated equilibrium potential for INaCa (−73 mV) under our experimental conditions. Both forward (3 Na+ in: 1 Ca2+ out) and reverse INaCa (3 Na+ out: 1 Ca2+ in) were inhibited by PLM overexpression (group effect, P ≤ 0.0001), consistent with our previous finding that PLM overexpression resulted in reduced reverse INaCa in rat cardiac myocytes (31). In addition, the group × voltage interaction effect was highly significant (P < 0.0001), indicating that the differences in INaCa between GFP and PLM myocytes are amplified at more positive voltages. Both S68E (group effect, P < 0.0001; interaction effect, P < 0.0001) and S63A (group effect, P < 0.0001; interaction effect, P < 0.075) overexpression resulted in inhibition of INaCa, with S68E perhaps more effective than wild-type PLM. S68A overexpression, however, had no effects on INaCa (group effect, P ≤ 0.10; interaction effect, P < 0.9995) in adult rat myocytes.
Figure 8B shows the effects of Ser68 mutants on maximal contraction amplitudes measured at [Ca2+]o of 0.6 mM, expressed as a percentage of their respective GFP controls. It can be appreciated that the effects of PLM mutants on contractility are directly correlated with their ability to inhibit INaCa (Fig. 8A). It should also be noted that the lack of effect of S68A on myocyte contractility, [Ca2+]i homeostasis, and INaCa was not due to lack of correct expression of the PLM mutant (Figs. 6 and 9B).
Association of PLM and its mutants with NCX1.
We have previously demonstrated that PLM associated with NCX1 in adult rat cardiac myocytes (18). One plausible explanation for the lack of effect of S68A on myocyte contractility, [Ca2+]i homeostasis, and INaCa is that mutation of Ser68 to Ala resulted in a loss of association of S68A to NCX1, despite correct targeting of S68A to sarcolemma, t-tubules, and intercalated discs (Fig. 9B). To test the validity of this hypothesis, we performed coimmunoprecipitation experiments in HEK-293 cells coexpressing exogenous rat NCX1 and PLM or its Ser mutants. We chose HEK-293 cells because they are devoid of endogenous PLM and NCX1 (Fig. 9A). It should be pointed out that PLM in cardiac myocytes may form oligomers, as suggested by the formation of taurine-permeable ion channels by PLM in lipid bilayers (3). Therefore, an apparent observation that S68A coimmunoprecipitated with NCX1 in cardiac myocytes may be an artifact because NCX1, rather than associated with S68A, actually formed a NCX1-PLM/S68A complex. The lack of background endogenous PLM in HEK-293 cells avoids this ambiguity. Figure 9A demonstrates that PLM (full effect on myocyte contractility, [Ca2+]i homeostasis, and INaCa), S63A (partial effect), and S68A (no effect) were all capable of coimmunoprecipitating exogenous rat NCX1, indicating the lack of effect of the S68A mutant was not due to a loss of association with NCX1.
The first major finding of the present study is that the monoclonal B8 antibody specifically recognized the NH2-terminus of dog but not rat PLM (Figs. 2–4). This serendipitous finding allowed us to unambiguously detect expression of exogenous dog PLM and its COOH-terminal mutants in adult rat cardiac myocytes (Figs. 3–5 and 9). Indeed, confocal microscopy demonstrated that cellular localization of exogenous dog PLM (and its COOH-terminal mutants) was similar to that of endogenous rat PLM, with little to no accumulation of exogenous dog PLM in intracellular organelles (Fig. 5). This important observation suggests that the inhibitory effects of overexpressed PLM on NCX1 (31) and Na+-K+-ATPase (32) in adult rat myocytes were not due to sequestration of these ion transporters by overexpressed PLM in intracellular compartments.
Overexpression (23) or downregulation (18) of PLM in adult rat myocytes resulted in an altered dynamic range of contraction and [Ca2+]i transient amplitudes in response to increasing [Ca2+]o, indicating that one of the physiological functions of PLM is the regulation of cardiac contractility. Deletion of the COOH-terminus of PLM resulted in a complete loss of activity (TM43; Table 1), suggesting either the entire COOH-terminus or the four potential phosphorylation sites contained within it mediated PLM's effects on cardiac contractility. Mutating four phosphorylation sites to Ala (3S1T; Table 1) or deleting the two more distal phosphorylation sites (TM65; Table 1) also resulted in a loss of PLM's effect on myocyte contractility, thereby narrowing the focus to Ser68 and/or Thr69. Because Ser68 is the common target for phosphorylation by PKA and PKC (27), we next mutated Ser68 and observed the effect on myocyte contractility. Mutating Ser68 to Ala (S68A; Table 1) resulted in a complete loss of PLM's effect on contractility. By contrast, preserving Ser68 but mutating Ser63 (a target for PKC; Ref. 27) to Ala (S63A; Table 1) partially maintained PLM's modulating effect on contraction. Surprisingly, placing a negative charge at Ser68 by mutating it to glutamic acid (S68E; Table 1) resulted in a mutant whose effect on contractility was indistinguishable from that of wild-type PLM. These observations, when taken together, indicate that Ser68 is critical in mediating PLM's effect on myocyte contractility. It should be reemphasized that the lack of effect of the S68A mutant on cardiac contractility was not due to incorrect targeting (Fig. 9B) or a loss of association with NCX1 (Fig. 9A).
[Ca2+]i occupies a central role in cardiac excitation-contraction coupling. The mechanism by which PLM regulated cardiac contractility is most likely mediated by modulating Ca2+ fluxes, as we have previously proposed (18, 23). Indeed, in rat cardiac myocytes in which PLM was overexpressed, both systolic [Ca2+]i and [Ca2+]i transient amplitudes were higher at 0.6 mM [Ca2+]o but lower at 5 mM [Ca2+]o (23), thereby mirroring the changes in contraction amplitudes (Tables 2 and 3). Focusing on Ser68 mutants, mutating Ser68 to Ala (S68A) resulted in [Ca2+]i transient characteristics indistinguishable from those of control GFP myocytes (Table 3). By contrast, both S68E and S63A PLM mutants altered [Ca2+]i transient behavior in a manner similar to that observed for wild-type PLM (Table 3). The agreement between contraction and [Ca2+]i transient data in myocytes overexpressing PLM or its Ser68 mutants supports our hypothesis that PLM exerts its effect on contractility by regulating Ca2+ fluxes and that Ser68 is intimately involved in mediating PLM's effect on [Ca2+]i transients and contractility.
Of the many ion transport pathways known to modulate Ca2+ fluxes during excitation-contraction, to date only Na+-K+-ATPase and NCX1 are known to be influenced by PLM in the adult heart (10, 18, 31, 32), although other members of the FXYD family have been shown to regulate Na+-K+-ATPase activity in noncardiac tissues (6, 7, 16). On the basis of the observed differential effects of high versus low [Ca2+]o on [Ca2+]i transient and contraction amplitudes in myocytes in which PLM was either overexpressed or downregulated, we have previously argued that inhibition of NCX1, rather than Na+-K+-ATPase, likely explained PLM's effect on cardiac contractility and [Ca2+]i homeostasis (18, 23, 31). To reiterate, inhibition of Na+-K+-ATPase in PLM-overexpressed myocytes (10, 32) would be expected to elevate [Na+]i, resulting in decreased forward Na+/Ca2+ exchange but increased reverse Na+/Ca2+ exchange. This would result in increased contraction and [Ca2+]i transient amplitudes at both low and high [Ca2+]o, similar to what one would observe with digitalis glycoside, a well-known Na+-K+-ATPase inhibitor. This was not observed. Rather, inhibition of NCX1, which mediates both Ca2+ influx and efflux during a cardiac cycle (1, 4), could theoretically explain the observed PLM's effect on cardiac contractility and [Ca2+]i transients (18, 23, 31). For this reason, we focused on the effects of PLM and its Ser68 mutants on Na+/Ca2+ exchange activity. Both forward and reverse INaCa were lower in PLM-overexpressed myocytes (Fig. 8), consistent with our previous observation that reverse INaCa was decreased in rat myocytes overexpressing dog PLM (31). Overexpressing the S68A mutant had no effect on INaCa, whereas overexpressing S68E or S63A mutants inhibited INaCa. The fact that inhibition of INaCa by PLM, S68E, and S63A mutants correlated with their effects on contractility and [Ca2+]i transients and that lack of inhibition of INaCa by S68A mutant correlated with absence of effects on [Ca2+]i transients and contractility (Fig. 8) lends strong support to our hypothesis that the effects of PLM on contractility is mediated, at least in part, by its inhibition of NCX1.
Ser68 in PLM is phosphorylated by both PKA and PKC (27). NCX1 activity is known to be modulated by α-adrenergic stimulation (2), presumably mediated via PKC (14). Interestingly, altered NCX1 activity by PKC did not require direct phosphorylation of NCX1 (13). It is at present controversial as to whether INaCa is affected by isoproterenol treatment (2, 12, 29). With respect to Na+-K+-ATPase, its activity in guinea pig ventricular myocytes was increased by PKA (11) and α-adrenergic stimulation, the latter presumably mediated by PKC (28). Recent evidence suggested that the putative PKA phosphorylation site (Ser943) in the α1-subunit of Na+-K+-ATPase (9) was internalized and not accessible to protein kinases (24). Indeed, increased Na+-K+-ATPase activity after transient cardiac ischemia was found to be mediated by PKA but not associated with enhanced phosphorylation of the α1-subunit (10). Rather, it was proposed that increased PLM phosphorylation by PKA during ischemia lessened its inhibition of Na+-K+-ATPase, thereby enhancing Na+ pump activity (10). These observations, when taken together, suggest a model in which the regulatory influences of PKA and PKC on Na+-K+-ATPase and NCX1 are mediated by phosphorylating PLM and not the individual ion transporters. Viewed in this context, Ser68 in PLM occupies a strategic role in the regulation of NCX1 and Na+-K+-ATPase activity. Our present data on PLM Ser68 mutants are consistent with this model in that conformational changes at Ser68 (by point mutation) had profound influences on INaCa and cardiac contractility. In addition, PLM phosphorylation at Ser68 was reported to be important in mediating forskolin-induced relaxation of vascular smooth muscle, presumably due to enhanced Na+-K+-ATPase activity with resultant increased Ca2+ efflux via Na+/Ca2+exchange (17).
Despite the fact that the present study did not directly address the effects of phosphorylation of PLM on INaCa, contractility, and [Ca2+]i transients, some inferences may be gleaned from results obtained with the PLM Ser mutants. Assuming that S68A and S68E mutants mimic PLM with 100% dephosphorylated and phosphorylated Ser68, respectively, our data on INaCa are consistent with a model in which phosphorylation at Ser68 results in inhibition of NCX1 activity. Because the degree of inhibition of INaCa by wild-type PLM was between that observed for S68A (no inhibition) and S68E (presumably maximal inhibition), the data also suggest that Ser68 in wild-type PLM was partially phosphorylated in the basal state.
On the basis of the INaCa data measured at +100 mV (Fig. 8A), the percent inhibition of INaCa by overexpressed PLM, S68A, and S68E can be estimated to be 27.2, 7.1, and 51.1, respectively, compared with GFP control (INaCa = 4.52 ± 0.12 pA/pF). Assuming that 1) wild-type PLM and its serine mutants are essentially equally expressed in myocytes; 2) Ser63 and Ser68 are the only physiologically relevant phosphorylation sites in cardiac PLM (27); 3) S68A and S68E mutants mimic 100% dephosphorylated and phosphorylated Ser68, respectively; and 4) the degree of phosphorylation of PLM (especially at Ser68) is linearly proportional to degree of inhibition of INaCa, the fraction of Ser63 (x) and that of Ser68 (y) in PLM that is phosphorylated in the basal state can be estimated to be 0.16 and 0.46, respectively. This estimate can be compared with that obtained from 32P incorporation studies into PLM in intact guinea pig ventricles (20). The incorporation of 32P into PLM (15-kDa protein) was 84.4 ± 7.4 pmol/mg protein under basal conditions and increased to 218.4 ± 14.8 pmol/mg protein with isoproterenol stimulation (20). From these data, and assuming that 1) isoproterenol treatment resulted in 100% phosphorylation of Ser68, and 2) Ser63 and Ser68 are the only physiologically relevant phosphorylation sites in cardiac PLM, the term (x + y)/(x + 1) could be estimated to be ∼0.4, where x and y are the fractions of Ser63 and Ser68 phosphorylated under basal conditions. This value of ∼0.4 for (x + y)/(x + 1) determined from 32P incorporation studies compares reasonably well with the value of ∼0.5 for (x + y)/(x + 1) estimated from experiments on the inhibition of INaCa by PLM Ser mutants.
There are limitations to the present study. The first is that we did not measure [Ca2+]i transient and contraction amplitudes in control GFP and PLM myocytes and myocytes overexpressing PLM mutants at an intermediate [Ca2+]o of 1–2 mM. Our previous results indicate that the effects of overexpressing (23) or downregulating PLM (18) on [Ca2+]i transient and contraction amplitudes were sufficiently characterized with measurements at 0.6 and 5.0 mM [Ca2+]o. The second limitation is that we did not evaluate the effects of the PLM mutants on Na+-K+-ATPase activity. This is because we favor inhibition of NCX1 by PLM as an explanation for PLM's effect on cardiac contractility. Nevertheless, the important regulatory influence of PLM on Na+-K+-ATPase activity (10, 32) would argue for evaluation of the effects of these PLM mutants on Na+-K+-ATPase activity in a future study. An important limitation is that we did not study the effects of PKA and PKC stimulation (by forskolin and phorbol esters, respectively, or by hormones such as isoproterenol, angiotensin II, and endothelin-1) on contractility and [Ca2+]i transients. These agents affect many pathways involved in cardiac excitation-contraction coupling, e.g., L-type Ca2+ current, SERCA2 activity, myosin phosphorylation, etc. It would be difficult to unambiguously ascribe the effect (e.g., enhanced contractility) of these agents to increased PLM phosphorylation alone. A systematic discussion of the approach to study the effects of PLM phosphorylation on cardiac contractility and [Ca2+]i homeostasis (e.g., by using PLM-null mice) is beyond the scope of the present study. Finally, we have totally ignored the osmoregulatory properties of PLM (8). Whole cell capacitance, however, was not changed in myocytes overexpressing PLM or its Ser68 mutants. This observation suggests that steady-state myocyte membrane surface area, and by extrapolation cell volume, was not affected by PLM (or its Ser68 mutants) overexpression.
In summary, we demonstrated that PLM overexpression affected cardiac contractility and both forward and reverse INaCa. Using deletion and substitution mutants, we identified Ser68 as critical in mediating PLM's effects on contractility, [Ca2+]i transients, and Na+/Ca2+ exchange activity. On the basis of the percent inhibition of INaCa by PLM Ser68 mutants and making some simplifying assumptions, we estimated that ∼16% of Ser63 and ∼46% of Ser68 in PLM were phosphorylated in the basal state. Expression of SERCA2, NCX1, the α-subunit of Na+-K+-ATPase, and calsequestrin was not affected by overexpression of PLM or its mutants. Because Ser68 in PLM is the common phosphorylation target for PKA and PKC, and because regulation of ion transport does not necessarily require direct phosphorylation of NCX1 by PKA or PKC, we speculate that the effects of PKA and PKC on Na+/Ca2+ exchange in cardiac myocytes may involve PLM Ser68 phosphorylation.
This study was supported in part by National Institutes of Health Grants HL-58672 (to J. Y. Cheung), DK-46678 (to J. Y. Cheung, coinvestigator), GM-46691 (to L. I. Rothblum), HL-70548 (to J. R. Moorman), GM-64640 (to J. R. Moorman), and HL-69074 (to A. L. Tucker); American Heart Association, Pennsylvania Affiliate, Grants-In-Aid 0265426U (to X.-Q. Zhang) and 0355744U (to J.Y. Cheung); American Heart Association, Pennsylvania Affiliate, Postdoctoral Fellowship 0425319U (to B. A. Ahlers); and grants from the Geisinger Foundation (to J. Y. Cheung and L. I. Rothblum).
We thank Caitlin Custer for assistance in preparation of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society