INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Phospholamban is a 52-amino acid integral membrane protein of cardiac sarcoplasmic reticulum (SR), which is the principal regulator of the Ca-ATPase in this membrane system. The Ca-ATPase is a 110-kDa integral membrane protein of cardiac SR that transports Ca²⁺ ions into the SR to promote cardiac muscle relaxation (reviewed by Simmerman and Jones, 1998). Phospholamban interacts with the Ca-ATPase at specific cytoplasmic binding sites (Toyofuku et al., 1994a,b), but transmembrane interactions between the two proteins also appear to be intimately involved in the inhibition process (Kimura et al., 1996). It has been well demonstrated that dephosphorylated phospholamban inhibits Ca-ATPase activity by decreasing enzyme sensitivity to Ca²⁺ for activation of ATP hydrolysis (MacLennan et al., 1992; Voss et al., 1994; Autry and Jones, 1997). The inhibition of the Ca-ATPase is relieved by phosphorylation of phospholamban at Ser¹⁶ or Thr¹⁷ or by the binding of a monoclonal antibody against the cytoplasmic domain of phospholamban (Autry and Jones, 1997), resulting in a marked increase in Ca-ATPase activity. The majority of studies have shown that phospholamban does not alter Ca-ATPase V_max significantly; it alters only the [Ca²⁺] required to produce maximum activity (Kranias, 1985; MacLennan et al., 1992; Jones and Field, 1993; Cantilina et al., 1993; Reddy et al., 1996; Autry and Jones, 1997; Chu et al., 1998), although this point is still contested (Antipenko et al., 1997a; Simmerman and Jones, 1998).

Despite our knowledge of the functional effects of phospholamban on Ca-ATPase activity, fundamental questions remain about the mechanism by which phospholamban exerts its inhibitory interaction (Scheme 1). A number of investigators have previously studied the effect of phospholamban on the partial reactions of the Ca-ATPase cycle (Shigekawa et al., 1976; Jones et al., 1978; Sumida et al., 1978, 1980; Tada et al., 1979, 1980; Kranias et al., 1980; Cantilina et al., 1993; Hughes et al., 1994; Antipenko et al., 1997a, 1999). Unfortunately, these studies have provided conflicting results about which Ca-ATPase partial reaction(s) are affected by phospholamban, and thus a unified model has been slow to emerge. Several studies (Tada et al., 1979; Antipenko et al., 1997a, 1999) have suggested that phospholamban controls Ca-ATPase activity by decreasing by twofold the rate of phosphoenzyme decomposition. Kinetics studies of Ca-ATPase in skeletal muscle SR by Froehlich and Taylor (1975, 1976) have shown that Ca-ATPase phosphoenzyme decomposition is a two-step process (steps 5 and 6, Scheme 1), in which E2P is first hydrolyzed to E2·P_i, followed by the release of inorganic phosphate (P_i) from the enzyme forming the E2 intermediate. The relative rates of these two steps are modulated by a variety of factors, including ionized [Ca²⁺], suggesting that the inhibition of Ca-ATPase phosphoenzyme decomposition by phospholamban can account for the effect of phospholamban on the [Ca²⁺] dependence of Ca-ATPase activity (Simmerman and Jones, 1998). In contrast, however, a number of investigators (Cantilina et al., 1993; Negash et al., 1996; Antipenko et al., 1997a) have provided evidence that phospholamban inhibits Ca-ATPase cycling by increasing the activation energy of a slow Ca²⁺-dependent conformational change, which is required for ATP-dependent phosphoenzyme formation, and which occurs during the E1 + 2Ca²⁺ to E1·Ca₂ transition (step 2, Scheme 1). Previously, Inesi et al. (1980) showed that the cooperative binding of two Ca²⁺ ions to the Ca-ATPase occurs in two discrete steps separated by a conformational change (Scheme 2). The binding of the first Ca²⁺ ion to the Ca-ATPase E1 intermediate (forming E1·Ca) stimulates a conformational change in the enzyme (E1·Ca to E1'·Ca), which increases the affinity of the enzyme for binding the second Ca²⁺ ion, forming E1'·Ca₂. Subsequently, Cantilina et al. (1993) proposed that phospholamban decreases by 10-fold the forward and reverse rate constants for the E1·Ca to E1'·Ca transition. While this model can account for the shift in the [Ca²⁺] dependence of Ca-ATPase activity induced by phospholamban, the proposal remains to be tested quantitatively. Clearly, these mechanistic issues have to be resolved to provide a better understanding of the inhibition of Ca-ATPase by phospholamban.

View larger version (13K): [in this window] [in a new window]   Scheme 1

View larger version (11K): [in this window] [in a new window]   Scheme 2

In the present work, we have carried out kinetics studies of the cardiac Ca-ATPase to test the hypotheses that phospholamban inhibits Ca-ATPase cycling by decreasing the rate of the E1·Ca to E1'·Ca transition and/or the rate of phosphoenzyme hydrolysis. To assist in these studies, we have made use of the baculovirus-Sf21 insect cell (Spodoptera frugiperda insect cell) expression system to study Ca-ATPase-phospholamban functional interactions. An advantage of this system is that sufficient expressed material is produced to allow detailed kinetics characterizations, including phosphoenzyme formation and decomposition studies. Expression systems previously used to study the regulatory effects of phospholamban on the Ca-ATPase have yielded much lower levels of protein (discussed in Autry and Jones, 1997), precluding the type of kinetics studies presently reported. In addition, this expression system facilitates the study of the phosphorylation and dephosphorylation kinetics of the cardiac Ca-ATPase expressed by itself, independently of the influence of phospholamban, compared directly with parallel kinetics studies of the Ca-ATPase coexpressed with wild-type pentameric phospholamban (the natural inhibitor) or with the monomeric L37A-phospholamban mutant (a potent superactive mutant or "super shifter") (Simmerman et al., 1996; Autry and Jones, 1997; Kimura et al., 1997). Any effect of wild-type phospholamban on Ca-ATPase kinetics would be expected to be amplified in the presence of the more potent L37A-phospholamban and absent when the Ca-ATPase is expressed alone without phospholamban. Finally, to further document the specific effect of phospholamban on Ca-ATPase kinetics, we took advantage of our monoclonal antibody 2D12 raised against phospholamban (Sham et al., 1991; Briggs et al., 1992), which selectively reverses phospholamban inhibitory effects on Ca-ATPase, like phospholamban phosphorylation (Cantilina et al., 1993; Antipenko et al., 1997b). Use of the anti-phospholamban antibody allowed large amounts of membrane material to be utilized for kinetics studies, removing the need for protein kinase activation and inhibition of phosphatase activity during lengthy kinetics experiments. Our results show that phospholamban acts by decreasing Ca-ATPase affinity for Ca²⁺, not by slowing the rate of the E1·Ca to E1'·Ca conformational transition (Cantilina et al., 1993; Negash et al., 1996; Antipenko et al., 1997a) or the rate of dephosphorylation of the E2P intermediate (Tada et al., 1979; Antipenko et al., 1997a, 1999), as previously proposed by others.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials

[-³²P]ATP was purchased from ICN, and ¹²⁵I-labeled protein A was purchased from DuPont NEN. Prestained molecular weight standards and polyvinylidene difluoride (PVDF) membranes were purchased from BioRad. Sf21 insect cells were purchased from Invitrogen, and the BaculoGold system was obtained from Pharmigen.

Protein expression and isolation

Recombinant baculoviruses containing cDNA inserts for canine cardiac Ca-ATPase (SERCA2a) or canine phospholamban were prepared using the BaculoGold system as recently described (Autry and Jones, 1997). Wild-type canine SERCA2a and canine phospholamban (wild type and mutant) were expressed in Sf21 cells grown in suspension (1.5 × 10⁶ cells/ml) at 27°C in Grace's insect cell medium (Life Technologies) supplemented with 10% fetal bovine serum (Atlanta Biologicals) and containing 0.1% Pluronic F-68 (Life Technologies). Microsomes were isolated from insect cells harvested 48 h after infection with baculoviruses. For expression of the Ca-ATPase alone, a multiplicity of infection of 10 (viruses per cell) was used. For coexpression of the Ca-ATPase and phospholamban, a multiplicity of infection of 15 was used for SERCA2a and 5 for phospholamban (wild type and mutant). Virus-infected Sf21 cells in 600 ml of suspension (9 × 10⁸ cells) were sedimented and then washed twice with an ice-cold solution containing 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.4 mM KH₂PO₄ (pH 7.4), by centrifugation for 5 min at 1500 rpm at 4°C in an IEC GP8R refrigerated centrifuge. The washed cells were resuspended in 100 ml of ice-cold 10 mM NaHCO₃ containing 10 µg/ml aprotinin, 2 µg/ml leupeptin, 1 µg/ml pepstatin A, and 0.1 mM pefabloc, and then homogenized for 90 s with a Brinkman Polytron (~50% full speed) in a cold room. The homogenate was centrifuged for 20 min at 9000 rpm (10,000 × g), 4°C, in a Sorvall SS34 rotor. The supernatant was collected and the Sf21 insect cell microsomes were pelleted by centrifugation for 35 min at 26,000 rpm (70,000 × g), 4°C, in a Beckman Ti45 rotor. The pellets were resuspended to ~5 mg/ml in 250 mM sucrose, 30 mM histidine (pH 7.4) and stored in small aliquots at 50°C. Protein concentrations were determined by the method of Lowry et al. (1951), using bovine serum albumin (Sigma) as a standard. The average yield per 600 ml of infection was 30 mg of microsomal protein. Additional details of the purification and characterization of Sf21 insect cell microsomes are provided elsewhere (Autry and Jones, 1997).

Electrophoresis and immunoblotting

Before electrophoresis, samples were solubilized at 37°C for 5 min in a dissociation medium that contained 62.5 mM Tris (pH 6.8), 5% glycerol, 5% sodium dodecyl sulfate (SDS), 40 mM dithiothreitol, and 0.0025% bromphenol blue. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (BioRad Mini-Protean II System) was conducted by the method of Porzio and Pearson (1977), using 8% polyacrylamide. Kaleidoscope prestained molecular weight markers (BioRad) were used as standards. Gels were stained using GelCode Blue Stain reagent (Pierce), or proteins were transferred (BioRad Mini-Transblot System) to PVDF membranes (BioRad) for immunoblotting. The transfer protocol was carried out according to the instructions provided by the manufacturer, except that methanol was excluded from the transfer buffer. We found that the use of methanol in the transfer buffer resulted in a significant amount of protein being retained in the polyacrylamide gel after transfer. In addition, a second membrane sheet was used for each blot, to capture protein that migrated through the first sheet without binding. The PVDF membranes were probed with anti-SERCA2a monoclonal antibody 2A7-A1 for detection of SERCA2a or with anti-phospholamban monoclonal antibody 2D12 for detection of phospholamban (Movsesian et al., 1994). Antibody-binding proteins were visualized using [¹²⁵I]-protein A, and labeling intensities from both membranes were quantified and summed using a Molecular Dynamics Phosphorimager SI. For quantitative immunoblotting, SERCA2a was purified from canine cardiac SR vesicles, and recombinant canine phospholamban was purified from Sf21 insect cells by 2A7-A1 or 2D12 monoclonal antibody affinity chromatography, respectively. These samples were used as standards to quantify the amount of SERCA2a and phospholamban in the Sf21 insect cell microsomes (described below). We found that for SERCA2a, 80% of the blot intensity appeared on the first membrane, and the remaining 20% on the second membrane. For pentameric phospholamban, more than 99% of the blot intensity appeared on the first membrane, with only a faint trace of blot intensity on the second membrane. For monomeric phospholamban, 85% of the blot intensity appeared on the first membrane, with the remaining 15% of the intensity localized to the second membrane. No GelCode Blue-stained protein bands in the vicinity of SERCA2a or phospholamban were retained on the polyacrylamide gel after transfer to the PVDF membranes.

SERCA2a ATPase assay

[Ca²⁺]-dependent ATPase activity of SERCA2a in the Sf21 insect cell microsomes was measured colorimetrically at 37°C, using a malachite green-ammonium molybdate assay (Lanzetta et al., 1979; Mahaney et al., 1995). SERCA2a incubation tubes contained 0.05 mg/ml protein in 50 mM 3-[N-morpholino]propanesulfonic acid (MOPS) (pH 7.0), 3 mM MgCl₂, 100 mM KCl, 1 mM EGTA, and 0-1.0 mM CaCl₂, to give the desired ionized [Ca²⁺], as previously determined (Autry and Jones, 1997). To initiate the ATPase reaction, 5 mM MgATP was added to the incubation tube. After 10 min of reaction at 37°C, a 50-µl aliquot of the incubation mixture was transferred into an assay tube containing 1.6 ml of malachite green-ammonium molybdate reagent at room temperature. After 30 s, the colorimetric reaction was quenched by the addition of 200 µl of 34% sodium citrate (Sigma) into the assay tube. For the determination of phosphate, a standard curve was constructed using aliquots of a 0.65 mM phosphate standard solution (Sigma), assayed in a similar fashion. After 30 min of color development, the absorbance of the malachite green reagent was measured at 660 nm. SERCA2a samples were pretreated with the Ca²⁺ ionophore A23187 (CalBiochem) (20 µg/mg of total protein) before addition to the incubation tubes. Before preparation of the incubation tubes, the SERCA2a samples were incubated for 20 min on ice without or with affinity-purified anti-phospholamban monoclonal antibody 2D12, at an antibody-to-total protein weight ratio of 1:1 (Autry and Jones, 1997).

ATP-dependent phosphoenzyme formation

Ca-ATPase phosphoenzyme formation experiments were carried out using ice-cold solutions. Mixing was conducted in a cold room (2 ± 1°C) to prevent significant sample warming during additions and vortexing. Before phosphoenzyme formation, the SERCA2a-containing Sf21 insect cell microsomes were permeabilized by the addition of Ca²⁺ ionophore A23187 (20 µg/mg total protein). Next, 0.25 ml of a solution containing 0.2 mg Sf21 microsomes/ml in 50 mM MOPS (pH 7.0), 3 mM MgCl₂, 100 mM KCl, 1 mM EGTA was placed in a 7-ml glass scintillation vial (VWR) and set on ice. The EGTA preincubation was designed to remove all traces of Ca²⁺ from the medium and to stabilize SERCA2a in a Ca²⁺-free state. To initiate phosphoenzyme formation, 0.25 ml of an ice-cold solution containing 50 mM MOPS (pH 7.25), 3 mM MgCl₂, 100 mM KCl, 1 mM EGTA, varying CaCl₂, and 2 µM [-³²P]ATP (20,000 cpm/nmol) was added rapidly (less than 1 s) to the microsome-containing vial with intermittent vortexing. The final conditions after mixing were 50 mM MOPS (pH 7.0), 3 mM MgCl₂, 100 mM KCl, 1 mM EGTA, and either 0.5 mM CaCl₂ (ionized [Ca²⁺]_final = 268 nM), 0.7 mM CaCl₂ (ionized [Ca²⁺]_final = 625 nM), or 1 mM CaCl₂ (ionized [Ca²⁺]_final = 15 µM). The reaction was allowed to proceed for different times before quenching by the rapid addition of 0.5 ml of ice-cold 9% perchloric acid + 6 mM H₃PO₄, followed by vigorous vortexing. Blank tubes were prepared by first adding 0.5 ml of quench solution to the 0.25-ml microsome-containing solution, vortexing vigorously, then adding 0.25 ml of the ATP-containing solution. A 25-µl aliquot of 10 mg/ml bovine serum albumin was added to each quenched sample to act as a carrier protein during the processing of the sample vials. The quenched samples were pelleted by centrifugation for 10 min at 3000 × g, 4°C, in an IEC GP8R refrigerated centrifuge, and then washed three times by similar centrifugation using an ice-cold solution of 5% trichloroacetic acid, 6% polyphosphoric acid, 4 mM H₃PO₄, and 5 mM nonradioactive ATP. Pellet recovery after washing was greater than 95%, determined by protein assay. The final pellets were dissolved in 5 ml 1 N NaOH, and the [³²P]phosphoenzyme was assayed by counting the Cerenkov radiation.

Separation and analysis of ³²P_i in the presence of [-³²P]ATP

[³²P]P_i liberation during phosphoenzyme formation was assayed according to Mahaney et al. (1995). After the first sedimentation of the quenched phosphorylation sample, 0.5 ml of the supernatant was added to 0.5 ml of a 2% aqueous charcoal slurry. The sample was vortexed and centrifuged for 10 min at 3000 × g, 4°C. A 0.5-ml aliquot of the supernatant was added to another 0.5 ml of a 2% aqueous charcoal slurry, vortexed, and centrifuged as before. A 0.5-ml aliquot of the supernatant from this second centrifugation was added to 0.75 ml of water saturated with 1:1 isobutanol:xylene contained in a 16 × 100 mm borosilicate assay tube and vortexed. To each tube was added 0.4 ml of 5% ammonium molybdate (in 4 N HCl) reagent, and the samples were vortexed briefly. Complex formation between P_i and the molybdate (yellow color development) was allowed to proceed for 15 min. After the incubation, 2 ml of water-saturated 1:1 isobutane:xylene was added to each tube and vortexed 3 × 10 s with 10 s between vortexing periods. The phases were allowed to separate for 10 min, after which 1 ml of the organic (upper) phase was pipetted into 5 ml of 1 N NaOH contained in a 7-ml glass scintillation vial. The vials were capped and shaken until the yellow color disappeared, and ³²P was assayed by counting the Cerenkov radiation. All steps involving organic solvents were carried out in a fume hood, and tubes containing solvents were covered when not in use, to minimize solvent loss due to evaporation.

Dephosphorylation of the phosphoenzyme by EGTA

SERCA2a samples were phosphorylated with 1 µM [-³²P]ATP at 0°C for 30 s as described above, followed by the rapid addition of 0.25 ml of solution containing 50 mM MOPS (pH 7.0), 3 mM MgCl₂, 100 mM KCl, variable CaCl₂ (as defined above), and 15 mM EGTA. The final [EGTA] after mixing was 5 mM, which was sufficient to reduce the ionized [Ca²⁺] to less than 1 nM. The reaction was allowed to proceed for various time periods before quenching by the rapid addition of 0.75 ml of ice-cold 9% perchloric acid + 6 mM H₃PO₄, followed by vigorous vortexing. A zero-time sample was obtained by quenching the phosphorylation reaction at 30 s and adding 0.5 ml of the EGTA chase solution to the quenched sample. Sample blanks were prepared as described above, except for the addition of 0.5 ml of the EGTA-containing solution to the quenched sample. The quenched samples were processed and assayed for ³²P-containing enzyme as described above.

Curve fitting and kinetic modeling

[Ca²⁺]-dependent ATPase activity data and phosphoenzyme kinetics data were fit using the program KFIT written by N. C. Millar. The amplitude and rate parameters were constrained to positive numbers and allowed to vary without bound during the fits. The best fits of the data were chosen on the basis of optimization of the determination coefficient, R², and/or minimization of the sum-of-squares error, ². Simulation of the time courses of phosphoenzyme formation, P_i liberation, and EP decay were carried out using the program KINSIM (Barshop et al., 1983), the reactions of Scheme 3, and the rate constants presented in Table 6 (Results). Because the rate of the E1P-to-E2P transition (Scheme 1, step 4) is significantly faster (>300 s¹; Mahaney et al., 1995) than any of the other steps under the conditions of our experiments (described below), the phosphorylated Ca-ATPase was treated as a single species in the simulations. Initial values for the rate constants were obtained in the present study (see Results) and from Cantilina et al. (1993). The rate constants reported by Cantilina et al. define the Ca-ATPase cycle at 25°C; thus these constants were reduced by a factor of 5 for the present simulations (0°C), based on the general assumption that the reaction rates decreased by a factor of 2 for each 10°C decrease in temperature (i.e., Q₁₀ of 2). The EGTA + Ca²⁺ equilibration, which occurred upon the simultaneous addition of Ca²⁺ and ATP to the microsomes preequilibrated in EGTA, was not modeled in the simulations, because this equilibration is rapid (Smith et al., 1984; Harrison and Bers, 1989) relative to the reaction steps of the SERCA2a cycle under our experimental conditions. The amplitude and rate parameters were constrained to positive numbers and allowed to vary without bound during the simulations. The goodness of fit of the simulation to the data was judged by maximizing the R² value for the simulation.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Protein expression, characterization, and assay

For this study, SERCA2a was expressed in Sf21 insect cells alone, coexpressed with wild-type phospholamban, or coexpressed with monomeric mutant L37A-phospholamban (Autry and Jones, 1997), and isolated as Sf21 insect cell microsomes. SDS-PAGE (Fig. 1 A) and immunoblotting (Fig. 1 B) of the Sf21 microsomes indicated that Sf21 cells produced similar amounts of recombinant Ca-ATPase and phospholamban in the expressed samples. Furthermore, the gel and the blots of the microsomal samples (Fig. 1, A and B, lanes 1-3) confirmed that there was almost no high-molecular-weight SERCA2a aggregate in the expressed samples. The amount of SERCA2a and phospholamban in the Sf21 microsome samples was determined by quantitative immunoblotting (Fig. 1, B-D), using purified SERCA2a and phospholamban samples as standards. As shown in Fig. 1, the purified proteins (empty symbols) provided linear standard curves that showed SERCA2a microsomes contained 8.3% SERCA2a and no phospholamban, SERCA2a + wild-type phospholamban microsomes contained 3% SERCA2a and 0.26% phospholamban, and SERCA2a + L37A-phospholamban microsomes contained 4.6% SERCA2a and 0.36% phospholamban (Table 1). The blot data for the expressed samples shown in Fig. 1 (C and D) was obtained using 10 µg of microsomal protein. Similar experiments conducted using 5 µg of microsomal protein (not shown) provided nearly identical results. Thus, even though the SERCA2a blot intensity of the SERCA2a alone sample is greater than that of the standards used (Fig. 1 C, diamonds), the extrapolated slope of the standard curve provides a reliable value for the amount of SERCA2a in this expressed sample. Using the molecular masses for canine SERCA2a (110,000 Da) (Autry and Jones, 1997) and phospholamban (6080 Da) (Fujii et al., 1987), we determined that the molar ratio of SERCA2a:phospholamban was 1:1.6 for the SERCA2a + wild-type phospholamban sample and 1:1.5 for the L37A-phospholamban sample. The similarity of the SERCA2a-to-phospholamban ratio in the two phospholamban-containing samples facilitated direct and quantitative comparison of the effects of wild-type versus L37A-phospholamban on SERCA2a phosphorylation kinetics.

View larger version (33K): [in this window] [in a new window]   FIGURE 1   SDS-PAGE and immunoblotting of canine SERCA2a and phospholamban expressed in Sf21 microsomes. (A) GelCode Blue-stained SDS gel (10 µg protein per lane) showing Sf21 microsomes containing SERCA2a alone (lane 1), SERCA2a + wild-type phospholamban (lane 2), and SERCA2a + L37A-phospholamban (lane 3). (B) Corresponding immunoblot of the SDS gel shown in A, developed with SERCA2a antibody 2A7-A1 (top) or phospholamban antibody 2D12 (bottom). PLBp denotes pentameric phospholamban and PLBm denotes monomeric phospholamban. Lanes A-E contain 0.2-1.0 µg purified SERCA2a (top) or 0.02-0.1 µg purified phospholamban (bottom). Lanes 1-3 contain 10 µg of Sf21 microsomal protein per lane and correspond to SERCA2a alone, SERCA2a + wild type phospholamban, and SERCA2a + L37A-phospholamban, respectively. (C and D) Standard curves (empty symbols) constructed using purified SERCA2a and phospholamban, respectively, showing the relationship between blot intensity and amount of protein loaded. The filled symbols correspond to the blot intensities obtained using Sf21 microsomes containing SERCA2a alone (), SERCA2a + wild type phospholamban (), and SERCA2a + L37A-phospholamban (), which were used in conjunction with the standard curves to deduce the amount of SERCA2a or phospholamban in each of the sample types studied (Table 1). An identical experiment conducted using 5 µg of Sf21 microsomal protein per lane (not shown) produced equivalent results.

To test the activity of the expressed SERCA2a and the functional coupling between SERCA2a and phospholamban in Sf21 cell microsomes, ATPase assays were conducted at 37°C at a series of [Ca²⁺] in the presence and absence of anti-phospholamban monoclonal antibody 2D12, which reverses the inhibitory interaction between phospholamban and SERCA2a (Briggs et al., 1992; Sham et al., 1991). The ATPase activity of the expressed SERCA2a displayed the usual sigmoidal [Ca²⁺] dependence, and the maximum steady-state activities per mg of microsomal protein for the three sample types were SERCA2a alone, 0.6 µmol·mg¹·min¹; SERCA2a + wild-type phospholamban, 0.2 µmol·mg¹·min¹; and SERCA2a + L37A-phospholamban, 0.3 µmol·mg¹·min¹. Differences in the maximum activities correlated with differences in Ca-ATPase expression levels between the different samples deduced from the SDS-PAGE and immunoblot analysis. That is, when corrected for the amount of SERCA2a in each sample type, the maximum activity for each sample was ~0.75 µmol P_i produced per minute per nmol SERCA2a (Table 1). For comparison, a similar analysis was carried out using dog cardiac SR (CSR) vesicles. The ATPase activity of the CSR vesicles per mg of microsomal protein measured under identical conditions was 2.1 µmol·mg¹·min¹. When corrected for the amount of SERCA2a in this preparation of CSR, which was reported as 30% (Jones et al., 1979; Autry and Jones, 1997), the maximum activity of the sample was 0.77 µmol·nmol SERCA2a¹·min¹, which is identical to that of the expressed samples (Table 1). Thus there were no differences in SERCA2a specific activity between the expressed samples, which were identical to that in dog cardiac SR.

SERCA2a expressed without phospholamban had a high apparent Ca²⁺ affinity (ionized [Ca²⁺] giving half-maximum activation of the Ca-ATPase, K_Ca = 200 nM), which was unaffected by treatment of the sample with phospholamban monoclonal antibody (Table 1). When coexpressed with wild-type phospholamban, the SERCA2a activity curve was shifted to the right relative to that of SERCA2a expressed alone, resulting in an increase in K_Ca to 420 nM. When coexpressed with the monomeric mutant L37A-phospholamban, a potent super shifter (Autry and Jones, 1997; Kimura et al., 1997), the SERCA2a activity curve was shifted to the right to a greater extent, resulting in an increase in K_Ca to 800 nM. For both phospholamban-containing samples, treatment with anti-phospholamban antibody 2D12 shifted the Ca²⁺ activation curve to the left, resulting in twofold decreased K_Ca values. The K_Ca value for the antibody-treated SERCA2a + wild-type phospholamban sample was equivalent to that obtained for SERCA2a in the absence of phospholamban, whereas the K_Ca value for the antibody-treated SERCA2a + L37A-phospholamban sample was greater than that of the SERCA2a alone sample (Table 1). This characteristic of the SERCA2a + L37A-phospholamban sample has been reported previously (Autry and Jones, 1997), and it is a common feature of the super-shifting monomeric phospholamban mutants containing mutations in the leucine-isoleucine zipper domain (Autry and Jones, 1998). Note that the absolute amount of the K_Ca shift by the antibody for the L37A-phospholamban mutant (0.4 µM shift) was two times greater than the shift for wild-type phospholamban (0.2 µM shift). This demonstrates that the antibody effect on the L37A-phospholamban mutant is amplified compared to the antibody effect on wild-type phospholamban, even though the shift in Ca affinity by L37A-phospholamban on the Ca-ATPase is not reversed completely by the antibody. This suggests that the antibody is acting in a mechanistically realistic way on L37A-phospholamban/Ca-ATPase interactions, even though it is not capable of totally reversing the inhibitory effect of this potent mutant on Ca-ATPase activity. High ionized [Ca²⁺], however, is capable of reversing the inhibitory effect completely (see below). For comparison, the K_Ca values measured for dog cardiac SR were 400 nM and 170 nM for the sample without and with pretreatment with anti-phospholamban antibody, respectively, similar to the expressed SERCA2a + wild-type phospholamban sample. Ca-ATPase activities were identical between all three sets of microsomal membranes when measured at saturating ionized [Ca²⁺], and under this V_max condition the monoclonal antibody produced no effect (Table 1). Lack of effect of the monoclonal antibody on the Ca-ATPase V_max in native cardiac SR vesicles is well known (Simmerman and Jones, 1998). The results show that phospholamban was functionally coupled to Ca-ATPase in the two phospholamban-coexpressed samples, similar to that observed in native cardiac SR vesicles. Furthermore, the results confirm earlier observations that the monomeric L37A-phospholamban mutant is a more potent inhibitor of steady-state Ca-ATPase activity than wild-type phospholamban, but even the more potent L37A-phospholamban does not affect the maximum ATPase activity (Autry and Jones, 1997; Kimura et al., 1997).

SERCA2a phosphoenzyme formation

The first goal of our kinetics studies was to test the hypothesis that phospholamban inhibits Ca-ATPase cycling by decreasing by 10-fold the rate of the E1·Ca-to-E1'·Ca transition (Cantilina et al., 1993). To test this hypothesis, we measured the effect of phospholamban on ATP-dependent SERCA2a phosphoenzyme (EP) formation, starting with the Ca-ATPase in a Ca²⁺-free state. SERCA2a-containing microsomes were preincubated at 0°C in 1 mM EGTA in the experimental buffer to remove all traces of Ca²⁺ from the medium and to stabilize the Ca-ATPase in a Ca²⁺-free state (Scheme 1). At time 0, an equal volume of the same ice-cold experimental buffer containing [-³²P]ATP (1 µM final) and Ca²⁺ (various levels) was added to the enzyme solution, and the reaction was allowed to proceed at 0°C for serial times before acid quenching and determination of phosphoenzyme formed. Samples were preincubated either without or with anti-phospholamban monoclonal antibody 2D12 before the initiation of phosphoenzyme formation. The low temperature and low [ATP] were used to slow the rate of EP formation and thus allow its measurement by hand mixing techniques over a 20-30-s time interval after the addition of ATP. The EGTA preincubation was included to allow us to measure the effect of phospholamban on the E1·Ca-to-E1'·Ca transition, which limits the rate of EP formation under these conditions (Cantilina et al., 1993; Tada et al., 1980). Specific control experiments of the expressed enzyme (not shown) showed that the EGTA preincubation had no deleterious effect on Ca-ATPase stability, even after a 1-h incubation on ice.

In the first set of experiments, we measured the effect of phospholamban on the [Ca²⁺] dependence of steady-state EP levels formed by 1 µM ATP at 0°C for SERCA2a expressed in Sf21 cell microsomes (Fig. 2 and Table 2). The steady-state EP level of each sample type displayed a sigmoidal dependence on [Ca²⁺], and the K_Ca values (Table 2) of the curves were nearly identical to those obtained from the [Ca²⁺]-dependent ATPase activity of three SERCA2a samples measured at 37°C at saturating [ATP] (Table 1). The maximum steady-state EP level was similar for the three sample types, in terms of nmol EP formed per mg total Sf21 microsomal protein: 0.055 nmol EP/mg for SERCA2a alone, 0.02 nmol EP/mg for SERCA2a + wild type phospholamban, and 0.03 nmol EP/mg for SERCA2a + L37A-phospholamban. When the raw EP levels were corrected for Ca-ATPase expression levels in the different samples (Table 1), the three samples showed virtually identical maximum EP levels of ~0.075 nmol/nmol SERCA2a (Fig. 2 and Table 2). For comparison, similar experiments using dog cardiac SR (data not shown) revealed a maximum EP level of 0.21 nmol/mg SR, which was equivalent to 0.073 nmol/nmol SERCA2a. This result shows that the phosphorylation capacity of SERCA2a in the expressed samples was identical to that of native SERCA2a in cardiac SR. Thus the relatively low maximum EP levels reported here arose from the low, subsaturating [ATP] concentration utilized in the experiment (cf. Froehlich and Taylor, 1975; Tada et al., 1980) and not from inactive SERCA2a in the expressed samples.

View larger version (25K): [in this window] [in a new window]   FIGURE 2   [Ca2+] dependence of steady-state phosphoenzyme levels of SERCA2a in Sf21 microsomes formed by 1 µM ATP at 0°C. Steady-state SERCA2a phosphoenzyme (EP) levels were measured by phosphorylating Sf21 microsomes (0.1 mg/ml) preincubated in 1 mM EGTA with 1 µM [-32P]ATP and various [Ca2+] levels for 30 s at 0°C, followed by acid quenching and processing as described in Materials and Methods. (A) SERCA2a expressed alone; (B) SERCA2a coexpressed with wild-type phospholamban; (C) SERCA2a coexpressed with L37A-phospholamban. The steady-state EP levels of SERCA2a in Sf21 microsomes pretreated with phospholamban monoclonal antibody, 2D12, are denoted by empty circles (), and the control, untreated microsomes are denoted by the empty squares (). The data were simulated using the reactions of Scheme 3 and the rate constants in Table 6. Simulation of the antibody-treated data is denoted by the filled circles (), and simulation of the control, untreated data is denoted by the filled squares (). The curves represent fits of the data to the Hill equation, which provided the KCa and maximum EP values (Table 2). Empty symbols represent the average of triplicate measurements for a single preparation of each sample type. Error bars denote the high and low values in each average. Additional experiments using different preparations provided equivalent results.

View this table: [in this window] [in a new window]   TABLE 2   Inhibitory effects of phospholamban (wild-type or L37A-PLB) on the [Ca2+] dependence of steady-state SERCA2a phosphoenzyme levels at 0°C using 1 µM Mg[-32P]ATP

Ca-ATPase expressed without phospholamban displayed a K_ca of 200 nM for steady-state EP formation, which was unaffected by treatment of these microsomes with 2D12 (Table 2). When SERCA2a was coexpressed with wild-type phospholamban, the K_ca value increased to 460 nM (K_Ca = 210 nM relative to the antibody-treated sample). When coexpressed with the monomeric mutant L37A-phospholamban, the K_ca value increased to a greater extent to 800 nM (K_Ca = 400 nM relative to the antibody-treated sample). For both phospholamban-coexpressed samples, treatment with 2D12 shifted the Ca²⁺ activation curve to the left, resulting in twofold decreased K_Ca values. As observed with measurement of steady-state ATPase activity at 37°C, the K_Ca value for the antibody-treated SERCA2a + wild-type phospholamban sample was not significantly different from that obtained for expression of SERCA2a in the absence of phospholamban. In contrast, the K_Ca value for the antibody-treated SERCA2a + L37A-phospholamban sample was greater than that of the SERCA2a alone sample, as was also observed for the 37°C steady-state activity measurements (Table 1). Similar to the effects of phospholamban on SERCA2a ATPase activity, treatment of each sample with anti-phospholamban antibody had no effect on the maximum steady-state EP levels formed, which were identical between all three sets of microsomes. Thus the results obtained for formation of steady-state EP levels at 0°C, using 1 µM ATP, mirrored the results obtained for measurement of steady-state ATP hydrolysis at 37°C using 5 mM ATP. Furthermore, the results show that the monomeric L37A-phospholamban mutant is a more potent inhibitor of SERCA2a steady-state phosphoenzyme formation than is wild-type phospholamban, but, in analogy to measurements of ATP hydrolysis (Autry et al., 1997; Kimura et al., 1997), even the more potent L37A-phospholamban does not affect the maximum SERCA2a steady-state EP level observed at saturating [Ca²⁺].

In the next set of experiments, we measured the effect of phospholamban on the time course of EP formation by SERCA2a in Sf21 cell microsomes at 0°C (Fig. 3 and Table 3). These measurements were carried out for 268 nM [Ca²⁺]_free, which is near the K_Ca value for each sample, and at a saturating Ca²⁺ level (15 µM [Ca²⁺]_free), where phospholamban has no effect on Ca-ATPase activity (Fig. 2 and Table 1). At 268 nM [Ca²⁺]_free (Fig. 3, left), the EP formation time course for SERCA2a expressed alone (top left) exhibited a brief lag (with a time constant of 5 s¹) followed by a monoexponential increase with a rate of 0.21 s¹, and the steady-state phosphoenzyme level was 0.036 nmol EP/mg total protein (0.048 nmol EP/nmol SERCA2a). The presence of a lag phase in the EP formation time course is consistent with the fact that a series of kinetic steps precedes phosphoryl transfer from ATP to the enzyme (Frost and Pearson, 1953). Pretreatment of the sample with anti-phospholamban monoclonal antibody had no effect on the phosphoenzyme formation time course (Table 3). When coexpressed with wild-type phospholamban, the steady-state Ca-ATPase phosphoenzyme level (0.007 nmol EP/mg protein = 0.026 nmol EP/nmol SERCA2a) was decreased 33% relative to samples pretreated with anti-phospholamban monoclonal antibody (0.011 nmol EP/mg protein = 0.040 nmol EP/nmol SERCA2a), but neither the lag preceding EP formation (time constant of 1.8 s¹) nor the apparent rate of EP formation (0.18 s¹) was changed relative to the antibody-treated samples (0.18 s¹). When coexpressed with the more potent monomeric L37A-phospholamban mutant, the steady-state Ca-ATPase phosphoenzyme level (0.006 nmol EP/mg protein = 0.016 nmol EP/nmol SERCA2a) was decreased by 45% relative to the sample pretreated with anti-phospholamban antibody (0.011 nmol EP/mg protein = 0.029 nmol EP/nmol SERCA2a), but again neither the lag preceding EP formation (time constant of 0.5 s¹) nor the apparent rate of EP formation (0.22 s¹) was changed relative to samples pretreated with anti-phospholamban monoclonal antibody (0.22 s¹).

View larger version (39K): [in this window] [in a new window]   FIGURE 3   Time course of phosphoenzyme formation by SERCA2a in Sf21 microsomes by 1 µM ATP at 0°C. SERCA2a phosphorylation time courses were measured by adding 1 µM [-32P]ATP and either 0.5 mM CaCl2 (268 nM ionized Ca2+, left panels) or 1.0 mM CaCl2 (15 µM ionized Ca2+, right panels) to Sf21 microsomes (0.1 mg/ml) preincubated in 1 mM EGTA (final conditions). The reaction was allow to proceed at 0°C for serial times as indicated, followed by acid quenching and processing as described in Materials and Methods. (Top) SERCA2a expressed alone; (middle) SERCA2a + wild-type phospholamban; (bottom) SERCA2a + L37A-phospholamban. Phosphorylation of SERCA2a in Sf21 microsomes pretreated with phospholamban monoclonal antibody, 2D12, are denoted by filled circles (), and the control, untreated microsomes are denoted by the empty circles (). The solid curves represent simulation of the EP formation time course, using the reactions of Scheme 3 and the rate constants in Table 6. The dotted curves represent simulations where the values of k3 and k3 (Table 6) were each decreased by a factor of 10. Symbols represent the average of triplicate measurements for a single preparation of each sample type. Error bars represent high and low values in each average. Additional experiments using different preparations provided equivalent results.

At 15 µM [Ca²⁺]_free (Fig. 3, right), the apparent rate of EP formation (0.38 s¹) and the steady-state EP level (0.076 nmol EP/nmol SERCA2a) were similar for each of the samples (Table 3). In addition, there was no discernible lag phase preceding EP formation for any of the expressed SERCA2a samples. Before the steady-state EP levels were adjusted for the amount of SERCA2a in each sample type, the steady-state EP levels were 0.055 nmol EP/mg protein for SERCA2a expressed alone, 0.021 nmol EP/mg protein for SERCA2a + wild-type phospholamban, and 0.030 nmol EP/mg for SERCA2a + L37A-phospholamban. Pretreatment of the three samples with anti-phospholamban monoclonal antibody had no effect on the rate of EP formation or the steady-state EP level (Table 3). Additional EP formation experiments, carried out at 625 nM [Ca²⁺]_free (not shown), produced results entirely similar to those obtained at 268 nM ionized [Ca²⁺]. Phospholamban had no effect on the apparent rate of EP formation (0.26 s¹ for each sample, ± antibody treatment), but treatment of the phospholamban-containing samples with anti-phospholamban antibody resulted in a stimulation of the steady-state EP level of the wild-type phospholamban sample by 30% and of the more potent L37A-phospholamban sample by 70%. In summary, the results of the phosphoenzyme formation studies show that phospholamban does not affect the apparent rate of EP formation. Because the apparent rate of SERCA2a phosphorylation is directly controlled by the rate of the Ca-ATPase E1·Ca-to-E1'·Ca transition under the conditions of our experiments (Tada et al., 1980; Cantilina et al., 1993), the results do not support the hypothesis that phospholamban inhibits this transition in the SERCA2a cycle.

SERCA2a P_i liberation and phosphoenzyme decomposition

The second goal of our kinetics studies was to test the hypothesis that phospholamban inhibits Ca-ATPase cycling by decreasing the rate of EP decomposition (step 5, Scheme 1). To test this hypothesis, we measured the effect of phospholamban coexpression on the time course of P_i release (Fig. 4 and Table 4), concomitant with our EP formation experiments. This was accomplished by collecting an aliquot of the EP reaction mixture for each sample after the quenched membranes were pelletted. The unreacted radiolabeled ATP was removed from the sample by charcoal extraction, and the ³²P_i in the aliquot was separated by complexation with ammonium molybdate and extraction into organic solvent. The P_i liberation time courses for all samples at both 268 nM and 15 µM [Ca²⁺]_free exhibited a brief lag (~5-7 s), which coincided with the build-up of SERCA2a EP, followed by a linear increase in P_i over time during the steady-state phase of SERCA2a turnover. At 268 nM ionized [Ca²⁺], the rate of P_i liberation for SERCA2a expressed alone was 3.6 × 10³ nmol P_i·mg protein¹·s¹ (6.6 × 10³ nmol P_i·nmol SERCA2a¹·s¹), and this rate was not affected by pretreatment of the sample with anti-phospholamban monoclonal antibody. When coexpressed with wild-type phospholamban, the rate of P_i liberation by SERCA2a (1.3 × 10³ nmol P_i·mg protein¹·s¹ or 4.5 × 10³ nmol P_i·nmol SERCA2a¹·s¹) was decreased 33% relative to the sample pretreated with anti-phospholamban monoclonal antibody (2.0 × 10³ nmol P_i·mg protein¹·s¹ or 7.0 × 10³ nmol P_i·nmol SERCA2a¹·s¹). When coexpressed with the more potent monomeric L37A-phospholamban mutant, the rate of P_i liberation by SERCA2a (1.9× 10³ nmol P_i·mg protein¹·s¹ or 4.5 × 10³ nmol P_i·nmol SERCA2a¹·s¹) was decreased by 45% relative to the sample pretreated with anti-phospholamban antibody (3.3 × 10³ nmol P_i·mg protein¹·s¹ or 8.0 × 10³ nmol P_i·nmol SERCA2a¹·s¹). At 15 µM [Ca²⁺]_free, the rate of P_i liberation by SERCA2a in each of the expressed samples was unaffected by pretreatment with anti-phospholamban monoclonal antibody. When the raw P_i liberation rates (4.4 × 10³ (SERCA2a), 5.3 × 10³ (SERCA2a + wild-type phospholamban), and 5.9 × 10³ (SERCA2a + L37A-phospholamban) nmol P_i·mg protein¹·s¹) were corrected for differences in Ca-ATPase expression levels in the different samples (Table 1), the three samples showed essentially the same rates: 0.010 (SERCA2a alone), 0.011 (SERCA2a + wild-type phospholamban) and 0.015 (SERCA2a + L37A-phospholamban) nmol P_i·nmol SERCA2a¹·s¹ (Table 4). Thus, even for the enzyme at 0°C with micromolar ATP concentration, inhibition of overall ATP hydrolysis by phospholamban occurs only at subsaturating ionized Ca²⁺ concentrations, and there is no apparent V_max effect. Additional P_i liberation experiments, carried out at 625 nM [Ca²⁺]_free (not shown), produced results entirely similar to those obtained at 268 nM ionized [Ca²⁺]. Treatment of the phospholamban-containing samples with anti-phospholamban antibody 2D12 stimulated the rate of P_i liberation by the wild-type phospholamban sample by 30% and that of the more potent L37A-phospholamban sample by 45%. Taken together, the results show that phospholamban inhibits the apparent rate of product formation by SERCA2a.

View larger version (35K): [in this window] [in a new window]   FIGURE 4   Time course of phosphate liberation by SERCA2a in Sf21 microsomes in the presence of 1 µM ATP at 0°C. SERCA2a phosphorylation time courses were measured as described in the legend of Fig. 3. Inorganic phosphate in each sample was isolated and processed as described in Materials and Methods. (Top) SERCA2a expressed alone; (middle) SERCA2a + wild-type phospholamban; (bottom) SERCA2a + L37A-phospholamban. Phosphorylation of SERCA2a in Sf21 microsomes pretreated with phospholamban monoclonal antibody, 2D12, is denoted by filled circles (), and the control, untreated microsomes are denoted by the empty circles (). The solid curves represent simulation of the Pi liberation time course, using the reactions of Scheme 3 and the rate constants in Table 6. The dotted curves represent simulations where the values of k3 and k3 (Table 6) were each decreased by a factor of 10. Symbols represent the average of triplicate measurements for a single preparation of each sample type (Table 4). Error bars represent high and low values in each average. Additional experiments using different preparations provided equivalent results.

Correlation of the P_i liberation and EP formation data showed that phospholamban (both wild-type and L37A-phospholamban) inhibited the rate of P_i liberation to the same extent as the inhibition of the steady-state EP level in each sample. Because product formation (P_i) depends on the rate of EP decomposition and the concentration of EP in the steady state (d[P_i]/dt = k × [EP]), it appeared that the phospholamban-mediated reduction in the rate of P_i production was due to the decreased steady-state level of EP induced by phospholamban. Alternatively, phospholamban could have inhibited the rate of EP decomposition, resulting in the reduced rate of P_i liberation. To distinguish further between these two possibilities, we measured the effect of phospholamban on the rate of SERCA2a phosphoenzyme decomposition (Fig. 5). In this experiment, SERCA2a was phosphorylated with 1 µM ATP for 30 s at 0°C, which was sufficient time for the system to attain steady-state EP levels (Fig. 3). At 30 s, 5 mM EGTA (final concentration) was added to the reaction mixture to remove all traces of ionized Ca²⁺ from the reaction mixture and thus to stop SERCA2a phosphorylation by ATP. The phosphoenzyme present at the time of the EGTA addition decayed with an exponential time course, which was followed by quenching of the decay reaction at serial times after EGTA addition.

View larger version (37K): [in this window] [in a new window]   FIGURE 5   Time course of phosphoenzyme decomposition by SERCA2a in Sf21 microsomes after a 5 mM EGTA chase at 0°C. SERCA2a in Sf21 microsomes (0.1 mg/ml) preincubated in 1 mM EGTA was phosphorylated by 1 µM [-32P]ATP and either 0.5 mM CaCl2 (268 nM ionized Ca2+, left panels) or 1.0 mM CaCl2 (15 µM ionized Ca2+, right panels) at 0°C (final conditions). After 30 s of phosphorylation, 5 mM EGTA (final) was added to the reaction to halt subsequent EP formation, and the decay of radiolabeled EP at 0°C was monitored by acid quenching the sample at serial times as indicated. The samples were processed as described in Materials and Methods. (Top) SERCA2a expressed alone; (middle) SERCA2a + wild-type phospholamban; (bottom) SERCA2a + L37A-phospholamban. Phosphoenzyme decomposition of SERCA2a in Sf21 microsomes pretreated with phospholamban monoclonal antibody, 2D12, is denoted by filled circles (), and the control, untreated microsomes are denoted by empty circles (). The solid curves represent simulation of the EP decay time course, using the reactions of Scheme 3 and the rate constants in Table 6. Symbols represent the average of triplicate measurements for a single preparation of each sample type. Error bars represent high and low values in each average. Additional experiments using different preparations provided equivalent results.

Fig. 5 shows the effect of phospholamban on the time course of EP decomposition by SERCA2a in Sf21 cell microsomes at 0°C. The decomposition time course for all samples was best fit by a monoexponential function. For each sample at each [Ca²⁺], the EP level at the zero time of the EP decay time course corresponded to the steady-state EP level measured in the EP formation experiments (Fig. 3 and Table 3). For phosphoenzyme formed at 268 nM [Ca²⁺]_free, the apparent rate of EP decomposition in zero ionized Ca²⁺ was similar for all of the SERCA2a samples, 0.12 s¹ for SERCA2a expressed alone, 0.15 s¹ for SERCA2a + wild-type phospholamban, and 0.14 s¹ for SERCA2a + L37A-phospholamban (Table 5). Pretreatment of the samples with anti-phospholamban monoclonal antibody had no effect on the rate of phosphoenzyme decay. Additional experiments in which phosphoenzyme was formed at 15 µM [Ca²⁺]_free (Fig. 5 and Table 5) or 625 nM [Ca²⁺]_free (not shown), but at which phosphoenzyme decay proceeded at zero ionized Ca²⁺, produced similar results. The EP decomposition rate was nearly identical for all samples (0.15 s¹ at 625 nM [Ca²⁺]_free and 0.20 s¹ 15 µM [Ca²⁺]_free), and the rate was not affected by pretreatment with anti-phospholamban antibody. These results show that phospholamban does not alter the rate of SERCA2a phosphoenzyme decomposition occurring in the absence of significant ionized Ca²⁺. Thus we conclude that phospholamban does not inhibit the rate of EP decay. Therefore, phospholamban inhibits P_i liberation as a consequence of the phospholamban-dependent reduction of steady-state SERCA2a EP levels, rather than by directly inhibiting the rate of EP decay.

Simulation of SERCA2a kinetics data

To determine the kinetic step(s) in the Ca-ATPase cycle affected by phospholamban that would result in a decreased steady-state EP level but no change in the rate of EP formation or decay, the [Ca²⁺]-dependent steady-state EP level and the time courses of EP formation, P_i liberation, and EP decay for each sample were simulated using the reactions of Scheme 3 and the rate constants presented in Table 6. The details of the simulations are presented in Materials and Methods. We first simulated the EP formation and P_i liberation time courses measured for the SERCA2a alone sample (Figs. 3 and 4, top, solid lines), to generate a set of rate constants that described the SERCA2a reaction cycle in the absence of phospholamban (Table 6). The best simulation of the data was obtained using a catalytic site density of 0.08 nmol/mg protein (0.11 nmol/nmol SERCA2a), based on the maximum amount of EP formed at saturating Ca²⁺ by the subsaturating, 1 µM ATP concentration used in the experiments. In the simulations, the rate of EP formation was dependent primarily on the forward rate of the E1·Ca-to-E1'·Ca transition (step 3), whereas the steady-state EP level was dependent both on the forward and reverse rates (i.e., the equilibrium constant) of the first Ca²⁺ ion binding to E1 (step 2) and the forward rate of the E1·Ca-to-E1'·Ca transition (step 3). The rate of the E2-to-E1 transition (step 1) had only minor effects on the EP formation time course, provided the rate of k₁ remained greater than 0.1 s¹. At lower values of k₁, the simulated EP formation did not fit the experimental data. The rate of the second Ca²⁺ ion binding to E1'·Ca (step 4), the rate of ATP binding to E1'·Ca₂ (step 5), and the rate of phosphoryl transfer from ATP to the enzyme (step 6) had little influence on the simulations, because the rates of these steps were much faster than the adjacent slow steps (steps 3 and 7). The rates of EP decomposition and P_i release (steps 7 and 8) were selected to allow the phosphoenzyme to rise to its steady-state level without passing through an "overshoot," while simultaneously matching the lag and linear phases of the P_i liberation time course. Some minor adjustments in steps 7 and 8 were required to produce a more precise simulation of the steady-state phase of EP formation and P_i liberation at each ionized [Ca²⁺] studied, but the selected rates for steps 7 and 8 were verified by simulating the SERCA2a alone EP decay data at each Ca²⁺ level (Fig. 5, top, solid lines). By carrying out the same simulation at a series of ionized [Ca²⁺] between 0 and 15 µM, the [Ca²⁺] dependence of the steady-state EP levels were accurately reproduced (Fig. 2, solid symbols). This result validated Scheme 3 and the kinetic constants of Table 6 as an appropriate model for simulating SERCA2a kinetics under our experimental conditions and for elucidating the effect of phospholamban on the SERCA2a reaction cycle.

View larger version (9K): [in this window] [in a new window]   Scheme 3

Simulation of the kinetics data from the SERCA2a + wild-type phospholamban and SERCA2a + L37A-phospholamban microsomes (Figs. 3-5, solid lines) was carried out using a catalytic site density of 0.04 nmol/mg for the SERCA2a + wild-type phospholamban microsomes and 0.07 nmol/mg for the SERCA2a + L37A-phospholamban microsomes. The antibody-treated SERCA2a + wild-type phospholamban and SERCA2a + L37A-phospholamban kinetics data (Figs. 3 and 4, solid lines) was simulated using the same kinetics constants used for the expressed SERCA2a microsomes, except that for the SERCA2a + L37A-phospholamban microsomes it was necessary to reduce by twofold the forward rate (k₂) of step 2 (Ca²⁺ binding to E1) to 1 × 10⁸ M¹ s¹, to properly reproduce the rate and amplitude of the EP formation and P_i liberation time courses. Similar to the SERCA2a alone sample, some minor adjustments for steps 7 and 8 were required to produce a precise simulation of the EP formation and P_i liberation time courses measured at the various ionized [Ca²⁺] studied, but the selected rates for steps 7 and 8 were verified by simulating the EP decay data at each Ca²⁺ level (Fig. 5, solid lines). Finally, by carrying out the simulations at a series of ionized [Ca²⁺] levels between 0 and 15 µM, the [Ca²⁺] dependence of the steady-state EP levels for both antibody-treated samples were accurately reproduced (Fig. 2, solid symbols).

Two approaches were used to simulate the effect of coexpressed phospholamban on SERCA2a kinetics data obtained at 268 nM ionized [Ca²⁺] (Figs. 3 and 4). First, we tested the proposal by Cantilina et al. (1993) that phospholamban decreases the forward and reverse rates of the E1·Ca-to-E1'·Ca transition (step 3, Scheme 3) by 10-fold. Changing k₃ and k₃ from 0.4 s¹ to 0.04 s¹ resulted in EP formation (Fig. 3, left, dashed lines) and P_i liberation (Fig. 4, left, dashed lines) time courses that did not fit the data obtained from the two phospholamban-containing samples. Similar results were obtained when the untreated kinetics data obtained at 625 nM (not shown) were simulated, and even at saturating [Ca²⁺] (also not shown), where there should be no difference between untreated and antibody-treated data (Figs. 3-5). Therefore, the results of this approach suggest that phospholamban does not inhibit SERCA2a EP formation by inhibiting the E1·Ca-to-E1'·Ca transition.

We next focused on Ca²⁺ binding to E1 (step 2, Scheme 3) as a target for phospholamban inhibition, because our simulations showed that the kinetic constants for this step had substantial effects on the steady-state EP level but only minor effects on the rate of EP formation in the simulations (see above). For both the SERCA2a + wild-type phospholamban and the SERCA2a + L37A-phospholamban kinetics data, a twofold decrease in the equilibrium constant (K_eq) for step 2 was sufficient, by itself, to simulate the untreated EP formation and P_i liberation time courses at both 268 nM (Figs. 3 and 4, left, solid lines) and 625 nM (not shown) ionized [Ca²⁺]. For the SERCA2a + wild-type phospholamban microsomes, k₂ was reduced to 1 × 10⁸ M¹ s¹, whereas for the SERCA2a + L37A-phospholamban microsomes, k₂ was reduced to 5 × 10⁷ M¹ s¹. (The same result was achieved by holding k₂ constant and increasing k₂ by a factor of 2 to 240 s¹ for both samples.) At 15 µM ionized Ca²⁺, the twofold decrease in k₂ had no effect on the simulated kinetics; the EP formation, P_i liberation, and EP decay time courses generated were identical to the simulations of the antibody-treated data. This result was validated by conducting the simulation at a series of ionized [Ca²⁺], which accurately reproduced the [Ca²⁺] dependence of the steady-state EP levels measured for both untreated samples (Fig. 2, solid symbols). Thus the results of the simulations show that wild-type phospholamban inhibits SERCA2a by decreasing by twofold the equilibrium binding of the first Ca²⁺ to E1 relative to the SERCA2a alone sample. L37A-phospholamban is more potent than wild-type phospholamban, because it decreases equilibrium Ca²⁺ binding to E1 by fourfold relative to the SERCA2a alone sample. Nevertheless, high ionized [Ca²⁺] is capable of reversing the inhibitory effect of phospholamban on SERCA2a completely, even the inhibition by the more potent L37A-phospholamban mutant.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In the present study, we have measured the effects of phospholamban on the time course of Ca-ATPase phosphorylation by ATP, to test the hypothesis (Cantilina et al., 1993) that phospholamban inhibits Ca-ATPase cycling by decreasing by 10-fold the rate of the E1·Ca-to-E1'·Ca transition (Scheme 2). Likewise, we have measured the effects of phospholamban on the time courses of inorganic phosphate (P_i) liberation and phosphoenzyme (EP) decomposition, to test the hypothesis that phospholamban inhibits steady-state ATPase activity by decreasing the rate of SERCA2a EP decay. To assist in this work, we have made use of the baculovirus-Sf21 insect cell expression system (Autry and Jones, 1997) to produce microsomes containing 1) Ca-ATPase alone, 2) Ca-ATPase coexpressed with wild-type phospholamban, and 3) Ca-ATPase coexpressed with the monomeric L37A-phospholamban mutant (L37A-phospholamban). To further define the specific effects of phospholamban on SERCA2a kinetics, we used a monoclonal antibody against phospholamban to reverse phospholamban inhibitory effects. The expressed SERCA2a samples had the same specific ATPase activity as SERCA2a in dog cardiac SR, and the SERCA2a in the expressed samples was phosphorylated to the same extent as SERCA2a in dog cardiac SR. The inhibitory effects of wild-type phospholamban on SERCA2a in the expressed sample were nearly identical to those measured for cardiac SR, and pretreatment with anti-phospholamban antibody relieved the inhibition to the same extent in the two samples. Because of the relatively high levels of SERCA2a and phospholamban expression provided by the baculovirus-Sf21 insect cell expression system, we were able to measure the effects of phospholamban on the phosphorylation and dephosphorylation kinetics of the expressed SERCA2a for the first time. Furthermore, the