Dissociation of phospholamban regulation of cardiac sarcoplasmic reticulum Ca2+ATPase by quercetin.

Quercetin had a biphasic effect on Ca2+ uptake and calcium-stimulated ATP hydrolysis in isolated cardiac sarcoplasmic reticulum (SR). Stimulation of Ca2+ATPase was observed at low quercetin concentrations (<25 μM) followed by inhibition at higher concentrations. The effects were dependent upon the SR protein concentration, the MgATP concentration, and intact phospholamban regulation of cardiac Ca2+ATPase. Only the inhibitory effects at higher quercetin concentrations were observed in skeletal muscle SR which lacks phospholamban and in cardiac SR treated to remove phospholamban regulation. Stimulation was additive with monoclonal antibody 1D11 (directed against phospholamban) at submaximal antibody concentrations; however, the maximal antibody and quercetin stimulation were identical. Quercetin increased the calcium sensitivity of the Ca2+ATPase like that observed with phosphorylation of phospholamban or treatment with monoclonal antibody 1D11. In addition, low concentrations of quercetin increased the steady-state formation of phosphoenzyme from ATP or Pi, but higher quercetin decreased phosphoenzyme levels. Quercetin, even under stimulatory conditions, was a competitive inhibitor of ATP, but appears to relieve the Ca2+ATPase from phospholamban inhibition, thereby, producing an activation. The subsequent inhibitory action of higher quercetin concentrations results from competition of quercetin with the nucleotide binding site of the Ca2+ATPase. The data suggest that quercetin interacts with the nucleotide binding site to mask phospholamban's inhibition of the SR Ca2+ATPase and suggests that phospholamban may interact at or near the nucleotide binding site.

Quercetin had a biphasic effect on Ca 2؉ uptake and calcium-stimulated ATP hydrolysis in isolated cardiac sarcoplasmic reticulum (SR). Stimulation of Ca 2؉ ATPase was observed at low quercetin concentrations (<25 M) followed by inhibition at higher concentrations. The effects were dependent upon the SR protein concentration, the MgATP concentration, and intact phospholamban regulation of cardiac Ca 2؉ ATPase. Only the inhibitory effects at higher quercetin concentrations were observed in skeletal muscle SR which lacks phospholamban and in cardiac SR treated to remove phospholamban regulation. Stimulation was additive with monoclonal antibody 1D11 (directed against phospholamban) at submaximal antibody concentrations; however, the maximal antibody and quercetin stimulation were identical. Quercetin increased the calcium sensitivity of the Ca 2؉ ATPase like that observed with phosphorylation of phospholamban or treatment with monoclonal antibody 1D11. In addition, low concentrations of quercetin increased the steady-state formation of phosphoenzyme from ATP or P i , but higher quercetin decreased phosphoenzyme levels. Quercetin, even under stimulatory conditions, was a competitive inhibitor of ATP, but appears to relieve the Ca 2؉ ATPase from phospholamban inhibition, thereby, producing an activation. The subsequent inhibitory action of higher quercetin concentrations results from competition of quercetin with the nucleotide binding site of the Ca 2؉ ATPase. The data suggest that quercetin interacts with the nucleotide binding site to mask phospholamban's inhibition of the SR Ca 2؉ ATPase and suggests that phospholamban may interact at or near the nucleotide binding site.
In cardiac and skeletal muscle, the sarcoplasmic reticulum (SR) 1 Ca 2ϩ ATPase is responsible for the reuptake of calcium into the SR to allow relaxation. Ca 2ϩ ATPase transports two Ca 2ϩ ions across the SR membrane against a large concentration gradient by hydrolytic coupling with one ATP molecule. Structural studies (Toyoshima et al., 1993;Stokes et al., 1994) suggest that Ca 2ϩ ATPase molecules can exist in monomeric and oligomeric forms with enzymatic activity being regulated by its oligomeric state (Squier et al., 1988;Voss et al., 1991Voss et al., , 1994Kutchai et al., 1994). Large aggregates of enzyme have low or no activity, and dissociated enzyme (possibly dimers) has high activity. Ca 2ϩ ATPase exists in two chemically equivalent conformations that differ in their calcium affinity, E 1 (high Ca 2ϩ affinity) and E 2 (low Ca 2ϩ affinity). E 1 binds Ca 2ϩ cooperatively in a pH-dependent manner and E 2 binds Ca 2ϩ in a pH-independent and negatively cooperative manner (Nakamura and Tajima, 1995). ATP can modulate the intermolecular interactions to alter the relative populations of E 1 and E 2 (Scofano et al., 1979). There appears to be a single nucleotide binding site per Ca 2ϩ ATPase (McIntosh and Boyer, 1983;Cable et al., 1985;Bishop et al., 1987;Lacapere and Guillain, 1993), which serves two functions, catalytic and regulatory. ATP concentrations above micromolar, which saturate the catalytic nucleotide binding site, serve a regulatory role to produce further increases in Ca 2ϩ ATPase activity. Lastly, Ca 2ϩ binding to E 1 is accompanied by a protein conformational change which is an integral part of the cooperative binding mechanism and of enzyme activation by Ca 2ϩ . This conformational change is accelerated when millimolar MgATP is bound to the regulatory nucleotide binding site (Inesi et al., 1980;Stahl and Jencks, 1984;McIntosh and Davidson, 1984).
Phospholamban (PLB), an integral membrane protein present in cardiac SR but not in fast-twitch skeletal muscle SR, reduces the calcium sensitivity of the Ca 2ϩ ATPase. Studies in PLB-deficient mice demonstrate the crucial role PLB plays in regulating basal contractile parameters and the heart's responsiveness to hormonal activators (Luo et al., 1994). The phosphorylation domain and the nucleotide/hinge domain of the Ca 2ϩ ATPase, and the cytoplasmic N terminus of PLB have been shown to be required for a PLB-regulated Ca 2ϩ ATPase complex (Toyofuku et al., 1992(Toyofuku et al., , 1993(Toyofuku et al., , 1994. A recent model postulates that PLB inhibits the Ca 2ϩ ATPase by aggregating it into a kinetically unfavorable associated state (Voss et al., 1994) through electrostatically controlled protein-protein interactions. Co-expression studies (Toyofuku et al., 1994), however, indicate that both charged and hydrophobic residues between amino acids 2 and 18 in PLB are essential for interaction with the Ca 2ϩ ATPase. Relief of PLB inhibition can be achieved in one of several ways, viz. phosphorylation of PLB by protein kinases (Tada et al., 1974;Kranias, 1985), certain monoclonal antibodies directed against the N terminus of PLB (Cantilina et al., 1993;Mayer et al., 1996), mild trypsinization of PLB to remove its hydrophilic N terminus (Lu et al., 1993), nonsolubilizing concentrations of C 12 E 8 (Lu and Kirchberger, 1994), solubilization with deoxycholate and reconstitution of cardiac SR (Kim et al., 1990), treatment with heparin and related compounds (Xu and Kirchberger, 1989) or charged detergents (Chiesi and Schwaller, 1989), and the polyphenol tannin (Chiesi and Schwaller, 1994). These treatments all invoke the disruption of the interaction between PLB and the Ca 2ϩ ATPase, presumably by direct effects on PLB, its lipid environment, or alterations in electrostatic protein-protein in-teractions. No treatment that directly interacts with the Ca 2ϩ ATPase to relieve it from PLB inhibition has been reported.
Quercetin (Q), a bioflavanoid, acts as a inhibitor of numerous enzymes involved in energy conversion reactions (Racker, 1986), phosphodiesterase (Kuppasamy and Das, 1992), and numerous protein kinases (Vlahos et al., 1994). Furthermore, it is an inhibitor of skeletal muscle Ca 2ϩ ATPase (Shoshan and MacLennan, 1981;Fischer et al., 1987) with an IC 50 of 12 M in 45 Ca 2ϩ uptake experiments. Quercetin competitively inhibits ATP binding to the Ca 2ϩ ATPase at low ATP concentrations (Shoshan and MacLennan, 1981) and has an affinity for the E-Ca 2 (fully ligated with respect to calcium at the exterior high affinity calcium sites, unligated with respect to ATP) conformational state that is approximately 10-fold greater than for other conformational states in the hydrolytic cycle (Fischer et al., 1987). However, more detailed study of quercetin's effects on the partial reactions of the Ca 2ϩ ATPase (Shoshan and MacLennan, 1981) indicate that the major effect of quercetin may not lie in its ability to inhibit nucleotide binding but possibly another step in the reaction sequence.
In this report, evidence suggesting that quercetin's biphasic effects result from its interaction with the nucleotide binding site on cardiac SR Ca 2ϩ ATPase is presented. Stimulation appears to result from the disruption or masking of PLB's inhibitory effect on the Ca 2ϩ ATPase, while the inhibitory effect at higher concentrations is like that observed in skeletal muscle SR.

EXPERIMENTAL PROCEDURES
Materials-[␥-33 P]ATP and [ 33 P]P i were obtained from Amersham Corp. and 45 CaCl 2 from DuPont NEN. Quercetin and several common analogs and phosphoenol pyruvate were purchased from Sigma. Quercetin and similar compounds were dissolved in dimethyl sulfoxide or, in some experiments, in ethanol. Pyruvate kinase was from Boehringer Mannheim and A23187 from Calbiochem. Bovine pancreatic trypsin and soybean trypsin inhibitor were obtained from Worthington. Stimulatory anti-PLB mAb 1D11 was prepared as described by Mayer et al. (1996).
Preparation of Sarcoplasmic Reticulum-Canine cardiac sarcoplasmic reticulum was obtained as described previously (Mayer et al., 1996). Rabbit skeletal muscle sarcoplasmic reticulum was isolated from a discontinuous sucrose density gradient. Briefly, rabbit back and hindleg skeletal muscle was cleaned and minced. Muscle (100-g aliquots) were placed in 300 ml of buffer I (20 mM MOPS/KOH, pH 7.1, 0.3 M sucrose, and 1 mM phenylmethylsulfonyl fluoride) then homogenized in a Waring blender, two times, 30 s each. The homogenate was centrifuged at 2,600 ϫ g for 30 min and the resulting supernatant was poured through cheesecloth. The pellet was resuspended in 200 ml of buffer I, and the homogenization and centrifugation steps were repeated. The two supernatants were combined and centrifuged at 10,000 ϫ g for 30 min. The pellet was resuspended in 25 ml of buffer II (20 mM MOPS/KOH, pH 7.1, 0.3 M sucrose, and 0.6 M KCl) and incubated on ice for 1 h. Next, the sample was centrifuged at 165,000 ϫ g for 1 h, and the resulting pellet was resuspended in 4 ml of buffer III (20 mM MOPS/KOH, pH 7.1, 0.3 M sucrose, and 0.2 M KCl). The sample was applied to a discontinuous sucrose density gradient comprised of 10 ml of 45% sucrose, 15 ml of 30% sucrose, and 7 ml of 20% sucrose in buffer III. The gradient was centrifuged in a SW-28 rotor at 120,000 ϫ g for 4 h. Skeletal muscle SR, collected from the interface between 30 and 45% sucrose, was diluted in buffer III to fill a Ti45 centrifuge tube. Following centrifugation at 165,000 ϫ g for 1 h, the membrane pellet was resuspended in 3 ml of buffer III (per 100 g of muscle). Membranes were stored at Ϫ70°C and protein determination was measured by the Pierce BCA protein assay with bovine serum albumin as a standard.
Trypsin Treatment of Cardiac SR-The procedure described by Lu et al. (1993) was adapted. Briefly, cardiac SR diluted to 1 mg/ml in buffer III was treated with 10 g/ml trypsin for 5 min at 25°C. Control SR was prepared similarly except that trypsin was not added. The SR was diluted into 2 ml of ice-cold buffer III containing 0.12 mg/ml soybean trypsin inhibitor and centrifuged at 100,000 ϫ g for 30 min in a Beckman TLX tabletop ultracentrifuge. The resulting pellet was resuspended in buffer III to yield a 2 mg/ml protein concentration. Trypsin cleavage of PLB was verified by testing the effect of mAb 1D11 in the ATPase activity assay.
Deoxycholate Solubilization and Reconstitution of Cardiac SR-Cardiac SR was solubilized with 0.7 mg of deoxycholate/mg of SR protein as described by Kim et al. (1990). Deoxycholate was removed by addition of 0.5 g/ml Bio-Beads SM-2. The mixture was stirred for 1 h at room temperature. At the end of the incubation, the sample was recovered using a narrow bore pipette. Control cardiac SR was treated in a similar manner except deoxycholate was not added. As with the trypsinized samples, the loss of PLB-regulation was verified by testing for the effect of mAb 1D11 on ATPase activity.

45
Ca 2ϩ Uptake Measurement and ATPase Activity Microtiter Plate Assay-These measurements were performed as described by Mayer et al. (1996).
Phosphoenzyme Formation-Phosphoenzyme was formed from [ 33 P]ATP or [ 33 P]P i using standard procedures. After preincubating 400

FIG. 1. Concentration dependence of the effects of quercetin on cardiac (q) and skeletal muscle (E) SR Ca 2؉ ATPase activity.
Ca 2ϩ -stimulated ATP hydrolysis at 0.178 M Ca 2ϩ , 0.5 mM MgATP, and 5% Me 2 SO was measured at 25°C. Shown are the mean of triplicate measurements of pooled canine cardiac SR (roughly 50 dog hearts) and rabbit skeletal muscle SR (four rabbits). Basal cardiac SR Ca 2ϩ ATPase activity was 55 nmol of P i /mg/min, and the maximal activity (100%) was 120 nmol of P i /mg/min. Basal skeletal muscle Ca 2ϩ ATPase activity (100%) was 320 nmol of P i /mg/min).

FIG. 2. Protein dependence of the effects of quercetin on cardiac SR Ca 2؉ ATPase activity.
Ca 2ϩ -stimulated ATP hydrolysis was measured as in Fig. 1, except with a fixed concentration of 10 M quercetin and the protein concentration and reaction time were varied. The Ca 2ϩ ATPase activities were normalized to matched Me 2 SO solvent controls, and the stoichiometry of quercetin to mg SR protein is plotted on the x axis.
g/ml cardiac SR in 150 mM KCl, 1 mM MgCl 2 , 5 mM NaN 3 , 0.5 mM EGTA, 20 mM imidazole, pH 7.0, with various concentrations of CaCl 2 and quercetin for 20 min, formation of E 1 ϳP from 20 M [ 33 P]ATP was measured for 15 s at 0°C. Following preincubation for 20 min of 800 g/ml cardiac SR in 40 mM Mes/Tris, pH 6.0, 10 mM MgCl 2 , 20% dimethyl sulfoxide, and varying amounts of quercetin, E 2 ϳP formation from 100 M [ 33 P]P i was measured after 10 min at room temperature. The reaction was quenched with 0.25 M perchloric acid containing 4 mM P i . Acid precipitated membranes were washed three times with 0.125 M perchloric acid, 4 mM P i and pelleted in a clinical centrifuge. Radioactive pellets were dissolved in 0.1 M NaOH, 2% Na 2 CO 3 , 1 mM P i , and 2% sodium lauryl sulfate then counted in a scintillation counter using Ready Safe scintillation mixture.

RESULTS
Low micromolar concentrations of quercetin markedly increased calcium-dependent ATP hydrolysis in cardiac SR vesicles at submaximal free Ca 2ϩ . Stimulation of Ca 2ϩ ATPase is highly dependent on the ratio of quercetin to SR protein. In Fig.  1, at a cardiac SR protein concentration of 10 g/ml and 0.5 mM MgATP (solid circles), maximal stimulation occurred at 3.16 M quercetin which corresponds to 0.316 nmole of Q/g of SR protein. If cardiac SR protein contains 1.12 pmol of Ca 2ϩ ATPase/g of SR protein as determined by maximal phosphoenzyme formation (data not shown), then the maximal stimulatory ratio is about 282 nmol of Q/nmol of Ca 2ϩ ATPase. A precipitous decline in Ca 2ϩ ATPase activity below the control rate occurred at quercetin concentrations above 10 M. At 20 M quercetin, cardiac SR Ca 2ϩ ATPase activity was about the same as control cardiac SR. At saturating free calcium concentrations above 1 M, quercetin did not stimulate Ca 2ϩ ATPase activity, but the inhibitory effect at concentrations above 10 M quercetin was still observed (data not shown). The half-maximal inhibitory concentration was about 20 M quercetin at 10 g/ml SR and 0.5 mM MgATP or 2.0 nmol of Q/g of SR (1786 nmol of Q/nmol of Ca 2ϩ ATPase). The quercetin concentration response curve (not shown) on 45 Ca 2ϩ uptake into cardiac SR vesicles was congruent with the curve shown in Fig. 1 except the protein concentration was 100 g/ml and the MgATP concentration was 3.16 mM. The optimal stimulatory ratio of quercetin to SR protein was 223 nmol of Q/mg of SR protein.
Quercetin did not stimulate Ca 2ϩ ATPase activity in rabbit skeletal muscle SR (Fig. 1, open circles) which lacks PLB. However, like cardiac SR, quercetin (Ͼ10 M concentration) markedly inhibited the skeletal muscle SR Ca 2ϩ ATPase as previously reported (Shoshan and MacLennan, 1981;Fischer et al., 1987). Half-maximal inhibition occurred at 40 M for 5 g/ml skeletal muscle SR which corresponds to a ratio of 8.0 nmol of Q/g of SR protein. Using an estimate of 4.5 nmol of Ca 2ϩ ATPase/mg of skeletal muscle SR (Shoshan and MacLennan, 1981;Bishop et al., 1987), then half-maximal inhibition of skeletal muscle SR Ca 2ϩ ATPase occurred at a ratio of 1778 nmol of Q/nmol of Ca 2ϩ ATPase, which is identical to the value calculated above for cardiac SR. A value of 89 nmol of Q/nmol of Ca 2ϩ ATPase was calculated using a half-maximal inhibitory concentration of 12 M quercetin and 30 g/ml protein as re-ported by Shoshan and MacLennan (1981) and 740 nmol of Q/nmol of Ca 2ϩ ATPase from the values of 10 M quercetin and 3 g/ml SR protein reported by Fischer et al. (1987). This disparate range of values exemplifies the complexity of the interaction of quercetin with the Ca 2ϩ ATPase protein. Fig. 2 shows the biphasic effect of protein concentration on the response of cardiac SR Ca 2ϩ ATPase to a fixed quercetin concentration of 10 M. Rates of calcium-stimulated ATP hydrolysis were normalized to matched solvent controls for each protein concentration and the relative rate versus the ratio of quercetin to SR protein are plotted. The free calcium concentration was 0.178 M and the MgATP concentration was 0.5 mM. At low quercetin/protein ratios (Ͻ400 nmol of Q/mg of SR protein), a steep increase in Ca 2ϩ ATPase activity was observed until an optimal ratio was achieved. At higher ratios, the stimulatory effect diminishes eventually leading to inhibition relative to the control Ca 2ϩ ATPase activity. The optimal stimulatory ratio of quercetin to cardiac SR protein for this data was 400 nmol of Q/mg of SR protein.
Chemical structures of various flavanoids and their activity in modulating cardiac SR Ca 2ϩ ATPase at 0.178 M free Ca 2ϩ are shown in Table I. mAb 1D11, typically, stimulated control Ca 2ϩ ATPase activity around 300%. Each flavanoid (25 M) was present 20 min prior to, and during the 30 min incubation with 0.5 mM MgATP. A wide range of Ca 2ϩ ATPase activities were measured, some of these compounds tested at lower concentrations gave greater stimulation, e.g. 6 M quercetagetin produced a 275% stimulation. This reversed dose dependence for stimulation of Ca 2ϩ ATPase activity by quercetagetin suggests a similar biphasic dose-response profile, but with a different optimal stoichiometry of flavanoid to SR protein. A logical SAR for this group of flavanoids would require construction of complete dose-response curves for each compound to determine the optimal stoichiometry. The effects of some of these compounds at 25 M were assayed on skeletal muscle SR Ca 2ϩ ATPase under conditions identical to those used for cardiac SR, and the activities relative to the solvent control were fisetin (25%), robinetin (7%), quercetin (50%), myricetin (0%), apigenin (73%), and quercetagetin (5%).
The effect of saturating anti-PLB mAb 1D11 (Mayer et al., 1996) and an optimal stimulatory amount of quercetin on the apparent calcium sensitivity of cardiac SR Ca 2ϩ ATPase activity were identical ( Fig. 3A; Tables II and III). A leftward shift in the pCa curve indicating an apparent increased calcium sensitivity was observed. Furthermore, combining these treatments was not additive. Thus, quercetin appears to mimic mAb 1D11 in relieving the cardiac SR Ca 2ϩ ATPase from the endogenous PLB inhibition. In another set of experiments, cardiac SR was solubilized with deoxycholate, and the detergent was removed with Bio-Beads to reform intact vesicles that no longer exhibit PLB regulation (Kim et al., 1991). The loss of PLB inhibition was verified by the inability of the anti-PLB mAb 1D11 to stimulate the Ca 2ϩ ATPase activity. The deoxycholatetreated SR Ca 2ϩ ATPase activity versus pCa curve shows a leftward shift similar to that obtained following mAb 1D11 treatment. A submaximal quercetin dose was used to show an intermediate shift in apparent calcium sensitivity. Again, there was no additivity with quercetin and deoxycholate-treatment.
Stimulation by quercetin appears to require a PLB-regulated Ca 2ϩ ATPase. Complete quercetin dose-response curves at a submaximal free calcium of 0.178 M were constructed following two different procedures to eliminate PLB regulation (Fig.  4). First, cardiac SR was solubilized with deoxycholate and reconstituted as above. Deoxycholate-treatment stimulated Ca 2ϩ ATPase activity almost 3-fold and abolished the stimulatory effect of quercetin. Quercetin's inhibitory effect was not  (Table II). The V max values are about three times higher than those shown below which were obtained at 25°C. B, effects of 10 M quercetin on cardiac (q, E) and deoxycholate-solubilized (f, Ⅺ) SR Ca 2ϩ ATPase activity. Ca 2ϩ -stimulated ATP hydrolysis was measured in triplicate over a range of free Ca 2ϩ in the absence (q, f) and presence of 10 M quercetin (E, Ⅺ). Curve fits were performed and the fitted lines are shown. The results of the fit are summarized in Table III. altered. An alternative treatment using mild trypsinization of cardiac SR (Lu et al., 1993) which results primarily in the loss of the cytoplasmic domain of PLB was used. Proteolysis of PLB was nearly complete, as treatment of trypsinized cardiac SR with the stimulatory anti-PLB mAb 1D11 produced only a small stimulation (about 20%) of Ca 2ϩ ATPase activity and only trace amounts of intact PLB following Western blot analysis were observed (data not shown). Again, quercetin no longer exerted a stimulatory effect, while its inhibitory effect was unchanged.
The ability of quercetin to mimic deregulation of PLB inhibition is best illustrated by the quercetin dose-response curves (Fig. 5) constructed in the presence of increasing amounts of mAb 1D11 ranging from none to a saturating amount of 10 g/ml. Each curve peaks at the same quercetin concentration of 25 M. Moreover, the maximal Ca 2ϩ ATPase activity approaches the same rate which is identical to the maximal mAb 1D11 effect. The additivity of quercetin and mAb 1D11 at submaximal concentrations is clear-cut, but the mechanism of quercetin's effect needed clarification.
A direct interaction of stimulatory amounts of quercetin and other flavanoids with PLB was checked. None of the compounds at 25 M concentration altered the mAb 1D11 antibody binding in a competitive enzyme-linked immunosorbent assay (data not shown). The effect of short cytoplasmic PLB-derived peptides on quercetin-induced stimulation of cardiac SR Ca 2ϩ ATPase activity was assayed in the 45 Ca 2ϩ uptake assay (Fig. 6). None of the peptides at concentrations up to 100 M affected the basal Ca 2ϩ ATPase activity. Neither the PLB 7-17 peptide nor the synthetic protein kinase A substrate, LR-RASLG (Kemptide), had any effect on quercetin's stimulatory action. Synthetic PLB peptide 1-25 at 50 M produced about a 40% reduction of the maximal quercetin stimulation, yet 5 M PLB 1-25 completely abolished the maximal stimulation by mAb 1D11 or by protein kinase A catalyzed phosphorylation (data not shown). Furthermore, 1% polyvinyl pyrrolidine, a nonspecific chelator of phenolic compounds, produced a similar 40% inhibition of quercetin-induced stimulation. Therefore, it does not appear that quercetin has a specific interaction with PLB.
Quercetin is well known to affect ATP-dependent enzymes (Racker, 1986) and is an inhibitor of skeletal muscle SR Ca 2ϩ ATPase (Shoshan and MacLennan, 1981). The MgATP dependence of cardiac SR Ca 2ϩ ATPase under conditions of quercetin stimulation was characterized and compared with the stimulation produced by mAb 1D11 (Fig. 7A). At low MgATP (5.62 M), 10 M quercetin had no apparent effect on Ca 2ϩ ATPase activity, and yet at saturating MgATP (500 M), this same quercetin concentration rivaled the maximal mAb 1D11 response. The data are replotted in Fig. 7B and Table IV as an Eadie-Scatchard plot. The K m for MgATP under control Ca 2ϩ ATPase assay conditions (1% Me 2 SO) is 6.7 M, and the V max is 58.5 nmol of P i /mg/min. Addition of mAb 1D11 had no effect on the K m for MgATP (6.9 M), but the V max was almost doubled to 111.9 nmol of P i /mg/min. Thus, mAb 1D11 is a noncompetitive activator of cardiac SR Ca 2ϩ ATPase or, more accurately, PLB is a noncompetitive inhibitor of cardiac SR Ca 2ϩ ATPase. Quercetin likewise doubled the V max to 121.8 nmol of P i /mg/min, but it also significantly increased the K m for MgATP to 31.3 M. Even under optimal stimulating conditions, quercetin acts as a competitive inhibitor for MgATP activation of cardiac SR Ca 2ϩ ATPase. However, stimulation is observed because quercetin removes the inhibition by PLB.
MgATP dependence curves were performed in the presence of varying quercetin concentrations producing both stimulatory and inhibitory effects. A linear Dixon plot of apparent K m(MgATP) versus each quercetin concentration was obtained (Fig. 8). A K i estimate of 26.4 M quercetin was determined. This value is in the same range as the optimal stimulatory quercetin concentration in the presence of saturating MgATP. Taken together these experiments suggest that quercetin interacts with the nucleotide binding site of Ca 2ϩ ATPase and this site is associated with the PLB binding site.
The effect of quercetin on cardiac SR Ca 2ϩ ATPase phosphoenzyme formation in the forward and reverse modes was studied. E 1 ϳP is formed in the presence of Ca 2ϩ and ATP, while E 2 ϳP is formed from P i in the absence of ATP. At a submaximal Ca 2ϩ concentration of 0.1 M (Fig. 9), low stoichiometries of quercetin to Ca 2ϩ ATPase (1-100 nmol of quercetin/mg of SR protein) moderately increased the steady-state formation of E 1 ϳP (20 -25% increase) and E 2 ϳP (40% increase). In contrast, mAb 1D11 produces a 100% increase of steady-state E 1 ϳP level and an identical 40% increase in E 2 ϳP 2 at this same calcium concentration. The discrepancy in the stimulation of E 1 ϳP formation observed with quercetin and mAb 1D11 may be explained by quercetin's interaction with the nucleotide binding site at the low ATP concentration (20 M) used in the E 1 ϳP phosphoenzyme experiments. At high stoichiometries of quercetin to Ca 2ϩ ATPase, formation of phosphoenzyme is greatly diminshed in both modes. Presumably,

DISCUSSION
Quercetin's Interaction with Ca 2ϩ ATPase-The results show that the enzymatic activity of cardiac SR Ca 2ϩ ATPase, but not of skeletal muscle SR Ca 2ϩ ATPase, is stimulated at submicromolar Ca 2ϩ by low concentrations of quercetin. Stimulation is not observed at saturating Ca 2ϩ (Ͼ1 M). The stimulatory action is attenuated at higher quercetin concentrations followed by inhibition of enzyme activity. The stimulatory effect requires intact PLB regulation of the cardiac Ca 2ϩ ATPase, as following treatments that disrupt PLB regulation only the inhibitory range of quercetin activity remains. Furthermore, the maximal stimulation is identical to that observed when PLB regulation is removed and quercetin stimulation is additive with submaximal amounts of mAb 1D11. The inhibitory concentration range of quercetin in cardiac SR is the same as in skeletal muscle SR (which lacks PLB) and in cardiac SR treated to remove PLB regulation. The inhibitory effects of quercetin on skeletal muscle SR Ca 2ϩ ATPase have been well characterized (Shoshan and MacLennan, 1981;Fischer et al., 1987). Quercetin inhibits the binding of ATP and ADP to Ca 2ϩ ATPase, but not the binding of calcium (Shoshan and MacLennan, 1981). Indeed, quercetin had preferential affinity for the enzyme fully ligated with respect to exterior high affinity Ca 2ϩ binding sites and unligated with respect to ATP (Fischer et al., 1987). However, as noted by Shoshan and MacLennan (1981) and shown in the phosphoenzyme experiments in Fig. 9, quercetin may stabilize a conformational intermediate that cannot energize the release of Ca 2ϩ into the SR lumen nor donate P i back to ADP. Indeed, quercetin did not reduce the steady-state level of phosphoenzyme over the drug concentration range that inhibits ATP hydrolysis, implying that quercetin can bind to phosphorylated forms of the enzyme that have an unoccupied regulatory nucleotide binding site. Thus, in cardiac SR, quercetin appears to obfuscate the inhibitory effects of PLB, thereby producing an apparent stimulation (i.e. quercetin is not a direct activator), prior to producing inhibitory effects as described for skeletal muscle SR Ca 2ϩ ATPase.
The effects of quercetin and other bioflavanoids were variable and appear to be dependent on the assay conditions. There is a complex relationship between the Ca 2ϩ ATPase, PLB, and substrates which must be considered when interpreting the results. Quercetin has a biphasic protein-dependence in cardiac SR (Fig. 2), which may be explained by a necessary stoichiometry of quercetin to Ca 2ϩ ATPase molecules to disrupt PLB FIG. 4. Concentration dependence of the effects of quercetin on cardiac (q), trypsin-treated cardiac (E), and deoxycholatesolubilized followed by Bio-Bead detergent removal reconstituted cardiac (f) SR. Ca 2ϩ -stimulated ATP hydrolysis was measured as in Fig. 1. Shown are the mean of triplicate measurements.
FIG. 5. Concentration dependence of the effects of quercetin on cardiac SR Ca 2؉ ATPase in the presence of increasing amounts of mAb 1D11; 0 g/ml (f), 1.25 g/ml (Ⅺ)), 2.5 g/ml (q), 5 g/ml (E), and 10 g/ml (å). Ca 2ϩ -stimulated ATP hydrolysis was measured in triplicate as described for Fig. 1. regulation. This stoichiometry is a complex function of quercetin and ATP concentrations due to competitive binding at the nucleotide site.
Both the stimulatory and inhibitory effects of quercetin appear to result from direct interaction with the nucleotide binding site of Ca 2ϩ ATPase. No evidence for a direct quercetin-PLB interaction or quercetin antioxidative activity on the SR phospholipids was found. Quercetin acts as a powerful antioxidant against lipid peroxidation when lipid bilayers are exposed to aqueous oxygen radicals (Terao et al., 1994). It is most likely to be localized near the surface of phospholipid bilayers. No alterations in 45 Ca uptake rates in control and quercetin-treated cardiac SR were found (data not shown) when measured in the absence of oxygen and the presence of ascorbate. The ATP dependence data (Fig. 7, A and B) demonstrate quercetin's competitive inhibition for nucleotide binding in cardiac SR, whereas PLB acts as a noncompetitive inhibitor.
The data clearly shows that stimulation by quercetin requires a PLB-regulated Ca 2ϩ ATPase. Under conditions where PLB regulation has been disrupted, e.g. deoxycholate treatment, limited trypsinization, in micromolar Ca 2ϩ , or in saturating anti-PLB mAb 1D11, only inhibition by quercetin is observed. The maximal stimulation elicited by quercetin corresponds to the complete disruption of PLB regulation. Moreover, quercetin stimulation is additive with submaximal amounts of anti-PLB mAb 1D11 (Fig. 5) and the V max(MgATP) using mAb 1D11 or quercetin were the same (Fig. 7B). Quercetin interacts with the nucleotide binding site of Ca 2ϩ ATPase and removes the inhibitory effect of PLB. This results in an increase in the number of active Ca 2ϩ ATPase units. Thus, the anti-PLB or "phospholamban inhibitor" effect of quercetin causes the stimulation of Ca 2ϩ ATPase activity.
Phospholamban's Interaction with Ca 2ϩ ATPase: Recruitment Hypothesis-PLB regulation involves the ability of the Ca 2ϩ ATPase to be activated by Ca 2ϩ and ATP without altering the affinity of substrate binding. Calcium and nucleotide binding to cardiac SR Ca 2ϩ ATPase has an affinity and cooperativity identical to skeletal muscle SR (Cantilina et al., 1993). Recent work (Voss et al., 1994) suggests that PLB regulation of the cardiac Ca 2ϩ ATPase is related to critical changes in protein FIG. 7. A, MgATP dependence of the effect of 10 g/ml mAb 1D11 and 25 M quercetin on Ca 2ϩ -stimulated ATP hydrolysis. Ca 2ϩ ATPase activity was measured in triplicate as described in Fig. 1 in the presence of an ATP-regenerating system. The rates of Ca 2ϩ -stimulated ATP hydrolysis in control (q) and mAb 1D11-treated (ࡗ) and quercetintreated (f) cardiac SR were determined in the presence of 3.16 M to 3.16 mM MgATP ([ATP]). B, Eadie-Scatchard plot of V/[ATP] versus V, showing the effect of mAb 1D11 10 g/ml and 25 M quercetin on the V max(MgATP) and the K m(MgATP ). The data shown in A were transformed to yield the V/[ATP] versus V plot. The x intercept is the V max and the slope ϭ Ϫ1/K m . The results from a linear regression fit are summarized in Table IV. FIG. 8. Dixon plot for a competitive inhibitor of MgATP activation of Ca 2؉ ATPase activity. The effect of quercetin concentrations ranging from 5 to 100 M on the MgATP concentration dependence of Ca 2ϩ ATPase was measured. For each quercetin concentration, Eadie-Scatchard plots were constructed, and the K m(ATP)app was determined. The x intercept equals ϪK i(quercetin) , and the y intercept is the K m(MgATP) for control cardiac SR. The linear regression fit of the data is shown yielding a K i(quercetin) of 26 M. dynamics and protein-protein interactions. According to their hypothesis, PLB inhibits Ca 2ϩ ATPase by aggregating it into large oligomeric complexes, and phosphorylation of PLB activates by dissociating these large aggregates. Alternatively, mAb 1D11 or saturating micromolar Ca 2ϩ produced the same effect. We would like to extend their hypothesis to state that the PLB-aggregated Ca 2ϩ ATPase molecules are enzymatically inactive. Disruption of PLB-aggregated Ca 2ϩ ATPase allows enzymatic activity to occur. In effect, disruption of PLB-regulation increases the availability of active Ca 2ϩ ATPase units, i.e. recruitment of more pumps from a PLB-aggregated reserve. The recruitment hypothesis is supported by experimental evidence here as well as in numerous publications. For years, researchers have argued that the shift in apparent calcium sensitivity of cardiac SR Ca 2ϩ ATPase observed following PLB phosphorylation or treatment with anti-PLB mAb was solely a kinetic effect as described by Cantilina et al. (1993). An alternative explanation is that unphosphorylated PLB aggregates and inactivates a population of Ca 2ϩ ATPase molecules. PLBphosphorylation or anti-PLB mAb treatment releases the inactive pool of Ca 2ϩ ATPase causing an increase in the number of active pump units resulting in a leftward shift in the pCa curve (Fig. 3, A and B). Quercetin produced a concentration-dependent increase in the steady-state level of phosphoenzyme intermediate (Fig. 9), i.e. an increase in the number of active pump units. Similarly, Cantilina et al. (1993) (see their Fig. 7) measured the formation of phosphoenzyme intermediate in the absence and presence of anti-PLB mAb and at varied Ca 2ϩ concentration. Their contention was that disruption of PLBregulation mimicked the effect of micromolar Ca 2ϩ increasing the rate of phosphoenzyme formation and hence produced a kinetic effect similar to other reports. However, they failed to recognize the importance of an increase in steady-state EϳP level at each subsaturating Ca 2ϩ which can be due only to an increase in number of Ca 2ϩ ATPase units. The steady-state phosphoenzyme level reflects the availability of Ca 2ϩ ATPase molecules to undergo catalysis and is dependent upon the amount of Ca 2ϩ bound to the Ca 2ϩ ATPase. Thus, more Ca 2ϩ ATPase units turning over per unit of time produces an apparent increase in the rate of phosphoenzyme formation.
The MgATP dependence of the cardiac SR Ca 2ϩ ATPase shown in Fig. 7, A and B, clearly demonstrates an increase in the number of active Ca 2ϩ ATPase units upon addition of anti-PLB mAb 1D11. There was no effect on the K m for nucleotide, but the V max , which reflects the number of active Ca 2ϩ ATPase units, increased. Similar results were obtained following partial tryptic digestion of PLB in cardiac SR vesicles (Lu et al., 1993) (Fig. 4). The observation that quercetin also increases the V max suggests that it is increasing the number of active pumps by disrupting PLB-Ca 2ϩ ATPase interaction. Furthermore, nonsolubilizing concentrations of C 12 E 8 (Lu and Kirchberger, 1994) or C 12 E 9 3 produce equivalent effects on V max . Low concentrations of C 12 E 8 partially disaggregate Ca 2ϩ ATPase in cardiac SR providing a physical correlate for the functional effects (Shi et al., 1996). Hence, we would expect quercetin to have similar effects on Ca 2ϩ ATPase aggregation.
Recent studies with transgenic PLB-deficient mice and comparisons with their heterozygous, wild-type, and PLB overexpressing relatives add supporting evidence to the recruitment hypothesis. The SR Ca 2ϩ ATPase of PLB-deficient mice has a high apparent Ca 2ϩ sensitivity compared to the wild-type counterpart, while cardiac SR from the heterozygous mice (40% of wild-type) have an intermediate Ca 2ϩ sensitivity (Luo et al., 1994). In transgenic mice overexpressing PLB relative to wildtype mice, there is an additional rightward shift in the pCa curve (Kadambi et al., 1996), EC 50 ϭ 0.27 M (wild-type) versus 0.48 M (overexpressed). The rightward shift in the pCa curves with increasing PLB expression suggests a progressive inactivation of Ca 2ϩ ATPase units. It follows that the role of PLB in heart muscle is to maintain a reserve pool of Ca 2ϩ ATPase units for periods of increased contractile activity and not merely to inhibit the Ca 2ϩ pumping activity. Data were normalized to the control steady-state levels of 0.92 Ϯ 0.08 and 1.00 Ϯ 0.12 nmol of phosphoenzyme/mg of SR protein, respectively. mAb 1D11 increased the forward steady-state phosphoenzyme about 40% and the reverse reaction by 100%.