Maximal Inhibition of SERCA2 Ca 2 1 Affinity by Phospholamban in Transgenic Hearts Overexpressing a Non-phosphorylatable Form of Phospholamban*

Phospholamban is a phosphoprotein in the cardiac sarcoplasmic reticulum (SR) which regulates the apparent Ca 2 1 affinity of the SR Ca 2 1 -ATPase (SERCA2). To determine the levels of phospholamban which are associated with maximal inhibition of SERCA2, several lines of transgenic mice were generated which expressed increasing levels of a non-phosphorylatable form of phospholamban (S16A,T17A) specifically in the heart. This mutant form of phospholamban was chosen to prevent phosphorylation as a compensatory mechanism in vivo . Quantitative immunoblotting revealed increased phospholamban protein levels of 1.8-, 2.6-, 3.7-, and 4.7-fold in transgenic hearts compared with wild types. There were no changes in the expression levels of SERCA2, calsequestrin, calreticulin, and ryanodine receptor. Assessment of SR Ca 2 1 uptake in hearts of transgenic mice indicated increases in the inhibition of the affinity of SERCA2 for Ca 2 1 with increased phospholamban expression. Maximal inhibition was obtained at phospholamban expression levels of 2.6-fold or higher. Transgenic hearts with functional saturation in phospholam-ban:SERCA2 ( > 2.6:1) exhibited increases in b -myosin heavy chain expression, associated with cardiac hypertrophy. These findings demonstrate that overexpression of a non-phosphorylatable were used for the following studies. Site-directed Mutagenesis— The PCR methodology by Bowman et al . (27) was used to incorporate the site-specific mutation S16A,T17A (TC- CACT 3 GCCGCT) into the PLB cDNA. A 0.85-kb Sal I fragment containing the PLB cDNA and the SV40 polyadenylation signal se- quence (PLB cDNA-SV40-poly(A)) was released from the a -myosin heavy chain promoter ( a -MHC)-PLB-SV40 fusion gene, used previously to generate transgenic mice overexpressing wild-type PLB (25, 27). This Sal I PLB cDNA-SV40-poly(A) fragment was then subcloned into a pBluescript SKII( 2 ) vector (Stratagene), which has both T3 and T7 primer sites flanking the insert. PCR mutagenesis was performed by two consecutive PCR amplifications, using two different sets of primers, as described previously (28). In the first PCR amplification, a 5 9 -end mutant primer (5 9 -CT ATC AGG AGA GCC GCC GCT ATT GAA ATG CC-3 9 ), corresponding to nucleotides 32–62 of the PLB coding sequence, and a 3 9 -end T7 primer were used to generate the desired mutant PLB cDNA minor product. In the second PCR amplification, an aliquot of the first PCR product and the T3 and T7 primers was used to amplify the full-length insert, which contained the desired mutation in the PLB cDNA. The final product was cut with Sal I, gel purified, and resub-cloned into the Sal I site of a second pBluescript SKII( 2 ) vector, which was then transformed into XL1-Blue competent cells. Colonies from the transformed cells containing the desired mutant PLB cDNA were iden- tified by DNA sequencing. The mutated PLB cDNA-SV40-poly(A) sequence was excised by Sal I from the pBluescript SKII( 2 ) vector and ligated into the Sal I site of the 5.5-kb calsequestrin calreticulin ryanodine receptor (1:500), b -MHC and a -actin (1:2,000) and visualized with either 35 S-labeled (2 3 5 or peroxidase-labeled secondary (Amersham Pharmacia Biotech). The degree of labeling was determined using a PhosphorImager and the ImageQuant software. For immunodetection of PLB phosphorylation sites, polyclonal antibodies raised against a PLB peptide (residues 9–199) phosphorylated at Ser 16 (PLB-phospho- serine 16) or at Thr 17 (PLB-phosphothreonine 17) were used. The samples were separated by 15% SDS-PAGE and transferred onto 0.05- m m nitrocellulose membrane. The membranes were incubated with PLB-phosphoserine 16 (1:10,000 dilution) and PLB-phosphothreonine 17 (1:5,000 dilution) antibodies and visualized with peroxidase-labeled secondary antibodies (Amersham Pharmacia Biotech). The degree of labeling was determined using a PhosphorImager and the ImageQuant

Phospholamban (PLB), 1 a 52-amino acid phosphoprotein, has been shown to interact with and regulate the apparent Ca 2ϩ affinity of the sarcoplasmic reticulum (SR) Ca 2ϩ -ATPase (1). The mechanism of action and functional significance of PLB have been well characterized in cardiac muscle because of the abundant expression of this protein in cardiac SR (2). Low levels of PLB expression have also been detected in slow twitch skeletal muscle (3,4), smooth muscle (5), and a non-muscle tissue, the vascular endothelium (6), although the role of PLB in these tissues is not well characterized at present. In cardiac muscle, dephosphorylated PLB inhibits the apparent affinity of the SR Ca 2ϩ -ATPase (SERCA2) for Ca 2ϩ (7)(8)(9)(10)(11), and phosphorylation of PLB, in response to ␤-adrenergic stimulation, removes its inhibition of SERCA2 (12,13). In vitro and in vivo studies have shown that PLB is phosphorylated at Ser 16 by cAMP-dependent protein kinase and at Thr 17 by Ca 2ϩ /calmodulin-dependent protein kinase (14 -16). Phosphorylation at each of these sites is associated with stimulation of the initial rates of SR Ca 2ϩ transport, especially at low or diastolic Ca 2ϩ concentrations (7,9,11,17). The stimulatory effects of PLB phosphorylation at these two sites can be reversed by a cardiac SR-associated type 1 protein phosphatase, which is also subject to cAMP-dependent phosphorylation of its inhibitor protein (18,19).
The apparent affinity of the SERCA2 for Ca 2ϩ is not only regulated by the phosphorylation state of PLB, but is also modulated by changes in the PLB:SERCA2 ratio. Alterations in the stoichiometric ratio of PLB to SERCA2, associated with alterations in SR Ca 2ϩ transport, have been implicated as important determinants of depressed left ventricular function in physiological and pathophysiological conditions. In hypothyroidism, increases in the PLB:SERCA2 ratio reflect decreases in the rates of SR Ca 2ϩ transport and relaxation; in hyperthyroidism, decreases in this ratio are associated with increases in the rates of SR Ca 2ϩ transport and relaxation (20,21). In murine atrial muscle, the PLB:SERCA2 ratio is shown to be 4-fold lower than ventricular muscle, and this has been suggested to reflect the enhanced rates of contraction and relaxation in this muscle (22). Furthermore, transgenic mice, either deficient in PLB or expressing reduced levels of PLB (PLBheterozygous), exhibited increased rates of SR Ca 2ϩ transport and enhanced cardiac ventricular function compared with wildtype littermates (23). A direct linear correlation was obtained between the relative levels of PLB:SERCA2 and the apparent affinity of SERCA2 for Ca 2ϩ as well as the rates of contraction and relaxation in isolated beating hearts or isolated ventricular cardiomyocytes from wild-type, PLB-heterozygous, and PLBdeficient mice (23). Thus, the functional stoichiometry of PLB: SERCA2 in cardiac muscle plays an important role in modulating myocardial contractility by regulating the rate of Ca 2ϩ sequestration into the SR lumen.
The molar stoichiometry of PLB:SERCA2 in native membranes of cardiac SR is presently unclear because different ratios of oligomeric and monomeric forms of PLB and SERCA2 have been reported in the literature. In vitro studies using 32 P labeling of PLB and SERCA2 reported a 1:1 ratio of PLB to SERCA2, assuming that PLB (23 kDa) was a heterodimer and that the functional unit of SERCA2 was a dimer (existing in its phosphorylated (EP) and unphosphorylated (E) state) (11). The use of calmodulin affinity labeling of PLB suggested a ratio of one PLB monomer to one SERCA2 monomer (24), whereas the use of a monoclonal antibody to detect the PLB-phosphorylated intermediates indicated a relationship of 2 mol of PLB monomer to 1 mol of SERCA2 monomer (17). More recently, in vivo studies showed that overexpression of wild-type PLB, either in the hearts of transgenic mice or adenoviral transfected cardiomyocytes, resulted in depressed SR and left ventricular function, suggesting that there is a fraction of Ca 2ϩ pumps in the native SR which is not functionally regulated by PLB (25,26). Furthermore, there was a close linear correlation observed between the relative levels of PLB:SERCA2 and the EC 50 values of SERCA2 for Ca 2ϩ in PLB 2-fold overexpression, wildtype, PLB-heterozygous, and PLB-deficient hearts, indicating that in transgenic hearts the overexpressed PLB was functionally coupled to SERCA2. However, it was unclear from these results whether all of the spare, unregulated pumps in the SR were saturated by the overexpressed PLB. Thus, to determine the functional stoichiometric ratio of PLB to SERCA2, which is associated with maximal inhibition of the affinity of SERCA2 for Ca 2ϩ , transgenic mice overexpressing a mutant form of PLB (S16A,T17A) were generated. The use of this mutant PLB, which cannot become phosphorylated, assured the lack of any compensation occurring at the level of PLB phosphorylation, to relieve its inhibitory effects in vivo. Assessment of SR Ca 2ϩ uptake in the transgenic hearts revealed increased inhibition of the affinity of SERCA2 for Ca 2ϩ with increased expression of PLB. Saturation of the PLB:SERCA2 ratio was obtained at PLB expression levels greater than 2-fold. Furthermore, cardiac hypertrophy was observed in transgenic hearts whose PLB:SERCA2 stoichiometry reached saturation, suggesting a compensatory response to the inhibitory effects of PLB in vivo.

EXPERIMENTAL PROCEDURES
The ethics committee of the University of Cincinnati approved the handling and maintenance of the animals in this study. 10 -12-week-old mice of either sex were used for the following studies.
Site-directed Mutagenesis-The PCR methodology by Bowman et al. (27) was used to incorporate the site-specific mutation S16A,T17A (TC-CACT 3 GCCGCT) into the PLB cDNA. A 0.85-kb SalI fragment containing the PLB cDNA and the SV40 polyadenylation signal sequence (PLB cDNA-SV40-poly(A)) was released from the ␣-myosin heavy chain promoter (␣-MHC)-PLB-SV40 fusion gene, used previously to generate transgenic mice overexpressing wild-type PLB (25,27). This SalI PLB cDNA-SV40-poly(A) fragment was then subcloned into a pBluescript SKII(Ϫ) vector (Stratagene), which has both T3 and T7 primer sites flanking the insert. PCR mutagenesis was performed by two consecutive PCR amplifications, using two different sets of primers, as described previously (28). In the first PCR amplification, a 5Ј-end mutant primer (5Ј-CT ATC AGG AGA GCC GCC GCT ATT GAA ATG CC-3Ј), corresponding to nucleotides 32-62 of the PLB coding sequence, and a 3Ј-end T7 primer were used to generate the desired mutant PLB cDNA minor product. In the second PCR amplification, an aliquot of the first PCR product and the T3 and T7 primers was used to amplify the full-length insert, which contained the desired mutation in the PLB cDNA. The final product was cut with SalI, gel purified, and resubcloned into the SalI site of a second pBluescript SKII(Ϫ) vector, which was then transformed into XL1-Blue competent cells. Colonies from the transformed cells containing the desired mutant PLB cDNA were identified by DNA sequencing. The mutated PLB cDNA-SV40-poly(A) sequence was excised by SalI from the pBluescript SKII(Ϫ) vector and ligated into the SalI site of the 5.5-kb mouse ␣-MHC promoter, also contained in the pBluescript SKII(Ϫ) vector.
Generation and Identification of Mutant Mice-The entire expression construct was contained in the pBluescript SKII(Ϫ) vector as an SpeI-KpnI fragment, which was composed of the cardiac-specific ␣-MHC promoter (5.5 kb), the PLB coding region with S16A,T17A (0.6 kb), and the SV40-poly(A) signal sequence (0.25 kb). The SpeI-KpnI fragment was released from the plasmid vector, gel purified, and used for pronuclear microinjection of fertilized eggs from FVB/N mice to generate transgenic mice according to standard procedures (29). Transgenic mice harboring the mutated PLB transgene were identified using PCR methodology and Southern analysis of genomic DNA isolated from tail biopsies, as described previously (30,31). The transgene expression, driven by the cardiac-specific ␣-MHC promoter, was determined by Northern analysis of total RNA from transgenic mouse hearts (32). Two different lines of hemizygous transgenic mice overexpressing 1.8-fold and 1.9fold mutant PLB were mated to generate transgenic offspring that would overexpress higher levels of mutant PLB. Transgenic offspring, expressing either one transgene or both transgenes from each parent, were identified by Southern blot analysis using genomic DNA obtained from tail biopsies. Briefly, genomic DNA was digested with BamHI and EcoRI overnight, separated by gel electrophoresis, and transferred onto a nitrocellulose membrane. 32 P-Labeled PLB cDNA was hybridized to the membrane, and the copy number of the transgene was determined relative to the endogenous PLB gene, using a PhosphorImager and ImageQuant analysis system. Transgenic offspring exhibiting greater transgene levels than either of their transgenic parents were chosen to be studied. In these offspring, the transgene levels were similar from mating to mating.
Western Blot Analysis-Quantitative immunoblotting of cardiac homogenates and microsomes enriched in SR membranes was carried out as described previously (33). Briefly, a pool of three to six hearts was prepared from either wild-type or transgenic mice and homogenized at 4°C in buffer A, pH 7.0, containing (in mmol/liter) 10 imidazole, 300 sucrose, 1 dithiothreitol, 1 sodium metabisulfite, and 0.3 phenylmethylsulfonyl fluoride. These cardiac homogenates were used to assess the levels of PLB, SERCA2, calsequestrin, calreticulin, ryanodine receptor, ␤-myosin heavy chain, and ␣-actin in wild-type and transgenic mouse hearts. To determine if the overexpressed mutant form of PLB was inserted into the SR membrane, preparations of microsomes enriched in SR membrane were prepared by differential centrifugation of the cardiac homogenate. Homogenates were centrifuged at 8,000 ϫ g (20 min), and the pellets were rehomogenized in buffer A and centrifuged as above. The supernatants from the two spins were combined, 4.0 M NaCl was added to a final concentration of 0.6 M and centrifuged at 100,000 ϫ g (60 min). The resulting pellet was washed in buffer A and recentrifuged at 100,000 ϫ g (60 min). The final pellet was resuspended in buffer A and stored at Ϫ80°C. The protein concentrations of homogenates and enriched microsomes were determined by the Bio-Rad method using bovine serum albumin as a standard. The homogenates and microsomes were incubated with equal volumes of loading buffer (20% glycerol, 2% ␤-mercaptoethanol, 4% SDS, 0.001% bromphenol blue, and 130 mmol/liter Tris-Cl, pH 6.8). Cardiac homogenates were separated by 8% SDS-PAGE (ryanodine and ␤-MHC) or 13% SDS-PAGE (PLB, SERCA2, calsequestrin, calreticulin, and ␣-actin) and transferred to nitrocellulose membranes (0.05 m for PLB; 0.22 m for SERCA2, calsequestrin, calreticulin, ryanodine receptor, ␤-MHC, and ␣-actin (Schleicher & Schuell)). The membranes were incubated with PLB (1:1,000 dilution), SERCA2 (1:500), calsequestrin (1:2,500), calreticulin (1:10,000), ryanodine receptor (1:500), ␤-MHC (1:2,500), and ␣-actin (1:2,000) antibodies and visualized with either 35 S-labeled (2 ϫ 10 5 cpm/ml) or peroxidase-labeled secondary antibodies (Amersham Pharmacia Biotech). The degree of labeling was determined using a PhosphorImager and the ImageQuant software. For immunodetection of PLB phosphorylation sites, polyclonal antibodies raised against a PLB peptide (residues 9 -199) phosphorylated at Ser 16 (PLB-phosphoserine 16) or at Thr 17 (PLB-phosphothreonine 17) were used. The samples were separated by 15% SDS-PAGE and transferred onto 0.05-m nitrocellulose membrane. The membranes were incubated with PLBphosphoserine 16 (1:10,000 dilution) and PLB-phosphothreonine 17 (1:5,000 dilution) antibodies and visualized with peroxidase-labeled secondary antibodies (Amersham Pharmacia Biotech). The degree of labeling was determined using a PhosphorImager and the ImageQuant software.
SR Ca 2ϩ Uptake Assay-Mouse hearts were excised, frozen in liquid nitrogen, and stored at Ϫ80°C. The frozen hearts were powdered and homogenized in 50 mM KH 2 PO 4 , pH 7.0, 10 mM NaF, 1 mM EDTA, 0.3 mM sucrose, 0.3 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol. The initial rates of Ca 2ϩ uptake in whole-heart homogenates were obtained and calculated as described previously (34).
In Vitro Phosphorylation-Cyclic AMP-dependent protein kinase or Ca 2ϩ /calmodulin-dependent protein kinase phosphorylation was performed as described previously (28) in cardiac homogenates of wild-type and transgenic mice. In the non-radioactive phosphorylation experiments, 4.0 mM ATP was used in place of the 0.1 mM [␥-32 P]ATP in the phosphorylation assay buffer.
Materials-Generous gifts of materials included mouse ␣-MHC promoter from Dr. J. Robbins (Children's Hospital Medical Center, Cincinnati, OH), rabbit polyclonal anti-calsequestrin affinity-purified antibody from Dr. L. R. Jones (Indiana University, Indianapolis), and mouse monoclonal anti-␤-MHC antibody from Dr. J. J. Leger (Pharmacie INSERM Unité, 300 LPM2, Montpellier, France). The SERCA2 polyclonal antibody was generated in rabbits using the 192-205 amino acid sequence portion of the SERCA2. The mouse anti-PLB and antiryanodine monoclonal antibodies were obtained from Affinity BioReagents, Inc. The rabbit anti-calreticulin polyclonal antibody was obtained from Stressgen Biotech, Inc. The mouse anti-␣-actin monoclonal antibody was obtained from Sigma Chemical Co. The rabbit anti-PLBphosphoserine 16 and phosphothreonine 17 antibodies were obtained from PhosphoProtein Research Inc.
Data Analysis-Data were plotted, and curve fits were obtained using KaleidaGraph by Abelbeck Software. The KaleidaGraph program uses the Levenberg-Marquardt algorithm for non-linear curve fitting. Data were weighted using the reciprocal of the weighting factor () calculated from the equation ϭ (S.E.) 2 , where S.E. is the standard error of the mean of replicate measures of the EC 50 (n ϭ 3-9).
Statistical Analyses-Data are expressed as mean Ϯ S.E. Statistical analyses were performed using Student's t test for unpaired observations. Values of p Ͻ 0.05 were considered statistically significant.

Generation of Transgenic Mice Expressing Mutant
Phospholamban (S16A,T17A) in the Heart-To determine the saturation point at which SERCA2 is inhibited maximally by PLB, several lines of transgenic mice were generated which expressed increasing levels of PLB. A mutant form of PLB in which both phosphorylation sites, Ser 16 and Thr 17 , were mutated to Ala 16 and Ala 17 , respectively, was used to ensure that the overexpressed PLB would not become phosphorylated in vivo, resulting in attenuation of its inhibitory effects. Previous studies in expression systems have shown that site-directed mutagenesis of Ser 16 or Thr 17 to Ala in PLB (35) does not alter the inhibitory interaction between the mutant PLB and SERCA2. Therefore, both Ser 16 and Thr 17 were mutated to Ala 16 and Ala 17 , respectively (TCCACT 3 GCCGCT) in the mouse PLB cDNA, and cardiac-specific expression of mutant PLB was driven using the ␣-MHC promoter. 15 founder mice were identified by PCR and Southern blot analyses, and these were bred for further characterization studies. Northern blot analysis (Fig. 1) of total RNA from hearts of wild-type and transgenic mice revealed the presence of two endogenous PLB transcripts at 2.8 and 0.7 kb, as described previously (25). 4 of the 15 transgenic lines also demonstrated strong signals of the transgenic transcript migrating at ϳ1.0 kb. These lines were bred and propagated for further characterization at the protein level.
To quantitate the levels of PLB protein expression in the hearts of the four transgenic lines, cardiac homogenates from transgenic and wild-type mice were processed in parallel for Western blot analysis. Quantitative immunoblotting revealed (a) 1.8-fold increase in one transgenic line (line 78); (b) 1.9-fold increase in a second transgenic line (line 72); (c) 3.7-fold increase in a third transgenic line (line 86); and (d) 4.7-fold increase in a fourth transgenic line (line 38) in the levels of PLB expression compared with control wild-type hearts (1.0-fold) (Fig. 2). These increases in PLB levels were similar utilizing samples that were either non-boiled (PLB pentamers and monomers) or boiled (PLB monomers) prior to SDS-PAGE. To determine whether the overexpressed PLB was incorporated into the SR membranes, enriched SR microsomal preparations were isolated from transgenic lines 78 and 38 and wild-type hearts. The SR preparations along with their respective homogenates were processed in parallel for quantitative immunoblotting. The levels of PLB overexpression in the microsomes were similar to the levels of overexpression in crude cardiac homogenates from each of the two transgenic lines (data not shown). These results indicate that the overexpressed PLB was incorporated into the SR membrane.
To obtain an additional level of PLB overexpression, two separate transgenic lines (78 and 72), which overexpressed PLB by 1.8-fold and 1.9-fold and were hemizygous for the mutant PLB transgene, were mated. Offspring identified by Southern blot analyses (see "Experimental Procedures") exhibited a 2.6-fold increase in the levels of PLB expression in their heart (Fig. 2B).
In Vitro Phosphorylation of Phospholamban-Cardiac homogenates from transgenic and wild-type mice were phosphoryl- ated in the presence of [␥-32 P]ATP and protein kinase A catalytic subunit or Ca 2ϩ /calmodulin and then processed for SDS-PAGE and autoradiography. The degree of 32 P incorporation in PLB was similar in transgenic and wild-type hearts, indicating that only the endogenous PLB could become phosphorylated in these hearts (Fig. 3A). To verify these findings further, in vitro phosphorylation assays were performed in the presence of nonradioactive ATP and then processed for SDS-PAGE and Western blot analysis. PLB site-specific phosphoserine and phosphothreonine polyclonal antibodies were used to detect PLB phosphorylated at either Ser 16 by cAMP-dependent protein kinase or Thr 17 by Ca 2ϩ /calmodulin-dependent protein kinase. Similar levels of Ser 16 -and Thr 17 -phosphorylated PLB were detected in wild-type and PLB mutant hearts (Fig. 3B). No alterations in Ser 16 or Thr 17 PLB phosphorylation were observed in the PLB-overexpressing mutant hearts, indicating that there was no effect on the expression of the endogenous PLB by the overexpressed mutant form of PLB. Furthermore, these results confirm that the overexpressed mutant form of PLB could not become phosphorylated by either cAMP-dependent protein kinase or Ca 2ϩ /calmodulin-dependent protein kinase in vitro.
Sarcoplasmic Reticulum Ca 2ϩ Uptake Rates-The effect of increasing levels of the PLB mutant on SERCA2 EC 50 values for Ca 2ϩ was evaluated by examining the initial rates of ATPdependent, oxalate-facilitated SR Ca 2ϩ uptake over a wide range of Ca 2ϩ concentrations, using cardiac homogenates from transgenic and wild-type mice. The incubation conditions in cardiac homogenates, which restrict Ca 2ϩ uptake to SR vesicles, have been defined previously (36,37). Ca 2ϩ uptake rates by transgenic hearts were significantly lower than those by wild-type hearts, especially at low Ca 2ϩ concentrations (Fig. 4), whereas there was no significant change in the maximum velocity of Ca 2ϩ uptake (V max ) (Table I). Furthermore, 1.8-and 2.6-fold increases in the levels of PLB were associated with progressive increases in the EC 50 values of SERCA2 for Ca 2ϩ (Table I), indicating that the overexpressed mutant form of PLB was capable of interacting with and inhibiting SERCA2. However, further increases (3.7-and 4.7-fold) in PLB levels did not result in any further increase in the SERCA2 EC 50 values for Ca 2ϩ in transgenic hearts, suggesting that saturation in the apparent affinity of SERCA2 for Ca 2ϩ was reached in the 2.6-fold PLB overexpression hearts (Table I).
Sarcoplasmic Reticulum Ca 2ϩ -handling Proteins and Compensatory Mechanisms-To determine whether overexpression of PLB and increased inhibition of the affinity of SERCA2 for Ca 2ϩ were associated with alterations in the expression of other SR proteins, the levels of SERCA2, calsequestrin, calreticulin, and ryanodine receptor were assessed, using quantitative immunoblotting. There was no significant difference in the levels of SERCA2 expression in PLB mutant hearts compared with wild-type hearts (Table II). Thus, because the levels of SERCA2 were not altered, increases in the expression of PLB in transgenic hearts resulted in increases in the relative PLB: SERCA2 ratio. Furthermore, the protein expression levels of calsequestrin, calreticulin, and ryanodine receptor were not altered significantly in transgenic hearts compared with wild types (Table II), indicating no compensatory responses by the major SR Ca 2ϩ -handling proteins in the PLB mutant hearts.
To examine whether any additional compensation had occurred in the PLB mutant hearts, the protein expression levels of ␣-actin and ␤-MHC were assessed by quantitative immunoblotting. No significant changes in the protein expression levels of ␣-actin were detected (105 Ϯ 13, 95 Ϯ 6, 99 Ϯ 4, 107 Ϯ 6 in 1.8-, 2.6-, 3.7-, and 4.7-fold PLB mutant hearts, respectively); however, increases in ␤-MHC protein expression were detected in transgenic hearts (Fig. 5). A small increase in the ␤-MHC protein levels was observed in the 1.8-fold transgenic hearts (1.3-fold Ϯ 0.1 increase in ␤-MHC), whereas greater increases were detected in the 2.6-, 3.7-, and 4.7-fold transgenic (Fig. 5). Because ␤-MHC has been reported previously to be a marker of hypertrophy (38), gravimetric analysis of some transgenic hearts was performed. Hearts from transgenic mice overex- pressing 4.7-fold PLB revealed a significant increase (18%) in the heart:body weight ratio (4.92 Ϯ 0.07 mg/g; n ϭ 9), although there was no significant difference in this ratio in the 1.8-fold PLB-overexpressing hearts (4.28 Ϯ 0.09; n ϭ 7) compared with wild-type controls (4.17 Ϯ 0.07 mg/g; n ϭ 14). Correlation between Relative PLB:SERCA2 Levels and the EC 50 Values of SR Ca 2ϩ Uptake for Ca 2ϩ -Previous studies, using transgenic mice overexpressing wild-type PLB, suggested that spare Ca 2ϩ pumps exist in the SR which are not regulated by PLB under basal conditions (25,26). To determine whether SERCA2 was inhibited maximally by PLB in any of our models, the relative protein levels of PLB:SERCA2 in wildtype, 1.8-, 2.6-, 3.7-, and 4.7-fold PLB mutant hearts were plotted against their respective SR Ca 2ϩ uptake EC 50 values (Fig. 6). In addition, the relative protein levels of PLB:SERCA2 and respective EC 50 values obtained in PLB-deficient (0.11 M) and PLB-heterozygous (0.18 M) hearts (23) were incorporated (Fig. 6). Ablation or reduction of PLB had no effect on SERCA2 protein expression levels (23,25). Thus, the relative ratio of PLB to SERCA2 was set as 1.0 in wild-type hearts; 0 in PLBdeficient hearts (34); 0.4 in PLB-heterozygous hearts (23); and 1.8, 2.6, 3.7, and 4.7 in the respective PLB mutant hearts. A four-parameter logistic fit was used to calculate the EC 50 value at which saturation of the relative PLB:SERCA2 ratio occurred. The maximal EC 50 value obtained from the fitted data was 0.63 Ϯ 0.02 M, which was similar to the EC 50 values obtained in the 2.6-, 3.7-, and 4.7-fold PLB-overexpressing hearts. To estimate the relative PLB:SERCA2 ratio at which saturation of SERCA2 inhibition by PLB occurs, we extrapolated the "fitted EC 50 value of saturation" (0.63 M) to the linear portion of the saturation curve (y ϭ 0.206x ϩ 0.097 r ϭ 0.999) and calculated the corresponding "functional PLB: SERCA2 ratio" as 2.6:1. Thus, these data suggest that the relative PLB:SERCA2 ratio, set as 1:1 in wild-type hearts, corresponds to a "functional stoichiometry" of 0.4:1 or that ϳ40% of the SR Ca 2ϩ pumps are functionally regulated by PLB in native mouse SR membranes.

DISCUSSION
This study presents the first in vivo evidence that maximal inhibition of the affinity of SERCA2 for Ca 2ϩ by PLB is obtained at PLB expression levels that are 2.6-fold or higher than those in wild-type hearts, indicating that the functional stoichiometry of PLB:SERCA2 is approximately 0.4:1 in vivo. The generation of transgenic models with cardiac-specific overexpression of various levels of a non-phosphorylatable form of PLB in its native phospholipid environment allowed us to examine the effects of alterations in PLB:SERCA2 ratio on SR function. Cardiac-specific overexpression of PLB harboring the S16A,T17A mutation was achieved using the ␣-MHC promoter, which is developmentally and hormonally regulated in vivo (25). The mutation of S16A,T17A in PLB was chosen because recent studies in transgenic mice suggested that increased phosphorylation of PLB may constitute an important compensatory mechanism in the heart (39). Such increased PLB phosphorylation would attenuate the inhibitory effects of PLB overexpression on the affinity of SERCA2 for Ca 2ϩ and prevent estimates of PLB:SERCA2 ratios. Furthermore, previous studies showed that replacing Ser 16 by Ala or Thr 17 by Ala in PLB did not compromise its inhibitory effects in expression systems (35), indicating that these amino acid substitutions did not alter the interaction between PLB and SERCA2. Quantitative immunoblots of cardiac homogenates and enriched SR preparations from transgenic mice revealed 1.8-, 2.6-, 3.7-, and 4.7fold increases in PLB protein levels compared with wild-type littermates and confirmed that the SR membrane was capable of accommodating increased PLB levels. Thus, the PLB-overexpressing mice provided an attractive system for further elucidation of the PLB regulatory effects on SERCA2. Biochemical analysis of the SR Ca 2ϩ transport system indicated that the EC 50 of SERCA2 for Ca 2ϩ was increased significantly by PLB overexpression. However, the maximal velocity of Ca 2ϩ transport was similar in PLB-overexpressing and wild-type hearts. These findings together with our previous observations in PLBheterozygous and PLB-deficient hearts (28,34) show that PLB is not a modulator of the maximal velocity of the SERCA2 pump. When the relative levels of PLB in our mouse models with reduced or overexpressed PLB were plotted against the Ca 2ϩ transport EC 50 values, there was a close linear correlation up to 1.8-fold PLB overexpression. Maximal increases in EC 50 were observed in hearts overexpressing PLB by 2.6-fold or higher, suggesting a "functional saturation" of SERCA2 by PLB. Extrapolation between the EC 50 values and the PLB levels in the genetically engineered mouse models indicated that approximately 40% of the SR Ca 2ϩ pumps are functionally regulated by PLB in native SR membranes.
The functional stoichiometry of PLB:SERCA2 was shown previously to be a key regulator of cardiac contractile parameters in PLB-deficient, PLB-heterozygous, and PLB wild-type hearts (23). Furthermore, the relative ratio of PLB to SERCA2 was observed to remain constant throughout murine postnatal development (40), indicating that strict regulation of the relative PLB and SERCA2 levels is critical for maintaining proper cardiac function. However, the functional stoichiometry of PLB:SERCA2 in native membranes has been reported to be less than 1:1 and up to 2:1 (11, 17, 24 -26), reflecting the difficulties in assessing the levels of these two proteins in SR membranes. Overexpression of PLB in transgenic hearts (25) or cardiac myocytes (26) revealed inhibition of the affinity of SERCA2 for Ca 2ϩ , suggesting that the PLB:SERCA2 stoichiometry is less than 1:1, and a fraction of the SR Ca 2ϩ pumps is not regulated by PLB in the native SR (25,26). To determine the magnitude of this fraction of unregulated SR Ca 2ϩ pumps in vivo, we generated a series of transgenic lines with increasing levels of PLB expression in the heart and assessed the degree of inhibition of SR Ca 2ϩ transport rates by PLB. This two-prong approach allowed us to determine the level of PLB required to "saturate" inhibition of the SR Ca 2ϩ pumps and assess indirectly the native stoichiometry of PLB:SERCA2.
Several studies have reported previously that increases in the relative PLB:SERCA2 ratio may be associated with pathophysiological conditions. An increase in the PLB:SERCA2 ratio was observed in the hearts of hypothyroid rats and mice (1.82:1 and 1.93:1 PLB:SERCA2, respectively), and this alteration resulted in decreased SR Ca 2ϩ transport and depressed left ventricular function (20,21). A comparison of the mouse hypothyroid PLB:SERCA2 ratio with the saturating ratio obtained in this study (2.6:1, PLB:SERCA2) indicates that there was still a fraction of Ca 2ϩ pumps which was not regulated by PLB in hypothyroidism. Furthermore, in human heart failure, some studies have reported an increase in the relative PLB:SERCA2 ratio and suggested that this may contribute to the deteriorated cardiac function (41). Recent studies in failing human hearts have also revealed reduced levels of PLB phosphorylation at Ser 16 (42) and increased mRNA expression and activity of a type 1 protein phosphatase (43), indicating that a higher fraction of PLB is in the dephosphorylated state and contributes to greater inhibition of SERCA2. Thus, changes in the relative PLB:SERCA2 ratio and/or changes in the levels of PLB phosphorylation may be important in the regulation of Ca 2ϩ handling in cardiac function and dysfunction. Consistent with these findings, we observed that increases in the PLB:SERCA2 ratio higher than 2.6-fold resulted in induction of a fetal gene program associated with increased expression of ␤-MHC protein. This hypertrophic response may constitute an important compensatory mechanism in the transgenic hearts with overexpression of a non-phosphorylatable form of PLB. The molecular mechanisms underlying the regulatory effects of PLB overexpression on SERCA2 are not clear. Previous studies have shown that monomeric PLB and SERCA2 have the ability to form different oligomeric complexes in the SR membrane (44 -47). Wild-type PLB has been proposed to be 20 -30% monomeric, based on SDS-PAGE or fluorescence energy transfer measurements (48 -50). SERCA2 has also been shown to consist of highly dynamic monomers as well as large stationary aggregates and slow rotating oligomers in SR vesicles, which, upon PLB phosphorylation, disassociate and become more active (47). In addition, electron paramagnetic resonance and fluorescence energy transfer measurements have revealed that (a) wild-type PLB depolymerizes in the presence of SERCA2; (b) SERCA2 prefers to bind to PLB monomers and small PLB oligomers (having less than 5 subunits); and (c) phosphorylation of PLB is associated with increases in PLB oligomerization (46,50). This reciprocal relationship between PLB oligomerization upon its phosphorylation and activation of SERCA2 is consistent with the increased inhibition of SERCA2 by monomeric PLB mutants in expression systems (48). Thus, the monomeric form of PLB is the more effective inhibitor of the SR Ca 2ϩ pump, and alterations in the equilibrium between PLB pentamers and monomers, caused by PLB phosphorylation/dephosphorylation, may influence the calculation of the functional PLB:SERCA2 stoichiometry. In our study, we correlated total PLB protein expression with the EC 50 of SERCA2 transport, assuming that the inserted mutations did not alter the pentamer:monomer ratio or the affinity of PLB for SERCA2 compared with wild-type PLB. Thus, the saturating stoichiometry of 2.6:1 for PLB:SERCA2 represents a "functional estimate" based on SERCA2 uptake measurements and a relative corresponding ratio of 1:1 in wild-type hearts.
In summary, our findings demonstrate that overexpression of a non-phosphorylatable form of PLB in transgenic mouse hearts resulted in saturation of the functional PLB:SERCA2  5. Western blot analysis of ␤-MHC protein levels in wildtype (WT) and PLB mutant (MT) transgenic hearts. Panel A, representative immunoblot of ␤-MHC from wild-type and PLB mutant transgenic hearts (4.7-fold PLB mutant). Increasing amounts of cardiac homogenates (3, 6, 9 g for wild-type; and 1, 2, 3 g for 4.7-fold PLB mutant) were subjected to SDS-PAGE and immunoblotting, as described under "Experimental Procedures." Panel B, quantification of ␤-MHC protein expression levels in hearts from wild-type and PLB mutant transgenic mice. Values represent the mean Ϯ S.E. of three to four determinations. Three to six hearts were pooled from each group.
FIG. 6. Relation between relative PLB/SERCA2 protein levels and the affinity of SERCA2 for Ca 2؉ in PLB-deficient (PLB KO), PLB-heterozygous (PLB HZ), wild-type (WT), and PLB mutant (MT) transgenic hearts. The EC 50 of SR Ca 2ϩ uptake for each model was plotted against its respective PLB:SERCA2 ratio. Values represent mean Ϯ S.E. of three to nine determinations. The broken line represents the linear fit obtained from PLB-deficient, PLB-heterozygous, wildtype, and 1.8-fold PLB mutant transgenic hearts (y ϭ 0.0968 ϩ 0.2059x; r ϭ 0.999). The solid line represents the four parameter logistic fit obtained from fitting the data points from all models; y ϭ [maximum Ϫ minimum]/(1 ϩ [K/x] n ϩ minimum; maximum inhibition ϭ 0.634 Ϯ 0.019; minimum inhibition ϭ 0.100 Ϯ 0.002; chi square ϭ 6.91. ratio, which was associated with inhibition of the affinity of SERCA2 for Ca 2ϩ and induction of cardiac hypertrophy. Functional saturation was obtained at a relative ratio of 2.6:1 for PLB:SERCA2, indicating that approximately 40% of the SR Ca 2ϩ pumps are functionally interacting with and regulated by PLB in native SR. Future studies involving crystallization of PLB and SERCA2 in the plane of the SR membrane will provide more direct structural information regarding the important interaction and modulation of SERCA2 with PLB.