Impaired Cardiac Performance in Heterozygous Mice with a Null Mutation in the Sarco(endo)plasmic Reticulum Ca 2 1 -ATPase Isoform 2 (SERCA2) Gene*

The sarco(endo)plasmic reticulum Ca 2 1 -ATPase isoform 2 (SERCA2) gene encodes both SERCA2a, the cardiac sarcoplasmic reticulum Ca 2 1 pump, and SERCA2b, which is expressed in all tissues. To gain a better understanding of the physiological functions of SERCA2, we used gene targeting to develop a mouse in which the promoter and 5 * end of the gene were eliminated. Mating of heterozygous mutant mice yielded wild-type and heterozygous offspring; homozygous mutants were not observed. RNase protection, Western blotting, and biochemical analysis of heart samples showed that SERCA2 mRNA was reduced by ; 45% in heterozygous mutant hearts and that SERCA2 protein and maximal velocity of Ca 2 1 uptake into the sarcoplasmic reticulum were reduced by ; 35%. Measurements of cardiovascular performance via transducers in the left ventricle and right femoral artery of the anesthetized mouse revealed reductions in mean arterial pressure, systolic ventricular pressure, and the absolute values of both positive and negative d P /d t in heterozygous mutants. These results demonstrate that two functional copies of the SERCA2 gene are required to maintain normal levels of SERCA2 mRNA, protein, and Ca 2

The SERCA2 1 gene encodes two Ca 2ϩ -transporting ATPases, SERCA2a and SERCA2b, which differ in their C-terminal sequences as a result of alternative splicing (1)(2)(3). SERCA2a is expressed at highest levels in heart, where it plays a central role in cardiomyocyte Ca 2ϩ handling required for excitation/ contraction coupling (reviewed in Ref. 4). Contraction is initiated by an increase in Ca 2ϩ concentrations around the myofibrils, which occurs as Ca 2ϩ is released from the SR or enters the cell via channels in the sarcolemma. SERCA2a pumps Ca 2ϩ out of the cytosol and into the SR, thereby contributing to the low diastolic Ca 2ϩ levels required for relaxation and replenishing Ca 2ϩ stores needed for the next contraction. SERCA2a serves a similar function in slow twitch skeletal muscle and is also expressed in some smooth muscles (5). In contrast to the limited tissue distribution and organ-specific function of SERCA2a, SERCA2b is expressed in all tissues and it has been suggested that it plays an essential housekeeping role (1), although it undoubtedly serves some organ-specific functions as well.
An important role for SERCA2a in cardiac function is well established. Recent studies have shown that the levels of SERCA2a are decreased in several animal models of cardiac hypertrophy and human heart failure (reviewed in Refs. 6 and 7). However, it is currently unknown whether, and to what extent, the reductions in SERCA2a levels contribute to altered contractile function. In addition, it is unclear whether there are homeostatic mechanisms within the cardiac myocyte that are capable of sensing perturbations in the levels of pump activity and adjusting its expression accordingly. These are important issues, given the observed reductions in SERCA2a mRNA and protein expression in the failing heart, which correlate well with decreased myocardial Ca 2ϩ transport and contractile function. However, it is unclear whether the reduction in the level of SERCA2 expression and activity are a cause or a consequence of impaired cardiac function.
To obtain an animal model that could be used to examine the physiological functions of SERCA2 in cardiovascular and other tissues, we used gene targeting to prepare a mouse carrying a null mutation in the SERCA2 gene. Because homozygous SERCA2 null mutants were not observed, we examined cardiovascular function and SERCA2a expression and activity in hearts of heterozygous mutants, which display no outward manifestations of a disease phenotype. These studies show that SERCA2 mRNA, protein, and activity levels are reduced in heterozygous mutants carrying a single functional copy of the SERCA2 gene, and that these reductions have a significant effect on cardiac performance.

EXPERIMENTAL PROCEDURES
Preparation of Targeting Construct and Generation of Mutant Animals-A genomic clone containing the 5Ј end of the mouse SERCA2 gene was isolated from a strain 129/SvJ phage library and characterized by restriction mapping and sequence analysis. The targeting vector, pMJKO, was the same as that used previously to prepare a SERCA3 targeting construct (8). Two restriction fragments, a BglII-SacI fragment containing 2.5 kb of 5Ј flanking sequence and terminating 53 base pairs 5Ј of the transcription start site and a 2.8-kb SpeI restriction fragment beginning in exon 3 and terminating in intron 3, were inserted on either side of the neo gene. The orientation of the neo gene was opposite that of the SERCA2 gene. Electroporation of ES cells, selection of targeted ES cells, blastocyst-mediated transgenesis, and mating of chimeric mice with Black Swiss mice was carried out as described previously (8). DNA from targeted ES cells and tail biopsies were analyzed by Southern hybridization using probes from outside the region used to prepare the construct. The 3Ј probe was a 2.4-kb EcoRV-SpeI fragment from intron 3 that began 3.3 kb downstream from exon 3. The 5Ј probe was a 360-bp PstI-EcoRI fragment beginning ϳ3.6 kb upstream of the transcription initiation site.
Genotype Analysis-Genotyping was performed by Southern blot analysis of tail DNA using the 3Ј outside probe or by PCR analysis using a combination of three primers that amplify both wild-type and mutant alleles in the same reaction. For PCR analysis the following primers were used: primer 1 (5Ј-CGGCCTTCTAGAATTGCCGGCTG-3Ј), corresponding to genomic sequences near the 3Ј end of the SacI-SpeI fragment that was deleted in the targeted allele; primer 2 (5Ј-CTTACTA-AAGATATACATGCTGCCAGCAG-3Ј), complementary to sequences just 3Ј to the SpeI site in exon 3 that was used to prepare the construct; and primer 3 (5Ј-CTGACTAGGGGAGGAGTAGAAGG-3Ј), corresponding to sequences in the promoter region of the neo cassette. PCR conditions for simultaneous amplification of wild-type (221-bp product) and mutant (112-bp product) alleles were: denaturation at 94°C for 3 min, annealing for 45 s at 52°C, and extension for 45 s at 72°C.
Ribonuclease Protection Assays-DNA fragments used for generation of riboprobes for mouse SERCA2, phospholamban, and the L32 ribosomal subunit were subcloned into pBluescript. The subcloned fragments were: for SERCA2, a 451-bp NotI-XhoI fragment from the mouse SERCA2 gene, which contains 278 bp of exon 1 and 173 bp of intron 1; for phospholamban (9), a 70-bp fragment corresponding to nucleotides Ϫ181 to Ϫ111, relative to the initiation methionine codon; and for L32 (10), a 204-bp fragment corresponding to nucleotides Ϫ18 to ϩ186, relative to the initiation methionine codon. The pBluescript subclones were linearized, and 32 P-labeled riboprobes were synthesized using T7 RNA polymerase and the MAXIscript TM in vitro transcription kit (Ambion, Inc., Austin, TX). 5 g of total heart RNA (from 12-14-week-old mice, n ϭ 7 for each genotype) were analyzed using the RPA II TM ribonuclease protection assay kit (Ambion, Inc.), with all three probes included in each reaction. The protected fragments were separated by electrophoresis in a 5% denaturing polyacrylamide gel, analyzed by autoradiography, and quantitated by PhosphorImager analysis (Molecular Dynamics, Wayzata, MN). Levels of SERCA2 and phospholamban mRNAs were determined relative to L32 mRNA.
Western Blot Analysis-SERCA2a levels in cardiac homogenates from 15-16-week-old mice (n ϭ 3 males and 3 females for each genotype) were determined by Western blot analysis, essentially as described previously (11,12). Proteins were separated in SDS-polyacrylamide gels, transferred to nitrocellulose membranes, and probed with a polyclonal anti-SERCA2a antibody (13) or a monoclonal anti-␣-actin antibody (Sigma). Binding of the primary antibody was detected by peroxidase-conjugated secondary antibody and enhanced chemilumi-nescence (Kirkegaard and Perry Laboratories, Gaithersburg, MD). The signal obtained for each heterozygous mutant sample (5 g of protein; duplicate lanes for each sample) was compared with a standard curve prepared by plotting the signal intensities obtained with increasing amounts of wild-type cardiac homogenates (1-5 g of protein from pooled samples of 6 wild-type mice). Quantitation of the signals was performed by densitometry using a UMAX Astra 1200 s scanner and analyzed using NIH Image version 6.1 software.
Ca 2ϩ Uptake Assays-Hearts were removed from 15-16-week-old wild-type and heterozygous mutant mice, rinsed in phosphate-buffered saline, and stored in liquid nitrogen. Preparation of whole heart homogenates and analysis of both the maximum rate of Ca 2ϩ uptake and the Ca 2ϩ dependence of Ca 2ϩ uptake were performed as described by Luo et al. (14). The use of whole heart homogenates for the analysis of Ca 2ϩ uptake by SR vesicles has been discussed and validated in previous studies (14 -17). Duplicates of each sample (n ϭ 3 males and 3 females of each genotype) were analyzed and the values presented are means Ϯ S.E. The V max and Ca 2ϩ concentration required for half-maximal activation (K 0.5 ) were calculated by non-linear regression analysis using MicroCal Origin software. Statistical analyses were performed by unpaired two-tailed Student's t test.
Analysis of Left Ventricular Function and Blood Pressure-Anesthesia, surgical procedures, and collection of data for analysis of cardiac performance in the intact closed-chest mouse were carried out essentially as described previously (18,19). 12-14-week-old male wild-type (n ϭ 11) and heterozygous mutant (n ϭ 10) mice were analyzed. Each anesthetized mouse was placed on a thermally controlled surgical table, and a tracheotomy was performed. The right femoral artery was cannulated and connected to a COBE CDXIII pressure transducer (COBE Cardiovascular, Arvada, CO) for measurement of arterial blood pressure. The right femoral vein was cannulated and used for infusion of the ␤-adrenergic agonist, dobutamine (Abbott Laboratories, North Chicago, IL) or the ␤-adrenergic antagonist, propranolol (Sigma). The right carotid artery was cannulated with a Millar Mikro-Tip transducer with an outer diameter of 0.47 mm (model SPR-612, Millar Instruments, Houston, TX), which was then advanced into the left ventricle. After the animal had stabilized for ϳ30 min, measurements of ventricular function in the absence of stimulation were performed. Cardiovascular performance in response to ␤-adrenergic stimulation was examined by infusing increasing concentrations of dobutamine (1-32 ng/min/g of body weight) over a 3-min period, with peak responses measured during the final 30 s of the infusion period. At the end of each experiment, a bolus of propranolol (100 ng/g of body weight) was infused in order to reevaluate cardiac function at the end of each experiment in the absence of ␤-adrenergic stimulation. Pressure signals from both the COBE and Millar transducers were recorded using a MacLab 4/s data acquisition system (AD Instruments, Milford, MA) and a Macintosh 7100/80 computer as described previously (18,19). The analysis soft-

FIG. 1. SERCA2 gene targeting strategy and Southern blot analysis.
A, targeting strategy. Top, restriction map and organization of the 5Ј end of the wild-type gene. Exons 1-3 and location of the 5Ј and 3Ј probes used for Southern blot analysis are indicated. Middle, targeting construct in which 1.5 kb of SERCA2 genomic sequence, containing exons 1 and 2 and part of exon 3, was replaced by the neo gene. The herpes simplex virus thymidine kinase gene was placed 5Ј to the SERCA2 genomic sequence. Bottom, targeted SERCA2 allele, with BamHI restriction fragments identified by 5Ј and 3Ј probes indicated below. Restriction enzyme sites: B, BamHI; Bg, BglII, Sp, SpeI; S, SacI. B, Southern blot analysis of tail DNA from offspring of heterozygous matings. The samples were digested with BamHI and hybridized with the 5Ј and 3Ј probes, which identify a 14.3-kb fragment in the wild-type allele and 8.1-and 6.3-kb fragments, respectively, in the mutant allele. Only wildtype (ϩ/ϩ) and heterozygous (ϩ/Ϫ) mice were observed.
ware permits direct determination of arterial systolic and diastolic pressure, mean arterial pressure, heart rate, left ventricular systolic pressure, developed pressure, and both positive (dP/dt max ) and negative (dP/dt min ) dP/dt. Data were analyzed using a mixed, two-factor analysis of variance with repeated measures on the second factor. When necessary, post hoc comparisons were performed by single degree-of-freedom contrasts.

Preparation of Mutant Mice and Gross
Phenotype-To introduce a null mutation into the SERCA2 gene, we used a target-ing construct in which part of the promoter, the transcription initiation site, the first two exons, and part of the third exon were deleted (Fig. 1A). We anticipated that homologous recombination of this construct with the SERCA2 gene would eliminate transcription of the gene, thereby allowing unambiguous determination of mRNA and protein levels produced by the single remaining functional copy of the SERCA2 gene. After electroporation of ES cells, 3 of 105 colonies that survived the selection procedure were found to be correctly targeted when FIG. 2. RNase protection analysis of SERCA2 and phospholamban mRNAs in wild-type and heterozygous mutant hearts. A, total heart RNA (5 g/ lane) was hybridized with riboprobes for SERCA2, phospholamban (PLB), and the L32 ribosomal subunit. The samples were then treated with RNase A and RNase T1, separated by polyacrylamide gel electrophoresis, and visualized by autoradiography. Each pair of lanes, L-1-L-7, contains samples from 12-14-week-old, age-and sex-matched wild-type (W) and heterozygous mutant (H) mice; L-3 and L-5 were female pairs, and the others were male. B, the autoradiogram in panel A was examined by PhosphorImager analysis and the hybridization signals for SERCA2 and phospholamban mRNAs in hearts from wild-type (ϩ/ϩ) and heterozygous (ϩ/Ϫ) mice were quantitated relative to that of the L32 ribosomal subunit. The data are expressed as means Ϯ S.E. *, significantly different (p Ͻ 0.001) from wild-type as determined by Student's t test. analyzed by Southern blot hybridization. Fifteen male chimeric mice were obtained by blastocyst-mediated transgenesis. When bred with Black Swiss females, six of these chimeras yielded offspring carrying the mutant allele.
SERCA2 heterozygous mice were mated and the genotypes of offspring were determined by Southern blot analysis (Fig. 1B). Among 116 offspring, 41 were wild-type (35%) and 75 (65%) were heterozygous, with a similar percentage of males and females (53% and 47%, respectively). Homozygous mutants were not observed, and the ratio of wild-type:heterozygous offspring was very close to the 1:2 Mendelian ratio that would be expected if loss of homozygous mutants was occurring during embryonic development. Heterozygous mutants were viable, appeared healthy, and exhibited no outward manifestations of a disease phenotype. There was no evidence of cardiac hypertrophy in the mutant mice. The heart weight/body weight (mg/g) ratios of the 15-16-week-old mice used below for analysis of SERCA2a levels and Ca 2ϩ uptake activity in heart was 4.33 Ϯ 0.18 for wild-type mice and 4.33 Ϯ 0.16 for heterozygous mutants.
SERCA2 mRNA Levels Are Reduced in Hearts of Heterozygous Mutant Mice-To determine whether the loss of one copy of the SERCA2 gene causes a reduction in the levels of SERCA2 and phospholamban mRNAs, RNase protection analysis was performed. The SERCA2 probe does not distinguish between SERCA2a and SERCA2b mRNAs. However, because previous nuclease protection studies have shown that ϳ95% of the SERCA2 mRNA in heart encodes SERCA2a (5), only a small fraction of the mRNA detected in these experiments would be due to SERCA2b. As shown in Fig. 2A, there appeared to be a marked reduction of SERCA2 mRNA, but not phospholamban mRNA, in the hearts of heterozygous mutants. Quantitation of the hybridization signal intensities by PhosphorImager analysis and normalization with an internal control (Fig. 2B) showed that SERCA2 mRNA levels were reduced to 55 Ϯ 2% of wildtype levels, whereas phospholamban mRNA levels were the same in both genotypes.
SERCA2a Protein and Ca 2ϩ Uptake Activity Are Reduced in Hearts of Heterozygous Mutant Mice-Western blot analysis of whole heart homogenates was performed using an anti-SERCA2a antibody, and the levels of ␣-actin were analyzed as a control. SERCA2a in heterozygous mutant hearts was reduced to 63 Ϯ 6% of wild-type levels (p Ͻ 0.05), whereas ␣-actin levels were the same in mutant and wild-type hearts (Fig. 3).
Analyses of both the maximum rate and Ca 2ϩ dependence of Ca 2ϩ uptake activity were performed using the same cardiac homogenates used for Western blot analysis. The maximum velocity of Ca 2ϩ uptake (Fig. 4A) was significantly reduced (p Ͻ 0.01) in cardiac samples from heterozygous mutants (to ϳ66% of wild-type levels), with a V max of 46.5 Ϯ 2.5 and 30.6 Ϯ 1.7 nmol/mg/min in wild-type and heterozygous hearts, respectively. In contrast, the Ca 2ϩ dependence of Ca 2ϩ uptake activity (Fig. 4B) was essentially the same in both sets of samples, with a K 0.5 of 0.238 Ϯ 0.009 M in wild-type hearts and 0.242 Ϯ 0.003 M in heterozygous hearts.
Cardiovascular Function in Wild-type and Heterozygous Mutant Mice-In these experiments, measurements of cardiac function were recorded using a micro-tip transducer that was placed in the left ventricle of anesthetized mice via the right carotid artery. Blood pressure was recorded using a transducer that was inserted in the right femoral artery, and dobutamine, a ␤-adrenergic receptor agonist, was infused through a catheter in the right femoral vein. Representative tracings of arterial blood pressure, left ventricular systolic pressure, and left ventricular dP/dt under non-stimulated conditions are shown in Fig. 5.
To determine whether there were any differences in cardiovascular function between the two genotypes under control conditions and at increasing levels of ␤-adrenergic stimulation, dobutamine dose-response relationships were analyzed. Under non-stimulated conditions and at the lower doses of dobutamine, the mean value for heart rate rate was slightly reduced in heterozygous mutants (Fig. 6A); however, the differences were not statistically significant. Mean arterial blood pressure (Fig. 6B) and left ventricular systolic pressure (Fig. 6C) were significantly lower in heterozygous mutants than in wild-type mice under non-stimulated conditions and at the two lower FIG. 4. Ca 2؉ uptake by sarcoplasmic reticulum in cardiac homogenates from wild-type and heterozygous mutant mice. Ca 2ϩ uptake assays were performed using whole heart homogenates from 15-16-week-old mice (n ϭ 6 for each genotype). A, Ca 2ϩ uptake rates (nmol of Ca 2ϩ /mg of protein/min) at varying Ca 2ϩ concentrations. B, Ca 2ϩ dependence of Ca 2ϩ uptake; mean values at each Ca 2ϩ concentration are from the experiment shown in panel A and are expressed as a percentage of Ca 2ϩ uptake rates at pCa 5.5. doses of dobutamine, but not at higher levels of ␤-adrenergic stimulation. The absolute values of both positive and negative dP/dt were significantly lower in heterozygous mutants under non-stimulated conditions and at all levels of ␤-adrenergic stimulation (Fig. 7). A significant dobutamine dose response (p Ͻ 0.001) for each of these measurements of cardiac performance was observed in both wild-type and heterozygous mutant mice.

DISCUSSION
Our primary objective was to analyze the relationship between cardiac performance and the levels of SERCA2 expression and activity in heart. Because we anticipated that homozygous mutants would not survive and that it would be necessary to perform our studies using heterozygotes, a major concern, and an important issue in its own right, was whether SERCA2 levels in heart were critically dependent on the presence of two functional copies of the SERCA2 gene. Homozygous mutants were not observed, demonstrating that one or both variants of SERCA2 serve functions that are essential for life, and that other pumps cannot provide sufficient compensation to rescue the null mutant. The 2:1 genotype ratio of heterozygous and wild-type offspring indicated that embryonic development of heterozygous mutants was normal. Both young and adult heterozygous mutant mice appeared healthy, grew normally, and were indistinguishable from wild-type mice in both appearance and behavior. To gain a better understanding of the in vivo functions of SERCA2, we conducted biochemical and physiolog-ical analyses of heterozygous mutants, with the expectation that SERCA2 mRNA, protein, and activity levels would be reduced and that this would produce an informative cardiac phenotype.
There are now many examples of mouse knockout models in which mRNA and protein levels are reduced in heterozygous mutants, with a corresponding alteration of physiological function. A particularly relevant example is the phospholambandeficient mouse (14,19,20), in which heterozygotes display an ϳ50% decrease in phospholamban mRNA and protein levels and an increase in cardiac contractility and relaxation. Nevertheless, for proteins that play a central role in critical physiological processes, such as the cardiac SR Ca 2ϩ pump in excitation-contraction coupling, it is not unreasonable to anticipate that homeostatic mechanisms within the cell might sense perturbations and adjust its levels of expression. In fact, recent studies have shown that transgenic mice with a 2.6-fold increase in SERCA2a mRNA in heart have only a 20% increase in SERCA2a protein (21) and that transgenic overexpression of SERCA1 in mouse heart leads to a ϳ50% reduction in the levels of endogenous SERCA2a (12). The results of those studies indicate the existence of mechanisms that can limit SERCA2a expression in cardiac myocytes when Ca 2ϩ pump activity is greater than normal. The development of the SERCA2-deficient mouse has allowed an investigation of whether the amount of SERCA2a mRNA and protein produced from a single copy of the gene can be up-regulated in response to a deficiency in Ca 2ϩ pump activity.
To allow unambiguous determination of the levels of mRNA and protein produced from the remaining functional copy of the SERCA2 gene in heterozygous mutants, we prepared the model using a targeting construct in which 1.5 kb from the 5Ј end of the gene was replaced with the neomycin resistance gene. Homologous recombination between the construct and the SERCA2 gene eliminated the TATA box, both the transcription and translation initiation sites, and sequences encoding the N-terminal cytoplasmic domain of the protein. RNase protection analysis revealed that steady-state levels of SERCA2 mRNA in hearts of heterozygous mice were ϳ55% of wild-type levels. These data indicate that the expression of normal levels of SERCA2 mRNA in heart is dependent on the presence of two functional copies of the SERCA2 gene, and that transcriptional up-regulation of the SERCA2 gene in response to a deficit in Ca 2ϩ sequestering activity, if it occurs at all, is very limited.
SERCA2a protein, however, was ϳ65% of wild-type levels, indicating an ϳ30% increase in the amount of protein produced from the single functional copy of the gene in heterozygous mutants. The results of these experiments show that both copies of the SERCA2 gene are required to maintain normal SERCA2 mRNA and protein levels, although some compensatory up-regulation of SR Ca 2ϩ pump expression from the remaining gene does occur in heterozygotes. The mechanisms underlying the modest compensatory up-regulation are unclear, but may occur at several levels, including increased transcription and/or mRNA stability and increased translation and/or a reduction in the rate of SERCA2a protein degradation.
When compared with the levels in wild-type mice, the V max of Ca 2ϩ uptake activity and SERCA2a protein in heart samples from heterozygous mutants were reduced to similar extents, confirming that there is a direct relationship between SERCA2a levels and maximum rates of Ca 2ϩ sequestration.  7. Dobutamine dose-response relationships for minimum and maximum dP/dt. The first derivative of the left ventricular pressure wave of the same wild-type and heterozygous mutant (knockout) mice analyzed in Fig. 6 was determined at all levels of ␤-adrenergic stimulation. A, positive dP/dt (dP/dt max ). B, negative dP/dt (dP/dt min ). †, significantly different from wild-type at all dose levels (p Ͻ 0.01), as determined by analysis of variance.
The Ca 2ϩ dependence of Ca 2ϩ uptake activity was the same in mutant and wild-type mice, suggesting that under basal conditions there is no compensatory up-regulation of pump activity in mutant hearts via an increased affinity for Ca 2ϩ . Such compensation could occur, for example, by phosphorylation of phospholamban (which represses cardiac contractility and relaxation when it is in its non-phosphorylated state), thereby relieving its inhibition of pump activity (14,19,20). The lack of an alteration in Ca 2ϩ dependence of Ca 2ϩ uptake activity was somewhat surprising, as phospholamban mRNA levels were the same in wild-type and heterozygous mutant hearts. If the levels of phospholamban protein and its phosphorylation state are unchanged in heterozygous hearts, which were not examined in this study, then the increase in phospholamban/ SERCA2 ratios in heterozygous mutants would be expected to cause a reduction in the affinity of the pump for Ca 2ϩ , as observed in hearts of transgenic mice in which phospholamban was overexpressed (22). Additional studies will be needed to determine whether phospholamban protein levels and its phosphorylation state under basal conditions are altered in heterozygous mutant hearts.
A clear cardiovascular phenotype was observed in heterozygous mutants, demonstrating that the ϳ35% reduction in SERCA2a levels and Ca 2ϩ uptake activity resulting from the loss of one functional copy of the SERCA2 gene has a significant physiological effect. Under basal conditions, mean arterial blood pressure and left ventricular systolic pressure were significantly lower in heterozygotes than in wild-type mice. However, during maximum ␤-adrenergic stimulation, a condition in which inhibition of the pump by phospholamban is relieved (14,19,20), both genotypes exhibited similar arterial and left ventricular pressures. The reduced levels of SERCA2a protein and activity would be expected to lead to both a reduction in SR Ca 2ϩ stores, which would decrease the amount of Ca 2ϩ available for excitation-contraction coupling, and a reduction in the rate of Ca 2ϩ sequestration, which would reduce the relaxation rate. Consistent with this expectation, the absolute values of both maximum and minimum dP/dt were significantly lower in mutant mice under basal conditions and at all levels of ␤-adrenergic stimulation. These results, along with those of previous studies showing that overexpression of SERCA2a (21) or SERCA1 (12) in transgenic mouse hearts increases cardiac contractility and relaxation, demonstrate that SR Ca 2ϩ pump levels are a major determinant of cardiac muscle contractility and relaxation under both basal and stimulated conditions. Studies of human heart disease have revealed reductions in the levels of SERCA2a mRNA, protein, and activity (6,7,(23)(24)(25)(26)(27)(28), and similar observations have been made in animal models of heart disease (29 -33). In the failing human myocardium, Ca 2ϩ transients are markedly prolonged in both Ca 2ϩ release and uptake phases (34), and there is evidence that the reduction in SERCA2a levels is involved in the altered force-frequency relationship (27). On the basis of these and other studies, it has been suggested that SR dysfunction resulting from decreased levels of the SR Ca 2ϩ pump contributes to the pathogenesis of heart disease (6,(23)(24)(25). The experiments described here have shown a direct relationship between SERCA2a levels and Ca 2ϩ uptake activity, and have also shown that a reduction in pump activity impairs cardiac contractility and relaxation. However, the 12-16-week-old mutant mice that we analyzed appeared to be healthy, and overt evidence of heart disease, such as cardiac hypertrophy, was not observed. This indicates that an ϳ35% deficit in Ca 2ϩ sequestering activity, such as that resulting from the loss of one copy of the SERCA2 gene, is not a sufficient perturbation to cause heart disease in the relatively young adult mice analyzed here. However, it remains to be determined whether heterozygous mutants may be more prone to heart disease as they age or when subjected to rigorous exercise or other experimental protocols.