Transgenic approaches to define the functional role of dual site phospholamban phosphorylation.

Phospholamban is a critical regulator of the sarcoplasmic reticulum Ca2+-ATPase activity and myocardial contractility. Phosphorylation of phospholamban occurs on both Ser16 and Thr17 during isoproterenol stimulation. To determine the physiological significance of dual site phospholamban phosphorylation, we generated transgenic models expressing either wild-type or the Ser16 --> Ala mutant phospholamban in the cardiac compartment of the phospholamban knockout mice. Transgenic lines with similar levels of mutant or wild-type phospholamban were studied in parallel. Langendorff perfusion indicated that the basal hyperdynamic cardiac function of the knockout mouse was reversed to the same extent by reinsertion of either wild-type or mutant phospholamban. However, isoproterenol stimulation was associated with much lower responses in the contractile parameters of mutant phospholamban compared with wild-type hearts. These attenuated responses were due to lack of phosphorylation of mutant phospholamban, assessed in 32P labeling perfusion experiments. The lack of phospholamban phosphorylation in vivo was not due to conversion of Ser16 to Ala, since the mutated phospholamban form could serve as substrate for the calcium-calmodulin-dependent protein kinase in vitro. These findings indicate that phosphorylation of Ser16 is a prerequisite for Thr17 phosphorylation in phospholamban, and prevention of phosphoserine formation results in attenuation of the beta-agonist stimulatory responses in the mammalian heart.

the Ca 2ϩ -ATPase and enhanced myocardial performance. Furthermore, the stimulatory effects to ␤-adrenergic agonists were more pronounced in the PLB-overexpressing hearts, whereas these effects were attenuated in the PLB-knockout hearts compared with wild types (1,2). These studies suggested that PLB plays a prominent role in the heart's responses to ␤-agonists. However, PLB is phosphorylated on both Ser 16 and Thr 17 during isoproterenol stimulation (3) and the relative contribution of each site in the altered contractile responses of the heart is not presently well known. In vitro studies have shown that Ser 16 is phosphorylated by cAMP-dependent protein kinase, whereas Thr 17 is phosphorylated by Ca 2ϩ -calmodulin-dependent protein kinase (4). Phosphorylation of each site occurs in an independent manner, although it is not presently clear whether the stimulatory effects of the two phosphorylations on SR Ca 2ϩ transport are additive (5)(6)(7)(8)(9).
In vivo studies have shown that phosphorylation/dephosphorylation of PLB by Ca 2ϩ -calmodulin-dependent protein kinase has as a prerequisite the phosphorylation/dephosphorylation by cAMP-dependent protein kinase (10 -15). Furthermore, elevation of intracellular [Ca 2ϩ ] to higher levels than those attained by isoproterenol, which resulted in higher peak tension than that elicited during ␤-adrenergic stimulation, failed to phosphorylate PLB (3,16). However, a recent study, using phosphorylation site-specific antibodies for PLB, indicated that cAMP-dependent and Ca 2ϩ -calmodulin-dependent phosphorylation of PLB can occur in an independent manner and their effects may be additive in vivo (17). Thus, the functional role of dual site phosphorylation of PLB is not clear.
The availability of the PLB knockout mouse in conjunction with site-specific mutagenesis technology have provided us with an excellent opportunity to further examine the interaction between the cAMP-dependent and the Ca 2ϩ -calmodulindependent pathways of PLB phosphorylation in the regulation of basal and ␤-agonist stimulated cardiac contractility in vivo. The aims of the present study were to: 1) determine whether reintroduction of PLB in the null background is feasible, and whether it is able to reverse the hyperdynamic cardiac phenotype of the PLB knockout mouse; and 2) elucidate the physiological role of Thr 17 phosphorylation in PLB in the absence of Ser 16 (Ser 16 3 Ala) by directing cardiac-specific expression of mutant PLB in the knockout background.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-The site-specific mutation of Ser 16 3 Ala (TCC 3 GCC) was introduced into PLB cDNA by polymerase chain reaction (PCR) methodology. Briefly, a 0.9-kb SalI fragment containing PLB cDNA and the SV40 polyadenylation signal sequence (PLB cDNA-SV40-poly(A)) was released from the ␣-MHCp-PLB-SV40 fusion gene (1). This SalI PLB cDNA-SV40-poly(A) fragment was then subcloned into a pBluescript SKII(Ϫ) vector (Stratagene), which has T3 and T7 primer sites flanking the insert. Polymerase chain reaction mutagenesis was performed by two consecutive PCR amplifications using * This work was supported by National Institutes of Health Grants HL26057, HL22619, and HL52318. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  two different sets of primers. For the first PCR amplification, 100 pg of the subcloned plasmid DNA containing the 0.9-kb SalI fragment was used as template, along with a 5Ј end mutant primer (5Ј-CT ATC AGG AGA GCC GCC ACT ATT GAA ATG CC-3Ј) corresponding to nucleotides 32-62 of the PLB coding sequence, and a 3Ј end T7 primer, to generate a desired mutant PLB cDNA minor product. Subsequently, an aliquot of the first PCR product as well as the T3 and T7 primers was used for the second PCR. The final amplified product was excised and resubcloned into the SalI site of a second pBluescript SK II(Ϫ) vector, which was then transformed into XL1-Blue-competent cells. The mutated PLB cDNA-SV40-poly(A) sequence was resubcloned into the SalI site of the parent PLB overexpression vector pIBI 31 (1).
Generation and Identification of the Transgenic Mice-The expression fragment containing the cardiac-specific ␣-myosin heavy chain (␣-MHC) promoter, the mutated PLB cDNA, and the SV40 poly(A) signal sequence was used for pronuclear microinjection of fertilized eggs derived from the intercrossing of male PLB knockout (PLB-KO) and female FVB/N mice. The first generation of transgenic mice was PLBheterozygous. Mice harboring the mutant PLB transgene were identified using PCR methodology and Southern blot analysis (1). Breeding the transgene-positive mice with PLB-KO mice generated offspring expressing the mutant PLB transgene in the PLB-KO background. The ␣-MHCp-PLB-SV40 fusion gene was also used to generate wild-type PLB (PLB-WT) transgenic mice in a similar manner as the PLB Ser 16 3 Ala mutant (PLB-MU) transgenic mice.
Langendorff Perfusions-Hearts from wild-type and transgenic mice were subjected to retrograde aortic perfusion with modified Krebs-Henseleit buffer as described previously (19). After a 30 -40-min stabilization period, cumulative concentrations of isoproterenol (0.1 nM to 1 M) were administered into the buffer flow line at intervals of 7 min.
In Vivo Phosphorylation-Mouse hearts were perfused in a recirculating system containing 2 mCi of [ 32 P]orthophosphate for 30 min. At the end of this labeling period, isoproterenol (0.1 M) was administered into the perfusion system and hearts were stimulated for 2 min (19). After stimulation, preparations were freeze-clamped and homogenized in phosphate buffer (50 mM KH 2 PO 4 , 10 mM NaF, 1 mM EDTA, 0.3 M sucrose, and 0.3 mM phenylmethylsulfonyl fluoride, pH 7.0) containing 0.5 mM dithiothreitol and 1 M okadaic acid. Microsomal fractions enriched in SR membranes and myofibrillar proteins were prepared as described previously (20).
In Vitro Phosphorylation-Cyclic AMP-dependent protein kinase (PKA) phosphorylation of the cardiac homogenates (60 g) was carried out at 30°C in 30 l of reaction mixture containing 50 mM K ϩ phosphate buffer (pH 7.0), 10 mM MgCl 2 , 5 mM NaF, 0.5 mM EGTA, 0.1 mM ATP, 20 Ci of [␥-32 P]ATP, and 45 units of the PKA-catalytic subunit. For endogenous Ca 2ϩ -calmodulin-dependent protein kinase (Ca 2ϩ / CAM) phosphorylation of the cardiac homogenates, 0.5 mM CaCl 2 , 2 M calmodulin, and 1 M protein kinase inhibitor peptide-(5-24) amide were added to the above reaction mixture. Reactions were terminated with 30 l of SDS sample buffer after 2 min (PKA) or 5 min (Ca 2ϩ /CAM) incubation, which was associated with optimal phosphate incorporation in PLB. Thirty g of protein was subjected to 15% SDS-PAGE and autoradiography.
Statistical Analysis-Data are expressed as mean Ϯ S.E. Statistical analysis was performed using Student's t test for unpaired observations and analysis of variance followed by Bonferroni's t test for multiple comparisons. Values of p Ͻ 0.05 were considered statistically significant.

Reintroduction of Wild-type PLB into the PLB Knockout
Mouse Hearts-To determine whether the hyperdynamic cardiac function of the PLB-KO mice can be reversed by reintroduction of the missing PLB gene, we used the ␣-MHC promoter to direct expression of mouse PLB cDNA in the cardiac compartment of the PLB knockout mouse. Four PLB-WT transgenic lines were identified by PCR and Southern blot analyses of mouse genomic DNA. Northern blot analysis of total RNA isolated from the hearts of the PLB-WT transgenic mice revealed expression of the transgene, which migrated at 1.0 kb (1). This 1.0-kb message was not present in either control or PLB-KO mouse hearts. However, the control hearts showed the presence of the endogenous PLB messages migrating at 2.8 and 0.7 kb (data not shown).
To assess the protein levels of PLB in the transgenic hearts, we performed quantitative Western blot analysis of cardiac homogenates from PLB-WT in parallel with wild-type control mice. The levels of PLB were 0.7-fold in one line and 2.0-fold in three transgenic lines compared with control hearts. The line, which expressed PLB levels (0.7-fold) closer to those present in control hearts, was chosen for evaluation of the physiological effects of PLB reintroduction in the knockout background. Sarcoplasmic reticulum membranes isolated from the transgenic hearts indicated that the reintroduced PLB was inserted in the SR and there were no alterations in the SR Ca 2ϩ -ATPase levels compared with control hearts (data not shown).
Reversal of the PLB Knockout Hyperdynamic Cardiac Function by Reinsertion of Wild-type PLB-To determine whether the reintroduced PLB was capable of reversing the hyperdynamic cardiac function associated with PLB deficiency, hearts from PLB-WT and PLB-KO mice were subjected to Langendorff perfusion in parallel with control hearts. The PLB-KO hearts exhibited significantly enhanced myocardial performance compared with wild-type controls, as characterized by significant increases in the maximal rates of cardiac contraction (ϩdP/dt) and relaxation (ϪdP/dt) ( Table I). Reinsertion of PLB in the knockout background (PLB-WT) was associated with significant depression of the contractile parameters (Table I). However, since the levels of reintroduced PLB were 0.7-fold of those present in control hearts, reversal of the enhanced cardiac contractile parameters was not complete, or to the levels observed in control hearts (Table I). It is interesting to note that when the relative levels of PLB or PLB/SR Ca 2ϩ pump in the three animal models were plotted against ϩdP/dt and ϪdP/dt, a close linear correlation was observed (Fig. 1), consistent with our previous observations in PLB-KO, PLB-heterozygous, and control hearts (21).
Generation of Transgenic Mice Expressing Mutant (Ser 3 Ala) PLB in the Heart-The reversal of the hyperdynamic contractile parameters of the PLB-KO hearts by reintroduction of the wild-type PLB demonstrated the feasibility of reinserting various PLB mutants in the knockout background and assessing their functional relevance in vivo. In the present study, mutation of Ser 16 to Ala in PLB was performed by PCR sitedirected mutagenesis, and expression of the mutated PLB was driven by the ␣-MHC promoter in the PLB-KO mice, in a manner identical to that for wild-type PLB described above.

Phospholamban Phosphorylation in Transgenic Mice
Three lines of transgenic mice were generated, which were identified by PCR and Southern blot analyses. The levels of PLB were 0.7-fold in two lines and 2.0-fold in one line, compared with control hearts. One of the lines expressing 0.7-fold PLB, which was similar to the PLB levels expressed in PLB-WT hearts (Fig. 2), was selected for breeding and further studies. Analysis of the SR Ca 2ϩ -ATPase levels showed that there was no alteration upon introduction of mutant PLB in the mouse heart ( Fig. 2A). Langendorff perfusion indicated that the basal contractile parameters, end-diastolic pressure, left ventricular systolic pressure, and heart rate of the PLB-mutant (PLB-MU) hearts were similar to those of the PLB-WT hearts. To determine the functional importance of Ser 16 phosphorylation in the ␤-adrenergic responses, hearts from PLB-KO, PLB-WT, and PLB-MU mice were perfused in a Langendorff mode and subjected to increasing concentrations of isoproterenol. Isoproterenol stimulated ϩdP/dt, ϪdP/dt, and heart rate in PLB-WT, PLB-MU, and PLB-KO hearts in a dose-dependent manner (Fig. 3). Under maximal isoproterenol stimulation, the contractile parameters were similar between the PLB-KO and PLB-WT groups, whereas those of the PLB-MU hearts were much lower (Fig. 3,  A and B). It is interesting to note that, when the stimulatory responses were considered as absolute increments (mmHg/s) over the values observed in non-stimulated hearts, the degrees of stimulation in either the ϩdP/dt or ϪdP/dt values were similar between PLB-KO and PLB-MU hearts. The rate of contraction (ϩdP/dt) increased ϳ1000 mmHg/s in either PLB-MU (3357 Ϯ 193 to 4600 Ϯ 195 mmHg/s, n ϭ 7) or PLB-KO (5616 Ϯ 165 to 6497 Ϯ 143 mmHg/s, n ϭ 6) hearts (Fig. 3A). However, this increase was ϳ3000 mmHg/s in PLB-WT (2782 Ϯ 120 to 5958 Ϯ 206 mmHg/s, n ϭ 7) and in non-transgenic control (3610 Ϯ 140 to 6587 Ϯ 238 mmHg/s, n ϭ 6) hearts. Similarly, the rate of relaxation (ϪdP/dt) increased ϳ2000 mmHg/s in PLB-MU (2334 Ϯ 133 to 4228 Ϯ 191 mmHg/s, n ϭ 7) and PLB-KO (3297 Ϯ 118 to 5439 Ϯ 88 mmHg/s, n ϭ 6) hearts (Fig. 3B), whereas an increase of ϳ3000 mmHg/s was observed in PLB-WT hearts (2136 Ϯ 90 to 5396 Ϯ 322 mmHg/s, n ϭ 7) and in non-transgenic control hearts (1711 Ϯ 116 to 4641 Ϯ 170 mmHg/s, n ϭ 6).
In Vivo 32 P Incorporation in PLB-The observation that the degree of changes in contractile parameters was similar be-tween PLB-MU and PLB-KO hearts during isoproterenol stimulation, suggested that mutation of Ser 16 to Ala in PLB attenuated or abolished its regulatory role in the responses to ␤-agonists. This might be due to lack of phosphorylation of Thr 17 in the mutated PLB form during ␤-adrenergic stimulation. To determine whether Thr 17 was phosphorylated in vivo, PLB-WT and PLB-MU hearts were perfused in parallel with buffer containing [ 32 P]orthophosphate. The hearts were stimulated with 0.1 M isoproterenol as described previously (19). Cardiac myofibrillar and SR-enriched membrane preparations were isolated and subjected to SDS-gel electrophoresis and autoradiography (Fig. 4). Examination of the degree of 32 P labeling of the proteins in the SR-enriched membrane fraction indicated that phosphorylation of PLB was pronounced only in the PLB-WT hearts and incorporation of [ 32 P]phosphate in this protein was barely detectable in the PLB-MU hearts (Fig. 4A). However, the degree of 32 P-incorporation in troponin I and C-protein in the myofibrillar fraction, isolated from the same hearts, was similar between PLB-MU and PLB-WT mice (Fig.  4B). Furthermore, PLB-WT and PLB-MU hearts were perfused in parallel with non-radioactive buffer and stimulated with isoproterenol, as described above. The SR-enriched membrane preparations were subjected to SDS-gel electrophoresis and immunoblotting using antibodies specific to Ser 16 -or Thr 17phosphorylated PLB peptides. Both phosphoserine (Fig. 5A) and phosphothreonine (Fig. 5B) were detected in isoproterenolstimulated PLB-WT hearts. However, in PLB-MU hearts, there was a very low degree of phosphothreonine formation (Fig. 5B) but no phosphoserine detection (Fig. 5A), consistent with mutation of this site to Ala.
In Vitro Phosphorylation of PLB-The lack of phosphorylation of Thr 17 in the perfused PLB-MU hearts could be due to structural alterations in PLB upon mutation of Ser 16 to Ala, rendering the adjacent Thr 17 residue inaccessible to protein kinase. To determine whether this mutation in PLB alters its ability to become phosphorylated by the Ca 2ϩ -calmodulin-dependent protein kinase, cardiac homogenates from PLB-WT and PLB-MU mice were incubated with Ca 2ϩ and calmodulin in the phosphorylation assay buffer and processed for SDSpolyacrylamide gel electrophoresis and autoradiography. The degree of 32 P incorporation in PLB was similar between PLB-WT and PLB-MU hearts (Fig. 6A), suggesting that Thr 17 in PLB-MU hearts could be phosphorylated in vitro. Formation of phosphothreonine in PLB-MU hearts was also verified by immunodetection, using a polyclonal antibody raised to a PLB peptide phosphorylated at Thr 17 (data not shown). These data indicate that mutation of Ser 16 did not prevent phosphorylation of the adjacent Thr 17 residue in PLB by the Ca 2ϩ -calmodulindependent protein kinase. However, incubation of the same cardiac homogenates with the protein kinase A catalytic subunit under optimal phosphorylation conditions indicated that only the PLB in PLB-WT hearts could be phosphorylated, consistent with the lack of the Ser 16 (Ser 16 3 Ala) site in PLB-MU hearts (Fig. 6B). DISCUSSION Our results are the first to demonstrate that a cardiac phenotype, generated by gene targeting, can be reversed by reinsertion of the missing gene in the null background. In previous studies, we showed that ablation of PLB resulted in enhanced basal cardiac contractile parameters assessed at the cellular, organ, and intact animal levels (2,18,22,23). In this study, we used the ␣-myosin heavy chain promoter to direct cardiacspecific expression of wild-type PLB in the knockout background and observed reversal of the hyperdynamic function of the PLB-deficient hearts. The degree of inhibition of the contractile parameters was proportional to the expression levels of PLB, in agreement with our previous studies in PLB-heterozygous and PLB-homozygous mice (2,21). The success of PLB transgenesis in the genetically altered background, accompanied by reversal of the knockout phenotype, indicated that the PLB-deficient mouse provides an attractive model system for expression of various PLB mutants in the heart and elucidation of structure-function relationships in vivo. With the first mutant we studied, we sought to elucidate the functional significance of dual site PLB phosphorylation during ␤-adrenergic stimulation. A site-specific mutation was introduced into the PLB coding region, converting Ser 16 to Ala, and mutant PLB expression was directed in the knockout background. Quantitative immunoblotting of cardiac homogenates and SR-enriched microsomal preparations showed that all the mutant PLB was inserted in the SR membranes. Transgenic mice, expressing similar levels of wild-type or mutant PLB in the heart, exhibited similar contractile parameters under basal conditions, indicating that the mutant PLB was capable of modulating contractility in a manner similar to that for wildtype PLB. However, isoproterenol stimulation was associated with much lower enhancement of the rates of contraction and relaxation in the PLB-mutant hearts compared with PLB-wildtype hearts, whereas the heart rate responses were similar between these groups. It is interesting to note that the maximal increases in contractile parameters of the PLB-mutant hearts were similar to those of PLB-knockout hearts under ␤-adrenergic stimulation. These findings suggest that mutation of Ser 16 in PLB compromised the contribution of this phosphoprotein in the stimulatory responses of the heart to isoproterenol. Actually, when [ 32 P]orthophosphate was included in the perfusate buffer, there was no 32 P labeling of mutant PLB observed, even under maximal isoproterenol stimulation. However, in the same hearts, the degree of phosphorylation of troponin I and C-protein in the myofibrils was similar between hearts expressing mutant or wild-type PLB. Thus, cardiac phosphoproteins other than PLB were responsible for mediating the attenuated responses of the PLB-mutant hearts. To exclude the possibility that the lack of phosphothreonine in vivo was due to mutation of the adjacent Ser 16 to Ala, cardiac homogenates from PLBwild-type and PLB-mutant mice were incubated in vitro under optimal phosphorylation conditions for the Ca 2ϩ -calmodulindependent protein kinase. The degree of mutant PLB phosphorylation was similar to that of wild-type PLB and phosphothreonine formation was verified using the phospholamban phosphorylation site-specific antibody, indicating that the mutant PLB form was capable of being phosphorylated on Thr 17 .
The role of PLB phosphorylation by cAMP-dependent and Ca 2ϩ -calmodulin-dependent protein kinases has been the subject of several studies. Reports by Raeymaekers et al. (7), Tada et al. (8), and Kranias (6) indicated that the stimulatory effects of the two protein kinases on sarcoplasmic reticulum Ca 2ϩ transport can be additive, whereas a study by Colyer and Wang (9) suggested that maximal stimulation of the Ca 2ϩ pump occurs by PLB phosphorylation at a single site and that additional phosphorylation of the other site does not further stimulate pump activity. Furthermore, all the in vitro studies agree that phosphorylation of PLB by cAMP-dependent and Ca 2ϩcalmodulin-dependent protein kinases occurs in an independent manner, whereas in vivo findings (3,13,24) indicate that phosphorylation/dephosphorylation of Thr 17 occurs only subsequent to phosphorylation/dephosphorylation of Ser 16 during ␤-adrenergic stimulation. Our findings in transgenic animals, demonstrated that: (a) Thr 17 in PLB cannot be phosphorylated in the absence of Ser 16 phosphorylation, even under maximal isoproterenol stimulation of intact, beating hearts; and (b) phosphorylation of Thr 17 in PLB does not require prior phosphorylation of Ser 16 in in vitro experiments. Thus, phosphorylation of Thr 17 occurs independently of Ser 16 phosphorylation in vitro, whereas phosphorylation of the adjacent Ser 16 residue appears to be a prerequisite for in vivo phosphothreonine formation during ␤-agonist stimulation. This apparent discrepancy between in vivo and in vitro findings may be due to differences in the levels of calcium available to activate the SR Ca 2ϩ -calmodulin-dependent protein kinase. In vitro conditions generally include optimal calcium concentrations. However, in in vivo studies, phosphorylation of Ser 16 may be required to occur first and enhance the SR Ca 2ϩ uptake rates and, thus, SR Ca 2ϩ load. This would lead to increased Ca 2ϩ levels released by the SR, activation of the Ca 2ϩ -calmodulin-dependent protein kinase, and phosphorylation of Thr 17 in PLB. The phosphorylation and activation of the sarcolemmal Ca 2ϩ channels may also contribute to the increased Ca 2ϩ levels required for in vivo phosphorylation of Thr 17 in PLB (17). However, the inhibition of the SR-associated protein phosphatase 1 activity, which was suggested to be an important determinant for Thr 17 phosphorylation (17), did not appear to play any role in our transgenic experiments, even under maximal isoproterenol stimulation.
The functional significance of Ca 2ϩ -calmodulin-dependent phosphorylation of PLB has been previously examined in intact cardiac myocytes (15,25). Phosphorylation of PLB by Ca 2ϩcalmodulin-dependent protein kinase II was suggested to increase the V max of the SR Ca 2ϩ -ATPase, whereas phosphorylation by protein kinase A increased the Ca 2ϩ affinity of the pump (15). Furthermore, inhibition of Ca 2ϩ -calmodulindependent protein kinase II was shown to slow-down sarcoplasmic reticulum Ca 2ϩ uptake and the decline of [Ca 2ϩ ] even after inhibition of protein kinase A (25), suggesting that the phosphothreonine in PLB may be important in regulation of diastolic Ca 2ϩ and prevention of cytosolic Ca 2ϩ overload especially under pathophysiological conditions (17,26,27).
In summary, our findings indicate that the hyperdynamic PLB knockout phenotype can be reversed by reintroduction of PLB in the null background, and demonstrate the potential power of this technology in performing PLB structure-function studies in vivo. Expression of mutant PLB in which Ser 16 was replaced by Ala in the knockout background indicated that the phosphorylation of Thr 17 in PLB requires prior phosphorylation of Ser 16 during ␤-adrenergic stimulation. In the absence of Ser 16 phosphorylation, the degree of the stimulatory effects by ␤-agonists was similar to that obtained in PLB knockout hearts, suggesting that cardiac phosphoproteins other than PLB mediate these responses. Future studies using transgenic mice harboring the Thr 17 3 Ala or Ser 16 -Thr 17 3 Ala-Ala mutations in PLB will further delineate the interrelationship of dual site phosphorylation in PLB and elucidate the functional relevance of each phosphorylation site under physiological and pathophysiological conditions.