Differential Phospholamban Gene Expression in Murine Cardiac Compartments
Molecular and Physiological Analyses
Abstract Phospholamban, the regulator of the Ca2+ pump in cardiac sarcoplasmic reticulum, is differentially expressed between murine atrial and ventricular muscles. Quantitative analyses of RNA isolated from atrial flaps and ventricular apices indicated that the phospholamban gene transcript copy number is 2.5-fold higher in the ventricle compared with the atrium of the FVB/N mouse and 6-fold higher in the ventricle compared with the atrium of the B6D2/F1 mouse strain. These findings were corroborated by in situ hybridization studies of cardiopulmonary sections from both murine strains, and phospholamban transcripts were also observed in pulmonary myocardia of both strains. Analyses of phospholamban transcript levels relative to α-myosin heavy chain (α-MHC) revealed a 3-fold higher phospholamban abundance in the ventricle compared with the atrium of the FVB/N murine strain. However, the relative mRNA level of Ca2+-ATPase (ratio of sarcoplasmic reticulum Ca2+-ATPase [SERCA2] to α-MHC) in the ventricle was 80% of that in the atrium. Consequently, the relative ratio of phospholamban to SERCA2 mRNA was 4.2-fold lower in the atrium than in the ventricle. The lower transcript ratio of phospholamban to SERCA2 in the atrium was associated with significantly shortened times to half-relaxation (17.40±0.71 milliseconds for atrium versus 30.58±2.04 milliseconds for ventricle), assessed in isolated superfused cardiac tissue preparations recorded at maximum length tension. Contraction times, measured as times to peak tension, were also significantly shortened in atrial muscle (27.36±0.82 milliseconds) compared with ventricular muscle (44.60±2.55 milliseconds), assessed in the same tissue preparations. These findings suggest that phospholamban gene expression is differentially regulated in murine atrial and ventricular muscles and that this differential expression may be associated with differences in the contractile parameters of these cardiac compartments.
Phospholamban is a regulatory SR phosphoprotein, which is physiologically important in modulating cardiac contraction and relaxation.1 This complex phosphoprotein regulates the rate of Ca2+ uptake into cardiac SR. Phosphorylation of phospholamban is associated with activation of the Ca2+ pump,2 3 4 5 and this activation is reversed by dephosphorylation of phospholamban.6 Therefore, the SR Ca2+ pump appears to be under reversible regulation by phosphorylation/dephosphorylation reactions on phospholamban. The phospholamban protein comprises 52 amino acids and is proposed to consist of a hydrophobic α-helical domain (AA, 31 to 52) and a cytoplasmic domain (AA, 1 to 30) containing three distinct sites of phosphorylation by three different protein kinases.7 Two of these sites, serine 16 and threonine 17, have been shown to be phosphorylated in intact beating hearts upon isoproterenol stimulation.8 This phosphorylation of phospholamban is accompanied by an increased rate of SR Ca2+ uptake, which has been proposed to be responsible for the increased rate of myocardial relaxation (−dP/dt) observed during isoproterenol stimulation of the heart.9 10
The significant contribution of phospholamban to the regulation of myocardial contractility and relaxation has recently been established through the development of a phospholamban-deficient mouse by using gene-targeting methodology.11 Hearts from phospholamban-deficient mice exhibited increases in the affinity of the SR Ca2+ pump for Ca2+, and this was associated with significant enhancement of the contractile parameters. As phospholamban has been shown to be a major regulator of myocardial contractility,11 defining the regulation of phospholamban gene expression in the heart becomes an important goal. Studies of phospholamban gene expression are ideally suited to the murine system because the murine phospholamban gene has been isolated and cloned,12 a murine model of phospholamban ablation has been developed,11 and a regulatory role for phospholamban has been established in the intact murine heart.11 12
Although phospholamban is known to be expressed and to be functionally important in the intact mammalian heart, little is known about the relative levels of its expression in different myocardial compartments. Recently, murine atrium was reported to be devoid of phospholamban,13 whereas phospholamban mRNAs were detected in atrial and ventricular muscle in the rabbit.14 Furthermore, the phospholamban gene transcripts were shown to be regulated in a similar manner between atrial and ventricular muscles of the rabbit in response to thyroxine treatment.14 Although the significance of changes in phospholamban mRNA levels is not currently understood, decreases in myocardial phospholamban mRNA levels have been observed during the deterioration of cardiac function in cardiomyopathic disease.15 16 Therefore, further studies are required to delineate the relation between alterations in the levels of phospholamban gene transcripts and alterations in myocardial physiology and pathophysiology.
This report initiates studies on phospholamban gene expression in murine atrium and ventricle. The studies have been designed to test the hypothesis that phospholamban gene transcripts are differentially expressed among these compartments and that differences in phospholamban transcript expression between atrium and ventricle may reflect alterations in contraction and relaxation rates and responses to isoproterenol in these cardiac muscles. Our findings indicate that the phospholamban gene is expressed at both the message and protein levels in murine atrium and ventricle, and in situ analysis reveals the presence of phospholamban transcripts in murine pulmonary myocardium. Analyses of phospholamban transcript levels indicate differential phospholamban gene expression in murine atrial and ventricular muscles, although the abundance of the SERCA2 mRNA appears similar in these compartments relative to α-MHC transcripts. The differences in the relative gene expression levels of phospholamban to SERCA2 transcripts appear to be associated with differences in contraction and relaxation parameters, assessed in isolated superfused atrial and ventricular tissues. These findings indicate that differential regulation of phospholamban gene expression between murine atrium and ventricle may be an important determinant of the contractile properties of these muscles, providing valuable insights into the functional analyses of transgenic animal models.
Materials and Methods
Genomic DNA isolated from FVB/N mice was digested with the restriction endonuclease Pst I, electrophoresed in agarose gel, and blotted onto Gene Screen nylon membrane (NEN Dupont) according to the manufacturer’s protocol. The membrane was probed with a 32P–end-labeled oligonucleotide (see “RNA Dot-Blot Analysis”), antisense to the coding region of the murine phospholamban gene. Hybridized membranes were washed, sealed in plastic, and exposed to x-ray film for autoradiography.
Heart tissues were dissected from 8- to 12-week-old female FVB/N or B6D2/F1 mice. Total RNA was extracted from right and left atrial flaps (auricles), ventricular apices comprising left and right ventricular muscle, and whole hearts. The excised tissues were rinsed in cold (4°C) PBS and pooled for homogenization as follows on the basis of size: atrial flaps (n=20 to 30), ventricular apices (n=4 to 6), and whole hearts (n=2 or 3). Pooled tissues were homogenized by Tissumizer (Tekmar) in guanidinium thiocyanate, and the total RNA was extracted according to the method of Chomczynski and Sacchi.17 RNA was precipitated at −20°C with ethanol and isolated by centrifugation. Purified resuspended RNA was quantified and analyzed for protein contamination by UV spectrophotometry. Aliquots of quantified RNA from each sample, stained with ethidium bromide, were electrophoretically size-fractionated on denaturing agarose gels containing 2.2 mol/L formaldehyde and analyzed for degradation under UV light (data not shown).
RNA extracts were size-fractionated in denaturing agarose by electrophoresis and subsequently blotted onto nylon membranes according to the method outlined by Sambrook et al.18 The membranes were baked at 80°C under vacuum for 2 hours. Baked membranes were prehybridized for 1 to 2 hours and then hybridized overnight at 42°C with a 60-base oligonucleotide, antisense to a portion of the murine phospholamban gene coding region, which was 32P-labeled by using an end-labeling terminal transferase kit (NEN Dupont). After hybridization, membranes were washed at 61°C for 1.5 to 2 hours in a solution of 0.5% SDS and 0.5× SSC. Washed membranes were then sealed in plastic and exposed in cassettes to x-ray film for autoradiography. The amount of RNA blotted per sample was visualized by methylene blue staining of the hybridized membranes.
RNA Dot-Blot Analysis
Three to four separate RNA preparations were analyzed on separate membranes from pooled tissues (right and left atrial flaps, ventricular apices, and whole hearts) to obtain the phospholamban copy numbers. Each RNA preparation was quantified and then divided into three aliquots. The three aliquots were serially diluted and blotted onto the same nylon membrane using a Bio-Rad dot-blot filtration apparatus. The membranes were subsequently baked at 80°C for 2 hours under vacuum. Baked membranes were sequentially prehybridized at room temperature for 1 hour and at 42°C for 2 to 3 hours and then hybridized at 42°C for 15 to 18 hours with excess probe. Blots were hybridized first with a 60-base oligonucleotide, antisense to the murine phospholamban gene coding region, which was end-labeled with 32P by using T4 kinase enzyme and [γ-32P]ATP. Radiolabeling by this method was used for quantification, since it ensured that each molecule of probe carried only a single label. Hybridized blots were washed as described above for Northern blots, sealed in plastic, and exposed to x-ray film and to a PhosphorImager screen (24 hours). In order to normalize for RNA loading onto the blotted membranes, the hybridized blots were subsequently rehybridized by identical methodology with a 32P–end-labeled 60-base oligonucleotide, antisense to a portion of the murine 18S gene. The radiolabeled 18S probe was used as a tracer mixed with excess cold 18S oligonucleotide. 18S hybridized blots were washed, sealed in plastic, and exposed to both autoradiography film and a PhosphorImager screen (10 to 20 minutes). Quantification of phospholamban transcripts was accomplished with phospholamban standard curves generated for each experiment by blotting serially diluted linearized plasmid (pBS SK−) containing a 970-bp fragment of the murine phospholamban gene. The standard curve was generated by quantification of phospholamban hybridization using PhosphorImager analysis, and the relative intensity of each sample was quantified with an ImageQuant computer program (Molecular Dynamics). Linear regression analysis of standard curves was used for quantification of phospholamban transcripts in all sample dilutions, which were corrected for loading with respect to the relative intensity of 18S signal in whole hearts. Phospholamban mRNA copy numbers for atrial versus ventricular muscle within a strain are reported as mean±SEM per microgram RNA normalized for loading to whole-heart 18S and are compared for significance by paired t test.
RNA dot-blot quantifications of atrial and ventricular phospholamban and SERCA2 gene transcripts relative to α-MHC mRNA were also performed. Total RNA was extracted and pooled from 180 atrial flaps and from 30 ventricular apices. Total RNA was also extracted and pooled from eight whole hearts, and this was used in parallel with atrial and ventricular RNA to standardize for loading (see below). Serial dilutions of pooled RNA extracts were blotted, three consecutive times each, onto triplicate membranes. This procedure was performed twice. Thus, a total of six membranes were prepared. Each membrane contained three atrial and three ventricular blots of serially diluted RNA. Each of the six membranes was prepared for hybridization as outlined above and then hybridized with either the 32P–end-labeled phospholamban oligonucleotide, a 32P–end-labeled 60-base oligonucleotide (antisense to the 3′ untranslated region of the murine SERCA2 gene), or a 32P–end-labeled 51-base α-MHC oligonucleotide (antisense to the murine α-MHC coding region). Each blotted atrial and ventricular sample was standardized for membrane loading by using the 18S rRNA hybridization signal from whole-heart samples as an internal standard (described above).
Quantification of hybridized phospholamban, SERCA2, and α-MHC signal was performed by PhosphorImager analysis and expressed relative to the transcript hybridization signal in whole heart. Relative PhosphorImager pixel units were expressed per microgram total RNA for each of the transcripts probed. SERCA2 and phospholamban signal intensity was expressed relative to α-MHC transcript signal intensity as either the ratio of SERCA2 to α-MHC or phospholamban to α-MHC, yielding the relative expression of each transcript to the atrial or ventricular α-MHC transcripts.
In Situ Hybridization
In preparation for sectioning, whole heart/lung samples were excised from 9- to 12-week-old female FVB/N and B6D2/F1 mice, rinsed with PBS, fixed in 4% paraformaldehyde in PBS, cryopreserved in 30% sucrose in PBS, and embedded in O.C.T. Embedding Compound (Miles Diagnostics). Embedded tissues were sectioned by cryostat to 6- to 8-μm thickness, mounted onto prepared slides, permeabilized by proteinase K treatment, and postfixed with 4% paraformaldehyde for hybridization as described by Subramaniam et al.19 All solutions were prepared under RNase-free conditions as outlined by Sambrook et al.18 Adjacently sectioned tissues were hybridized with one of three 35S-labeled riboprobes: an antisense murine α-MHC riboprobe, previously described by Jones et al,20 an antisense murine phospholamban riboprobe (see below), and a sense murine phospholamban riboprobe (see below). Hybridizations were performed according to Cox et al,21 with modifications. Briefly, sections were hybridized overnight at 60°C (55°C for the α-MHC probe) with 105 cpm/μL of riboprobe diluted in an in situ hybridization buffer (NEN Dupont). For all three riboprobes, stringent washes included two posthybridization washes in 1× SSC, 50% formamide, and 10 mmol/L dithiothreitol at 65°C for 30 minutes, with an intervening RNase A treatment. After the washing, mounted sections were serially dehydrated by passage through ethanolic solutions of increasing concentration. Dehydrated slides were prepared for autoradiography as previously described20 21 and subsequently photographed by using the dark-field optics of an Olympus BHTU microscope. After dark-field microscopy, slides were rehydrated and prepared for phase-contrast microscopy by staining with hematoxylin and eosin.
Synthesis of Phospholamban Riboprobes
The phospholamban riboprobe construct was a polymerase chain reaction product, subcloned into pBS SK−, which was generated by using a 970-bp cDNA corresponding to a portion of the murine phospholamban gene as a template and specific primers based on the murine phospholamban sequence.12 The sequence of the phospholamban template corresponding to the antisense phospholamban riboprobe is 5′-TTGCTTCCCC CATGACGGAG TGCTCGGCTT TAAGCTGAGT TGGCATGTTG CAGGTCTGGA GTGGCGGCAC GTCTTCACAG AAGCATCACA ATGATGCAGA TCAGCA-3′. Double-stranded DNA sequence analysis of the phospholamban riboprobe template was performed by the dideoxynucleotide chain terminating method using a Sequenase II kit (USB Corp). By use of selected primers, both strands of the riboprobe construct were sequence-analyzed. After sequence verification, the specificity of the [32P]UTP-labeled riboprobes was demonstrated by their ability to hybridize to single fragments on a Southern blot of murine (FVB/N) genomic DNA digested with different restriction endonucleases (data not shown). The [32P]UTP-labeled riboprobes were also analyzed by electrophoresis on acrylamide gels. After autoradiography, both phospholamban sense and antisense riboprobes yielded single bands of appropriate size (data not shown).
Synthesis of riboprobes from linearized plasmid templates for in situ hybridization was carried out by using [35S]UTP; cold ATP, GTP, and CTP; and T7 RNA polymerase (sense riboprobe) or T3 RNA polymerase (antisense riboprobe) according to the protocol of Melton et al,22 with modifications. The cRNA riboprobe products were purified by extraction with RNAzol (Cinna/Biotecx), precipitated overnight at −20°C in ethanol, and collected by centrifugation. Pelleted cRNA products were reconstituted in 100 mmol/L dithiothreitol and stored at −80°C before use.
Western Blot Analyses
Excised atrial flaps, ventricular apices, and whole hearts from female FVB/N mice were pooled and homogenized according to previously reported protocols.23 Three independent homogenate preparations were analyzed for both atrial and ventricular muscles relative to a common whole-heart homogenate used as a relative standard on each blot. Homogenate protein was quantified spectrophotometrically by the Bio-Rad method, according to manufacturer’s recommended protocol. Because of the number of mice required for the harvest of atrial flaps (≈1 to 2 mg each), only the FVB/N strain was chosen for the protein expression assays, because it had the lowest phospholamban mRNA copy numbers. Serially diluted protein homogenates from atria and ventricles were electrophoresed on 15% polyacrylamide-SDS gels, with dilutions of whole-heart homogenates used as standards for each experiment. Separated proteins were electroblotted onto nitrocellulose membranes for 2 hours at 4°C. For detection of phospholamban, blotted proteins were reacted with a phospholamban monoclonal antibody (U.B.I.) (1:1000 dilution) and visualized with 35S-labeled anti-mouse secondary antibody (Amersham). For detection of Ca2+-ATPase, blotted proteins were reacted with a Ca2+-ATPase polyclonal antibody raised in rabbit2 (1:500 dilution) and were detected by using an 35S-labeled anti-rabbit secondary antibody (Amersham). The hybridized phospholamban or Ca2+-ATPase protein bands were visualized by using either autoradiography or a PhosphorImager screen (Molecular Dynamics).
Contractility Measurements for Atrial and Ventricular Tissues
Left atrial and right ventricular tissues were prepared from anesthetized heparinized (5 IU/g) FVB/N female mice according to the method of Grupp and Grupp.24 Right atria were excluded because of automatic pacemaker activity, and left ventricles were excluded on the basis of the thickness of the ventricular wall. Ventricular muscles containing the ventricular outflow tract were obtained via microdissection of the right ventricular free wall. Since the outflow tract is anatomically organized to eject the right ventricular volume toward the pulmonary artery, the orientation of the muscle fibers is identified for directional placement in the clamp holders. Two atrial or two ventricular tissues from different hearts were mounted in the same bath by using double–electrode clamp holders. For atrial muscle measurements, the whole left atrial appendage was used. This muscle was oriented by fixating points directionally, from the opening, proximal to the mitral valve, to the apex of the atrial appendage. Great care was taken so that the orientation of the tissues was identical between experiments. One end of each strip was clamped firmly on top of two punctate electrodes, used to stimulate the strip, while the other end of each strip was connected to the mechanical arm of a force transducer via monofilament suture. The resting tension for each tissue was set by use of a micrometer, and after a short equilibration at low resting tension, a length-tension curve was produced with each muscle. Final equilibration was achieved close to Lmax, and in the present study, all muscles were normalized for loading conditions. Normalization according to this method ensures identical loading for the muscles, irrespective of cross-sectional area.24 This resting tension was maintained throughout the experiment. Tissues mounted in baths were superfused with oxygenated warm (35°C) Krebs-Henseleit solution24 containing (mmol/L) NaCl 118, KCl 4.7, CaCl 2.5, MgSO4 1.2, KH2PO4 1.2, NaHCO3 25, and glucose 11 (pH 7.4). The muscles were stimulated with isoproterenol, which was added to the baths in a cumulative fashion over a concentration range of 10−9 to 10−5 mol/L. For each tissue, resting tension, developed force, +dF/dt, and −dF/dt were recorded. Measurements of TPT and RT1/2 were made for each tissue by using a digital micrometer. The RT1/2 and TPT values were expressed in milliseconds (mean±SEM).
The B6D2/F1 mice used in the present study were obtained from the Jackson Laboratory, Bar Harbor, Me. FVB/N mice were obtained from Harlan Laboratories, Madison, Wis. pBS SK− was obtained from Stratagene Inc. [γ-32P]ATP, [α-32P]dATP, [α-32P]UTP, and [α-35S]UTP were purchased from New England Nuclear. T4 kinase was purchased from Promega. Oligonucleotides used as probes and primers were synthesized at the University of Cincinnati. The 35S-labeled anti-mouse secondary antibody and 35S-labeled anti-rabbit secondary antibody were obtained from Amersham.
Quantitative Analyses of Steady State Phospholamban Gene Transcript Levels in Atrial and Ventricular Muscles
Two major forms of phospholamban mRNA have previously been detected in FVB/N mouse hearts, corresponding to 0.7 and 2.8 kb.12 To determine whether both of these messages are also present in the adult hearts of the B6D2/F1 mouse strain, Northern blot analysis was performed on total cardiac RNA extracts (Fig 1⇓, left and middle). Northern blots were probed with a 32P-labeled 60-base oligonucleotide, antisense to the murine phospholamban gene coding region. This probe recognized the presence of both major phospholamban messages in 20-μg size-fractionated cardiac RNA isolated from both FVB/N and B6D2/F1 murine strains (Fig 1⇓, left). The amount of RNA loaded per lane is depicted by methylene blue staining of the 18S and 28S rRNA species (Fig 1⇓, middle). On Northern blots, there was no apparent difference between strains as to the levels of expression for each phospholamban message when compared with the relative amounts of total RNA for each sample depicted by methylene blue staining (Fig 1⇓, middle).
Steady state phospholamban transcript copy numbers were quantified for atrial and ventricular muscles of both FVB/N and B6D2/F1 mouse strains by RNA dot-blot analyses. The 60-base oligonucleotide, which was shown to recognize both phospholamban transcripts (Fig 1⇑, left), was chosen as a probe for RNA dot blots. Before its use, the specificity of this oligonucleotide probe was also demonstrated by using Southern blots (Fig 1⇑, right). Southern analysis of digested murine genomic DNA, probed with the 60-base oligonucleotide, demonstrated recognition of only a single appropriately sized DNA band, containing the phospholamban gene (Fig 1⇑, right). In previous Southern analysis of murine genomic DNA, restricted with the endonuclease Pst I, a single band of 5.5 kb was detected by using randomly primed cDNA corresponding to the murine phospholamban coding region as a probe (data not shown). The fact that the end-labeled 60-base oligonucleotide detects an identical band in the Southern analysis of murine genomic DNA demonstrates that this probe is highly specific for phospholamban. Therefore, the characteristics of this oligonucleotide probe in both Northern and Southern analyses demonstrated its efficacy as a probe for dot-blot analysis, where the RNA/DNA samples are not separated by size fractioning.
Pooled RNA extracts from atrial flaps and ventricular apices were blotted in triplicate onto the same membrane and hybridized with the 60-base oligonucleotide probe described above. A representative blot containing RNA from the FVB/N mouse is shown in Fig 2⇓, left. Total pooled RNA extracts from whole hearts were also blotted in triplicate onto the same membrane and served as standards for the 18S rRNA signal, which was used to normalize each sample for RNA loading onto the membranes (Fig 2⇓, right). In addition, each blot contained several controls: (1) tRNA blotted at the same concentration as the cardiac RNA samples, serving as a negative control; (2) the diluent (SSC) blotted as a blank, demonstrating that there was no RNA carryover or contamination of the diluent used in serial dilutions; and (3) serial dilutions of linearized plasmid containing a portion of the murine phospholamban gene, generating a standard curve for quantifying phospholamban mRNA levels. The plasmid dilutions also served as a positive control for phospholamban oligonucleotide binding and as a negative control for binding of the 18S antisense oligonucleotide. Furthermore, Northern analysis of fractionated total murine cardiac RNA, probed with the antisense 18S oligonucleotide, revealed only one band, demonstrating the specificity of this probe (data not shown).
Standard curves were generated from each dot blot, and those having reliability coefficients of r>.97 were used to quantify the phospholamban transcript numbers. Copy numbers were calculated by linear regression analyses using triplicate determinations of three or four separate RNA preparations from each tissue. Each of the RNA preparations consisted of multiple pooled tissue samples (see “Materials and Methods”).
Our results indicate that phospholamban expression is significantly lower in the atrial flaps than in the ventricular apices of both mouse strains. In the FVB/N mouse, there was a 2.5-fold difference in phospholamban transcript abundance between these two cardiac compartments: 12±2 million molecules per microgram atrial RNA versus 30±2 million molecules per microgram ventricular RNA (P<.002). In the B6D2/F1 mouse, there is a sixfold difference between atrial flaps and ventricular apices: 15±4 million molecules per microgram atrial RNA versus 90±12 million molecules per microgram ventricular RNA.
To determine whether the reduced phospholamban gene expression in mouse atria could be an indicator of reduced SR gene expression, we analyzed the transcript levels of phospholamban and SERCA2 relative to α-MHC in this tissue. Only the FVB/N mouse strain was used in these studies, since a large number of animals (90 mice) was required for the atrial determinations alone and since the FVB/N mouse represents the strain of choice for the generation of transgenic animal models. For these studies, pooled RNAs extracted from 180 atrial flaps, 30 ventricular apices, and 8 whole hearts were used for triplicate determinations on each of six blots. Our data indicate that the relative ratio of phospholamban to α-MHC was 3.2-fold higher in ventricular apices than in atrial flaps (Table 1⇓). This ratio is similar to the ratio of ventricular to atrial phospholamban transcript copy number, which is described above for the FVB/N strain. However, the relative ratio of SERCA2 to α-MHC was 1.53 in the atrial flaps and 1.17 in the ventricular apices (Table 1⇓). From these ratios, the relative ratio of phospholamban to SERCA2 transcripts was obtained. This ratio appeared to be 4.2-fold higher in the ventricular apices than in atrial flaps (Table 1⇓).
Expression of Phospholamban Gene Transcripts in Intact Myocardial Compartments Visualized by In Situ Hybridization
The differential phospholamban expression in atrial and ventricular muscles, observed in our in vitro studies, prompted us to extend our studies to the in situ level. Intact, embedded heart/lung samples were sectioned and hybridized for both the FVB/N (Fig 3⇓) and B6D2/F1 (Fig 4⇓) mouse strains. Serial sections from each tissue were hybridized with the phospholamban antisense riboprobe, the phospholamban sense riboprobe, or the murine α-MHC antisense riboprobe (Figs 3⇓ and 4⇓). In situ hybridization studies performed in both the FVB/N mouse and the B6D2/F1 mouse strains demonstrated different intensities of the phospholamban antisense riboprobe hybridization signal between atrium and ventricle (Figs 3A⇓ and 4A⇓). Phospholamban mRNA expression appeared to be higher in the intact ventricle than in the atrium of the FVB/N strain, in agreement with the quantitative analyses described above, whereas the phospholamban mRNA expression in the B6D2/F1 ventricle appeared to have an even greater hybridization signal intensity, consistent with the dot-blot analyses. The atrial hybridization signal, although much lower than that in ventricular muscle, was well above the hybridization signal detected in the lung parenchymal tissue in each strain (Figs 3A⇓ and 4A⇓). The degree of nonspecific hybridization in these tissues for both strains was evaluated by using a phospholamban sense riboprobe (Figs 3B⇓ and 4B⇓).
The presence of phospholamban mRNA in the pulmonary myocardium of the FVB/N mouse is shown in Fig 3A⇑. The hybridization signal indicating phospholamban mRNA expression in this tissue appears similar to that observed in the atrium (Fig 3A⇑). Furthermore, this signal is above the levels detected in pulmonary parenchyma by the antisense riboprobe (Fig 3A⇑) and in pulmonary myocardium by the sense riboprobe (Fig 3B⇑), which depicts nonspecific binding. Likewise, the signal intensity from the hybridized phospholamban gene transcripts in the pulmonary myocardium is similar to that observed in the atrium of the B6D2/F1 heart (Fig 4A⇑). The similarity in phospholamban expression levels between atrial and pulmonary myocardial muscles in both strains is consistent with the hypothesis that pulmonary myocardium is embryologically derived from atrial muscle.20 Support for this hypothesis is also provided by our observations that the riboprobe signal intensity for α-MHC is greater than for phospholamban in the pulmonary myocardium (Figs 3A⇑, 4A⇑, 3C⇑, and 4C⇑), reflective of the atrial signal intensity for these two riboprobes.
In contrast to the differential expression of phospholamban gene transcripts in myocardial compartments of both mouse strains, the expression of α-MHC gene transcripts in these compartments appeared similar. This is demonstrated by hybridization of atrial, ventricular, and pulmonary myocardial muscles with the α-MHC antisense riboprobe (Figs 3C⇑ and 4C⇑). There is essentially no difference between the levels of expression of α-MHC mRNA between these three muscle types, which are all well above the nonspecific hybridization signal observed in the lung parenchyma (Figs 3C⇑ and 4C⇑), a tissue that is devoid of α-MHC protein.19 20 We have previously shown that hybridization of murine cardiac and pulmonary tissues with a sense riboprobe, complementary to the α-MHC riboprobe used in this study, demonstrated low nonspecific hybridization.20 To demonstrate that the tissue was preserved during the hybridization and washing procedures, the contractile elements of the myocardium were visualized after hematoxylin and eosin staining (Figs 3D⇑ and 4D⇑).
Expression of Phospholamban Protein in Atrium
Western analyses of cardiac protein homogenates were performed to demonstrate that the phospholamban mRNA was translated to protein in murine atria. Only the FVB/N mouse strain was used in these studies, since a large number of animals was required for atrial determinations and since this represents the mouse strain of choice for the production of transgenic animal models. A whole cardiac homogenate was prepared from pooled intact hearts and used in parallel with atrial and ventricular homogenates on each blot. Fig 5⇓ shows representative results obtained with three different homogenates prepared from pooled tissues. Our data indicate that the phospholamban protein could be detected in the atrium (Fig 5A⇓), although its levels appeared lower than in the ventricle (Fig 5B⇓), consistent with the mRNA data (Fig 2⇑ and Table 1⇑).
In a similar manner, Western blot analyses were performed on cardiac homogenates, demonstrating Ca2+-ATPase protein expression in atrial and ventricular muscles. The autoradiograms depicted in Fig 6⇓ show representative results from Western blot experiments, which were performed in triplicate by using three different protein homogenates prepared from pooled tissues. These data indicate that the Ca2+-ATPase protein expression levels are similar between murine atrium (Fig 6A⇓) and ventricle (Fig 6B⇓).
Contractility Measurements of Atrial and Ventricular Muscle Tissues
Since the phospholamban gene appears to be differentially expressed in the murine atrium compared with the ventricle and since phospholamban has recently been shown to be an important regulator of basal myocardial contractility,11 it was of interest to assess the contractile parameters of isolated murine atrial and ventricular muscles. For each tissue, resting tension, developed force, +dF/dt, and −dF/dt were recorded (Fig 7⇓). At the beginning of each experiment, length-tension relationships were established. At Lmax, +dF/dt and −dF/dt were measured for superfused atrial and ventricular muscle tissues. Mean±SEM of +dF/dt values are reported as (positive) milligrams per second (Table 2⇓), and mean±SEM of −dF/dt values are reported as (negative) milligrams per second (Table 2⇓). The average +dF/dt value for atrial muscle is significantly faster than that for ventricular muscle (P=.014, Table 2⇓), and the average −dF/dt value for atrial muscle is also significantly faster than that for ventricular muscle (P=.009, Table 2⇓).
Contraction and relaxation times were established by averaging triplicate determinations of TPT (Fig 8A⇓) and RT1/2 (Fig 8B⇓). For atrial muscle, the average TPT at Lmax was 27.36±0.82 milliseconds (mean±SEM, n=8) compared with an average TPT of 44.60±2.55 milliseconds for ventricular muscle (n=6). For these preparations, the atrial muscles exhibited significantly shorter TPT averages than did ventricular muscles (P=.0000, df=12) (Fig 8A⇓). In addition, the average RT1/2 values at Lmax differ significantly (P=.0000, df=12), demonstrating that the atrial muscles (RT1/2, 17.40±0.71 milliseconds [mean±SEM]) relax more quickly than do the ventricular muscle tissues (RT1/2, 30.58±2.04 milliseconds) (Fig 8B⇓).
Atrial and ventricular muscle tissues were also stimulated with isoproterenol (Fig 8⇑), which was added to the superfusion baths in a cumulative fashion over a concentration range of 10−9 to 10−5 mol/L. Dose-response relations showing the effects of isoproterenol on the rates of contraction and relaxation revealed that atrial muscle was more sensitive (ED50, 3.2×10−8 mol/L) than ventricular muscle (ED50, 2×10−7 mol/L) to isoproterenol stimulation. At maximal isoproterenol stimulation, the atrial TPT decreased by 15% compared with the respective nonstimulated level at Lmax (Fig 8A⇑), decreasing from 27.36±0.82 milliseconds to an average TPT of 23.31±0.66 milliseconds (P=.002, df=14). In comparison, the average ventricular TPT shortened by 28% at maximal isoproterenol stimulation compared with the respective average TPT value at Lmax (Fig 8A⇑), decreasing from 44.60±2.55 to 32.00±1.41 milliseconds (P=.0015, df=14). At maximal isoproterenol stimulation, the atrial muscles exhibited a significantly faster average TPT compared with the ventricular muscles (P=.0001, df=10); however, the extent of the isoproterenol-stimulated decrease in time was greater for the ventricular muscles than for the atrial muscles (28% versus 15%). The average atrial RT1/2 was also significantly faster than the average ventricular RT1/2, when this parameter was decreased in response to isoproterenol stimulation (P=.0000, df=12). At maximal isoproterenol stimulation, the atrial RT1/2 decreased by 26%, from 17.40±0.71 to 12.86±0.22 milliseconds (P=.0000, df=14) (Fig 8B⇑). In contrast, at maximal isoproterenol stimulation, the average ventricular RT1/2 decreased by 39%, from 30.58±2.04 to 18.57±0.96 milliseconds (P=.0003, df=10) (Fig 8B⇑).
The maximal response to isoproterenol for each strip was also evaluated by measuring the increases in +dF/dt and −dF/dt. The percent increase in +dF/dt at maximal isoproterenol response was 231±17% (n=8, P<.05) for atrial muscle and 318±68% (n=5, P<.05) for ventricular muscle. −dF/dt increased by 260±19% (P<.05) for atrial muscle and 368±92% (P<.05) for ventricular muscle at maximal isoproterenol stimulation.
In the present study, we report that phospholamban expression is significantly lower in murine atrial flaps than in ventricular apices and that the reduced phospholamban expression may be associated with enhancement of the basal contractile parameters of atrial muscle compared with ventricular muscle. This is the first report on the quantification of phospholamban expression in cardiac compartments and on the assessment of physiological parameters in murine atrial and ventricular muscles. Furthermore, the present study shows that although phospholamban, the regulator of the SR Ca2+ pump, is differentially expressed in atrial and ventricular muscles, expression of the SR Ca2+ pump is similar between these cardiac compartments. The differential phospholamban expression between atrial and ventricular muscles may be due to several factors, including (1) tissue-specific expression of trans-acting factors, which regulate phospholamban gene transcription; (2) differential phospholamban mRNA processing, based on 3′ untranslated region signaling; and (3) differential phospholamban mRNA stability between atrial and ventricular tissues. Currently, little is known about regulation of phospholamban gene transcription in the mouse. In the rat, a highly conserved upstream region important in the regulation of basal transcriptional activity has been identified in the phospholamban gene,25 and in the chicken, a portion of the 5′ region of the phospholamban gene has been suggested to be important in regulating phospholamban gene expression.26
Our in vitro analyses of steady state cardiac phospholamban gene transcripts were performed in two strains of mice, the B6D2/F1 and the FVB/N strains. Investigation of atrial phospholamban transcription levels in two distinct strains was important, since a previous report indicated that the murine atrium was devoid of phospholamban.13 In initial studies, we used the B6D2/F1 strain, because the AT-1 cell line derived from transgenic B6D2/F1 atrium was shown to lack phospholamban expression. We observed phospholamban transcripts in the B6D2/F1 mouse atrium, and in order to demonstrate that this did not represent an isolated strain-related phenomenon, we extended our studies to include the FVB/N strain, which is the murine strain of choice for the generation of transgenic mice. We observed phospholamban transcripts in the FVB/N murine atrium as well, and the reasons for the apparent discrepancies between our findings and previous reports13 are not presently known.
It is significant to note that by both in vitro and in situ analyses, phospholamban mRNA was detected at relatively high levels in the atria and pulmonary myocardia of both the B6D2/F1 and FVB/N murine strains. The functional significance of phospholamban gene expression in murine pulmonary myocardium is not currently understood. This tissue exists as a striated muscle layer situated between the tunica media of the pulmonary vein and the venous luminal epithelium.27 28 Recently, it has been demonstrated that the pulmonary myocardium consists of atrial cardiac muscle, which extends into the pulmonary veins.20 This demonstration is based on atrial-specific gene expression in the murine pulmonary myocardium, which includes the expression of α-MHC and atrial natriuretic factor.20 Our in situ analyses suggest that phospholamban is transcribed in both atrium and pulmonary myocardium at similar levels, further supporting this theory.
The function of the pulmonary myocardium as it relates to cardiopulmonary physiology and pathophysiology is not understood. However, one hypothesis is that this muscle serves as a pulmonary venous valve.20 According to this hypothesis, the pulmonary myocardium may have a role in regulating left atrial preload. It may also have a role in the pathophysiology of pulmonary hypertension.20 Interestingly, the presence of phospholamban in this tissue leads to the speculation that the contractile state of the pulmonary myocardium is in some way regulated by phosphorylation reactions and that it may exhibit sensitivity to adrenergic stimulation. This hypothesis is supported by the findings of MacLeod and Hunter,29 which demonstrated sensitivity of the rat pulmonary vein to catecholamines such as isoproterenol, epinephrine, and norepinephrine. Electron microscopic studies provide additional support for this hypothesis, demonstrating the location of adrenergic axons proximal to the cardiac muscle of the pulmonary veins in mice.30
The differential expression of phospholamban in murine atrial and ventricular muscles prompted us to examine the expression levels of the SR Ca2+ pump relative to α-MHC in these compartments. The ratio of SR Ca2+ pump to α-MHC transcripts was only slightly higher in the atrium relative to the ventricle. However, in the same preparations, the ratio of phospholamban to α-MHC transcripts was ≈3-fold higher in the ventricle than in the atrium, in close agreement with the 2.5-fold difference in phospholamban transcript copy numbers between these compartments, as reported in the present study. These findings indicate that the relative transcript abundance of phospholamban to the SR Ca2+ pump is 4-5 fold higher in ventricular than in atrial muscle. Thus, the genes encoding for the two SR proteins, the pump and its regulator, do not appear to be coordinately regulated in the mouse, in agreement with previous observations in other animal species.14 23 The higher relative ratio of phospholamban to the SR Ca2+ pump in ventricular muscle may reflect a greater number of the SR Ca2+ pumps in the inhibited state, resulting in lower contractile parameters.
To examine the physiological significance of the higher relative ratio of phospholamban to SR Ca2+ pump in ventricle compared with atrium, we studied the contractile parameters of these muscles in isolated tissue preparations during stimulated isometric contraction. The ability to measure contraction and relaxation parameters in murine cardiac tissue preparations enabled us to make functional comparisons of tension development between the atrium and the ventricle in this animal species. The atrial and ventricular tissues differed significantly in their contraction and relaxation parameters, expressed as rates of force development (+dF/dt and −dF/dt), TPT, and RT1/2. At Lmax and under equivalent resting tension and developed force, the atrial tissues exhibited significantly shorter TPT and RT1/2 parameters compared with ventricular tissues. Because the resting tension is analogous to preload in the intact myocardium, our data indicate that at a similar preload and at similar developed tensions, atrial +dF/dt values were faster than ventricular values. Likewise, at similar preloads and developed tensions, atrial −dF/dt values were faster than ventricular values. Thus, it is interesting to propose that reduced phospholamban expression in the atrial muscle would be associated with a reduced number of the SR Ca2+ pumps in the inhibited state, resulting in increases in the affinity of the SR Ca2+-ATPase for Ca2+ and enhanced myocardial −dF/dt. Furthermore, since more Ca2+ would be accumulated in the SR per unit time, more Ca2+ would be available for subsequent contractions leading to increased +dF/dt. These findings on enhanced contractile parameters, associated with reduced phospholamban expression levels, are consistent with our recent observations in a phospholamban knockout mouse.11 31 However, differences in other proteins involved in the excitation-contraction pathway may also contribute to the observed contractile differences in murine atrial and ventricular muscles. In addition to differences in proteins involved in the excitation-contraction pathway, there are also significant differences in the electrophysiology of atrial and ventricular muscles. The action potentials of atrial muscle cells are shorter in duration (faster) than those of ventricular cells. Similarly, abbreviated Ca2+ currents were exhibited in atrial compared with ventricular myocytes. Such factors may also play a role in the differences observed in the contractile parameters for atrial and ventricular muscles.
Phospholamban has also been implicated as an important phosphoprotein in mediating the isoproterenol responses in the intact heart.11 31 Thus, it was of special interest to subject mouse atrial and ventricular strips to isoproterenol and to assess the responses of their contractile parameters. Maximal isoproterenol stimulation was associated with significant increases in contraction and relaxation parameters of both atrial and ventricular muscle strips, although the atrial muscle appeared more sensitive than the ventricular muscle to isoproterenol. However, the present data also demonstrate that isoproterenol is more efficacious in mouse ventricular muscles compared with atria. This may be due to the fact that the ventricular muscles, which have threefold more phospholamban than atrial muscles, maintain more phospholamban, which can be phosphorylated. Therefore, the ventricles may simply require more isoproterenol in order to phosphorylate all the SR phospholamban, relieving its inhibitory effects over the Ca2+-ATPase enzyme. Although the differences in the isoproterenol responses of atria and ventricles could be due to differences in the degree of phosphorylation of phospholamban, troponin I,32 and the 15-kD sarcolemmal protein33 or even differences in the β-receptor signal transduction pathway, our findings are consistent with the hypothesis that phospholamban has a major role in the isoproterenol responses of cardiac muscle.11 31 However, the precise role of phospholamban in the responses of these cardiac compartments to β-agonist stimulation will be further elucidated by using transgenic animal models with altered phospholamban expression levels.
In summary, the present data provide the first evidence of differential phospholamban expression in murine atrial and ventricular myocardial muscles and provide a foundation for future studies designed to identify cis-acting and trans-acting elements responsible for this differential regulation and their role in heart disease. The differences in phospholamban expression between murine atrium and ventricle appeared to reflect alterations in the contractile parameters of these muscles. Although the size of the murine heart is very small, we were able to accurately record the contractile parameters and their alterations by β-adrenergic stimulation in isolated atrial and ventricular muscle tissue preparations. Thus, studies with genetically altered or diseased mouse atria and ventricles are now feasible and will provide us with a better understanding of the physiological and pathophysiological role of phospholamban in these cardiac compartments.
Selected Abbreviations and Acronyms
|α-MHC||=||α-myosin heavy chain|
|+dF/dt||=||rate of contraction|
|−dF/dt||=||rate of relaxation|
|pBS SK−||=||pBluescript plasmid|
|RT1/2||=||time to half-relaxation|
|SSC||=||standard saline citrate|
|TPT||=||time to peak tension|
This study was supported by a grant from the American Heart Association, Ohio Affiliate, Inc (AHA-SW9331-F), and National Institutes of Health grants HL-08901-01 (Dr Koss) and HL-26057 and HL-22619 (Dr Kranias). The authors would like to thank Drs Gunther Grupp and Alejandro Sanchez for helpful discussions and technical advice pertaining to some of the experimental techniques used in this study. We also thank Jay Slack for his kind gift of the mouse SERCA2 oligonucleotide. The authors would also like to thank Gilbert Newman and Mehtap Tosun for excellent technical assistance.
- Received November 22, 1994.
- Accepted April 24, 1995.
- © 1995 American Heart Association, Inc.
Kim HW, Steenaart NAE, Ferguson DG, Kranias EG. Functional reconstitution of the cardiac sarcoplasmic reticulum Ca2+-ATPase with phospholamban in phospholipid vesicles. J Biol Chem. 1990;265:1702-1709.
Kirchberger MA, Tada M, Katz AM. Adenosine 3′:5′:monophosphate dependent protein kinase-catalyzed phosphorylation reaction and its relationship to calcium-transport in cardiac sarcoplasmic reticulum. J Biol Chem. 1974;249:6166-6173.
Kranias EG. Regulation of Ca2+ transport by phosphoprotein phosphatase activity associated with sarcoplasmic reticulum. J Biol Chem. 1985;260:11006-11010.
Simmerman HKB, Collins JH, Theibert JL, Wegener AD, Jones LR. Sequence analysis of phospholamban: identification of phosphorylation sites and two structural domains. J Biol Chem. 1986;261:13333-13341.
Wegener AD, Simmerman HKB, Lindemann JP, Jones LR. Phospholamban phosphorylation in intact ventricles: phosphorylation of serine 16 and threonine 17 in response to β-adrenergic stimulation. J Biol Chem. 1989;264:11468-11474.
Lindeman JP, Jones LR, Hathaway DR, Henry BG, Watanabe AM. β-Adrenergic stimulation of phospholamban phosphorylation and Ca2+-ATPase activity in guinea pig ventricles. J Biol Chem. 1983;258:464-471.
Kranias EG, Garvey JL, Srivastava RD, Soloro RJ. Phosphorylation and functional modifications of sarcoplasmic reticulum and myofibrils in isolated rabbit hearts stimulated with isoprenaline. Biochem J. 1985;226:113-121.
Luo W, Grupp IL, Harrer J, Ponniah S, Grupp G, Duffy JJ, Doetschman T, Kranias EG. Targeted ablation of the phospholamban gene is associated with markedly enhanced myocardial contractility and loss of β-agonist stimulation. Circ Res. 1994;75:401-409.
Ganim JR, Luo W, Ponniah S, Grupp I, Kim HW, Ferguson DG, Kadambi V, Neumann JC, Doetschman T, Kranias EG. Mouse phospholamban gene expression during development in vivo and in vitro. Circ Res. 1992;71:1021-1030.
Arai M, Otsu K, MacLennan DH, Alpert NR, Periasamy M. Effect of thyroid hormone on the expression of mRNA encoding sarcoplasmic reticulum proteins. Circ Res. 1991;69:266-276.
Feldman AM, Ray PE, Silan CM, Mercer JA, Minrobe W, Bristow MR. Selective gene expression in failing human heart. Circulation. 1991;83:1866-1872.
Arai M, Alpert NR, MacLennan DH, Barton P, Periasamy M. Alterations in sarcoplasmic reticulum gene expression in human heart failure: a possible mechanism for alterations in systolic and diastolic properties of the failing myocardium. Circ Res. 1993;72:463-469.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning. New York, NY: Cold Spring Harbor Laboratory Press; 1989.
Subramaniam A, Jones WK, Gulick J, Wert S, Neumann J, Robbins J. Tissue-specific regulation of the β-myosin heavy chain gene promoter in transgenic mice. J Biol Chem. 1991;266:24613-24620.
Melton DA, Krieg PA, Redbagliati MR, Maniatis T, Zink K, Green MR. Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 1984;12:7035-7056.
Kiss E, Jakab G, Kranias EG, Edes I. Thyroid hormone–induced alterations in phospholamban protein expression: regulatory effects on sarcoplasmic reticulum Ca2+ transport and myocardial relaxation. Circ Res. 1994;75:245-251.
Grupp IL, Grupp G. Isolated heart preparations perfused or superfused with balanced salt solutions. Methods Pharmacol. 1984;5:111-128.
Toyofuku T, Zak R. Characterization of cDNA and genomic sequences encoding a chicken phospholamban. J Biol Chem. 1991;266:5375-5383.
Springall DR, Rhatnagar M, Wharton J, Hamid Q, Gulbenkian S, Hedges M, Meleagros L, Bloom SR, Polak JM. Expression of the atrial natriuretic peptide in the cardiac muscle of rat extrapulmonary and intrapulmonary veins. Thorax. 1988;43:44-52.
Lyons G, Schiaffino S, Sasson D, Barton P, Buckingham M. Developmental regulation of myosin gene expression in mouse cardiac muscle. J Cell Biol. 1990;111:2427-2436.
MacLeod DP, Hunter EG. The pharmacology of the cardiac muscle of the great veins of the rat. Can J Physiol Pharmacol. 1966;45:463-472.
Luo W, Kiss E, Koss KL, Grupp IL, Harrer J, Edes I, Jones WK, Kranias EG. Cardiac remodeling by alterations in phospholamban protein levels. In: Dhalla NS, Pierce GN, Panagia V, Blamish RE, eds. Heart Hypertrophy and Failure. Boston, Mass: Kluwer Academic Publishers. In press.
Garvey J, Kranias EG, Solaro RJ. Phosphorylation of C-protein, troponin I, and phospholamban in isolated rabbit hearts. Biochem J. 1988;245:709-714.
Presti CF, Jones LR, Lindemann JP. Isoproterenol-induced phosphorylation of a 15-kilodalton sarcolemmal protein in intact myocardium. J Biol Chem. 1991;266:11126-11130.