THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 280, No. 5, Issue of February 4, pp. 3500 –3506, 2005
Printed in U.S.A.
Impaired Ca2ⴙ Store Functions in Skeletal and Cardiac Muscle Cells
from Sarcalumenin-deficient Mice*
Received for publication, June 14, 2004, and in revised form, November 15, 2004
Published, JBC Papers in Press, November 29, 2004, DOI 10.1074/jbc.M406618200
Morikatsu Yoshida‡, Susumu Minamisawa§, Miei Shimura¶, Shinji Komazaki储, Hideaki Kume‡,
Miao Zhang‡, Kiyoyuki Matsumura‡, Miyuki Nishi‡, Minori Saito**, Yasutake Saeki‡‡,
Yoshihiro Ishikawa§, Teruyuki Yanagisawa**, and Hiroshi Takeshima‡§§
From the Departments of ‡Medical Chemistry and **Molecular Pharmacology, Tohoku University Graduate School of
Medicine, Sendai, Miyagi 980-8575, Japan, the Departments of §Physiology and ¶Internal Medicine, Yokohama City
University School of Medicine, Yokohama, Kanagawa 236-0004, Japan, the 储Department of Anatomy, Saitama Medical
School, Moroyama-machi, Saitama 350-0495, Japan, and the ‡‡Department of Physiology, Tsurumi University School of
Dental Medicine, Yokohama, Kanagawa 230-8501, Japan
Excitation-contraction coupling requires the physiological
role of the sarcoplasmic reticulum (SR)1 in muscle cells (1); SR
Ca2⫹ release triggers contraction, and relaxation is achieved by
Ca2⫹ uptake into the SR through Ca2⫹-Mg2⫹ ATPase (SERCA).
Therefore, Ca2⫹ sequestration in the SR is essential in muscle
contraction. Several Ca2⫹-binding proteins have been characterized in the SR and proposed to maintain luminal Ca2⫹ level.
Calsequestrin (CSQ) is a major Ca2⫹-binding protein in the
striated muscle SR and shows high capacity and low affinity
Ca2⫹ binding properties (2). CSQ is associated with the SR
terminal cisternae near the Ca2⫹ release site and likely modulates the opening of Ca2⫹ release channels/ryanodine receptors on the SR and store-operated Ca2⫹ channels on the plasma
membrane by regulating the luminal Ca2⫹ level (3, 4). Because
of its abundant expression in striated muscle, CSQ is thought
to greatly contribute to Ca2⫹-buffering effects in the SR. On the
other hand, sarcalumenin (SAR) and histidine-rich Ca2⫹-binding protein are also Ca2⫹-binding proteins in the SR, but their
functions are as yet unknown.
SAR isoforms are generated as 160- and 53-kDa glycoproteins by the alternative splicing of the primary transcript derived from the same gene and are specifically expressed in
skeletal and cardiac muscle cells (5). The expression levels of
SAR isoforms remarkably increase during muscle development, suggesting the involvement of the isoforms in mature SR
functions (6). The amino-terminal half of the 160-kDa isoform
is characterized by the juxtapositions of negatively charged
residues and forms a Ca2⫹-binding region (5). Indeed, the 160kDa SAR isoform binds ⬃35 Ca ions/molecule with a dissociation constant of ⬃0.6 mM and shows low affinity and high
capacity Ca2⫹ binding properties. Our search of the NCBI data
base found putative nucleotide-binding motifs for the P-loopcontaining ATPase/GTPase in the carboxyl-terminal region
shared by SAR isoforms (see Fig. 1). This observation may
suggest that SAR has an enzyme activity in the SR lumen in
addition to its role in Ca2⫹ buffering. However, there are as yet
no model animal systems, such as those with genetic diseases
at the locus, and the physiological role of SAR is still unknown
at present. By generating and analyzing a mouse model carrying the targeted mutation, we determine the important contribution of SAR to SR functions in skeletal and cardiac muscle
cells.
* This work was supported in part by grants from the Ministry of
Education, Science, Sports and Culture of Japan, the Ministry of Health
and Welfare of Japan, the Naito Foundation, the Kato Memorial Bioscience Foundation, the Mitsubishi Pharma Research Foundation, the
Japan Foundation for Applied Enzymology, the Mitsubishi Foundation,
the Vehicle Racing Commemorative Foundation, and the Nakatomi
Foundation. 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 U.S.C. Section 1734
solely to indicate this fact.
§§ To whom correspondence should be addressed. Tel.: 81-22-7178084; Fax: 81-22-717-8090; E-mail: [email protected].
1
The abbreviations used are: SR, sarcoplasmic reticulum; CSQ,
calsequestrin; SAR, sarcalumenin; SERCA, sarcoplasmic/endoplasmic
reticulum Ca2⫹-Mg2⫹ ATPase; LV, left ventricular; MOPS, 4-morpholinepropanesulfonic acid; ANOVA, analysis of variance.
Generation of Mutant Mice—A homology search using the rabbit SAR
sequence data previously reported (5) found mouse SAR cDNA sequences in the expressed sequence tag data bases. Based on the sequence data obtained, several SAR cDNA clones were isolated from a
mouse skeletal muscle cDNA library to determine the primary structure. A mouse genomic DNA library was screened with the cDNA
fragment to yield a phage clone carrying the 5⬘-terminal segment of the
SAR gene. Mutant mice carrying a targeted mutation were generated
essentially as described previously (7). As shown in Fig. 1, the targeting
vector was constructed using genomic DNA fragments, a synthetic
polylinker carrying NotI and BamHI sites, the 1.1-kb XhoI/SalI fragment from pMC1 Neo poly(A) (Stratagene), and the 0.7-kb XhoI/SalI
fragment from the pMC1 diphtheria toxin gene, and pBluescript SK(⫺)
(Stratagene). Of ⬃300 G418-resistant embryonic stem clones originally
Sarcalumenin (SAR), specifically expressed in striated muscle cells, is a Ca2ⴙ-binding protein localized in
the sarcoplasmic reticulum (SR) of the intracellular
Ca2ⴙ store. By generating SAR-deficient mice, we herein
examined its physiological role. The mutant mice were
apparently normal in growth, health, and reproduction,
indicating that SAR is not essential for fundamental
muscle functions. SAR-deficient skeletal muscle carrying irregular SR ultrastructures retained normal force
generation but showed slow relaxation phases after contractions. A weakened Ca2ⴙ uptake activity was detected in the SR prepared from mutant muscle, indicating that SAR contributes to Ca2ⴙ buffering in the SR
lumen and also to the maintenance of Ca2ⴙ pump proteins. Cardiac myocytes from SAR-deficient mice
showed slow contraction and relaxation accompanied
by impaired Ca2ⴙ transients, and the mutant mice exhibited a number of impairments in cardiac performance as determined in electrocardiography, ventricular
catheterization, and echocardiography. The results obtained demonstrate that SAR plays important roles in
improving the Ca2ⴙ handling functions of the SR in striated muscle.
MATERIALS AND METHODS
3500
This paper is available on line at http://www.jbc.org
Sarcalumenin-deficient Mice
derived from J1 cells, 12 clones showed the expected homologous mutation. The chimeric mice produced with the embryonic stem clones
numbered 299 and 374 were crossed with C57BL/6J mice, and the
targeted mutation was transmitted to the pups. The SAR-deficient and
wild-type mice obtained by crossing the heterozygous mutants were
used for the analysis in this study. To determine mouse genotypes, PCR
analysis was carried out using the synthetic primers PSAR-1 (GTAAAGTTCCCAGCACAGAACAGG), PSAR-2 (GAAAACCCACGACATCTGACCTTTG),and Pneo-5⬘a (GCCACACGCGTCACCTTAATATGCG),
as shown in Fig. 1. For immunoblot analysis, the mouse SAR cDNA
fragment encoding amino acid residues 438 – 890 was cloned into
pMALc-2 (New England Biolabs) to produce a maltose-binding fusion
protein. Wistar rats were repeatedly immunized with the purified fusion protein to yield an antiserum, and antibody to SAR was purified
using protein A-Sepharose (Amersham Biosciences). Total microsomal
protein from mouse hindlimb preparations was analyzed with the antibody to SAR as described previously (7).
Ultrastructural Analysis—Photo and electron microscopic analyses
were carried out as described previously (8). Hindlimb muscles from
adult mice were treated with a prefixative solution containing 2.5%
glutaraldehyde and 0.1 M sodium cacodylate (pH 7.4) and then postfixed with a buffer containing 1% OsO4 and 0.1 M sodium cacodylate (pH
7.4). The fixed samples were washed, dehydrated with alcohol and
acetone, and embedded in Epon. Thin sections were prepared, stained
with uranyl acetate and lead citrate, and observed under an electron
microscope (JEM-200CX; JEOL).
Muscle Contraction and Ca2⫹ Measurement—The contraction of skeletal muscle bundles was examined essentially as described previously
(9). Muscle preparations of the extensor digitorum longus and soleus
were dissected from mice and mounted on a transducer (LVS-50GA;
KYOWA Electronic Instruments); Krebs-Ringer solution (120 mM NaCl,
5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 1 mM NaH2PO4, 25 mM NaHCO3,
and 11 mM glucose) aerated with 95% O2 and 5% CO2 was used as the
bathing solution. The preparations were stretched with a resting tension of 0.5 ⫻ g at room temperature and were field-stimulated with a
supramaximal voltage at various frequencies for 300 ms.
Ventricular cardiac myocytes were isolated essentially as described
previously (10). In brief, the hearts were excised from anesthetized mice
and were mounted on a temperature-controlled perfusion system. After
perfusion with a Ca2⫹-free solution (heart medium) composed of 111 mM
NaCl, 5.4 mM KCl, 0.5 mM NaH2PO4, 12 mM NaHCO3, 10 mM KHCO3,
30 mM taurin, 2 mM carnitine, 10 mM HEPES (pH 7.4), 0.5 mM MgCl2,
and 20 mM glucose at 37 °C for 3 min, the hearts were digested for 12
min in the heart medium containing 1 mg/ml collagenase type II
(Worthington). The ventricular parts were then cut into pieces and
gently agitated to dissociate single cells. Contractile responses of the
isolated myocytes upon field stimulation (0.5 Hz, 5-ms duration) were
assessed in Tyrode’s solution (1.8 mM CaCl2) with or without 10⫺8 M
isoproterenol using a video-based edge-detection system (IonOptix Corporation). For Ca2⫹ measurements, isolated myocytes were loaded with
fura-2 acetoxymethyl (AM) (2 ␮M) for 30 min in Tyrode’s solution
containing 0.5 mM CaCl2, and Ca2⫹ transients were recorded with a
dual-excitation fluorescence photomultiplier system (IonOptix) at
26 °C. The myocytes in Tyrode’s solution (1.8 mM CaCl2) with or without
10⫺7 M isoproterenol were exposed to light emitted by a 75-W lamp and
passed through either a 360- or a 380-nm filter, and fluorescence emission
between 480 and 520 nm was detected by a photomultiplier tube. The
380-nm excitation scan was carried out at a 333-Hz sampling rate, and
the 360-nm excitation scan was examined at the start and end of the
protocol to normalize qualitative changes in the intracellular Ca2⫹ level.
Transthoracic Echocardiography—Mice were anesthetized with
Avertin (250 ␮g/g intraperitoneally) and subjected to transthoracic
echocardiography using a Toshiba Aplio instrument and a 14-MHz
linear transducer (PLT-1202S; Toshiba). Interventricular wall thickness, left ventricular (LV) posterior wall thickness, end diastolic LV
dimension (EDD), and end systolic LV dimension (ESD) were measured
on M-mode images. Ejection time (ET) was measured as the duration of
aortic valve opening. The percentage fractional shortening (FS) and the
heart rate-corrected mean velocity of circumferential fiber shortening
(mean Vcfc) were calculated as FS ⫽ (EDD ⫺ ESD)/EDD ⫻ 100 and
mean Vcfc ⫽ FS/(ET/公cycle length), respectively.
Electrocardiography and Hemodynamic Study—Mice anesthetized
with Avertin were used for the in vivo hemodynamic study. To record
the electrocardiogram, small wire electrodes were attached to each limb
of the mice. The mice were placed in the supine position, and a midline
cervical incision was made to expose the trachea and right carotid
artery. They were intubated with a blunt 21-gauge needle tube and
3501
then connected to a volume-cycled rodent ventilator (SN-480 –7; Shinano) with a tidal volume of 0.2 ml and a respiratory rate of 100
beats/min. A 1.4-F microtip catheter (SPR671; Millar Instruments) was
retrogradely inserted into the LV through the right carotid artery, and
real time data were collected with an analog-to-digital converter (Power
Laboratory; ADInstruments) and analyzed using commercial software
(Chart, version 4; ADInstruments). To evaluate ␤-adrenergic responsiveness, isoproterenol was administrated intraperitoneally. Hemodynamic data were obtained for at least 5 min.
SR Ca2⫹ Uptake and SERCA Enzyme Assay—SR vesicles were prepared from mice hindlimbs as described previously (11), and their Ca2⫹
uptake activities were measured by monitoring absorbance changes at
650 and 700 nm (BioSpec-1600, Shimadzu) using the Ca2⫹ indicator dye
arsenazo III as described previously (12). The assay mixture (final
volume, 1 ml) contained 50 mM KCl, 2 mM MgCl2, 10 mM MOPS-KOH
(pH 6.8), 0.6 mg protein/ml SR vesicles, 30 ␮M arsenazo III, 1 mM ATP,
and 30 ␮M CaCl2. To examine oxalate-facilitated Ca2⫹ uptake, 1 mM
potassium oxalate was added to the assay mixture. SERCA enzyme
activity was determined as described previously (13), using the reaction
mixture containing 50 mM MOPS/Tris (pH 7.0), 7 mM MgCl2, 0.1 M KCl,
0.6 mM CaCl2, 0.5 mM EGTA, 5 mM ATP, 1.33 ␮M A23187, and ⬃10
␮g/ml SR vesicles. Ca2⫹-ATPase activity was obtained by subtracting
the Ca2⫹-independent ATPase activity, which was determined in the
presence of 5 mM EGTA without added CaCl2. For the Northern blot
analysis of the SERCA isoforms, total RNA was prepared from hindlimb
muscle and was analyzed as described previously (8). The hybridization
probes used were PCR-amplified cDNAs derived from the 3⬘-noncoding
sequences of the SERCA isoforms (14).
RESULTS
Generation of SAR-deficient Mice—To determine the primary structure of mouse SAR isoforms, we isolated cDNA
clones from the mouse skeletal muscle library. Between mouse
and rabbit SAR primary structures, high conservation is observed in the carboxyl-terminal region, whereas moderate divergence is observed in the amino-terminal half (Fig. 1D).
Using the cDNA fragment as a probe, the mouse genomic DNA
segment carrying the SAR gene was isolated by screening a
phage library. In the vector used for introducing a targeted
mutation (Fig. 1A), the genomic DNA segment containing the
putative promoter and first protein-coding sequence is replaced
by the neomycin resistance gene. Embryonic stem cells were
transfected with the targeting vector, and the resulting G418resistant clones were screened by Southern blot analysis (Fig.
1B). Two lines of chimeric mice generated using embryonic
stem cells carrying the targeted mutation were crossed for the
generation of heterozygous mutants detected by PCR genotyping (Fig. 1C). In this report, 8 –13-week-old mice homozygous
for the targeted mutation were produced by crossing between
the heterozygous mutants and were used for analyses as described below. Two SAR isoforms are generated from the same
gene, and immunoblot analysis showed the absence of either
the long or short SAR isoform in mutant muscle (Fig. 1E). The
SAR-deficient mice thus generated exhibit no abnormalities in
health or reproduction under our conventional housing conditions, demonstrating that SAR is not essential for the fundamental development and functions of skeletal and cardiac muscle cells.
Abnormal SR Structure in SAR-deficient Skeletal Muscle—
First, we analyzed the structures of skeletal muscle from SARdeficient mice. Photo microscopy showed no significant differences in the morphology of the muscle tissues between wildtype and mutant mice (data not shown). Electron microscopy
revealed ultrastructural abnormalities in the SR from SARdeficient skeletal muscle but not in the other membrane systems or myofibril assemblies. Almost all of the SR was composed of a tubular meshwork in wild-type muscle (Fig. 2A), but
most of the SR surrounding the myofibril looked like a fenestrated plate in mutant muscle (Fig. 2B). Statistical examination revealed a significant reduction in the amount of the
tubular SR structure in the mutant muscle (Fig. 2C).
3502
Sarcalumenin-deficient Mice
FIG. 1. Generation of SAR-deficient mice. A, homologous recombination at the SAR locus. The restriction maps of the wild-type allele,
targeting vector, and predicted mutant allele are illustrated. Lines B and E indicate BamHI and EcoRI sites, respectively. The first exon, the
neomycin resistance gene (neo), and the diphtheria toxin gene (DTA) are indicated by boxes; the direction of transcription is indicated by the arrow.
B, detection of mutant gene in Southern blot analysis. Genomic DNAs from mice were analyzed after digestion with BamHI; the hybridization
probe for the detection of the mutant gene and the expected sizes of the DNA fragments are shown in A. The size markers are indicated in kilobase
pairs. C, detection of mutant gene in PCR. Genomic DNAs were used as templates and amplified DNA fragments were analyzed on an agarose gel;
synthetic primers and the expected sizes of amplified DNAs are shown in A. The size markers are indicated in base pairs. D, structural features
of SAR isoforms expressed in striated muscle. After processing the signal sequence (SP), both isoforms are localized in the SR lumen. The longer
isoform contains the low affinity and high capacity Ca2⫹-binding region and also carries the putative ATPase/GTPase domain, composed of
consensus motif sequences (G1–G4). The amino-terminal domains in rabbit and mouse 160-kDa SAR share a glutamate- and aspartate-rich
sequence, whereas their primary structures are rather divergent. E, immunoblot analysis for detection of SAR in skeletal muscle. Total microsomal
proteins from mouse hindlimbs were separated on an SDS-polyacrylamide gel and analyzed with antibody to SAR. The size markers are indicated
in kDa.
FIG. 2. Abnormal SR structures in skeletal muscle from SAR-deficient mice. A and B, electron micrographs showing typical ultrastructures of the SR in wild-type (A) or SAR-deficient muscle (B). Wild-type muscle predominantly contained a well organized tubular SR network,
whereas fenestrated plate structures were often observed in mutant muscle. Scale bar, 0.5 ␮m. In statistical analysis (C), three wild-type mice
(total 132 SR structures) and four mutant mice (144 structures) were examined, and the data indicate the means ⫾ S.D. Significant differences
were observed in the appearance of tubular SR and fenestrated SR between the genotypes. **, p ⬍ 0.01 in t test.
Abnormal Contraction Responses in SAR-deficient Skeletal
Muscle—To survey functional abnormalities in SAR-deficient
skeletal muscle, we performed isometric tension measurements
in muscle bundles. SAR-deficient extensor digitorum longus
and soleus muscle preparations retained normal levels in maximum force generation (Fig. 3A and D). In experiments using
the extensor digitorum longus bundle containing mainly fast
fibers, SAR-deficient muscle showed a slower force decrease
after contraction than wild-type muscle (Fig. 3B). Cytoplasmic
Ca2⫹ is predominantly returned into the SR by the SERCAmediated enzyme reaction during the relaxation phase, and the
relaxation curve after contractions fitted well with a single
exponential equation. In the comparison of time constants of
the tension decay, the SAR-deficient muscle showed larger
values in twitch and tetanus than wild-type muscle (Fig. 3C). A
similar slow relaxation was observed in soleus bundles containing predominantly slow muscle fibers at high frequency (⬎40
Hz) stimuli (Fig. 3E). These results indicate that SAR-deficient
skeletal muscle retains the normal mechanism for excitationcontraction coupling but bears partially impaired Ca2⫹ signaling after contraction.
Impaired SR Functions in SAR-deficient Muscle—To survey
functional abnormalities in the SAR-deficient SR, we next analyzed Ca2⫹ uptake into SR vesicles prepared from hindlimb
muscles by a spectrophotometric assay using the Ca2⫹ indicator arsenazo III. In this assay, the addition of ATP to the
reaction mixture initiates active Ca2⫹ uptake into the SR
through the action of SERCA. SR vesicles from SAR-deficient
muscle showed reduced Ca2⫹ accumulation compared with
those of wild-type muscle (Fig. 4A). Oxalate forms an insoluble
complex at high Ca2⫹ concentrations in the SR lumen, and
SERCA activity obtains obvious relief from the “back inhibition” caused by increasing luminal-free Ca2⫹ concentrations.
Thus, oxalate facilitates the increases in both the total amount
Sarcalumenin-deficient Mice
3503
FIG. 3. Abnormal contractile properties of skeletal muscle in SAR-deficient mice. Isometric tension at different frequencies was
determined in extensor digitorum longus bundles (A–C) and soleus bundles (D and E). The force-frequency relationships (A and D), typical trace
of a twitch response (B), and time constants of tension decay (C and E) are shown. The data indicate the means ⫾ S.D., and significant differences
between the genotypes are examined by t test. *, p ⬍ 0.05; **, p ⬍ 0.01.
and reaction rate of Ca2⫹ uptake (12). In the presence of oxalate, no difference was detected in maximum Ca2⫹ accumulation between the genotypes (Fig. 4B), suggesting that the reduced Ca2⫹ buffering effects were restored by oxalate in SARdeficient muscle. However, the data still suggest a slow Ca2⫹
uptake in the SAR-deficient SR, compared with that in the
control SR. Stain-all detects major Ca2⫹-binding proteins in
the muscle SR (15) and showed no change in the expression
level of CSQ or histidine-rich Ca2⫹-binding protein (HRC) between the genotypes (Fig. 4C). Taken together, the data suggest impaired Ca2⫹ buffering in the SR lumen and also imply
reduced Ca2⫹ pumping activity in SAR-deficient muscle.
We then examined the expression level and enzyme activity
of SERCA in the SR from SAR-deficient muscle. SERCA is the
most abundant protein in the SR and is easily detected as a
protein band on an SDS-polyacrylamide gel stained with Coomassie Blue (Fig. 4D). The staining density of the SERCA
protein was significantly reduced in the SAR-deficient SR (Fig.
4E), but we detected no significant changes of other protein
bands including CSQ and the ryanodine receptor. SERCA enzyme activity in SR vesicles is usually monitored by ATP hydrolysis in the presence of a Ca2⫹ ionophore. Total SERCA
activity was significantly reduced in SAR-deficient SR vesicles
(Fig. 4F); this reduction correlates well with the lowered
SERCA protein level. On the other hand, no obvious differences
were detected in the mRNA expression level of SERCA isoforms between the genotypes in Northern blot hybridization
(Fig. 4G). These biochemical results indicate the specific reduction in SERCA protein level in the mutant muscle SR.
Abnormal Cardiac Functions in SAR-deficient Mice—Next,
we focused on cardiac functions in SAR-deficient mice. The
ratios of left ventricular weight to body weight were similar
between wild-type (3.69 ⫾ 0.08 mg/g) and mutant mice (3.48 ⫾
0.09 mg/g), indicating no cardiac hypertrophy in the mutant
mice. Photo and electron microscopic analysis could not detect
any structural abnormalities of cardiac myocytes from SARdeficient mice. Transthoracic echocardiography (Table I) revealed that both the percentage of fractional shortening and
the heart rate-corrected velocity of circumferential fiber shortening, both indicators of systolic cardiac function, were significantly lower in the SAR-deficient mice than in wild-type mice.
LV chamber size and wall thickness were not different between
the genotypes.
Electrocardiography revealed that the QRS duration was
significantly longer in SAR-deficient mice and that the heart
rate and PQ interval were not different between the genotypes
(Table II). Furthermore, an in vivo hemodynamic study was
performed to evaluate the global cardiac function in mutant
mice. The maximal derivative of left ventricular pressure, an
indicator of systolic function, was significantly decreased, and
the apparent time constant in the left ventricular diastolic
phase, an indicator of diastolic function, was remarkably increased in the SAR-deficient mice (Table II). In response to the
isoproterenol application, the maximal derivative of left ventricular pressure and the time constant became similar between the genotypes, whereas the prolonged QRS was not
restored in the mutant mice. On the other hand, the mutant
mice showed no abnormalities in heart rate, left ventricular
systolic pressure, the minimal derivative of left ventricular
pressure, and left ventricular end diastolic pressure.
These data in heart functions, together with the defined
subcellular distribution of SAR, suggest that Ca2⫹ handling
3504
Sarcalumenin-deficient Mice
FIG. 4. Impaired Ca2ⴙ uptake properties in skeletal muscle SR from SAR-deficient mice. Ca2⫹ uptake activities in SR vesicles were
determined in the photometric measurement using arsenazo III in the absence (A) or presence of oxalate (B). Ca2⫹ uptake into SR vesicles
decreases free Ca2⫹ level in the assay solution, and induces absorbance changes of the Ca2⫹ indicator. Ca2⫹ accumulation into SR vesicles was
partially impaired in SAR-deficient muscle, and the abnormalities were obviously restored by oxalate application. Total proteins in SR vesicles
were separated on SDS-polyacrylamide gels and were stained with Stain-all (C) or Coomassie Blue (D). The size markers are indicated in kDa. The
relative levels of SERCA protein evaluated in Coomassie staining (E) and SERCA enzyme activity (F) in the SR vesicle preparations were
statistically analyzed. Northern blot analysis using total RNAs from hindlimbs detected no difference in SERCA-1 expression between the
genotypes (G). Skeletal muscle contains SERCA-2 as a minor Ca2⫹ pump component, and we also detected no difference in SERCA-2 mRNA level
between the genotypes (data not shown). The data indicate the means ⫾ S.D., and significant differences between the genotypes are examined by
t test. *, p ⬍ 0.05; **, p ⬍ 0.01.
TABLE I
Echocardiographic abnormalities in SAR-deficient mice
The data are presented as the means ⫾ S.E. from eight wild-type and
14 mutant mice, and the statistical differences between the genotypes
are indicated with footnotes. BW, body weight; HR, heart rate; LVEDD,
end diastolic left ventricular dimension; LVESD, end systolic left ventricular dimension; LVFS, left ventricular percent fractional shortening; IVSth, interventricular wall thickness; LVPWth, left ventricular
posterior wall thickness; ET, ejection time; Vcfc, corrected velocity of
circumferential fiber shortening.
BW (g)
HR (bpm)
LVEDD (mm)
LVESD (mm)
LV%FS (%)
IVSth (mm)
LVPWth (mm)
ET (ms)
Vcfc (circ/s)
a
b
Wild type mice
SAR-deficient mice
23.3 ⫾ 0.9
440 ⫾ 9
3.61 ⫾ 0.16
2.40 ⫾ 0.13
33.9 ⫾ 0.9
0.60 ⫾ 0.03
0.60 ⫾ 0.03
50.6 ⫾ 1.3
2.49 ⫾ 0.09
23.6 ⫾ 0.8
422 ⫾ 10
3.77 ⫾ 0.11
2.62 ⫾ 0.10
30.6 ⫾ 0.6a
0.57 ⫾ 0.02
0.58 ⫾ 0.01
55.8 ⫾ 1.2b
2.08 ⫾ 0.05a
p ⬍ 0.01 in t test.
p ⬍ 0.05 in t test.
functions in the SR are remarkably impaired in SAR-deficient
cardiac myocytes and that ␤-adrenergic stimulation can improve, in part, the irregular functions.
Abnormal Cardiac Myocyte Functions in SAR-deficient
Mice—We also examined the effects of SAR deficiency on contractility at the single cardiac myocyte level (Table III). SARdeficient myocytes showed impaired performance in the percentage of cell shortening, the shortening rate (⫺dL/dt), and
the relengthening rate (⫹dL/dt). Next, we evaluated the properties of intracellular Ca2⫹ transients in mutant myocytes (Fig.
5). The fura-2 measurements revealed that the peak amplitude
of Ca2⫹ transient and rates in both the rise and decline phases
were significantly impaired in the mutant myocytes (Table IV),
whereas we did not detect any abnormality in the basal diastolic Ca2⫹ level. The abnormalities detected were somewhat
improved but still remained abnormal when isoproterenol was
applied to mutant myocytes.
Cardiac muscle contains both SAR isoforms as in the case of
skeletal muscle. On the other hand, skeletal muscle predominantly contains SERCA-1, whereas cardiac muscle specifically
contains SERCA-2. Western blot analysis detected reduced
SERCA-2 protein level in SR preparations from the mutant
heart, but Northern blot showed similar SERCA-2 mRNA levels between wild-type and mutant hearts (data not shown).
This observation suggests that SAR also contributes to SERCA
stability in cardiac myocytes.
DISCUSSION
The irregular SR structures are impressive in SAR-deficient
skeletal muscle (Fig. 2). However, the fenestrated plate-like SR
structures were slightly detected in wild-type muscle, and normal tubular structures were barely retained in mutant muscle.
The tubular structures could transform into the fenestrated
structures under certain SR conditions. Previous studies demonstrated that Ca2⫹ overloading induces swollen SR structures
in striated muscle cells (16, 17), suggesting that luminal Ca2⫹
level is an important determinant of the SR ultrastructure. A
Sarcalumenin-deficient Mice
3505
TABLE II
Hemodynamic and electrocardiographic abnormalities in SAR-deficient mice
Hemodynamic parameters are examined before and after isoproterenol administration (2 mg/kg body weight, intraperitoneally). The data are
presented as the means ⫾ S.E. from five wild-type and six mutant mice, and statistical differences between the genotypes are indicated with
footnotes. BW, body weight; LVW, left ventricular weight; RVW, right ventricular weight; HR, heart rate; LVP, maximal left ventricular systolic
pressure; LV dp/dt max or min, maximal or minimal first derivate of left ventricular pressure; EDP, end diastolic pressure; Tau, apparent time
constant in left ventricular diastolic phase.
Basal
BW (g)
LVW (mg)
RVW (mg)
LVW/BW (mg/g)
HR (bpm)
PQ interval (ms)
QRS interval (ms)
LVP (mmHg)
LV dP/dt max (mmHg/ms)
LV dP/d min (mmHg/ms)
EDP (mmHg)
Tau (ms)
a
b
Isoproterenol
Wild-type mice
SAR-deficient mice
Wild-type mice
SAR-deficient mice
24.4 ⫾ 1.7
90.4 ⫾ 8.2
22.3 ⫾ 2.1
3.69 ⫾ 0.08
461 ⫾ 8
42 ⫾ 2
18.8 ⫾ 0.6
90 ⫾ 3
9221 ⫾ 294
⫺5115 ⫾ 116
2.0 ⫾ 0.5
6.5 ⫾ 0.9
26.6 ⫾ 1.1
92.2 ⫾ 4.2
23.8 ⫾ 1.0
3.48 ⫾ 0.09
450 ⫾ 12
43 ⫾ 1
22.3 ⫾ 0.5a
92 ⫾ 3
7331 ⫾ 628b
⫺5154 ⫾ 300
1.5 ⫾ 0.3
9.2 ⫾ 0.6b
627 ⫾ 10
37 ⫾ 1
17.1 ⫾ 0.1
94 ⫾ 3
11,895 ⫾ 1022
⫺7670 ⫾ 957
1.2 ⫾ 0.5
5.3 ⫾ 1.0
609 ⫾ 12
39 ⫾ 1
21.8 ⫾ 0.3a
101 ⫾ 4
11,306 ⫾ 534
⫺7495 ⫾ 546
1.0 ⫾ 0.5
6.8 ⫾ 0.7
p ⬍ 0.01 in t test or ANOVA.
p ⬍ 0.05 in t test or ANOVA.
TABLE III
Abnormal contractile properties of single cardiac myocytes from SAR-deficient mice
The data are presented as the means ⫾ S.E. from at least 29 cells, and statistical differences between the genotypes are indicated with footnotes.
Basal
Cell length (␮m)
Cell shortening (%)
Shortening velocity (mm/s)
Relengthening velocity (mm/s)
a
b
Isoproterenol
Wild-type mice
SAR deficient mice
Wild-type mice
SAR-deficient mice
127 ⫾ 4
7.5 ⫾ 0.4
⫺130 ⫾ 9
117 ⫾ 11
129 ⫾ 5
5.8 ⫾ 0.4a
⫺98 ⫾ 9b
78 ⫾ 8a
141 ⫾ 5
12.9 ⫾ 0.7
⫺284 ⫾ 24
231 ⫾ 17
133 ⫾ 5
11.0 ⫾ 0.6b
⫺234 ⫾ 18
185 ⫾ 17
p ⬍ 0.01 in t test or ANOVA.
p ⬍ 0.05 in t test or ANOVA.
FIG. 5. Weak Ca2ⴙ transients in single cardiac myocytes from
SAR-deficient mice. Single myocytes loaded with fura-2 were electrically stimulated, and typical Ca2⫹ transients recorded are shown.
reduced Ca2⫹ level, suggested by poor Ca2⫹ handling in the SR
(Fig. 4), could facilitate structural changes of the SR in SARdeficient skeletal muscle. On the contrary, it seems unlikely
that the structural abnormalities of the SR directly induce poor
Ca2⫹ handling and further contractile abnormalities in the
SAR-deficient muscle. Indeed, insufficient Ca2⫹ buffering and
pumping were detected in SR vesicles, in which membrane
structures should be deformed during the biochemical
preparation.
Previous studies have demonstrated that SAR is an SRluminal protein capable of binding Ca2⫹ at a high capacity and
low affinity (5, 6). The present biochemical and physiological
data indicate that SAR significantly contributes to Ca2⫹ buffering and the maintenance of the SERCA protein in the skeletal muscle SR. In skeletal and cardiac myocytes, CSQ isoforms, the most abundant Ca2⫹-binding proteins, are thought
to predominantly contribute to the reduction in the free Ca2⫹
level in the SR lumen (2). It is rather surprising that the
detectable imperfection of luminal Ca2⫹ buffering in SR vesicles can be induced by SAR deficiency, despite normal CSQ and
histidine-rich Ca2⫹-binding protein levels in mutant muscle.
CSQ is mainly localized in the junctional SR near the Ca2⫹
release site (18), whereas SAR is predominantly distributed
throughout longitudinal SR regions (19). It is an interesting
possibility that several Ca2⫹-binding proteins take part in luminal Ca2⫹ buffering and transfer in different SR subregionspecific manners.
It is important to note that SERCA protein level decreased in
the SAR-deficient skeletal muscle but that the normal expression of SERCA mRNAs was retained (Fig. 4). The data probably
suggest that SERCA protein is unstable in the mutant muscle,
because it is unlikely that the translation efficiency of SERCA
mRNA is specifically affected by SAR deficiency. Most proteins
localized in the endoplasmic reticulum/SR lumen contain the
well known retention signal (KDEL in the one-letter code) at
the carboxyl terminus of their primary structures, whereas
SAR isoforms do not carry this retention signal. Moreover, a
current report suggests the direct interaction between SERCA
and SAR (20), and our data base search found putative GTP/
ATPase motifs within the primary structures of SAR isoforms
(Fig. 1). Taken together, SAR might interact with SERCA for
its residency in the SR and might be involved in the protein
turnover of SERCA mediated by its presumable chaperon-like
activity. Because the reduction in SERCA protein level was
3506
Sarcalumenin-deficient Mice
TABLE IV
Impaired Ca2⫹ transients of single cardiac myocytes from SAR-deficient mice
The data are presented as the means ⫾ S.E. from at least 20 cells, and statistical differences between the genotypes are indicated with a footnote.
Basal
Diastolic ratio (F360/380)
Peak amplitude (⌬F360/380)
Maximal rate of [Ca2⫹]i rise (⌬F360/380/s)
Minimal rate of [Ca2⫹]i decline (⌬F360/380/s)
a
Isoproterenol
Wild-type mice
SAR-deficient mice
Wild-type mice
SAR-deficient mice
1.36 ⫾ 0.04
0.49 ⫾ 0.03
12.4 ⫾ 0.9
⫺2.5 ⫾ 0.2
1.32 ⫾ 0.04
0.36 ⫾ 0.03a
8.0 ⫾ 0.8a
⫺1.8 ⫾ 0.1a
1.16 ⫾ 0.04
0.83 ⫾ 0.05
19.8 ⫾ 1.2
⫺6.1 ⫾ 0.7
1.36 ⫾ 0.03a
0.45 ⫾ 0.03a
11.6 ⫾ 1.1a
⫺3.4 ⫾ 0.3a
p ⬍ 0.01 in t test.
also observed in cardiac muscle SR preparations, SAR isoforms
likely stabilize both SERCA-1 and SERCA-2.
SAR-deficient cardiac myocytes showed slow contraction responses (Table III) and weak Ca2⫹ transients (Fig. 5 and Table
IV), and several hemodynamic abnormalities were detected in
the SAR-deficient mice (Tables I and II). Based on the impaired
SR Ca2⫹ handling functions defined in skeletal muscle, the
SAR deficiency probably reduces total Ca2⫹ accumulation and
also induces a weak attenuation of SERCA back inhibition in
the SR. Therefore, it is reasonable to think that the poor SR
functions induce weak Ca2⫹ transients and impaired contraction in mutant cardiac myocytes and that the abnormal responses of individual myocytes likely underlie abnormal in vivo
parameters observed in hemodynamic measurements. These
defects in the mutant heart were somehow restored by the
application of isoproterenol. Because ␤-adrenergic stimulation
facilitates SERCA activity by the phospholamban-mediated
physiological regulation in cardiac myocytes (21), it may be
that weakened SR functions were slightly improved by isoproterenol in the mutant heart. In addition to abnormal hemodynamic parameters, mutant mice showed the prolonged QRS,
and this abnormality seems to be insensitive to ␤-adrenergic
stimulation (Table II). The QRS expansion might be caused by
impaired Ca2⫹ transferring from the longitudinal region to the
terminal region within the SR, and this process may be insensitive to the ␤-adrenergic regulation. Alternatively, the SARdeficient heart might bear an irreversible damage in the excitation conducting system. On the other hand, we found slightly
impaired relaxation phases but not slow contraction phases in
skeletal muscle from the SAR-deficient mice. However, SR
vesicles from mutant skeletal muscle showed significantly impaired Ca2⫹ handling. In skeletal muscle, excess amounts of
both SERCA and luminal Ca2⫹ may always be present under
basal conditions, and thus SAR deficiency might cause minimum physiological defects.
Acknowledgments—We thank Miyuki Kameyama and Yoko Shinoda
for technical assistance and Dr. Hiroshi Suzuki for providing the recipe
for the SERCA enzyme assay.
REFERENCES
1. Endo, M. (1985) Curr. Top. Membr. Transp. 25, 181–230
2. MacLennan, D. H. (2000) Eur. J. Biochem. 267, 5291–5297
3. Beard, N. A., Sakowska, M. M., Dulhunty, A. F., and Laver, D. R. (2002)
Biophys. J. 82, 310 –320
4. Shin, D. W., Pan, Z., Kim, E. K., Lee, J. M., Bhat, M. B., Parness, J., Kim, do
H., and Ma, J. (2003) J. Biol. Chem. 278, 3286 –3292
5. Leberer, E., Timms, B. G., Campbell, K. P., and MacLennan, D. H. (1990)
J. Biol. Chem. 265, 10118 –10124
6. Raeymaekers, L., Verbist, J., Wuytack, F., Plessers, L., and Casteels, R. (1993)
Cell Calcium 14, 581–589
7. Takeshima, H., Iino, M., Takekura, H., Nishi, M., Kuno, J., Minowa, O.,
Takano, H., and Noda, T. (1994) Nature 369, 556 –559
8. Takeshima, H., Komazaki, S., Nishi, M., Iino, M., and Kangawa, K. (2000) Mol.
Cell 6, 11–22
9. Nishi, M., Komazaki, S., Kurebayashi, N., Ogawa, Y., Noda, T., Iino, M., and
Takeshima, H. (1999) J. Cell Biol. 147, 1473–1480
10. Minamisawa, S., Gu, Y., Ross, J., Jr., Chien, K. R., and Chen, J. (1999) J. Biol.
Chem. 274, 10066 –10070
11. Eletr, S., and Inesi, G. (1972) Biochim. Biophys. Acta 282, 174 –179
12. Kim, D. H., Ohnishi, S. T., and Ikemoto, N. (1983) J. Biol. Chem. 258,
9662–9668
13. Suzuki, H., and Kanazawa, T. (1999) Biochemistry 12, 820 – 825
14. Wu, K-D., Lee, W-S., Wey, J., Bungard, D., and Lytton, J. (1995) Am. J.
Physiol. 269, C775–C784
15. Damiani, E., and Margreth, A. (1991) Biochem. J. 277, 825– 832
16. Takeshima, H., Komazaki, S., Hirose, K., Nishi, M., Noda, T., and Iino, M.
(1998) EMBO J. 17, 3309 –3316
17. Uehara, A., Yasukochi, M., Imanaga, I., Nishi, M., and Takeshima, H. (2002)
Cell Calcium 31, 89 –96
18. Nori, A., Nadalini, K. A., Martini, A., Rizzuto, R., Villa, A., and Volpe, P. (1997)
Am. J. Physiol. 272, C1420 –C1428
19. Leberer, E., Charuk, J. H., Green, N. M., and MacLennan, D. H. (1989) Proc.
Natl. Acad. Sci. U. S. A. 86, 6047– 6051
20. Dowling, P., Doran, P., and Ohlendieck, K. (2004) Biochem. J. 379, 479 – 488
21. Tada, M., Yabuki, M., and Toyofuku, T. (1998) Ann. N. Y. Acad. Sci. 853,
116 –129