THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 266, No. 5, Issue of February 15, pp. 3499-3506,1993 Printed in U. S.A.

Different Intracellular Localization of Inositol 1,4,5-Trisphosphate and Ryanodine Receptors in Cardiomyocytes*

(Received for publication, September 23, 1992)

Yoshiyuki KijimaSt, Akitsugu Saitoj, Thomas L. Jettonll)) , Mark A. Magnusonll, and Sidney Fleischert From the §Department of Molecular Biology, Vanderbilt University, Nashville, Tennessee 37235 and the llDepartment of Molecular Physiology and Biophysics, Vanderbilt Uniuersity Medical School, Nashville, Tennessee 37232

The ryanodine and inositol 1,4,5-trisphosphate (Ips) receptors have previously been found to be intracellu- lar Ca2+ release channels characterized by their large size and 4-fold symmetry. In this study, cardiomy- ocytes are found to have a different intracellular lo- calization for the two receptors. At the level of light microscopy, the Ips receptor is immunolocalized in rat ventricular cardiomyocytes at the region of the inter- calated discs. By contrast, immunoreactivity of the ryanodine receptor is observed as transverse bands throughout the length of the cardiomyocyte, coincident with the triad junction at the I-bands. At the level of electron microscopy, immunogold particles directed to the IPS receptor specifically decorate the intercalated discs of rat ventricular and atrial cardiomyocytes, preferentially at the fascia adherens. Binding of [3H] IP3 and [3H]ryanodine were measured in cardiac sub- cellular fractions. IPS binding is enriched in a fraction containing intercalated discs. Little or no Ips binding was detected in longitudinal sarcoplasmic reticulum (SR), junctional SR, sarcolemma, mitochondria, and submitochondrial vesicles. Ryanodine binding is the highest in junctional SR. We conclude that the Ips receptor is present in ventricular and atrial cardiom- yocytes and localized at the region of the intercalated discs. These results suggest a possible role of the IP3 receptor in Ca2+ entry through intercalated discs and/ or intercellular signaling between cardiomyocytes.

The ryanodine receptor and inositol 1,4,5-trisphosphate (IP3)’ receptor are intracellular Ca2+ release channels which represent a new class of ion channels characterized by their large size and 4-fold symmetry (1-3). In excitation-contrac- tion coupling of skeletal muscle and heart, the ryanodine

* This work was supported by National Institutes of Health Grants HL32711, Pol-HL46681 (to S. F.), DK42612 (to M. M.), by a Biomedical Research Support Grant from NIH (to S. F.) administered by Vanderbilt University, by a grant-in-aid from the National Science Committee of Vanderbilt University (to S. F.), and by a New Inves- tigator Award from the American Heart Association-Florida Affiliate (to Y. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$Awardee of a New Investigatorship from the American Heart Association-Florida Affiliate. This work was performed in part during the tenure from a Postdoctoral Fellowship of the Muscular Dystrophy Association.

11 Supported by Vanderbilt Molecular Endocrinology Training Pro- gram Grant DK07563.

’ The abbreviations used are: IP3, inositol 1,4,5-trisphosphate; ABC, avidin-biotin complex; DTT, dithiothreitol; MOPS, 3-(N-mor- pholino)-propane sulfonic acid; PBS, phosphate buffer saline; SR, sarcoplasmic reticulum; T-tubule, transverse tubule.

receptor serves as a Ca2+ release channel of SR and is mor- phologically identical to the foot structure spanning the gap between terminal cisternae of SR and sarcolemma/transverse tubules. The IPS receptor acts as a Ca2+ release channel of non-mitochondrial intracellular Ca2+ store(s) in smooth mus- cle and in non-muscle tissues (2, 4, 5 , 18).

There has been the suggestion that the IPS receptor may operate in heart: 1) the IPS cascade is present in heart (6), and hormones and drugs which activate it, e.g. al-sympatho- mimetics or endothelin, cause a positive inotropic effect on heart (7,8) with elevation of the myoplasmic calcium concen- tration (9, 10); 2) IP3 can cause contraction of skinned car- diomyofibers (1 1,12); 3) cardiac microsomes fused into planar lipid bilayers exhibit Ca2+ channel activity modulated by IP3 (13); and 4 ) positive Western and Northern blot analyses using whole heart tissue indicate that the IPS receptor is present in heart, albeit not necessarily localized in cardiom- yocytes (14-16). Nonetheless, IP3-induced Ca2+ release from cardiac SR has not been convincingly demonstrated and the role for the IP3 receptor in heart has remained controversial (43).

In this study, the IP3 and ryanodine receptors are both found to be present in rat cardiomyocytes by immunolocali- zation. The two receptors show different intracellular local- ization. Binding studies of 13H]IP3 and [3H]ryanodine to cardiac subcellular fractions support the immunocytochem- istry.

EXPERIMENTAL PROCEDURES

Antibodies-The antibody to the receptor is rabbit sequence- specific polyclonal serum produced against a synthetic peptide con- taining the carboxyl-terminal 19 amino acids of the mouse IP, recep- tor (17, 18). The antibodies to the cardiac ryanodine receptor are mouse monoclonal IgG (Ry-1, Ry-3, and Ry-5) raised against the ryanodine receptor purified from dog heart (19).

Western Blot Analyses-SDS-polyacrylamide gel electrophoresis was performed according to Laemmli (20) using either 7.5 or 6% polyacrylamide gel. Transelectrophoresis of protein onto an Immo- bilon-P membrane (Millipore) and subsequent immunostaining was carried out by the method of Towbin et al. (21) with minor modifi- cations. The transfer was performed overnight at room temperature using constant current (150 mv) in 0.01% (w/v) SDS, 25 mM Tris, 190 mM glycine, pH 8.3. An Immobilon-P membrane was blocked for 1 h a t room temperature with 5% (w/v) defatted milk in TBS (150 mM NaCI, 50 mM Tris-HCI, pH 7.4). The membrane was incubated

antibody in TBST (TBS with 0.05% (w/v) Tween 20) containing overnight a t 4 “C or for 3 h at room temperature with a primary

2.5% milk. The blot was washed with TBST (10 min X 3 times) to remove unbound antibodies and incubated for 1 h at room tempera- ture with a secondary antibody (goat anti-rabbit or anti-mouse IgG alkaline phosphatase conjugate, 1:5000 dilution, from Promega) in TBST containing 2.5% milk. After washing with TBST (10 min X 3 times), immunodetection was performed by color development using 5-bromo-4-chloro-3-indolylphosphate-toluidine saltlp-nitro-blue tet- razolium chloride as substrate reagents.

Light Microscope Immunohistochemistry-Male Sprague-Dawley

3499

3500 Localization of IPS and Ryanodine Receptors in Heart rats were decapitated, and the hearts were removed and rinsed in ice- cold PBS. The ventricular tissue was either quenched in liquid nitrogen-chilled isopentane for 5 min for cryostat-sectioning or im- mersion fixed in 4.0% paraformaldehyde in 0.1 M PBS overnight at 4 "C prior to conventional paraffin embedding. Immunostaining of the IP3 receptor was obtained on paraformaldehyde-fixed paraffin sections, whereas that of the ryanodine receptor was accomplished with cryostat sections.

Immunostaining of the IP3 receptor was performed on the paraffin sections (5 pm thickness) with avidin-biotinylated peroxidase com- plex. The sections were pretreated with 0.5% HzOz in absolute meth- anol to quench endogenous peroxidase activity. Endogenous avidin and biotin were blocked according to the manufacturer's kit recom- mendations (Vector Laboratories). Following equilibration in PBS, the sections were blocked with 5% normal donkey serum for 30 min. The anti-IP3 receptor serum or rabbit normal (non-immune) serum (1:lOO) was applied to alternate serial sections and incubated for 12 h at room temperature. After washing (15 min X 3 times) with 0.1% Triton X-lOO/PBS, the sections were incubated in biotinylated don- key anti-rabbit IgG (Jackson Immunoresearch) for 1 h. Following three more washes, the sections were then incubated with the ABC reagent from the standard peroxidase kit (Vector Laboratories) for 1 h. The peroxidase substrate consisted of 0.01% HZ02 and 0.003% 3- amino-9-ethylcarbazole in a 0.1 M sodium acetate buffer, pH 5.2.

For immunofluorescence detection of the ryanodine receptor, fro- zen tissue was allowed to equilibrate to cryostat temperature (-25 "C) and embedded in O.C.T. (Miles Scientific). Then 5-6-pm sections were cut and mounted on silanized slides. The sections were dried at room temperature for 2 min, then immersed in 4.0% paraformal- dehyde in PBS for 10 min. Following two rinses in PBS for 5 min each, the cardiac cryosections were permeabilized with 1.0% Triton X-lOO/PBS for 2 min and rinsed twice in PBS for 5 min. The cryosections were blocked for endogenous avidin and biotin using the Vector kit, and incubated with 5% normal donkey serum. The mixture of three anti-ryanodine receptor antibodies (Ry-1, Ry-3, and Ry-5) or mouse normal (non-immune) serum (1:IOO) served as the primary antibody on adjacent cryosections. Following incubation for 12 h at room temperature, the tissue sections were subjected to biotinylated donkey anti-mouse IgG (Jackson Immunoresearch) and incubated with streptavidin-Texas Red (Jackson Immunoresearch) for 1 h. Photomicrography was accomplished using a Leitz Laborlux S fluo- rescence photomicroscope and recorded on Kodak Plus-X film (125 ASA) for transmitted light and Tri-X film (400 ASA) for fluorescence.

Immunolocalization by Electron Microscopy-Cubes of rat heart tissue from ventricle and atrium (0.5 mm3) were fixed for 2 h at 4 "C with 0.5% glutaraldehyde (EM-grade, Polyscience), 3% formaldehyde (EM-grade, Electron Microscopic Science) in 0.1 M sodium phos- phate, pH 7.2, and subsequently embedded in Lowicryl K4M at 4 "C for 48 h as described previously (22). The thin sections were blocked for 10 min at 23 "C with 4% (w/v) defatted milk or 0.5% (w/v) ovalbumin in TBST and then incubated at 4 "C overnight with the anti-IP3 receptor antibody (1:50) in TBST containing 2% defatted milk or 0.5% ovalbumin. The sections were washed thoroughly with TBST and incubated for 1 h at 23 "C with goat anti-rabbit IgG conjugated with 10-nm gold particles (Ted Pella, Inc. CA) in TBST containing 0.5% ovalbumin. After further washing with TBST and HzO, the sections were stained with uranyl acetate and lead citrate and examined in a JEOL, JEM-IOOS electron microscope. In some experiments, protein A gold (10 nm, Sigma) was used instead of the secondary antibody.

Electron Microscopy-Fixation of the intercalated disc-containing fraction in 2% glutaraldehyde and 1% tannic acid for thin section electron microscopy was performed as described previously (23, 24). Representative samples were prepared by the Millipore-filtration method (25). Briefly, 0.4 ml of the fixed sample (200 pg of protein) was filtered through a Millipore filter (type VCWP, 0.1 pm of pore size). The sample trapped on the filter was additionally fixed with OsOl and stained with uranyl acetate and lead (23, 24).

Preparations of an Intercalated Disc-containing Fraction-An in- tercalated disc-containing faction was prepared from dog cardiac muscle by modification of the procedure described by Kensler and Goodenough (26). Ventricular muscle (-100 g) was ground by a meat grinder (4 mm of hole diameter) and divided into 3 X 30 g portions. Each portion was homogenized in 150 ml of 1 mM NaHC03 with a Polytron homogenizer (Brinkman Instruments Inc.) for 30 s at max- imum speed. The homogenate (from three blendings) was diluted with 1 mM NaHC03 to 1500 ml, left 10 min on ice, and filtered through two layers of cheesecloth. The filtrate was centrifuged three times as described previously (26). Lower portions of pellet of the

second and third spin were combined in 600 ml of KI-NaHC03 buffer (0.6 M KI, 6 mM Na2S203, 1 mM NaHC03), rehomogenized with the Polytron at dial 2 (20% of maximum) for 10 s, and stirred overnight to extract contractile elements.

Next day the KI homogenate was centrifuged for 40 min at 9,500 revolutions/minute in a Beckman JA-IO rotor. The pellet is resus- pended in 300 ml of KI-Tris solution (0.6 M KI, 6 mM Na2S203, 5 mM Tris-HC1, pH 8.3) using a Polytron for 10 s at dial 2, and centrifuged for 60 min at 40,000 revolutions/minute in a Beckman type Ti-45 rotor. The KI-insoluble pellet was resuspended in 120 ml of the KI- Tris solution using a Polytron for 10 s at dial 2. Half of the resus- pended sample was quick-frozen in liquid N, and kept at -80 "C until used. Another half (60 ml) was loaded on six tubes containing a sucrose step gradient (8 ml each of 20,25, 30, and 35% (w/w) sucrose solution in the KI-Tris solution), and centrifuged for 2 h a t 27,000 revolutions/minute in a Beckman SW28 rotor. The fraction at the interface of 25 and 30% sucrose was collected and diluted with 5 mM imidazole-HCI, pH 7.0, to reduce the sucrose concentration to less than 10%. The sample was centrifuged for 60 min at 40,000 revolu- tions/minute in a Beckman type Ti-45 rotor. The pellet was then resuspended in imidazole-sucrose solution (0.3 M sucrose, 5 mM imidazole-HC1, pH 7.0), quick-frozen, and kept at -80 "C until further purification.

The sample was further purified on a second sucrose gradient. Four ml of the sample (10.7 mg of protein) was loaded on a sucrose gradient containing 8 ml of each 27, 32, 38 and 45% (w/w) sucrose in 5 mM imidazole-HC1, pH 7.0, and centrifuged for 2.5 h at 49,000 revolu- tions/minute in a Beckman vertical rotor VTi-50. The fraction at an interface of 38 and 45% sucrose was collected, diluted with 5 mM imidazole, pH 7.0, and centrifuged for 30 min at 46,000 revolutions/ minute in a Beckman type Ti-70 rotor. The final pellet (the interca- lated disc-containing fraction) was resuspended in the imidazole- sucrose solution, quick frozen, and kept at -80 "C.

Preparation of Other Membrane Fractions-Cardiac junctional SR, longitudinal SR, and sarcolemma were prepared from fresh dog heart ventricles. Canine cardiac microsomes were prepared as described previously (27,28). The microsomes were subfractionated by selective calcium loading and sucrose density gradient centrifugation into five fractions as described previously (29). Fractions 1-5 were numbered from the top to the bottom of the gradient. In this study, fraction 1 was found to contain the highest (Na,K)-dependent ATPase activity (53 pmol Pi/mg.h, Table I) so that this fraction was used as sarco- lemma. Fraction 2 and fraction 5 were previously characterized as junctional and longitudinal SR, respectively (29). Calcium loaded into fraction 5 was removed by dilution of the vesicles (29) to reduce nonspecific IP3 binding.

Atrial microsomes were prepared from frozen dog atria. Dog heart atria were surgically separated from ventricles, cut into small pieces (5-8 mm), and quick frozen in liquid N,. The accumulated atrial tissue (-90 g from five to six dog hearts) was used for one preparation according to a protocol similar to that for ventricular microsomes (27, 28).

Rat cardiac membrane fractions were prepared by modification of the procedure for dog heart (27,28). Briefly, fresh rat heart ventricles (-6 g from six rat hearts) were cut into small cubes (3-5 mm) and homogenized in 50 ml of cold homogenization medium (0.29 M SU- crose, 0.5 mM DTT, 3 mM NaN3, 10 mM imidazole-HCI, pH 6.9). Homogenization in a 50-ml flask was for 15 s at dial 70 followed by 35 s at dial 100 (maximum speed) using a Virtishear homogenizer (Virtis Co., New York) with a rotor/stator (10-mm diameter). The homogenate was centrifuged in two tubes for 15 min at 5,000 revolu- tions/minute (3,030 X g,,,) in a Beckman JA-20 rotor. The super- natant was filtered through four layers of cheesecloth, adjusted to 70 ml with fresh homogenization medium, and centrifuged in two tubes for 15 min at 15,000 revolutions/minute (27,200 X g,,,) in a Beckman JA-20 rotor. The pellet was suspended in resuspension medium (0.65 M KCl, 0.5 mM DTT, 3 mM NaN3, 10 mM imidazole-HC1, pH 6.7) by hand homogenization in a Potter-Elvehjem tissue grinder with a Teflon pestle, and saved as HP (heavier pellet). The supernatant was filtered through four layers of cheesecloth and centrifuged for 2 h at 32,000 revolutions/minute (120,000 X g,,,) in a Beckman type Ti-45 rotor. The pellet was resuspended for salt washing in 17 ml of the resuspension medium using a Potter-Elvehjem tissue grinder and sedimented for 100 min at 50,000 revolutions/minute in a Beckman type Ti-70 rotor (260,000 X gmaX). The pellet was resuspended in the resuspension medium and saved as LP (lighter pellet). The HP and LP pellets were quick frozen and remained at -80 "C until used.

Intact mitochondria (30) and submitochondrial vesicles (31) were prepared from fresh bovine heart as described previously.

Localization of IPo and Ryartocline Receplors in Heurl 3501

Brain microsomes which exhibit the IP3-induced Ca2+ release were prepared from fresh dog brain as described previously (28, 32).

All of the membrane preparations were carried out at 4 "C or on ice. Protease inhibitors (1 pg/ml leupeptin, 0.5 pg/ml pepstatin A, 20 p M benzamidine, 20 p~ phenylmethylsulfonyl fluoride, 2.7 X TIU/ml aprotinin) were added to solutions for all of the preparations except mitochondrial preparations.

Tissue Extract from Kat Heart-Cardiac muscle extract was pre- pared from quick-frozen rat ventricle. Ventricular muscle was cut into small cubes (3-5 mm), frozen in liquid N2, and powdered with a pre-ooled mortar and pestle. The muscle powder was extracted for 60 min a t room temperature with an SDS extraction medium (10% SDS, 10 mM EDTA, 50 mM DTT, 0.1 mM phenylmethylsulfonyl fluoride, and 0.1 M Tris-HCI, pH 8.3) a t a ratio of 200 mg/ml (weight of frozen muscle/volume of the medium). During the incubation for 1 h, the sample was vigorously agitated several times with a vortex mixer. Finally, supernatant was collected and used for SDS-polyacrylamide gel electrophoresis followed by Western blot.

["H/IP:t Binding Assay-['H]IP, (0.59 p ~ , 17 Ci/mmol, -20,000 cpm/pmol, from Du Pont-New England Nuclear) was used to perform the binding assay according to Chadwick et al. (2) with slight modi- fications. Cardiac subcellular fractions (20-100 pg) were incubated for 10 min a t 4 "C in 0.1 ml of assay medium (1 mM DTT, 1 mM EDTA, 10 nM [:'H]lPR, 50 mM Tris-HCI, pH 8.3). The assay mixture was then centrifuged a t 4 "C for 10 min at 95,000 revolutions/minute (390,000 X g,,,*J with a Beckman TLA 100 rot,or. The supernatant containing free ligand was aspirated, and the pellet was rapidly washed twice with 0.2 ml of H20. The pellet was resuspended in 0.2 ml of H20, mixed with 5 ml of CytoScint scintillation mixture (ICN), and the radioactivity was measured in a Beckman LS 5000 TD scintillation counter. Nonspecific binding was determined in the presence of 10 p~ non-radioactive IP, (trilithium salt, from LC services Co.). In some cases nonspecific binding was determined in the presence of 1 mg/ml heparin (low molecular weight of bovine intestinal mucosa, from Sigma) which gave similar values to those determined in 10 p~ of cold IP,. Specific binding is the difference between total and nonspecific binding. Each data point was obtained from the average of three total minus one nonspecific binding.

/'H/Ryanodine Binding A.~ay-[~H]Ryanodine (-9400 cpm/pmol) was prepared as described previously (33). ["HlRyanodine binding was measured by the centrifugation method. Cardiac subcellular fractions (20-82 pg of protein) were incubated a t 23 "C for 30 min in 0.1 ml of assay medium (1 M NaCI, 25 p M CaCI2, 60 nM ['Hlryanodine, 50 mM Tris-HCI, pH 7.4). The assay mixture was then centrifuged a t 4 "C and the radioactivity of the pellet was measured as described above. Nonspecific binding was determined in the presence of 100 @ cold ryanodine (Lindhurst, New Jersey). Each data point was ob- tained from average of two total minus one nonspecific binding.

Na,K-dependent ATPase Activity Assay-Na,K-dependent ATP- ase activity was measured spectrophotometrically (at 340 nm of wave length) a t 37 "C by monitoring the oxidation of NADH in a coupling enzyme system. Cardiac subcellular fractions (0.5 mg protein/ml) were preincubated with 0.3 mg/ml SDS (SDS/protein = 0.6 (w/w)) for 10 min a t 37 "C to unmask the latent ATPase activity as described previously (34). The pretreated microsomes (50 pg of protein) were added to an assay medium (final volume = 1 ml) containing 1 mM ATP (Tris salt), 0.2 mM NADH, 2 mM phosphoenolpyruvate, 8.4 units of pyruvate kinase, and 12 units of lactate dehydrogenase in 3 mM MgC12, 100 mM NaCI, 10 mM KCI, 5 mM NaNI, 1 mM EGTA, 20 mM MOPS/Tris, pH 7.0. The "basal" ATPase, determined in the presence of 0.9 mM ouabain, was subtracted from the total ATPase activity to give the ouabain-sensitive ATPase activity.

Protein-Protein concentrations of the membrane fractions were determined by the method of Bradford (35) using bovine serum albumin as a standard.

RESULTS

Characterization of Antibodies to the IP3 and Ryanodine Receptors by Western Blot Analyses-The specificity of anti- bodies used in this study was examined by Western Blot analyses of rat cardiac preparations including membrane frac- tions and SDS extract from whole heart (Fig. la ) . The poly- clonal anti-IPS receptor antibody to the carboxyl terminus of the mouse cerebellar IP3 receptor (17, 18) cross-reacted with only a single band (250 kDa) in rat cardiac membrane frac- tions (Fig. IC). No immunoreactive band was observed in the SDS extract from whole heart (Fig. lb , lane 2) , indicating

a 1 2 3

r" Mr x -. -

200.0-

1 16.3,

97.4-

66.2-

200.0- & &

116.3- ~1 45.0- 97.4- L L

66.2- 1 p4 - b 1 2 c 1 2 3

- 200.0-

200.0- 116.3-

Mr x 10.'

Mr x lo3 .

97.4-

66.2- 116.3-

97.4- 45.0-

66.2-

d l 2 3 4 Mr x lo"

200.0-

1 16.3- 97.4-

66.2-

45.0-

FIG. 1. Character izat ion of antibodies to the IPS and ryano- dine receptor by Western blot analyses. Heavier pellet ( H P ) , lighter pellet (LP) , and SDS extract from whole heart used in this figure were prepared from rat heart ventricle. a, staining of SDS- polyacrylamide gel with Coomassie Brilliant Blue R-250. Fifteen p g of protein of H P (lane I ) and LP (lane 2) was loaded to each lane of a 7.5% SDS gel. Five pl of the whole heart extract (lane 3) was analyzed on a 6% SDS gel. b, anti-IP, receptor antibody (1:4000 dilution). Thirty p g of protein of dog brain microsomes as a control (lane I ) and 10 pl of the whole heart extract (lane 2) were analyzed on a 6% SDS gel. A 260 kDa band of the IPS receptor (double arrowhead) was detected in brain microsomes. The immunodetection was performed using a secondary antibody conjugated with alkaline phosphatase. c, anti-IP2 receptor antibody (1:4000). Thirty pg of protein of H P (lane I ) , LP (lane 2), and dog brain microsomes (lane 3 ) was analyzed on a 7.5% SDS gel. A strong positive band was found in dog brain microsomes (260 kDa) and rat heart H P (250 kDa). A very faint band of 250 kDa was detected in LP. d, anti-ryanodine receptor antibodies. Thirty pg of protein of LP (lane I ) and HP (lane 2 ) , 10 p l of the whole heart extract (lane 3), and 10 p g of protein of canine cardiac microsomes (lane 4 ) were analyzed on 7.5% gels. A strip containing lanes 1-3 was incubated with mixture of the mono- clonal antibodies (Ry-1, 3, and 5) (1:3000 of each), whereas that of lane 4 with Ry-5 (1:lOOO). Similar results were obtained in the canine cardiac microsomes using Ry-1 or Ry-3 (data not shown). The ryano- dine receptor (arrowhead) is specifically visualized as a doublet (400 and 390 kDa).

that the amount of the IPS receptor in the whole heart extract was below the threshold of detection. Strong immunoreactiv- ity was observed in the heavier pellet (HP) from rat heart (Fig. I C , lane 1 ) and dog brain microsomes (Fig. I C , lane 3), whereas the lighter pellet (LP) from rat heart showed only a faint reaction (Fig. IC, lane 2). The rat cardiac IP3 receptor exhibited a slightly faster electrophoretic mobility (250 kDa) than that of the dog brain IPa receptor (260 kDa), suggesting a subtle difference between the two receptors. Although dog

3502 Localization of IP3 and Ryanodine Receptors in Heart

brain microsomes gave a strong positive reaction on Western blot, the anti-IP, receptor antibody did not cross-react with dog cardiac fractions (data not shown), indicating that the carboxyl terminus of the dog cardiac IP3 receptor is immu- nologically distinct from that of the rat cardiac and dog brain receptors.

A mixture of three monoclonal anti-ryanodine receptor antibodies (Ry-l,3, and 5; Ref. 19) visualized specifically only a faint band a t 400 kDa in the SDS extract of rat whole heart (Fig. Id, lane 3 ) . In the membrane fractions (Fig. Id, lanes 1 and 2), immunoreactivity was found (stronger in HP than in LP) specifically as a high molecular mass doublet (400 and 390 kDa), suggesting proteolysis during the preparation as described previously (36). The apparent molecular weight of these bands was similar to that found in dog cardiac micro- somes (Fig. Id, lane 4 ) , indicating that the antibodies to the dog cardiac ryanodine receptor specifically recognize the ry- anodine receptor in rat heart. These results show that the immunoreactivity found with antibodies to the IP3 and ryano- dine receptors in tissue sections of rat heart (Figs. 2-4) is specific for the IP3 receptor (250 kDa) and the ryanodine receptor (400 kDa), respectively.

Immunohistochemistry by Light Microscopy-The IP3 receptor and the ryanodine receptor were immunolocalized in rat heart ventricle at the level of light microscopy. The immunoperoxidase reactivity of the IP3 receptor was found in rat ventricle cardiomyocytes and restricted to the region of the intercalated discs (Fig. 2, a and b). Immunoreactivity was observed in about 40% of all intercalated discs of more than 200 fields examined. No immunoreactivity was detected in adjacent sections with non-immune serum (Fig. 2, c and d).

The immunofluorescence of the ryanodine receptor was observed as transverse bands with periodicity along the full length of the cardiomyocyte (Fig. 3). The immunofluorescence bands were found to correspond to the I-bands using phase contrast optics (data not shown). In mammalian heart, the T-tubules are located at the region of the Z-discs of the sarcomere, that is in the center of the I-bands where the triad junction is located, i.e. the T-tubules are junctionally associ- ated with SR via the foot structures. This is compatible with the previous immunogold study performed on skeletal muscle (37). The immunoreactivity to the ryanodine receptor, at that time referred to as the “spanning protein,” was found adjacent to the T-tubules of the triad junction. These results indicate that the IP3 receptor is present in ventricular cardiomyocytes and that its intracellular localization is different from that of the ryanodine receptor.

Immunolocalitation by Electron Microscopy in Ventricular and Atrial Cardiomyocytes-The immunogold particles di- rected to the IP3 receptor were localized to the intercalated discs in rat ventricular cardiomyocytes (Fig. 4b). Intercalated discs are involved in junctional association of cardiomyocytes. They consist of gap junction (nexus) involved in intercellular signalling, desmosomes (macula adherens), and intercellular adherens junction (fascia adherens) (38). The immunoreactiv- ity of the IP3 receptor was found predominantly at or near the fascia adherens, localizing the IPS receptor to the inter- calated discs of the plasmalemma. The surface sarcolemma was negative (Fig. 4c). Immunoreactivity was not found in T- tubules, SR, mitochondria, or other intracellular membrane organelles. Similar results were obtained by using protein A gold instead of the secondary antibody (goat anti-rabbit IgG gold conjugate). Essentially no immunogold particles were found when using non-immune rabbit serum instead of the anti-IP3 receptor antibody (Fig. 4u). In atrial cardiomyocytes (Fig. 4d), immunoreactivity of the IP3 receptor were also localized to intercalated discs. The results of electron micros- copy confirm and extend those obtained by immunoperoxi-

FIG. 2. Immunolocalization of the IP3 receptor in rat ven- tricular cardiomyocytes at the level of light microscopy (X 978). a, paraformaldehyde-fixed paraffin section of cardiomyocytes stained with the anti-IPs receptor serum followed by the ABC per- oxidase technique. Arrowheads indicate immunoperoxidase reaction product associated with intercalated discs. A typical field is shown. b, same section as in a, but viewed by phase contrast optics. Arrowheads indicate phase-dense intercalated discs which were immunopositive in a. c, adjacent section stained with non-immune serum followed by the ABC-peroxidase technique. No immunoperoxidase product was found. The arrowhead points to an intercalated disc. d, same section as in c viewed with phase contrast. An interference filter (590 nm) was used in a and c.

dase at the level of light microscopy (Fig. 2), that is, the cardiac IP3 receptor appears to be localized in ventricular and atrial cardiomyocytes at the region of the intercalated discs.

Electron Microscopy of the Isoluted Intercalated Discs from Dog Heart-An intercalated disc-containing fraction was pre- pared from dog heart ventricles. Extraction of contractile elements by potassium iodide and sucrose gradient centrifu- gation were the two key steps. Electron microscopy shows that intact structure of fascia adherens is retained in this fraction (Fig. 5). Mitochondria are a major contaminant.

Subcellular Distribution of r3H]IP3 and [3H]Ryanodine Binding in Heart-Distribution of the IPS and ryanodine receptors was estimated by ligand binding to purified cardiac subcellular fractions (Table I). Binding of [3H]IP3 (10 nM of

Localization of IP3 and Ryanodine Receptors in Heart 3503

FIG. 3. Immunolocalization of the ryanodine receptor in rat ventricular cardiomyocytes at the level of light microscopy (X 1504). The cryosection was stained with mixture of the anti-ryanodine receptor antibodies (Ry-1, 3, and 51, followed by biutinylated anti- mouse IgG and streptavidin-Texas Red. The transverse banding pattern of the ryanodine receptor immunofluorescence can be seen along the full length of the cardiomyocytes consistent with the I-bands of the cardiac sarcomeres. No fluorescence was detected when non-immune serum was used (data not shown).

ligand concentration) and [,H]ryanodine (60 nM) was meas- ured to fractions isolated from heart ventricle: longitudinal SR, junctional SR, sarcolemma, an intercalated disc-contain- ing fraction, mitochondria, and submitochondrial vesicles. IP3 binding was highly enriched in the intercalated disc-contain- ing fraction (9.43 pmol/mg), compared with < 0.2 pmol/mg in other fractions. Ryanodine binding was highest in junc- tional SR (3.31 pmol/mg). In atrial microsomes, little IP3 binding was found (0.160 pmol/mg). These results support our immunolocalization findings (Figs. 3 and 4), i.e. the IP3 receptor is localized to the intercalated discs and essentially not present in SR.

DISCUSSION

This study provides the first definitive evidence that the IP3 receptor is present in cardiomyocytes and that it is local- ized to the intercalated discs. Both the IP3 receptor and ryanodine receptor are present in the cardiomyocyte, albeit with different subcellular localization. At the level of light microscopy, immunoperoxidase reactivity of the IP3 receptor appears to be located in the region of the intercalated discs of rat ventricular cardiomyocytes (Fig. 2). The immunofluores- cence of the ryanodine receptor is observed as transverse bands along the entire length of the ventricular cardiomy- ocytes corresponding to the I-bands (Fig. 3). This differential localization of the two receptors is decisively confirmed at the level of electron microscopy. That is, the immunoreactivity of the IP3 receptor is found specifically at intercalated discs of plasmalemma, preferentially at the fascia adherens (Fig. 4). The IP3 receptor is not present to any significant extent in SR, whereas the ryanodine receptor is localized to the terminal cisternae of SR.

These immunolocalization studies are supported by ligand binding to cardiac subcellular fractions. IP3 binding is highly

enriched in the intercalated disc-containing fraction and is essentially absent in longitudinal SR, junctional SR, and sarcolemma. Ryanodine binding is enriched in junctional SR (Table I) (29). The intercalated disc-containing fraction is contaminated mainly by mitochondrial membranes (Fig. 5). However, the IP, binding found in this fraction is not referable to mitochondria, since IP, binding to purified mitochondria and submitochondrial vesicles is negligibly small (Table I). These results strongly support the immunohistochemical studies. We conclude that the IPS receptor is associated with the intercalated discs, and its intracellular localization is different from that of the ryanodine receptor.

The presence of the IP, receptor in heart has previously been indicated from positive Western and Northern blot analyses using whole heart tissue (14-16). However, such studies do not necessarily mean that the IP3 receptor is localized within the cardiomyocytes, since ventricular myo- cytes account for only 20% of the cell population, albeit 80% of the mass of the heart (39). Heart also contains other types of cells, e.g. endothelial cells, vascular smooth muscle cells, neuronal cells, and fibroblasts. Immunohistochemistry is a more decisive way to study cellular and subcellular localiza- tion (Figs. 2-4).

The ryanodine receptor is the predominant intracellular Ca2+ release channel in heart ventricle and is involved in the cardiac excitation-contraction coupling. At the level of the light microscopy, strong immunofluorescence of the ryanodine receptor is observed as transverse banding corresponding to I-bands (Fig. 3), whereas immunofluorescence attributable to the IP3 receptor is only faintly observed at the intercalated discs (data not shown). With a more sensitive procedure than immunofluorescence, i.e. using immunostaining with peroxi- dase-generated reaction product (Fig. 2), immunoreactivity of the IP3 receptor could be better detected although sunwwhat

3504 Localization of IP3 and Ryanodine Receptors in Heart I". .. . ..- . 1 ,. . . ~" ~ ~

% .. ..""

FIG. 4. Immunolocalization of the IPS receptor in rat cardiomyocytes by electron microscopy. Thin sections of rat heart ventricular (a-c) and atrial ( d ) tissue embedded in Lowicryl K4M were immunolabeled with the anti-IPC3 antibody and goat anti-rabbit IgG gold conjugate (10-nm diameter). a, control (non-immune) serum (X 30,000). Immunogold particles are not observed. b, intercalated discs of plasmalemma were specifically decorated with immunogold (10 nm) (X 60,000). c, surface sarcolemma ( S L ) is negative, whereas the intercalated disc is positive (arrowhead) (X 60,000). d , atrial tissue stained with the anti-IPR receptor antibody (X 60,000). The immunogold particles are observed at the intercalated discs (arrowhead). Note the characteristic secretory granules (arrows) in the atrial cardiomyocytes. Surface sarcolemma of the atrial cardiomyocyte is negative (data not shown).

Localization of IP3 and Ryanodine Receptors in Heart 3505

FIG. 5. Electron Microscopy of the intercalated disc-containing fraction from dog heart ventricle. a, a complete top-to-bottom section was visualized by preparing a thin layer of the sample on a Millipore filter (25). Mitochondrial fragments can also be found in this fraction (X 18,000). b and c, higher magnification (X 60,000) of a. Positions of asterisks (*) correspond to those in a. Pairs of electron dense “mats” (arrowhead) are characteristics of fascia adherens.

less than half of intercalated discs were positive. In contrast to the level of light microscopy, immunogold particles directed to the IP3 receptor were consistently found associated with all of intercalated discs (Fig. 4). This difference in quantita- tion is likely due to the differences in fixation and embedding procedures used in light uersus electron microscopy (paraffin versus Lowicryl K4M): ie . 1) non-uniform penetration of the anti-IP3 receptor antibody into 5-pm thick paraffin sections versus <O.l-pm thick section in electron microscopy; and/or 2) alteration or attenuation of the epitope antigenicity in paraffin sections. Nonetheless, light microscopy allowed us to view large numbers of fields and showed generalized staining.

A number of immunohistochemical studies have been car- ried out on cerebellar Purkinje cells where the IP3 receptor is highly enriched (3, 17, 40). At the level of light microscopy using immunoperoxidase (3), the immunoreactivity of the IP3

receptor was reported to be observed not only in cytoplasm but also closer to plasma membranes of cerebellar Purkinje cells. However, other studies using immunogold electron mi- croscopy, which has superior spatial resolution, showed that the IP3 receptor is an intracellular Ca2+ release channel (17, 40). In cerebellar Purkinje cells, the IP3 receptor and ryano- dine receptor coexist in the same intracellular membrane organelles (17). We do not observe such subcellular colocali- zation of the two receptors in cardiomyocytes. Our results indicate that the cardiac IPS receptor has a different subcel- lular localization (in intercalated discs) from the ryanodine receptor (in junctional SR). We cannot preclude the possibil- ity that a small fraction of the total IPS receptor might exist in other intracellular organelles in cardiomyocytes such as SR or surface sarcolemma.

The intercalated discs of cardiomyocytes is composed of

3506 Localization of IP3 and Ryanodine Receptors in Heart

TABLE I Distribution of fH]IP, and fH]ryanodine binding in cardiac subcellular fractions

Suhfractions" [3H]IP3 binding* (3H]Ryanodine hindine'

(Na,K)-ATPase activitvd

Longitudinal SR Junctional SR Sarcolemma Intercalated disc-containing fraction Mitochondria Submitochondrial vesicles Atrial microsomes

pmol/mg 0.121 f 0.010' 0.52 0.163 f 0.006' 3.31 0.169 f 0.004' 1.67 9.43 f 0.22 0.22

0.064 f 0.004 NSf 0.052 f 0.003 NS 0.160 ? 0.010 0.84

pmol/mg. h 3.8

13.0 52.6 2.5 4.2 2.5

14.6 Longitudinal SR, junctional SR, sarcolemma, and the intercalated disc-containing fraction were prepared from canine heart ventricles.

Atrial microsomes were prepared from canine heart atria, and mitochondria and submitochondrial vesicles from bovine heart ventricles. *Values of [3H]IP3 binding (10 nM of ligand concentration) are the mean k S.D. of three measurements. Nonspecific binding to the

intercalated disc-containing fraction was -10% of the total binding. e Values of [3H]ryanodine binding (60 nM of ligand concentration) represent the average of the two measurements in which the difference

from the mean is less than 4% of the specific binding to junctional SR. Nonspecific binding to junctional SR was -50% of the total binding. (Na,Kj-dependent ATPase activity was measured a t 37 "C after unmasking the latent ATPase activity by preincubation with SDS. Nonspecific binding was determined in the presence of 1 mg heparin/ml. Negligibly small.

three distinct types of junctions, i.e, desmosome (macula 10. Vigne, P., Breittmayer, J.-P., Marsault, R., and Frelin, C. (1990) J. Biol.

adherens), intercellular adherens junction (fascia adherens), 11. Vites, A,"., and Pappano, A. (1992) Am. J. Physiol. 2 6 2 , H268-277

and gap junction (nexus) (38). The gap junction an 12. Kentish, J. C., Barsotti, R. J., Lea, T. J., Mulligan, I. P., Patel, J. R., and

important role in intercellular ekCtriCa1 Coupling, so that 13. Borgatta, L., Watras, J., Katz, A. M., and Ehrlich, B. E. (1991) Proc. Natl. heart is an syncytium* Fascia adherens and 14. Furuichi, T., Shiota, C., and Mikoshiba, K. (1990) FEBS Lett. 267,85-88 desmosomes have been thought to have a structural role in 15. Marks, A. R., Tempst, P., Chadwick, C. C., Riviere, L., Fleischer, S., and

association' In this study another function Of 16. Mignery, G. A,, Newton, C. L., Archer, B. T., 111, and Siidhof, T. C. (1990) fascia adherens is suggested, i.e. regulation of intracellular J. Biol. Chem. 2 6 5 , 12679-12685 [Ca"] in cardiomyocytes. T-lymphocyte activation, caZ+ 17. Walton, p. D., Airey, J. A., Sutko, J. L., Beck, C. F., Mignery, G . A.,

entry mediated by the plasma membrane IP, receptor is Siidhof, T. C., Deerinck, T. J., and Ellisman, M. H. (1991) J. Cell Bid. 113 , 1145-1157

believed to play an key role (41, 42). A possible role of the 18. Mignery, G. A., Siidhof, T. C., Takei, K., and DeCamilli, P. (1989) Nature

cardiac IPS receptor may be to mediate Ca2+ entry and/or to 19. Imagawa, T., Takasago, T., and Shigekawa, M. (1989) J. Biochem. (Tokyo) 342,192-195

modulate intercellular communication via the intercalated 20, Laemmli, u, K, (1970) Nature 227, 680-685 discs. 21. Towhin, H. T., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci.

receptor serves a key role as a ca2+ release channel of SR. 23. Saito, A., Wang, C.-T., and Fleischer, S. (1978) J. Cell Bid. 79,601-616 These studies show that the cardiac ~ p , receptor is localized 24. Saito, A,, Seiler, S., and Fleischer, S. (1984) J. Ultrastruct. Res. 8 6 , 277-

at intercalated discs in ventricular and atrial cardiomyocytes. 25. Palade, P., Saito, A., Mitchell, R. D., and Fleischer, S. (1983) J. Histoehem. Its precise role(s) remains to be defined. 26. Kensler, R. W., and Goodenough, D. A. (1980) J. Cell Biol. 8 6 , 755-764

27. Chamberlain, B. K., Levitsky, D. O., and Fleischer, S. (1983) J. Biol. Chem.

of T~~~~ Southwestern Medical Center) for the gift of the anti-Ip3 28. Kijima, Y., Ownbunmi, E., and Fleischer, S. (1991) J. Bid. Chem. 266 ,

receptor antibody, Drs. Toshiaki Imagawa and Munekazu Shigekawa 29. Inui, M., Wang, s., Saito, A,, and Fleischer, S. (1988) J. B ~ O L Chem. 2 6 3 , (National Cardiovascular Center Research Institute, Japan) for the 10843-10850

Dr. J. Oliver McIntyre (Vanderbilt for the gift of bovinl 31. Fleischer, S., Meissner, G., Smigel, M., and Wood, R. (1974) Methods gift of the monoclonal antibodies to the cardiac ryanodine receptor 30. Bock, H. o., and Fleischer, s. (1974) Methods EnzYmol. 31,374-391

mitochondrial fractions, and Dr. Phillip E. Williams (Vanderbilt 32. Palade, P., Dettharn, C., Volpe, P., Alderson, B., and Otero, A. S. (1989) University) for providing dog hearts and brains used in this study. Mol. Pharmacol. 36,664-672

33. Fleischer, S., Ogunhunmi, E. M., Dixon, M. C., and Fleer, A. M. (1985)

34. Jones, L. R., Besch, H. R., Jr., Fleming, J. W., McConnaughey, M. M., and

35. Bradford, M. M. (1976) Anal. Biochem. 72,248-254

Chem. 266,6782-6787

Ferenczi, M. A. (1990) Am. J. Physiol. 2 5 8 , H610-H615

Acad. Sci. U. S. A. 88,2486-2489

Nadal-Ginard, B. (1990) J. Biol. Chem. 265,20719-20722

106,342-348

In excitation-contraction coupling of heart, the ryanodine 22, Roth, J. (1989) Methods cellBiol, 31, 513-551 U. S. A. 76,4350-4354

293

Cytochem. 31,971-974

Acknowledgments-We thank Dr. Thomas C. Siidhof (University 268,6602-6609

22912-22918

Enzymol. 31,292-299

REFERENCES Proc. Natl. Acad. Sci. U. S. A. 8 2 , 7256-7259

1. Fleischer, S., and Inui, M. (1989) Annu. Reu. Biophys. Biophys. Chem. 18 , Watanabe, A. M. (1979) J. Biol. Chem. 254,530-539

2. Chadwick, C. C., Saito, A., and Fleischer, S. (1990) Proc. Natl. Acad. Sci. 36. Inui, M., Saito, A,, and Fleischer, 8. (1987) J. Biol. Chem. 2 6 2 , 15637-

3. Maeda, N., Niinobe, M., and Mikosbiba, K. (1990) EMBO J. 9 , 61-67 37. Kawamoto, R. M., Brunschwig, J.-P., Kim, K. C., and Caswell, A. H. (1986) 4. Berridge, M. J., and lrvine, R. F. (1989) Nature 341 , 197-205 5. Ferris, C. D., Huganir, R. L., Supattapone, S., and Snyder, S. H. (1989) 38. Severs, N. J. (1990) Int. J. Cardiol. 26,137-173

6. Brown, J. H., Buxton, I. L., and Brunton, L. L. (1985) Circ. Res. 5 7 , 532- 40. Satoh, T., Ross, C. A., Villa, A., Supattapone, S., Pozzan, T., Snyder, S. H.,

7. Scholz, J. , Schaefer, B., Schmitz, W., Scholz, H., Steinfath, M., Lohse, M., 41. Kuno, M., and Gardner, P. (1987) Nature 3 2 6 , 301-304

333-364

U. S. A. 87,2132-2136 15642

J. Cell Biol. 103 , 1405-1414

Nature 342,87-89 39. Wientzek, M., and Katz, S. (1991) J. Mol. Cell. Cardiol. 2 3 , 1149-1163

537 and Meldolesi, J. (1991) J. Cell Biol. 111,615-624

Schwahe, U., and Puurunen, J. (1988) J. Pharmacol. Exp. Ther. 2 4 6 , 42. Kahn, A. A,, Steiner, J. P., Klein, M. G., Schneider, M. F., and Snyder, S.

8. Vigne, P., Lazdunski, M., and Frelin, C. (1989) FEBS Lett. 249 , 143-146 43. Fahiato, A. (1990) in Recent Advances in Calcium Channels and Calcium 327-335 H. (1992) Science 267,815-818

9. Eckel, J., Gerlach-Eskuchen, E., and Reinauer, H. (1991) J. Mol. Cell. Antagonists (Yamada, K., and Shibata, S., eds) pp. 35-41, Pergamon Press, New York Cardiol. 2 3 , 617-625