Identification and functional reconstitution of the type 2 inositol 1,4,5-trisphosphate receptor from ventricular cardiac myocytes.

The inositol 1,4,5-trisphosphate receptor (InsP3R) is an intracellular Ca2+ release channel that mediates the rise in cytoplasmic calcium in response to receptor-activated production of InsP3. The InsP3R-mediated signaling pathway appears to be ubiquitous and is involved in many cellular processes including cell division, smooth muscle contraction, and neuronal signaling. Different regions of the heart also express InsP3 receptors. We report here that acutely dissociated ventricular myocytes from ferret and rat hearts express significant levels of InsP3R as indicated by immunoblotting with a receptor consensus antibody. InsP3 binding experiments (KD = 23.6 nM and Bmax = 0.46 pmol/mg) suggest the myocytes contain the high affinity type 2 InsP3 receptor. Exhaustive mRNA screening by polymerase chain reaction, RNase protection, and subsequent DNA sequencing positively identify the InsP3R as type 2. The type 2 receptor from ferret heart was then incorporated into planar lipid bilayers and formed Ca2+-selective, InsP3-activated, heparin-blocked ion channels. We conclude that the predominant InsP3 receptor isoform expressed in cardiac myocytes is type 2 and that it forms a functional InsP3-gated Ca2+ channel when reconstituted in planar lipid bilayers.

Inositol 1,4,5-trisphosphate (InsP 3 ) 1 is a well known second messenger mediating the regulated release of intracellular calcium and is produced through the action of phospholipase C. Phospholipase C is activated in response to the stimulation of cell surface receptors coupled to heterotrimeric G-proteins and tyrosine kinases resulting in the hydrolysis of phosphatidylinositol 4,5-bisphosphate to liberate InsP 3 and diacylglycerol. InsP 3 is a readily diffusible compound that binds to specific (InsP 3 R) receptors localized to the endoplasmic reticulum and results in the release of calcium from intracellular stores (see reviews Refs. 1-3).
The InsP 3 receptor consists of a family of 3-4 highly homol-ogous members. The primary structure of these InsP 3 Rs has been determined by cDNA cloning and sequencing (4 -7). The type 1 receptor is expressed at very high levels in cerebellar Purkinje cells and has been extensively characterized (4,6,8,9). When compared with the type 1 receptor, the type 2 and type 3 isoforms have an overall amino acid sequence identity of 69 and 64%, respectively (6,7). Even though the three principal types of receptor are very similar, they exhibit significantly different affinities for InsP 3 binding (designated T-2 Ͼ T-1 Ͼ Ͼ T-3), suggesting that the different receptor homologues have distinct functions within a cell (6,10). In cardiac muscle, the ryanodine receptor serves as the primary calcium release channel of the sarcoplasmic reticulum in excitation-contraction coupling (11,12). The role of InsP 3 -induced calcium release in cardiac cells is not well understood. Phosphatidylinositol 4,5-bisphosphate turnover coupled to ␣-adrenergic and muscarinic plasma membrane receptors has been shown to increase InsP 3 levels in cardiac muscle (13,14). Initial studies revealed that InsP 3 was capable of inducing a slow release of calcium from vesicular preparations as well as activate contraction in skinned ventricular rat muscle and chick heart preparations (15). Borgatta et al. (16) identified a low conductance, InsP 3 -sensitive, calcium release channel in sarcoplasmic reticulum vesicle preparations from canine heart. These channels were concentrated in the sarcoplasmic reticulum isolated from the ventricle septum that contains cells of the conducting system. Recently, biochemical and immunological approaches have been applied to identify the InsP 3 receptor in cardiac tissues. Moschella and Marks (17) observed InsP 3 R in rat heart using in situ hybridizations and immunocytochemical analysis and concluded that the receptor expressed in cardiac myocytes is structurally similar to the type 1 receptor. The expression level was approximately 50-fold lower than that of the cardiac ryanodine receptor. Gorza et al. (18,19) observed the highest levels of expression in Purkinje myocytes of the conduction system using type 1 (cerebellar) peptide antibodies and cRNA probes. The InsP 3 receptor was localized to the intercalated disc of cardiac myocytes using immunogold electron microscopy (20). These studies suggested that the receptor may be involved in cell-cell signaling or potentially calcium influx.
The biological role of this receptor family in cardiac tissues is unclear. Although the InsP 3 receptor is not the primary calcium release channel in cardiac tissue, InsP 3 may be involved in the regulation of Ca 2ϩ -induced Ca 2ϩ release. Increased InsP 3 and the resulting increase in cytoplasmic free calcium may be an important mechanism for controlling cardiac contractile force in response to hormones and pharmacological factors as well as in the diseased state. A greater than 2-fold increase in steady state InsP 3 receptor mRNA was observed during end-stage heart failure in human (21). This increase was accompanied by a 31% decrease in the ryanodine receptor * 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.
In this study, we have examined the expression of InsP 3 receptor in rat and ferret ventricular cardiac myocytes using immunological, molecular, and electrophysiological techniques. We show that in both rat and ferret myocytes the principal isoform of the InsP 3 receptor expressed is the type 2 homologue. Functional reconstitution of this receptor isoform into planar lipid bilayers resulted in the identification of a Ca 2ϩ release channel responsive to low levels of InsP 3  Membrane Preparation-Rat and ferret ventricular cardiac myocytes were isolated as described by Bassani et al. (22). Acutely dissociated ventricular myocytes were homogenized in 50 mM Tris-HCl, pH 8.3, 1 mM EDTA, 1 mM PMSF, 1 mM ␤-mercaptoethanol with 30 pestle strokes in a glass/Teflon Potter-Elvehjem tissue grinder. Large particulate was sedimented with a 2-min 325 ϫ g centrifugation. The resulting supernatant fraction was transferred to TLA 100.3 ultracentrifuge tubes (Beckman) and sedimented at 135,240 ϫ g max at 4°C for 10 min to yield crude microsomes. The microsomal pellet was resuspended in 50 mM Tris-HCl, pH 8.3, 0.6 M KCl, 1 mM EDTA, 1 mM PMSF, 1 mM ␤-mercaptoethanol and incubated on ice for 10 min. The microsomes were pelleted with a 10-min 135,240 ϫ g max centrifugation, resuspended, and washed in 50 mM Tris-HCl, pH 8.3, 1 mM EDTA, 1 mM PMSF, 1 mM ␤-mercaptoethanol three times. Washed microsomes were resuspended to approximately 5-10 mg/ml protein and either used immediately or stored at Ϫ80°C until use.
SDS-Polyacrylamide Gel Electrophoresis and Immunoblotting-SDS-polyacrylamide gel electrophoresis and immunoblotting were performed as described (23,24) using 5% SDS-polyacrylamide gels. Antibodies were directed against the 19 carboxyl-terminal (type 1-specific) (24) amino acid residues 2463-2476 (type 1 loop) (25) and the bacterially expressed amino-terminal 324 amino acids (receptor consensus) of the type 1 receptor (26). Peptide competition experiments for the type 1 carboxyl-terminal and type 1 loop antisera were performed by preincubation in specific peptide (1 mg/ml) for 1 h prior to incubation with the immunoblot.
InsP 3 Saturation Binding Measurement-InsP 3 binding to ferret myocyte membrane preparations was performed in the presence of increasing concentrations of [ 3 H]InsP 3 ranging from 0.42 to 50.5 nM. Each 100 l assay contained 100 g of protein in 50 mM Tris-HCl, 1 mM EDTA, 1 mM ␤-mercaptoethanol, pH 8.3, using the centrifugation binding assay as described previously (26). The binding assays were incubated on ice for 10 min in the presence of [ 3 H]InsP 3 and then microsomes were pelleted with a 10-min 289,000 ϫ g av centrifugation. The supernatants were removed from the assay tube, and the membrane pellet was solubilized in 1% SDS and counted in a Beckman liquid scintillation counter. All assays were performed in triplicate, and nonspecific 3 [H]InsP 3 binding for each concentration was determined in the presence of 10 M unlabeled InsP 3 . Similar results were obtained from three independent assays.
Reverse Transcription/PCR Analysis-The procedures for reverse transcription and PCR analysis are essentially as described by Newton et al. (10). Total RNA was isolated from rat and ferret ventricular myocytes, as well as rat left ventricular myocardium, whole heart, and cerebellum as described previously by Perin et al. (27). Five g of total RNA from rat and ferret samples was used as template for first strand cDNA synthesis. Each 50-l reaction contained 500 M dNTPs (Pharmacia Biotech Inc.), 20 units of RNase inhibitor (Promega), 1 g of pd(N) 6  Reactions were incubated at 37°C for 60 min and heat-inactivated for 5 min at 95°C.
Southern Blotting-PCR products derived from rat and ferret total RNA primed with the InsP 3 receptor consensus oligonucleotides were resolved on triplicate 1% agarose gels and blotted to individual nitrocellulose membranes as described by Sambrook et al. (28). Receptor type-specific control PCR products were amplified using the same common-primer pair and were generated from cDNA templates pI7, pI15, and pI924-2 corresponding to the type 1, 2, and 3 InsP 3 receptor, respectively (5,6,10). Blots were hybridized to receptor type-specific 32 P-labeled oligonucleotides and washed at high stringency. The receptor type-specific oligonucleotides used for the hybridizations correspond to sequences within the PCR product as follows: type 1, CGCATCGA-TGGTCTTGTTGGCCTCTTTGGATGGCTTCCT; type 2, CCTCCCTCA-CCGGCTGCATTCGAAGA; and type 3, AGGGCTTGCTTAGAATGT-CGC.
RNase Protection Assays-Ribonuclease protection assays were performed using a HybSpeed RPA kit (Ambion) as described by the manufacturer. Receptor type-specific transcription plasmids were constructed by PCR amplification of nucleotides 7664 -8186, 7448 -7959, and 7254 -7772 of the rat types 1, 2, and 3 isoforms, respectively, and insertion into EcoRI and BamHI digested pGEM-3Z (Promega). The oligonucleotide primers consisted of CGGGATCGGCTCTGATTCTGG-TTTACCTGTTCTC and CGGAATTCTTGTCCCTTTCCAAGCCACAA-ATGAAGCA for the type 1, CGGGATCCTGGCTCTTATCCTGGTCTA-CCTGTTCTC and CGGAATTCTTGTCCCTCTCCAAGCCACAGATGA-AGCA for the type 2, and CGGGATCCTGGCCCTCATCCTTGTCTAC-CTCTTCTC and CGGAATTCTTGTCCCTCTCCAGACCACAGATGAA-GCA for the type 3 receptors. InsP 3 R-specific antisense probes were generated from these plasmids, linearized with BamHI, using T7 phage RNA polymerase and the MAXIscript (Ambion) in vitro transcription kit in the presence of [ 32 P]UTP. The expected run-off transcription product sizes were 531, 526, and 528 nucleotides for types 1, 2, and 3 receptors, respectively. Transcription products were purified on 5% acrylamide, 8 M urea denaturing gels. Typical specific activities of the probes were greater than 1.3 ϫ 10 9 dpm/g and varied less than 5% between isoform-specific probes. The ␤-actin control antisense probe was derived from linearized pTRI-␤-actin-mouse plasmid DNA. Ribonuclease protection assays were performed in the presence of 80,000 cpm each InsP 3 R-specific and mouse ␤-actin antisense probes using 10 g of total RNA for all cardiac samples and with 2.5 g of total RNA from cerebellum. Products from the RNase protection reactions were resolved on 5% acrylamide, 8 M urea denaturing gels, dried, and autoradiographed or quantitated using a Packard Instant Imager running Packard Imager for Windows V2.03.
PCR Product Cloning and Sequencing-PCR products from rat and ferret RNA were digested with HindIII and ligated into similarly cut pBluescript SK. DNA sequencing was performed on double-stranded templates using 35 S-dATP and Sequenase version 2.0 sequencing reagents (U. S. Biochemical Corp.).
Reconstitution of InsP 3 Receptor into Liposomes-Microsomes from ferret ventricular cardiac myocytes were solubilized in 1% CHAPS and fractionated on 5-20% sucrose gradients as described previously (5,8). Gradient fractions containing the highest levels of receptor protein were identified by Western blotting and then reconstituted into proteoliposomes essentially as described by Ferris et al. (29).
Planar Lipid/Protein Bilayer Formation-Planar lipid bilayers were formed across a 220-m diameter aperture in the wall of a Delrin partition. Lipid bilayer-forming solution contained a 7:3 mixture of phosphatidylethanolamine and phosphatidylcholine dissolved in decane (50 mg/ml). Proteoliposomes were added to the solution on one side of the bilayer (defined as cis). The other side was defined as trans (virtual ground). Standard solutions contained 220 mM CsCH 3 SO 3 cis (20 mM trans), 10 mM HEPES, pH 7.4, and 1 mM EGTA ([Ca 2ϩ ] free ϭ 200 nM). The [Ca 2ϩ ] free was verified using a Ca 2ϩ electrode. A custom current/voltage conversion amplifier was used to optimize single-channel recording (12). Acquisition software (pClamp: Axon Instruments, Foster City, CA), an IBM compatible 486 computer, and a 12-bit A/D-D/A converter (Axon Instruments) were used. Single channel data were digitized at 5-10 kHz and filtered at 1 kHz. Channel sidedness was determined by InsP 3 sensitivity. The orientation of the channels studied was such that the InsP 3 -sensitive side (i.e. cytoplasmic side) was in the cis compartment.
The ion selectivity of previously reconstituted single InsP 3 R channels was nearly identical to that of single ryanodine receptor channels (30). Thus, a simple reconstitution strategy analogous to that successfully applied to study ryanodine receptor channels (31) was used here to examine single InsP 3 R channels. This strategy utilized a monovalent cation (Cs ϩ ) instead of a divalent cation as charge carrier. The use of a monovalent charge carrier improves experimental success rate and increases the single-to-noise ratio.

Expression of InsP 3 Receptor in Cardiac Ventricular
Myocytes-To define the expression of InsP 3 receptor(s) in ventricular cardiac myocytes, microsomal protein from both rat and ferret were electrophoresed on 5% SDS-polyacrylamide gels along with bovine cerebellar microsomal proteins as a control. Following electrotransfer to nitrocellulose membranes, the blots were either incubated with antibodies directed against the carboxyl-terminal 19 amino acids (24), amino acid residues 2463-2476 (type 1 loop) (25), or the amino-terminal 324 amino acids of the type 1 InsP 3 R. The antisera derived from the amino-terminal 324 amino acids is directed against the highly conserved ligand binding domain and reacts with all receptor types tested (10,26).
The type 1-specific carboxyl-terminal antisera elicited a strong signal in proteins from bovine cerebellar microsomes ( Fig. 1, 1st panel, 1st lane) but failed to detect the presence of any immunologically cross-reactive species of similar size in the myocyte preparations of ferret ( Fig. 1, 1st panel, 2nd lane). The signal observed in the cerebellar sample was generated from one-tenth (10 g) the amount of total membrane protein applied to the myocyte sample wells. Microsomes from rat and ferret cerebella were also immunoreactive (not shown). An immunoreactive signal was observed at M r Ӎ 97,700 in ferret myocytes which was eliminated by pre-absorbing the carboxylterminal antibody with its corresponding peptide (Fig. 1, 2nd  panel). To investigate further whether the immunoreactive signal (M r Ӎ 97,700) observed in the myocyte preparation with the carboxyl-terminal antisera represents a proteolytic fragment of the type 1 receptor, a second type 1-specific antibody directed against amino acid residues 2463-2476 (type 1 loop) was used for immunoblotting ( Fig. 1, 3rd panel). In this panel, the type 1 receptor of cerebellum is strongly immunoreactive with no apparent signal of similar mobility as observed in the myocyte sample. As was seen with the type 1 carboxyl-terminal antisera, a different band of lower apparent M r (Ӎ148,000) is seen which could be eliminated in peptide competition experiments ( Fig. 1, 4th panel). The band of M r Ӎ97,700 observed using the carboxyl-terminal antisera is not detected with this antibody and suggests that the cross-reacting species is not the type 1 receptor. This conclusion is based on the observation that the carboxyl-terminal antibody is reacting with a protein species of approximately 97.7 kDa, and if this was a proteolytic fragment of the type 1 receptor it would cross-react with the type 1 loop antibody which is directed against residues 2463-2476 only 267 amino acids NH 2 -terminal to the COOH-terminal antibody and thus should reside on the same proteolytic fragment. The identity of the protein species which was faintly detected by the type 1 "loop" antibody remains uncertain. The weak signals observed with this and the COOH-terminal antisera may be a consequence of the large amounts of myocyte membrane proteins loaded onto the SDS-PAGE gels.
When the Western blots were incubated with the consensus antisera derived from the receptor amino-terminal 324 amino acids, there was a significant signal observed at the expected apparent molecular weight (M r ϭ 260 ϫ 10 3 ) for the InsP 3 receptor in both cerebellum and ferret cardiac myocytes (Fig. 1,  5th panel). As judged by the signal intensity between the myocytes and cerebellar protein samples, these data suggest that there is a significant amount of InsP 3 receptor protein present in cardiac myocytes and that the predominant isoform expressed is not type 1. It is unlikely that this signal represents the type 3 InsP 3 R due to its similar mobility to the cerebellar type 1 receptor on SDS-PAGE, since the type 3 receptor has a significantly smaller apparent M r (10).
Inositol 1,4,5-Trisphosphate Binding to Cardiac Ventricular Myocyte Membrane Preparations-The ligand binding properties of the isoform expressed in cardiac myocytes was investigated using saturation binding measurements. Even though the various InsP 3 receptor family members exhibit a remarkable degree of sequence similarity, they have markedly distinct affinities for InsP 3 . The type 2 InsP 3 receptor has the highest InsP 3 affinity (K D ϭ 27 nM) of the three characterized isoforms with a relative order of, type 2 Ͼ type 1 Ͼ Ͼ type 3 (6, 10). In Fig.  2, [ 3 H]InsP 3 binding to ferret myocyte membrane preparations was measured in the presence of increasing concentrations of [ 3 H]InsP 3 ranging from 0.42 to 50.5 nM in 50 mM Tris-HCl, 1 mM EDTA, 1 mM ␤-mercaptoethanol, pH 8.3, using the centrifugation binding assay as described previously (26). All assays were performed in triplicate and nonspecific [ 3 H]InsP 3 binding for each concentration was determined in the presence of 10 M unlabeled InsP 3 . Similar results were obtained from three independent determinations.
Analysis of the binding data using the Scatchard transformation results in a binding isotherm with a calculated K D equal to 23.6 nM and a B max ϭ 0.46 pmol/mg (Fig. 2). These values are very similar to values previously reported (21 nM, 0.66 pmol/mg) with canine cardiac microsomes (32) and that of the cloned and expressed type 2 InsP 3 receptor ligand binding domain (27 nM) (6). Although significant variability exists for the reported K D values of the InsP 3 receptors due to differences in purity and experimental conditions, the values obtained in this study are consistent with those of previous studies (6) where membrane preparations and assays were performed us- ing similar methodologies. These results (Fig. 2), together with the Western blotting data (Fig. 1), suggest that the predominant species of InsP 3 R in cardiac ventricular myocytes is not the type 1 or type 3 InsP 3 receptor and is consistent with the hypothesis that it may represent the type 2 isoform.
Polymerase Chain Reaction Analysis of InsP 3 Receptor Expression-To establish the identity of the InsP 3 R isoforms expressed in heart, polymerase chain reaction (PCR) screening strategies were performed. Several studies (10, 33) have characterized the expression of the InsP 3 receptor in various tissues and cell lines using PCR strategies. Multiple homologues of InsP 3 receptor were detected in all tissues and cell types examined. Within a particular cell type, the receptors detected are usually the type 1 and type 3 isoforms. Newton et al. (10) observed that in many cultured cell lines, the type 2 receptor is not expressed at any significant level and proposed that its distribution may be limited as compared with the other receptor types.
To establish the pattern of the InsP 3 receptor mRNA expression in ventricular cardiomyocytes, total RNA was prepared from isolated cells essentially as described by Perin et al. (27). The total RNA was reverse-transcribed and used as templates in a PCR assay in which "consensus" oligonucleotides that recognize all three principal types of InsP 3 receptor cDNA sequences were used as primers. These amplification products, along with specific receptor type products from cloned rat sequences (5, 6, 10), were resolved on non-denaturing agarose gels, transferred to nitrocellulose, and subjected to hybridization with 32 P-labeled receptor type-specific oligonucleotide probes internal to the original primer pair. The results are illustrated in Fig. 3A and demonstrate that in both rat and ferret ventricular myocytes, the type 2 receptor isoform transcripts are the predominant InsP 3 receptor species present. Extended autoradiographic exposure of the filters resulted in weak signals detected in all lanes, including those of the control amplification products and was interpreted as nonspecific hybridization signal (data not shown).
To establish the expression pattern of the types 1-3 InsP 3 receptor species in cardiac tissues, total RNA from rat cardiac myocytes, left ventricular myocardium, whole heart, and cerebellum was reverse-transcribed and used as templates for PCR amplification with the consensus oligonucleotide primers as described above. As illustrated in Fig. 3B the types 1-3 receptor species can be amplified from left ventricular myocardium, whole heart, and cerebellum. The whole heart and cerebellum data are consistent with previously published results. Thus, these data show 1) the PCR primers are subtype-specific, 2) all of the receptor isoforms are expressed in whole heart, and 3) only the type 2 receptor was detected in isolated myocytes.

Analysis of InsP 3 R Isoform Expression Levels by
Ribonuclease Protection-To establish the relative levels and patterns of expression of the InsP 3 R gene family in myocytes and other cardiac tissues, ribonuclease protection assays were performed using InsP 3 R subtype-specific antisense RNA probes. InsP 3 receptor subtype-specific antisense probes corresponding to nucleotides 7664 -8186, 7448 -7959, and 7254 -7772 of the rat types 1-3 receptors, respectively, were synthesized using T7 phage RNA polymerase from linearized pGEM-3Z vector. These probes were hybridized to 10 g of total RNA from rat whole heart, left ventricular myocardium, acutely dissociated myocytes, and cerebellum (2.5 g of total RNA) and subjected to RNase A/T1 digestion. As illustrated in Fig. 4A, the type 1 probe strongly protects RNA from cerebellum comprising approximately 98.5% of the total InsP 3 R transcripts detected (see Table I). The type 2 and type 3 receptor accounts for approximately 1.3 and 0.2% of cerebellar transcripts, respectively. In whole heart and left ventricular myocardium, the type 1 receptor predominates the receptor steady state transcript pool (ϳ60%) with the type 2 (ϳ27%) and type 3 (ϳ13%) at lower levels ( Fig. 4A and Table I). The discrepancy between these quantitative assays and the PCR screening (Fig. 3B) most likely reflects affinity differences between the consensus oligonucleotides and the geometric nature of the PCR amplification. In either case, however, both assays clearly demonstrate the expression of all three isoforms in whole heart and myocardium.  3. Analysis of the mRNA expression and distribution of the type 1, type 2, and type 3 InsP 3 receptor in rat and ferret cardiac tissues as determined by PCR using primers that amplify all three receptor sequence types. A, PCR amplifications were performed on single-stranded cDNA prepared from rat and ferret ventricular myocyte RNA templates. B, PCR analysis of cDNA from rat ventricular myocytes, left ventricular myocardium (myocardium), whole heart, and cerebellum. PCR reaction controls consisted of a minus-RNA cDNA reaction (negative control) and cloned InsP 3 R cDNA type 1, type 2, and type 3 templates (amplification and hybridization controls). PCR reaction products were resolved on 1% agarose gels and Southern blotted with receptor type-specific 32 P-labeled oligonucleotide probes internal to the PCR primer pair as indicated.
In acutely dissociated myocytes, the predominant protected receptor species corresponds to the type 2 receptor (85%), with only minor contributions from the type 1 and type 3 receptor (1.9 and 13%, respectively) ( Fig. 4A and Table I). The high levels of the type 2 receptor in myocytes is consistent with the immunoblotting as well as the InsP 3 binding results and suggests that the other receptor isoforms are expressed at significant levels in other cell types of the heart such as smooth muscle and the conducting myocytes such as Purkinje cells (18,19). Fig. 4B illustrates RNase protection analysis at several stages of myocyte preparation. The experiments were intended to examine the fractionation of InsP 3 R subtypes during myocyte isolation. With incompletely dissociated myocardium the receptor distribution is essentially as that observed for whole heart and intact ventricular myocardium (Fig. 4A), with the type 1 receptor predominating at approximately 56% of the total receptor population. However, in unwashed myocytes (Ϫ4, 1 ϫ g sedimentations/platings in normal Tyrode buffer), the relative levels of receptor isoforms are dramatically shifted with the type 2 isoform predominating at greater than twice that observed in whole heart and myocardium (Fig. 4, A and B, and Table I). In washed myocytes (Fig. 4, A and B, identical panels) the type 2 receptor levels are further enriched and represent approximately 85% of the total receptor population with the type 1 receptor at almost undetectable levels and the type 3 receptor at 13%. It is possible that some RNA degradation occurs during the washing process; however, based on the integrity of the protected band and the ␤-actin signal, intensity between washed and unwashed myocytes in non-saturating exposures suggests that it is minimal. In addition, the levels of the type 3 receptor RNA remain almost constant throughout the isolation.
Sequence Analysis of PCR Products-To definitively establish that the type 2 receptor is the predominant isoform expressed in rat and ferret ventricular myocytes, products from the PCR reactions designed to amplify all InsP 3 receptor subtype mRNA were cloned and subjected to DNA sequence analysis.
The nucleotide sequences derived from four independent clones of the rat ventricular myocyte PCR products are identical to one another and reveal a very high degree of similarity (99.4%) with the previously reported type 2 InsP 3 receptor from rat brain (6) (Fig. 5). The three nucleotide changes observed between the published full-length rat type 2 sequence and those from rat cardiac myocytes are all silent changes and encode identical amino acid sequences.
A total of six ferret clones were sequenced and found to be very homologous to those of the rat type 2 sequences from brain and cardiac myocytes (Fig. 5). These clones exhibited an average similarity to the rat brain type 2 sequence of 99.2%. Only two of the ferret clones sequenced (F7 and F20) were identical and were 99% similar to the published rat sequence. The one notable difference between these two isolates and those of the other ferret or rat sequences is the presence of a G at nucleotide 548 (nt 7893 in X61677 ITPR2 (Genbank) Ref. 6) encoding an arginine. The other four ferret sequences characterized are very homologous to each other as well as the rat sequences and exhibit minimal sequence heterogeneity at the same nucleotide residues as in the full-length rat type 2 sequences.
Taken together, the rat and ferret sequences obtained from ventricular cardiac myocytes encode the type 2 InsP 3 receptor and have a high degree of similarity with the published rat type 2 receptor. The micro-heterogeneity at similar nucleotide positions observed in the ferret clones suggests that the type 2 receptor may be encoded by a small gene family of type 2 receptors or indicate that the ferret population tested was not as genetically homogeneous as that of rat.
Functional Reconstitution of Type 2 InsP 3 Receptors from Ferret Ventricular Myocytes-To examine the functional features of the ferret InsP 3 receptor channels, its large size was

FIG. 4. Ribonuclease protection analysis of the patterns and expression levels of the InsP 3 R in cardiac tissues. A, InsP 3 R isoform-specific antisense probes (designated T-1, T-2, and T-3) and
mouse ␤-actin were used in RNase protection assays with 10 g of total RNA from acutely isolated myocytes, left ventricular myocardium, whole heart, and cerebellum (2.5 g of total RNA) from rat (1st to 4th panels, respectively). Controls show InsP 3 R and ␤-actin antisense probes in the presence and absence of RNase (5th and 6th panels). End-labeled X174 RF DNA markers are included as standards. The calculated full-length protection products (nucleotides) for the three InsP 3 R and ␤-actin probes are indicated on the right. B, RNase protection analysis of InsP 3 R distribution during myocyte isolation. Receptor type-specific and ␤-actin antisense probes were reacted with 10 g of total RNA from acutely isolated myocytes (identical panel as in A), unwashed myocytes (panel 2), and incompletely dissociated myocardium (panel 3). Theoretical full-length protection products (nucleotides) for the three InsP 3 R and ␤-actin probes are indicated on the right.

TABLE I Relative abundance of InsP 3 R subtype mRNA in cerebellum and
cardiac samples Data are from the ribonuclease protection assays of Fig. 4 and were collected using a Packard Instant Imager. The data are expressed as the percent of the total ribonuclease resistant counts for each sample Ϯ S.D. (n ϭ 3). (Note, minor differences between the specific activity of the InsP 3 R probes were less than 5% and not taken into account for these comparisons.) RNA  exploited to enrich for InsP 3 receptor protein for reconstitution in planar lipid bilayers. Microsomes from ferret ventricular myocytes were solubilized in CHAPS detergent and subjected to sedimentation over 5-20% sucrose gradients. Following fractionation, the position of the receptor was established by immunoblotting using the NH 2 -terminal consensus antibody (not shown). Fractions with the greatest amounts of receptor protein were reconstituted into proteoliposomes using a method similar to Ferris et al. (29). The resulting liposomes were fused into planar lipid bilayers.
The conductance of Cs ϩ through single ryanodine receptor (RyR) channels in planar bilayers is quite large (500 -600 picosiemens; Ref. 12). The homology between RyR and InsP 3 R transmembrane segments (24) and the similar selectivity of single InsP 3 R channels in bilayers (30) led to the assumption that Cs ϩ conduction through single InsP 3 R channels may also be quite large. The function of individual type 2 InsP 3 R channels was defined by fusing proteoliposomes into planar lipid bilayers with Cs ϩ as the main cationic charge carrier. The free [Ca 2ϩ ] in the bathing solutions on both sides of the bilayer was buffered (1 mM EGTA) at 200 nM. The RyR channel, which is abundant in cardiac ventricular myocytes, will be closed at this relatively low [Ca 2ϩ ] (12). However, single InsP 3 R channels can be activated by InsP 3 at such low free [Ca 2ϩ ] (34).
Cationic single channels appeared in the bilayer following the incorporation of InsP 3 R-enriched proteoliposomes. Spontaneous infrequent single channel events were observed in the absence of added exogenous InsP 3 (Fig. 6A). Open events were brief with few events lasting longer than 20 ms and there were long periods (10 -20 s) during which no events were observed. The overall open probability (P o ) in the absence of added InsP 3 was 0.02 Ϯ 0.01 (n ϭ 4). Addition of 60 nM InsP 3 to the cis solution increased the P o to 0.23 Ϯ 0.14 (n ϭ 4; Fig. 6B). Thus, the Cs ϩ -conducting channel reconstituted from the type 2 InsP 3 R-enriched proteoliposomes was activated by InsP 3 .
The action of heparin on the InsP 3 -activated Cs ϩ -conducting channel was defined. Control channel activity was monitored in the presence of 60 nM InsP 3 in the cis solution (Fig. 7A). Hep- FIG. 5. Sequence analysis of PCR products from rat and ferret cardiac myocytes. PCR products were cloned into pBluescript vector and subjected to DNA sequence analysis. Sequences from rat cardiac myocytes (R) and ferret cardiac myocytes (F) are compared with the full-length rat cDNA (InsP3R2; Ref. 6) to generate a consensus. Variable nucleotide positions are indicated, otherwise positions of sequence identity are blank. The deduced amino acid sequence from the consensus and residues in ferret which depart from the full-length rat sequence are indicated. The sequences corresponding to the PCR primer pair are underlined and shaded amino acid residues indicate the putative fifth and sixth membrane spanning regions (2,9,31). arin (1.5 mg/ml) was added to the cis solution and stirred vigorously for 1 min. The frequency of single channel events dramatically decreased (Fig. 7B). Infrequent brief single channel events were observed in the presence of heparin. The P o in the presence of heparin was 0.03 Ϯ 0.04 (n ϭ 5). Thus, the InsP 3 -activated Cs ϩ -conducting channel reconstituted from the type 2 InsP 3 receptor-enriched proteoliposomes was blocked by heparin.
Channel experiments were performed in the presence of 200 nM free Ca 2ϩ to minimize the activity of RyR channels that may also be present in the proteoliposomes. The plant alkaloid ryanodine binds to the RyR protein with nanomolar affinity. Ryanodine binding dramatically alters single RyR channel activity in bilayers (11). To demonstrate that the Cs ϩ -conducting channel examined in this study was not a RyR, ryanodine (7 M) was added to the cis solution and vigorously stirred for 1 min (Fig. 8A). The addition of ryanodine did not change the P o , 0.14 Ϯ 0.09 versus 0.16 Ϯ 0.13 (n ϭ 3).
The conductance of the Cs ϩ -conducting channel was determined by plotting the amplitude of single channel events as a function of membrane potential (Fig. 8B). This analysis, however, was complicated by the existence of multiple conduction levels (see Figs. 7A and 8A). Data pooled from eight different single channel experiments are shown in Fig. 8B (filled circles). Data points clustered around two conductance levels (388 and 274 picosiemens). The fitted lines are linear regressions. The same two conduction levels were also commonly observed in channels pretreated with ryanodine (7 M; open circles) indicating that single channel current, like P o , was also unaffected by the presence of ryanodine.
A reliable method to define the reversal potential in multiple or subconducting channel experiments is to apply membrane voltage ramps. Voltage ramps were applied to the Cs ϩ -conducting channels to define reversal potential in the presence of 220 mM cis and 20 mM trans-CsCH 3 SO 3 (Fig. 8C, top record). Note that all conduction levels appear to reverse at the same point. In these solutions, the reversal potential of an ideal cationic selective channel would be near Ϫ60 mV. The measured reversal potential (Ϫ53 mV, see arrow) indicates that the InsP 3sensitive channel was nearly a perfect cationic selective channel. To determine if the channel was Ca 2ϩ -selective, 100 mM CaCl 2 was added to the trans solution (Fig. 8C, bottom record). A reversal potential shift would indicate that the channel was permeable to the added ion. In the presence of 100 mM CaCl 2 , the reversal potential shifted (Ϫ53 versus Ϫ15 mV, see arrows) toward the theoretical Ca 2ϩ equilibrium potential. These data indicate that this channel was indeed a Ca 2ϩ -selective channel. This channel is Ca 2ϩ -selective, InsP 3 -activated, inhibited by heparin, and thus is likely to be the InsP 3 R of cardiac ventricular myocytes. Since the identity of the InsP 3 R was established as type 2, this study is the first to examine the functional attributes of the type 2 InsP 3 -gated channel.

DISCUSSION
Identification of the Type 2 Receptor-The present study identifies the type 2 InsP 3 receptor as the predominant isoform of the InsP 3 receptor family expressed in rat and ferret ventricular myocytes. In addition, this study demonstrates that the isoform identified as a type 2 receptor encodes a functional calcium release channel activated by low levels of InsP 3 and inhibited by heparin.
In whole heart there is significant evidence that the type 1 receptor is expressed at high levels in the conduction system, particularly Purkinje myocytes (18,19). Initially, the expression of the type 1 InsP 3 receptor was reported in whole rat heart on Western blots using a type 1-specific antibody (5). However, interpretation of that study is complicated because it was uncertain as to whether or not the signal was derived from cardiac muscle, the associated vascular smooth muscle, or neural tissue. Recently, it has been reported by at least two groups that the type 1 receptor is indeed expressed in certain regions of the heart (17)(18)(19). Gorza and co-workers (18,19) demonstrated that the type 1 receptor is highly concentrated in the majority of Purkinje myocytes, distal elements of the conduction system, and in the atrio-ventricular node. This was achieved using a peptide antisera directed against the 19 carboxyl-terminal amino acids of the type 1 receptor isoform and via in situ hybridization using a 3Ј-cDNA fragment (pI2a, Ref. 5) of the type 1 receptor. Strong immunoreactivity was observed in the different regions of the conduction system with particularly intense signals observed in the Purkinje myocytes. Analysis at the RNA level revealed intense hybridization to conduction system myocytes probably corresponding to Purkinje bundles. Working ventricular myocytes on the other hand were only weakly reactive (18). These weak signals may result from cross-hybridization of type 2 mRNA, since the antisense probe used was from the 3Ј end of the type 1 cDNA which shares regions of sequence homology with the type 2 receptor. In a similar study, Moschella and Marks (17) reported that the InsP 3 receptor in cardiac myocytes is structurally most similar to the type 1 receptor of brain and vascular smooth muscle as judged by immunoreactivity and ribonuclease protection. This study demonstrated that the type 1 receptor is present in RNA obtained from whole heart but did discriminate the expression of the other InsP 3 R isoforms. The probe used in these experiments was directed against a rat aortic smooth muscle cDNA and was reported to be Ͼ90% similar to the mouse type 1 receptor. Immunolocalization studies using a InsP 3 receptorspecific anti-peptide antibody clearly demonstrated the expression of the receptor in cardiac myocytes (17). However, it is uncertain as to which receptor subtype was detected since the peptide-derived antibody was 92% similar, differing only by a single amino acid, between the type 1 and type 2 receptor homologues.
Using homogeneous preparations of myocytes (22), our type 1-specific antibodies failed to detect any cross-reacting species on Western blots (Fig. 1, 1st to 4th panels). Significant levels of receptor were detected using an antibody directed against the NH 2 -terminal 324 amino acids of the conserved ligand binding domain (Fig. 1, 5th panel). InsP 3 saturation binding assays performed with microsomes from ferret myocytes revealed that the receptor species present has a K D ϭ 23.6 nM and a B max of 0.46 pmol/mg. This value for the apparent K D is very similar to that obtained for the recombinant type 2 receptor (27 nM) (6) and that reported in canine heart (21 nM) (32). The type 2 receptor has been shown to exhibit the highest affinity for InsP 3 of the three receptor types, followed by the type 1 and type 3, respectively. Expression of the recombinant receptor subtypes in mammalian and bacterial systems have demonstrated that the type 2 receptor has approximately 3-fold greater affinity for InsP 3 than the type 1, which has nearly 10-fold higher affinity for ligand than that of the type 3 receptor (6,10). These data suggest that the consensus NH 2 -terminal antibody was detecting the type 2 InsP 3 R in the ventricular myocytes.
To investigate further the properties of the receptor expressed in cardiac myocytes, total RNA was prepared and polymerase chain reaction amplifications were performed using oligonucleotide primers that anneal to all three receptor subtypes. Southern hybridizations of the PCR products from both rat and ferret RNA using receptor type-specific probes indicated that the type 2 receptor is the primary species amplified from ventricular myocytes. In whole heart and left ventricular myocardium, however, all three of the receptor isoforms are detected.
The pattern and levels of expression for the individual InsP 3 R isoforms were examined using RNase protection with receptor subtype-specific antisense probes. The type 1 receptor was found to predominate in RNA derived from whole heart and left ventricular myocardium, whereas in myocytes the type 2 receptor was highly enriched comprising nearly 85% of the detectable signal. These results confirm the PCR screening results and suggest that in cardiac tissues all receptor isoforms are expressed, but their distribution and levels of expression are heterogeneous, possibly reflecting specialized roles for the InsP 3 R family members in various regions or cell types of the heart.
Sequence analysis of the PCR products of rat and ferret cardiac myocytes confirmed that the type 2 receptor sequence is present and highly conserved between rat and ferret. The ferret sequences show some micro-heterogeneity suggesting that the type 2 receptor may itself be encoded by a small highly homologous gene family of several distinct members.
Different homologues of the InsP 3 receptor are found in different cell types associated with the heart. The physiological role for the differential expression of distinct InsP 3 R isoforms in heart is unknown. However, differences between the patterns of expression for other calcium handling proteins, for example the sarcoplasmic reticulum-Ca 2ϩ channel, and Ca 2ϩ pump mRNA in conduction and working myocytes have been  (B, open circles). The reversal potential of the multi-or substating InsP 3 -sensitive channel was defined by applying voltage ramps. In the standard solution (220/20 Cs ϩ ), single channel current reversed at Ϫ53 mV (top record, arrow). Addition of 100 mM CaCl 2 to one side shifted the reversal potential (bottom record, arrow) and attenuated single unit current. Fitted lines were fit to randomly selected regions of both records to confirm reversal potential. Horizontal lines represent the zero current level for each record. noted (19). Conduction system myocytes exhibit preferential expression of the type 1 InsP 3 and type 3 ryanodine receptors, whereas the working myocytes show enhanced levels of SERCA2a as well as the type 2 ryanodine and InsP 3 receptors. It has been suggested that different InsP 3 receptor subtypes may have distinct biological roles within a cell (10). This hypothesis remains to be tested, but it is noteworthy that in whole heart, multiple types of InsP 3 receptors are expressed with the type 2 at significant levels in working myocytes and the type 1 predominating in conduction tissues.
Functional Reconstitution-Microsomal proteins from ferret ventricular myocytes were solubilized and sedimented on sucrose gradients to enrich for the InsP 3 receptor. These fractions were subsequently reconstituted into proteoliposomes and used to examine whether the type 2 receptors tetramers are functional Ca 2ϩ channels.
Cationic single channels appeared in the bilayer following the incorporation of proteoliposomes. These channels were Ca 2ϩ -selective, InsP 3 -activated, and blocked by heparin. Spontaneous infrequent single channel events were observed in the absence of added exogenous InsP 3 (Fig. 6). Spontaneous openings of an InsP 3 -activated channel in the absence of InsP 3 were also reported by Borgatta et al. (16). It is possible that these openings are in response to contaminant InsP 3 . Alternatively, the presence of multiple channels may lead to an over-estimation of the single channel P o in the absence of InsP 3 . Addition of 60 nM InsP 3 to the cis solution increased the P o dramatically. Despite the basal activity of the channel in the absence of InsP 3 , the channel was activated by a relatively low [InsP 3 ].
We conclude that the single channel activity represents the functional attributes of the type 2 InsP 3 R channel. This conclusion is supported by several lines of evidence. 1) The channels are derived from preparations that are significantly enriched for the type 2 receptor isoform. 2) The activation of channel openings at relatively low concentrations of InsP 3 (60 nM) (Fig. 6) suggests that the affinity for InsP 3 of this channel is higher than reported in previous studies. The IC 50 for maximal activation is approximately 57 nM as compared with 200 nM for the InsP 3 R of cerebellum (34). This difference may reflect the 3-4-fold higher InsP 3 binding affinity for InsP 3 of the type 2 receptors as compared with the type 1 receptor that predominates in cerebellar Purkinje cells. 3) The channel openings are inhibited by heparin, which is known to inhibit ligand binding to the InsP 3 receptor. 4) Single channel activity and current amplitude were not altered by ryanodine. 5) The channel is calcium-selective and permeable to monovalent cations, similar to the type 1 receptor of cerebellum (30).
Even though the type 2 InsP 3 receptor channels examined here exhibited a significantly lower IC 50 for maximal open probability compared with those studied from cerebellum, where the type 1 receptors predominate, similarities between the two channels are conspicuous. The similarities include multiple conductance states, Ca 2ϩ selectivity, permeability to monovalent cations, blockade by heparin, insensitivity to ryanodine, and activation by InsP 3 at low [Ca 2ϩ ]. The only clear difference appears to be the InsP 3 regulation of the channel. Currently, the biological significance of the similarities and differences is not clear. It is also unclear whether channel disruption due to experimental manipulations such as membrane isolation, solubilization, and reconstitution alters function. A detailed systematic characterization of type 1 and type 2 InsP 3 R channel function is currently being done. 2 In conclusion, this study demonstrates that the type 2 recep-tor is the principal isoform of the InsP 3 receptor family expressed in cardiac ventricular myocytes. The isoform expressed in both rat and ferret myocytes binds InsP 3 with similar affinity as the expressed recombinant protein. Polymerase chain reaction amplification, RNase protection experiments, and DNA sequence analysis clearly identify the expression of the type 2 receptor as well as the high degree of conservation of the type 2 receptor sequences between species. Reconstitution of the type 2 receptor into planar lipid bilayers demonstrates that the protein forms a functional calcium channel that is gated in response to InsP 3 and inhibited by heparin. This study, therefore, represents a significant step toward understanding the role of InsP 3 receptors in heart and determining if InsP 3 R function is type-specific.