Extended Junctional Sarcoplasmic Reticulum of Avian Cardiac Muscle Contains Functional Ryanodine Receptors*

The ryanodine receptor (RYR)/Ca2+ release channel of avian cardiac muscle was localized by immunocytochemical techniques and biochemically characterized using isolated membrane and receptor protein fractions. Monoclonal antibody C3-33 raised against the ca- nine cardiac RYR bound to the junctional sarcoplasmic reticulum of pigeon and finch hearts, both at peripheral couplings and at extended junctional sarcoplasmic reticulum (EJSR). Immunoblots of sarcoplasmic reticulum vesicles from pigeon and finch hearts showed this antibody recognized a single high molecular weight protein, which co-migrated with the canine M, 565,000 RYR/Ca2+ release channel polypeptide. The pigeon heart RYR bound [SH]ryanodine with high affinity in a CaB+-de- pendent manner, comparable to the canine cardiac RYR. Purification of the pigeon RYR yielded a 30 S protein complex, which bound the maximum calculated amount of [SHlryanodine ((440 60) pmoYmg protein), assuming one high affinity sitdtetrameric 30 S RYR comprised of M, 585,000 polypeptides. Autoradiography of isolated finch cardiac myocytes indicated [*H]ryanodine binding throughout the cells. These results suggest that avian heart contains a single population

T-tubules (interior couplings). "Junctional processes" (JP) (Sommer and Johnson (1969) or "feet" (Franzini-Armstrong, 1970) span the junctional gap, often incompletely. The geometry of couplings is propitious for E-C coupling by facilitating signal transmission both through direct contact (skeletal muscle; Schneider, 1981) and a diffusible transmitter such as Ca2+ (cardiac muscle ;Fabiato, 1983). Dihydropyridine receptors in the plasmalemma at the couplings are thought to mediate SR Ca2+ release by acting as voltage sensors in vertebrate skeletal muscle and as voltage-dependent Ca2+ channels in cardiac muscle (Rios and Pizarro, 1991). The mammalian cardiac Ca2+ release channel has been identified as the receptor for the plant alkaloid ryanodine, localized to the JPs which decorate the membranous envelopes of JSR, and purified as a 30 S ryanodine receptor (RYR) protein complex comprised of four M, 565,000 polypeptides (Inui et al., 1987;Anderson et al., 1989;Rardon et al., 1989), as determined by cDNA cloning and sequencing (Otsu et al., 1990;Nakai et al., 1990).
In avian hearts, JSR has a dual location; it is both attached to plasmalemma as part of the peripheral couplings (JSR proper) (-25% of JSR) and unattached in the form of EJSR (-75%) (for review, see Sommer et al., 1979see Sommer et al., , 1991. Although removed from the plasmalemma by several micrometers (avian hearts have no T-tubules), EJSR, like JSR proper, is decorated with JPs. A homologous, much less numerous organelle also occurs in mammalian cardiac muscle as so-called corbular SR (Dolber and Sommer, 19841, especially in cells without T-tubules (e.g. conduction cells). Anti-ryanodine receptor antibodies localized to the JPs have corroborated the structural homology of corbular SR with JSR proper in mammalian cardiac muscle (Jorgensen et al., 1993).
Here, we have characterized and localized avian cardiac RYRs biochemically and by immunocytochemical and radioligand binding techniques, respectively. Our results indicate that avian hearts display 1) a uniform RYR population, 2) localization of RYRs to EJSR, and 3) [3Hlryanodine binding properties similar to those established in mammalian hearts. These results support the idea that EJSR is a functional homologue of JSR proper. Some of the results of this study have been presented in abstract form (Junker et al., 1992).
Immunohistochemistry-Monoclonal antibody C3-33, developed against canine cardiac ryanodine receptor (Lai et al., 19921, was used to localize the RYR in finch, pigeon, and rat hearts. Freshly dissected whole hearts (finch) or pieces of hearts (rat and pigeon) were placed into a Tissue-Tek cryomold containing OCT embedding medium (Lab-Tek Division, Miles Laboratories, Naperville, IL), frozen in liquid nitrogencooled isopentane, and stored a t -80 "C. Eight-pm cryostat sections were cut, picked up on gelatin chrome-alum-coated glass slides, and stored at -80 "C. Immediately prior to staining, the slides were removed from the freezer, air dried, and fixed 5 min with 2 or 4% paraformaldehyde in 0.1 M phosphate buffer containing 10% sucrose. ABC peroxidase staining was developed using 3,3'-diaminobenzidine and HzOz and performed as previously described (Sar, 1985). C3-33 antibody was used at concentrations of 0.1 and 1 pg/ml. Normal mouse IgG (1 pg/ml) ("nonimmune serum") and preabsorbed antibody (0.1 pg of ' 23-33 + 6.8 pg of purified ryanodine receptodml) were used as controls. Stained sections were mounted in a medium containing nine parts glycerol to one part 0.01 M Tris, pH 7.6, and were photographed using a Leitz Ortholux microscope.
For electron microscopy, the coverslips and mounting medium were removed by soaking the slides in normal saline. The sections were fixed for 15 min in 2% glutaraldehyde, 0.1 M cacodylate buffer, pH 7.4, and postfixed by flooding the slides with 2% OsO, in water. The sections were dehydrated with alcohol and propylene oxide, embedded by inverting an Epon filled BEEM capsule over the section, cut on a Reichert OM-U3 ultramicrotome, mounted on copper grids, and examined on a Zeiss 10B electron microscope.
Preparation of Isolated Cells-Finch heart cells were isolated similarly as described (Bendukidze et al., 1985). Briefly, the heart of an anesthetized finch (60 pl of pentobarbital, 65 mg/ml, injected intraperitoneally) was excised from the thorax while still beating. Cell dissociation was begun by perfusing the coronary arteries for 5 min with an enzyme solution containing 1 mg/ml collagenase (Sigma, type I), 0.07 mg/ml Pronase (Sigma protease, type X N ) , 0.1 M NaCI, 0.05 M taurine, 0.01 M KCI, 0.0012 M KHzP04, 0.1 m~ CaCIz, 0.004 M MgSO,, 0.02 M dextrose, 0.01 M HEPES, pH 6.9 (with KOH). The ventricles were then chopped into small pieces and stirred in enzyme solution containing 0.2 m~ CaCl, a t 35 "C. Isolated cells were collected from the solution, placed into Eppendorftubes, and centrifuged for 5 s in a microfuge. The cells were resuspended and stored up to 1 h in the above solution with 0.2 mM CaCIz, but without enzymes. NaCl, 0.1 mM EGTA, 1.5 m~ CaCl,, 5 m~ dextrose, 20 m~ K-HEPES, pH Autoradiography-Isolated cells were rinsed in solution A (0.15 M 7.4). They were then incubated for 1 h a t room temperature in solution Acontaining either 50 rn [3Hlryanodine or 50 TIM [3H]ryanodine + 50 unlabeled ryanodine. Following incubation, the cells were rinsed three times with solution A. Aliquots (20 pl) were spread over dried emulsion precoated slides (Kodak NTB-2, Eastman Kodak, Rochester, NY) and stored in light-tight black dessicator boxes at -20 "C (Stumpf and Sar, 1975). After 30 days of exposure, the slides were developed in Kodak D-19 for 1 min, fixed with Kodak fixing solution for 5 min, and stained with methyl green-pyronin.
The slides were examined with an Olympus photomicroscope. To determine the percentage of cell clusters clearly labeled with C3H1ryanodine, randomly selected clusters were visually inspected. If the num- Fig. 1 were processed for electron microscopy. The antibody labels EJSR (arrows) located between myofibrils (brackets), and JSR proper a t FIG. 2. Electron microscopic localization of RYR using canine cardiac RYR monoclonal antibody C3-33. Sections such as those seen peripheral couplings (double arrows) located between myofibrils and plasmalemma, all at the level of 2 lines. Use of osmium tetroxide results in enhancement of the diaminobenzidine reaction product as part of the immunocytochemical procedure, as well as a weak background staining of Z lines (arrowheads), mitochondria (asterisks), and free SR (open arrows) useful for general orientation. Although these structures are visible in the controls (cf: Fig. 2, c and g), prominent staining of EJSR and JSR proper is not. For comparison, the appearance of EJSR and JSR proper prepared for conventional transmission electron microscopy after contrasting with osmium tetroxide, uranyl acetate, and lead citrate is shown in b. Solid circles = extracellular space. a, pigeon heart 0.1 pg C3-33/ml); b, pigeon heart-conventional transmission electron microscopy; c, pigeon heart-preabsorbed antibody control; d, rat heart (1 pg C333/ml); e, finch heart (1 pg C333/ml);f, finch heart (1 pg C3-33/ml): the EJSR membrane is stained, not its lumen; g, finch heart-nonimmune serum control (1 pg normal mouse IgG/ml). Bars = 1 pm.
ber of grains over the cells was visibly greater than background, the cluster was classified as labeled. Photographs of labeled clusters and controls were taken for grain counting, and the number of grains from two randomly selected regions over a cell cluster, each 90 pm2, was counted. Background levels were determined by counting the number of grains found in two regions of the surrounding extracellular space. The net grain count was determined by subtracting the background from the cluster grain count for each cluster. Statistical analysis was performed using a two sample test of means (SPCC System, Walmyr Publishing, Tempe,AZ).
Isolation of SR Vesicles and 30 S RYR Complex-Cardiac SR vesicles were prepared from pigeon and finch hearts similarly as described for mammalian heart (Meissner and Henderson, 1987). The hearts of three pigeons (-4 g each) were rapidly excised and immediately immersed in ice-cold 0.3 M sucrose. Minced hearts were added to 10 volumes of an ice-cold 20 m~ Tris-HCI, pH 7.4, buffer containing 0.3 M sucrose, 0.5 mM EDTA, and protease inhibitors (1 m~ diisopropyl fluorophosphate, 100 n~ aprotinin, 1 p~ leupeptin, 1 p~ pepstatin, 1 m~ benzamidine, 1 m~ iodoacetamide) and homogenized a t 4 "C for 60 s using an ultra torrex homogenizer (Tekmar). A crude SR membrane fraction was obtained as a 10,000-100,000 x g pellet which was resuspended in 0.4 M KC1 me-dium and layered at the top of a linear 20-40 % (w/w) sucrose gradient in 0.4 M KCI. f i r centrifugation for 2 h a t 240,000 x g in a Beckman SW41 rotor, membranes sedimenting a t 25433% (Fraction 1) and 35-40% (Fraction 2) sucrose were collected, diluted with two volumes of 0.4 M KCl, sedimented, resuspended in 0.3 M sucrose, 5 m~ potassium PIPES, pH 7.0, and stored in 0.25-ml aliquots at -135 "C. Finch hearts (-0.2 g) were frozen on dry ice following dissection, and a crude SR membrane fraction was prepared from five frozen hearts as described above.
The CHAPS-solubilized 30 S pigeon and canine heart RYRs were isolated by rate density gradient centrifugation in the presence of protease inhibitors as described (Anderson et al., 1989).
Wa2+ E m u Measurements-Pigeon heart SR vesicle fractions (5-10 mg protein/ml) were passively loaded with 5 m~ 45Ca2+ in a medium containing 0.1 M KC1 and 5 m~ 4sCa2+ (Meissner and Henderson, 1987). 4sCa2+ efflux was initiated by diluting vesicles 200-fold into isoosmolar efflux media and stopped by placing 0.4-ml aliquots at various times on 0.45-pm filters (Type HA, Millipore Co, Bedford, MA). Filters were washed with a quench solution containing 0.2 m~ EGTA, 10 mM Mg2+, and 20 p~ ruthenium red. Radioactivity remaining with the vesicles on the filters was determined by liquid scintillation counting.  PHlRyanodine Binding-[3HlRyanodine binding was determined as described . Unless otherwise indicated, membranes were incubated in the presence of protease inhibitors (1 m diisopropyl fluorophosphate or 0.2 m Pefabloc SC, 5 w of leupeptin) for 24 h at 12 "C in media containing 1 M NaCI, 20 m sodium PIPES, pH 7.2.100 1.1~ EGTA, 300 1.1~ Ca2+, 5 mM AMP, and 1-50 m [3Hlryanodine. Nonspecific binding was estimated using a 1000-fold excess of unlabeled ryanodine.

RESULTS
Immunohistochemistry-Immunocytochemical localization of the avian cardiac RYR using a mammalian cardiac muscle RYR monoclonal antibody (C3-33) showed multiple foci of immunoreactive product in cryostat sections of pigeon heart. Light microscopy revealed an antibody staining pattern con-preabsorbed antibody (C3-33 preincubated with purified ryanodine receptor) (Fig. l b ) or nonimmune serum (not shown) was used. Cryostat sections of finch and rat heart labeled with C3-33 yielded identical results (Fig. 1, c and dl. Electron micrographs of pigeon hearts showed high levels of RYR immunoreactivity in foci along the Z lines, especially in interfibrillar spaces (Fig. 2a). Comparison with conventional EM images (Fig. 2b) showed that immunostain was localized to both the JSR proper and EJSR. Free SR and mitochondria were not stained, In control sections incubated with preabsorbed antibody, no labeling was observed (Fig. 2c). Similar labeling patterns were observed for rat (Fig. 2 d ) and finch (Fig. 2 e ) cardiac muscle. Occasionally, the stain was clearly associated with the EJSR membrane (Fig. 2f). Control sections using nonimmune sera or preabsorbed monoclonal antibody C3-33 did not show any immunostain. As a n example, Fig. 2g shows the nonimmune serum control for finch heart.
Autoradiography-The presence of a functional RYR in the EJSR was tested by incubating isolated finch heart cells with t3H1ryanodine in the absence and presence of an excess of unlabeled ryanodine. In the autoradiographs, the cells were typically observed as small rectangular clusters comprised of two to three cells. The boxy appearance of the clusters (Fig. 31, as opposed to the elongated appearance of freshly isolated cardiac myocytes, indicated a state of contracture. Light microscopic examination showed that most clusters (131 out of 202) incubated with t3H]ryanodine alone were clearly labeled (Fig. 3u). By contrast, when cells were incubated with [3H]ryanodine in the presence of a 1000-fold excess of unlabeled ryanodine (Fig.

TABLE I1
PlRyanodine binding and 45Ca2+ efflux properties of membrane fractions fiom pigeon heart The fraction of Ca2+-permeable vesicles was obtained as shown in Fig.  4 by determining the amounts of 45Ca2+ remaining with the vesicles at 20 s in the quench (10 m~ Mg2+ and 20 p~ ruthenium red) and release (20 p~ free Ca2+) media. B,, and KD values of high-affinity L3H]ryanodine binding to sucrose gradient Fractions 1 and 2 were determined by Scatchard analysis. E,, value of [3H]ryanodine binding to crude SR membranes was estimated using a ligand concentration (50 m) that essentially fully occupied high-affinity binding sites. Results are given as the mean 2 S.D. of three preparations.  (m) 4.8 f 0.6 4.5 * 0.6 9.9 * 4.5 2.9 2 0.6 2.1 2 0.7 3b), over none of the cell clusters (0 of 130) was the number of grains greatly increased over that of background. The grains were fairly randomly distributed over the cells. They were present but not concentrated at the edges or boundaries between two neighboring cells, suggesting that both JSR proper and EJSR were binding [3Hlryanodine. However, the grains rarely formed a sarcomere-like pattern. This was probably due to their relatively large size which limited the spatial resolution of this technique to approximately 1 pm. Table I shows that the number of grain counts corrected for background was significantly higher ( p < 0.0001) over cells labeled with [3Hlryanodine alone than over control cells which were incubated with 13H]ryanodine in the presence of a n excess of unlabeled ryanodine.
45Ca2+ Release and PHlRyanodine Binding to SR-enriched Membrane Fractions-The functionality of the RYR in EJSR was further tested by determining the biochemical properties of the isolated avian EJSR and RYR. Since at present EJSR and JSR proper cannot be separated to test their individual activities, these studies were based on the observation that EJSR accounts for a large portion (-75%) of total JSR in birds (Bossen et al., 1978). Therefore, L3H]ryanodine binding properties of isolated membrane fractions should be reflected to a large extent by those of the EJSR. Because of their larger size, pigeon rather than finch hearts were used as the source for the isolation and characterization of avian cardiac RYR. A crude microsomal membrane fraction was prepared and subfractionated on a sucrose gradient (see "Experimental Procedures"). Table I1 shows the yields for the crude and 25-33 and 35-40% sucrose gradient membrane fractions.
The presence of a Ca2+-activated 45Ca2+ e m u pathway in pigeon heart membrane fractions was assessed by diluting passively loaded vesicles into efflux media that contained the two SR Ca2+ release channel inhibitors Mg2+ and ruthenium red, and either <lo-' M or 20 p~ free Ca2+. As previously observed for canine cardiac SR vesicles enriched in Ca2+ release activity (Meissner and Henderson, 19871, 45Ca2+ efflux was slow in the Ca2+ release channel inhibiting medium (Fig. 4). Omission of Mg2+ and ruthenium red from the <lo-* M Ca2+ efflux medium resulted in a n increased 45Ca2+ e m w rate. A further increase in the 45Ca2+ efflux rate was observed when vesicles were placed into a medium containing 20 p~ free Ca2+. About half of the radioactivity remained with the vesicles for longer times (Fig. 4), indicating the presence of a subpopulation of vesicles lacking a Ca2+-gated Ca2+ release pathway (Meissner and Henderson, 1987). Table I1 li'sts the fraction of vesicles that contained a Ca2+-gated release pathway in crude SR membrane and sucrose gradient Fractions 1 and 2.
L3H1Ryanodine binding experiments (see "Experimental Pro- cedures") showed that the drug bound with high affinity to pigeon heart SR membranes (not shown). Scatchard analysis of [3H]ryanodine binding data indicated the presence of a specific high-affinity site with a KO of 2-3 n~ (Table 11). On average, crude SR membranes and sucrose gradient Fractions 1 and 2 bound 4.8,4.5, and 9.9 pmoVmg protein, respectively (Table 111, as compared with 5-20 pmoVmg protein for mammalian skeletal and cardiac SR vesicles (see review by . In Fig. 5, the Ca2+ dependence of L3H1ryanodine binding to pigeon and canine cardiac SR membranes is compared in the presence of "physiological" concentrations of Mg2' (5 mM) and adenine nucleotide (5 mM) using the nonhydolyzable ATP analog AMP-PCP. An essentially identical Ca2+ dependence was observed. Specific [3Hlryanodine binding was half-maximal at about 3 J~M and reached a maximal value at about 30 p~ Ca2+.
SDS-PAGE and Immunoblot Analysis of SR-enriched Vesicle Fractions-SDS gel and immunoblot analysis showed that pigeon SR membrane fractions contained a high molecular weight protein band (Fig. 6, lanes 1 and 2) which co-migrated with the canine M, 565,000 RYR polypeptide (not shown). The high M , band of pigeon and rat SR membranes specifically reacted with a monoclonal antibody (C3-33) to the canine cardiac RYR (Fig. 6, lanes 3-5). Similarly, immunoblot analysis of crude finch SR membranes revealed a single high M, band that displayed a n identical mobility and cross-reacted with monoclonal antibody C3-33 (lane 6 of Fig. 6). No cross-reactivity was observed with a rat skeletal RYR antisemm  (not shown). Taken together, data of Table I1 and Figs. 5 and 6 suggest that avian and mammalian cardiac muscle ex- were electrophoresed through 3-12% gradient gels, electrophoretically transferred onto Immunobilon polyvinylidene difluoride membranes, probed with a canine cardiac muscle RYR monoclonal antibody (C3-33) and peroxidase-conjugated secondary antibody, and developed using diaminobenzidine and HzOz.
press functionally and immunologically related RYRs comprised of M, 565,000 polypeptides.
Purification of Pigeon Heart RyR-The [3Hlryanodine-labeled pigeon heart RYR was purified by rate density gradient centrifugation, as described for the canine cardiac RYR (Anderson et al., 1989). Pigeon and canine heart SR membranes were solubilized with CHAPS in the presence of a relatively high [3H]ryanodine concentration of 30 m to occupy essentially all high-affinity [3H]ryanodine-binding sites. CHAPS-solubilized, labeled pigeon and canine heart RYRs were placed onto parallel linear 5 2 0 % (w/w) sucrose gradients containing 1% CHAPS and 5 mg/ml phosphatidylcholine, and then centrifuged. Analysis of gradient fractions for 3H radioactivity and protein indicated a single peak of bound radioactivity which co-migrated with a small protein peak with an apparent sedimentation coefficient of 30 S (Anderson et al., 1989) in the bottom fractions PCa of the pigeon (Fig. 7A) and canine (not shown) gradients. Radioactivity was decreased close to background levels in the lower half of the gradients when the solubilized membranes were labeled in the presence of a 300-fold excess (10 p) of unlabeled ryanodine (not shown). SDS-polyacrylamide gel and immunoblot analysis revealed the presence of a >95% pure high molecular weight protein band (Fig. 7B, left lane) which was immunologically detected only in the 30 S RYR peak fraction (Fig. 7B, right lanes)

DISCUSSION
By stereology the total cell volume fraction of JSR in mouse cardiac myocytes (cell diameter -15 pm) and of JSR proper plus EJSR in finch cardiac myocytes (diameter -7 pm) is virtually identical (Bossen et al., 1978), suggesting functional equivalence of JSR proper and EJSR in avian heart. JSR proper is considered to be the site of Ca2+ release in mammalian skeletal and cardiac muscle. In skeletal muscle, the mechanical coupling hypothesis suggests that the SR RYR/Ca2+ release channel is physically linked to a voltage-sensing dihydropyridine receptor in the T-tubule (Rios and Pizarro, 1991). In contrast, in cardiac muscle E-C coupling is thought to be triggered by a diffusible transmitter molecule. A voltage-dependent dihydropyridine receptor/Ca2+ channel located in the surface membrane and T-tubules mediates the entry of Ca2+ ions that trigger SR Ca2+ release (Cannel1 et al., 1987;Nabauer et al., 1989). A close apposition of the surface membrane and SR Ca2+ channel proteins favors local control of the release channels by the trigger Ca2+, thereby supporting a rapid and graded translation of the action potential into SR Ca2+ release (Stern, 1992;Gyorke and Palade, 1992). The presence of putative Ca2+ release structures, micrometers away from any direct contact  Table I, 1.5 mg p r o t e i d d ) were solubilized with CHAPS (1.6%) in a medium containing 1.0 M NaCI, 20 m~ NaPIPES, pH 7.0,200 p~ Ca2+, 5 m~ A " , 5 mg/ml phosphatidyl choline, 1 m~ dithiothreitol, 1 m~ diisopropyl fluorophosphate, 5 p~ leupeptin, and 30 nM [3H]ryanodine. The solubilized proteins were loaded onto a linear 5-20% sucrose gradient in the above medium containing 1% CHAPS and centrifuged at 2 "C in a Beckman SW28 rotor for 16 h a t 26,000 revolutiondminute. Eighteen fractions of 2 ml each were collected and analyzed for :' H radioactivity. The arrow indicates the position of bound PHIryanodine peak fraction of gradients containing the canine cardiac RYR. B , SDS-PAGE and immunoblot (right lanes) analysis of selected gradient fractions. SDS gel was stained with Coomassie Blue (fraction 7, left lane). After transfer onto Immobilon polyvinylidene difluoride membranes, gradient fractions 3.5, 7,9,11,13,15, and 17 were probed with canine cardiac RYR monoclonal antibody C3-33 (right lanes).
with the cell surface membrane, suggests that the emphasis on both may play an important yet, perhaps, physiologically difa juxtaposition of Ca2+ trigger and release structures may need ferent role in regulating intracellular Ca2+ in cardiac muscle. to be reconsidered. Furthermore, the presence, especially in Several observations support the idea of a Ca2+-sensitive avian heart, of a large number of putative Ca2+ release organ-Ca2+ release mechanism by the EJSR of avian heart. First, elles with and without plasmalemmal contacts suggests that evidence for such a mechanism is provided by our observation that L3H1ryanodine binding to the avian RYRs is a Ca2+-dependent process. A similar ligand dependence of [3H]ryanodine binding and Ca2+ release channel activities has suggested that ryanodine is a sensitive ligand for probing mammalian release channel function (Chu et al., 1990;Meissner and El-Hashem, 1992). Second, mammalian cardiac muscle exhibits structural equivalents (corbular SR) (Dolber and Sommer, 1984;Jorgensen et al., 1993) that have been found to store Ca2+ and to contain calsequestrin, a Ca2+-binding protein characteristic of JSR (Jorgensen et al., 1988). Third, as recently reported (Airey et al., 1993) and confirmed in this study, avian heart appears to contain only a single RYR and not two RYRs as observed for avian skeletal muscle. It can be argued that Ca2+ ions move too slowly in cells (Allbritton et al., 1992;Nowycky and Pinter, 1993) to support a Ca2+-induced Ca2+ release mechanism by the EJSR. Birds such as finch and pigeon have a heart rate of about 5-10 beat&. Accordingly, the SR Ca2+ release phase during one contractionrelaxation cycle is probably limited to 25-50 ms. In the absence of Ca2+ buffers, this time is sufficient for Ca2+ to nearly equilibrate throughout a cell with a diameter of 10 pm (Nowycky and Pinter, 1993). On the other hand, the presence of intracellular Ca2+ buffers slows the movement of Ca2+ even as monentary increases in [Ca2+] can speed it up (Allbritton et al., 1992;Nowycky and Pinter, 1993). At present, precise quantitative information about parameters that determine the diffusion rate of Ca2+ such as the amounts of Ca2+ released by JSR and buffered by, or bound to, the cytosol, is largely missing.
Other conditions supporting a Ca2+-induced Ca2+ release mechanism may exist in avian hearts. One of these could involve the creation of a Ca2+ release wave expanding from JSR to EJSR. Such a mechanism would be attractive since the maintenance of a high Ca2+ concentration between adjacent release channels favors the rapid movement of trigger calcium. For example, the diffusion coefficient for Ca2+ increased from 13 to 65 pm2/s when the free [Ca2+] in the cytosolic extract of Xenopus oocytes was raised from about 90 n~ to 1 p~ (Allbritton et al., 1992). The structure of avian hearts enhances the potential effectiveness of a hypothetical Ca2+ release wave. EJSR typically occurs as a series of adjacent collars wrapping around, often partially, the myofibrillar compartments at the Z lines (Sommer et al., 1991). The collars are separated transversely from the JSR proper at the couplings and from each other only by very short segments (<0.20 p n ) of free SR. Such an arrangement of adjacent EJSRs can provide a pathway for Ca2+-induced Ca2+ release, initiated at the JSR proper at a peripheral coupling, to propagate through regenerative Ca2+ release rapidly across the myocyte from the nearest calcium release channels to the next along successive JPs decorating all EJSRs. Recent modeling studies suggest that, indeed, special geometric conditions may account for statistical recruitment of RYR clusters resulting in a graded response to the trigger Ca2+ (Stern, 1992). Graded force development is one of the fundamental properties of cardiac muscle, at least in mammals. Interesting evidence of how mammalian SR Ca2+ release channel activity may be modulated by Ca2+ has been recently described (Gyorke and Fill, 1993).
It should be kept in mind that in the absence of unambiguous proof of Ca2+-induced Ca2+ release by EJSR in avian cardiac muscle, other diffusible effectors must be considered such as cyclic ADP-ribose (Meszaros et al., 1993) or lipid metabolites (Sabbadini et al., 1992).
In conclusion, in this study we have described the first direct obsemations of a functional RYR without any plasmalemma1 contact. Radioligand labeling of isolated cells, SR membrane, and RYR fractions suggested that RYRs in avian heart were capable of binding [3Hlryanodine in a manner characteristic of mammalian JSR RYRs, implying the presence of an intrinsic Ca2+ channel activity in all RYRs of avian heart.