Proc. Nati. Acad. Sci. USA Vol. 88, pp. 2486-2489, March 1991 Physiology/Pharmacology

Regional differences in calcium-release channels from heart (sarcoplasmic reticulum/inositol 1,4,5-trisphosphate receptor/interventricular septum/ventricle)

Louis BORGATTA*, JAMES WATRAS*, ARNOLD M. KATZ*, AND BARBARA E. EHRLICH*tt Departments of *Medicine and tPhysiology, University of Connecticut, Farmington, CT 06030

Communicated by Jared M. Diamond, December 26, 1990

the suspension was centrifuged for 20 min at 6000 rpm (Beckman 35 Ti rotor). The pellets were rehomogenized in the same volume of buffer A and centrifuged for 20 min at 6000 rpm. The supernatant fluids were combined and filtered through two layers of cheesecloth and centrifuged for 15 min at 9000 rpm (Beckman 60 Ti rotor). The resulting supernatant fluid was filtered through eight layers of cheesecloth and centrifuged for 120 min at 16,000 rpm (Beckman 60 Ti rotor). The pellets from the latter spin were resuspended in buffer B (0.6 M KCI/5 mM NaN3/20 mM Na4P205/10 mM Hepes, pH 7.2 at 40C) and centrifuged for 20 min at 25,000 rpm. The pellets were rehomogenized and resuspended in a buffer containing 20 mM 3-(N-morpholino)propanesulfonic acid (Mops), 1o sucrose (pH 6.8). All buffers contained 20 ,M pepstatin, 20 ,uM leupeptin, aprotinin at 3 mg/liter, trypsin inhibitor at 10 mg/liter and N-tosyl-L-phenylalanine chloromethyl ketone (TPCK) at 100 mg/liter. Solutions were quickly frozen in liquid nitrogen and kept at -80°C. Vesicles were used within 4 weeks after preparation. To study the calcium-release channel these vesicles were then incorporated into preformed planar lipid bilayers as described (12, 13). Briefly, vesicles were added to one compartment after the membrane was formed, usually the cis compartment. Incorporation was monitored in the presence of a chloride and an osmotic gradient across the membrane [cis solution: 600 mM N-methyl-D-glucamine chloride/20 mM Hepes/10 mM CaCl2/0.2 mM EGTA, pH 7.3; trans solution: 250 mM Hepes/53 mM Ca(OH)2, pH 7.3]. The insertion of a chloride-permeable channel was the signal that a vesicle had fused with the bilayer. To monitor calciumrelease channels, the bilayer chamber was then perfused with a chloride-free solution. In all cases channel currents were measured under voltage-clamp conditions with 250 mM Hepes.Tris/1 mM EGTA/0.5 mM CaCl2 (0.1 ,uM free Ca), pH 7.3, on the cis side of the bilayer and 250 mM Hepes/53 mM Ca(OH)2, pH 7.3, on the trans side. Channel insertion and subsequent experiments were monitored under voltage-clamp conditions with a pair of Ag/AgCl electrodes contacting the solutions via CsCI junctions. The channel currents were amplified using a patch-clamp amplifier [Yale model MK-5 (Warner Instruments, Hamden, CT) or Axopatch lb (Axon Instruments, Burlingame, CA)] and recorded on chart (General Scanning, Watertown, MA) and tape recorders (Dagan Instruments, Minneapolis). Data were analyzed after filtering with an 8-pole Bessel filter (Frequency Devices, Haverhill, MA) to 300 Hz and digitizing at 1 kHz to transfer to a PDP 11/73 computer (Indec Systems, Sunnyvale, CA). Caffeine was purchased from Sigma. Lipids were purchased from Avanti Polar Lipids. All other reagents used were of analytical grade.

The heart is a heterogeneous tissue composed ABSTRACT of several cell types tailored for specialized functions. We found that intracellular channels also exhibit regional specialization. In cardiac and skeletal muscle these channels are called the calcium-release channel and are identified by activation with either calcium or caffeine and inhibition by the hexavalent cation ruthenium red. The calcium-release channel of the sarcoplasmic reticulum from the interventricular septum has a smaller conductance (31 pS vs. 100 pS) and has longer open and closed times when compared with the channel from leftventricular free wall. An additional calcium-permeable channel with an even smaller conductance (17 pS) was found in the septum, and this channel is similar to the inositol 1,4,5trisphosphate-gated channel from smooth muscle and different from the calcium-release channel (ryanodine receptor) from skeletal and cardiac muscle. The inositol 1,4,5-trisphosphateactivated channel may be derived from specialized conducting tissue that is relatively abundant in the septum, whereas the other calcium-release channels may be derived from regionally specialized myocardial cells in the septum and free wall.

The existence of myocardial cells specialized for impulse formation (the sinus node pacemaker), slow conduction (the atrioventricular node), and rapid conduction (the Purkinje fibers) has been recognized since the beginning of this century (1). More recently, it has become possible to relate the many morphologic, metabolic, and physiologic heterogeneities in the adult mammalian heart to specific molecular properties-for example, when comparing atria and ventricles (2-5) or ventricular epicardium and endocardium (6-8). Different types of voltage-dependent calcium channels of the outer membrane have been identified in heart and a variety of other cell types (e.g., refs. 9 and 10), but these channels do not appear to be regionally localized. In this paper we show that the intracellular calcium channels that initiate contraction exhibit a regional specialization. Evidence that the heart has calcium channels adapted to the specialized functions of specific regions may reflect a general property of organ tissues.

METHODS Sarcoplasmic reticulum vesicles from canine ventricular muscle were isolated by using a modification of existing methods (11). Specifically, hearts were excised from anesthetized dogs, placed in cold saline, and dissected free of fat, pericardium, and connective tissue; the atria and any portion of remaining aorta were discarded. The left-ventricular free wall and the high septum were dissected away from the rest of the myocardium, minced, and then homogenized separately with a Brinkmann Polytron at setting 6 (5 sec; four times) in 4 vol of buffer A (10 mM imidazole, pH 7.4 at 4°C). Another 4 vol of buffer A was added to the homogenates, and

Abbreviations: InsP3, inositol 1,4,5-trisphosphate. 1To whom reprint requests should be addressed at: Division of Cardiology, University of Connecticut, 263 Farmington Avenue, Farmington, CT 06030.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

2486

Physiology/Pharmacology: Borgatta et al. RESULTS When sarcoplasmic reticulum vesicles made from the free wall of canine left ventricle were incorporated into planar lipid bilayers, a 100-pS channel with an open probability (P.) >0.2 was usually observed. Addition of 5 mM caffeine further activated the channel (Fig. 1, top two traces). The slope conductance of the channel isolated from the free wall (Fig. 2) was comparable to values reported (13-16) from whole unselected ventricle. Kinetic analysis of the caffeineactivated channels demonstrated a mean open time of 17 + 5 ms and a mean closed time of 18 + 5 ms (n = 4). Addition of 20 AuM ruthenium red irreversibly inhibited channel activity', and subsequent addition of channel activators, such as calcium or caffeine, could not reactivate the channel. When sarcoplasmic reticulum vesicles made from the interventricular septum were incorporated into bilayers and analyzed under the same conditions used previously, a 31-pS channel was observed. We never detected a 100-pS channel when vesicles from septum were incorporated into the bilayer (Fig. 4 Left), suggesting that the 100-pS channel is absent, or if present, occurs at a density too low to be detected using the bilayer assay. Under control conditions (0.11uM free calcium) the probability of finding the septal channel open was markedly less than for the channel from the free wall (P. < 0.1). Although addition of 5-10 mM caffeine activated the septal channel, two striking differences were apparent. (i) The septal calcium-release channel currents were =3-fold lower in conductance than those from ventricular free wall (compare Fig. 1, top two and middle two traces and Fig. 2, closed circles and open triangles). (ii) The septal channels opened and closed for longer periods when compared with the channel from the free wall (compare Fig. 1, top two and middle two traces). In all cases the kinetics of the septal

free wall

septum

*

pA ~~~~~~~~~~5

FIG. 1. Comparison of single-channel recordings of calciumrelease channels from the left-ventricular free wall (top two traces) and the two channels from the septum (middle two and bottom two traces) in the presence of 0.1 jtM free calcium/5 mM caffeine at a holding potential of 0 mV. The bottom two traces were obtained in the presence of 3 ,uM InsP3. Zero current is indicated by dotted lines, and channel openings are upward deflections from zero current. Two channels are present in the recordings from the large-conductance septal channel (middle two traces); the other recordings are from experiments in which only one channel was present in the bilayer. Currents were filtered to 300 Hz.

Proc. Natl. Acad. Sci. USA 88 (1991)

2487

mV -40 -30 -20 -10

0

10

20

30

40

pA

FIG. 2. Comparison of the single-channel current-voltage relationship of the different calcium-release channels observed in heart. Data from this series of experiments are plotted with reported data (14) obtained with sarcoplasmic reticulum vesicles from whole, unselected ventricular muscle (A) (14). The slope from leftventricular free wall (v) and from whole, unselected ventricle is 100 pS. The slope conductances of the two channels isolated from septum are 31 pS (e) and 17 pS (o).

channels were slower than the free wall channels. Upon further analysis, the septal channels had a mean open time of 65 ms and mean closed time of 180 ms, compared with 17 ms and 18 ms for the free-wall channels. Only two experiments were suitable for kinetic analysis because in all our other experiments at least two septal channels were present in the bilayer. As it was usually possible to measure a single ventricular free-wall channel, the apparent clustering of septal channels may represent another difference between these regions of the canine heart. Like the channels obtained from the free wall, those from the septum were inhibited by ruthenium red. Despite the observed differences between the channels from the septum and free wall, vesicles from both regions of the heart bound ryanodine with similar affinity and maximum number of binding sites (Kd for septum, 3.2 ± 0.16 nM and for free wall, 2.7 ± 0.28 nM; Bm. for septum, 6.3 ± 0.36 pmol/mg and for free wall, 7.4 ± 2.2 pmol/mg (X ± SEM; five preparations; J.W. and B.E.E., unpublished observations). When septal vesicles were studied, a second, smaller channel was observed in =30% of the experiments (Fig. 1, bottom two traces). The slope conductance for these channels was 17 pS (Fig. 2, open circles). This smaller channel from septum was seen in the same membrane as the 31-pS channel (Fig. 3, bottom two traces) and in membranes lacking the 31-pS channel (Fig. 3, middle two traces) as well as after inhibition of the 31-pS channel with ruthenium red (data not shown). Although in the absence of inositol 1,4,5trisphosphate (InsP3) these channels were observed rarely (in 1 of 10 experiments), in four experiments InsP3 was added to determine whether in septum there was an InsP3-gated channel related to the InsP3-gated channel found in aortic smooth muscle (13) and in scallop muscle (17). In every case, addition of 1-3 ,uM InsP3 increased the activity of the 17-pS channel from a PO of much less than 1% (actually 0% in three experiments) to a PO of -5%. It was difficult to quantitate the increased activity after InsP3 addition due to the small size of the channel currents in comparison with the background current noise. It was not possible to determine whether the channel was inhibited by heparin because addition of heparin at 20 ,g/ml activated the 31-pS channel despite the presence of ruthenium red. The activating effect of heparin on the calcium-release channel has been described in skeletal sarcoplasmic reticulum (ref. 18; J.W. and B.E.E., unpublished observations). Thus, this 17-pS channel is more similar to the InsP3-gated channel of aortic smooth muscle and scallop in

Physiology/Pharmacology: Borgatta et al.

2488

Proc. Natl. Acad. Sci. USA 88 (1991) ferent regions of the heart have different calcium-release channels.

..

ON40volov"m 0 .

J.

." -.."

DISCUSSION

-

.W..

2n P

FIG. 3. Comparison of single-channel recordings from the channels isolated from septum in the presence of 0.1 ,AM free calcium, 5 mM caffeine, and 3 ,uM InsP3 taken from different experiments. The top two traces are recordings of the 31-pS channel. Two channels are present in this experiment. The middle two traces show the 17-pS channel alone, and the bottom two traces show the 31-pS and 17-pS channels together. Note that the current scale is 2.5 times more sensitive than in Fig. 1.

that it is not activated by caffeine or inhibited by ruthenium red but is activated by InsP3. The conductance is also similar to that of other InsP3-gated channels (13, 17). When sarcoplasmic reticulum vesicles made from whole, unselected ventricular muscle were studied, two classes of calcium-permeable channels were observed (Fig. 4 Middle). The larger channel was 100 pS, and this channel could be activated by caffeine and inhibited by ruthenium red, as described (13-16) for the calcium-release channel/ryanodine receptor. The smaller class of channels was 31 pS. This channel also could be activated by caffeine and inhibited by ruthenium red. We noticed that the two channel types were rarely observed in the same bilayer. When the current at 0 mV was plotted for all of the analyzed experiments with interventricular septum and left-ventricular free wall (Fig. 4 Left and Right, respectively), there is complete separation in the current amplitudes, supporting our hypothesis that dif-

pA

septum

unselected ventricle

free

wall"

FIG. 4. Comparison of the single-channel current at 0 mV for each of the experiments of sarcoplasmic reticulum channels from septum (Left), whole unselected ventricle from previous data (Middle; ref. 14), and left-ventricular free wall (Right). Each dot represents the mean current from a single experiment. Number of recorded experiments with the 31-pS channel (current of 1.9 pA) in the previous series of experiments (Middle) may be underestimated because we discarded many of these experiments to focus on a study of the 100-pS channel (current of 4.6 pA). Horizontal and vertical lines to right of dots indicate the means SDs of the single-channel current for all experiments in that group. **, Comparison of unitary currents of septal and free-wall channels showed currents to be significantly different (P < 0.001; Student's t test). ±

In this paper we have described two calcium-release channels that differ in single-channel conductance and in open and closed times. One channel has a conductance of 100 pS and appears to be localized to the free wall of the ventricle, whereas the other channel has a conductance of 31 pS and is found predominantly in the interventricular septum. Pharmacologically, however, the channels from both regions are similar, and the specificity of the agonists and antagonists support the hypothesis that both channels are ryanodine receptors (calcium-release channels). Recently, evidence has been presented to suggest that there are isoforms of the calcium-release channel in avian skeletal muscle (19, 20). These two channel proteins differ in molecular weight, immunological cross-reactivity, and peptide maps (19); moreover, the expression of these two channel proteins changes during development (20). A third channel protein isoform may also be present in these avian skeletal muscles (20). Thus, there is a precedent for the existence of calcium-channel isoforms, although little is known about the extent of the functional and/or structural differences between the channel isoforms. The observation of an InsP3-gated channel from heart was unexpected because InsP3-induced calcium release was not detected in vesicles made from whole, unselected ventricle (11, 13) despite reports of InsP3-induced contractions in permeabilized or skinned ventricular muscle (e.g., refs. 21 and 22). However, the probability of observing an InsP3gated channel in bilayer experiments was low, suggesting that few of the septal vesicles had InsP3-gated channels, and hence these channels would be difficult to detect by using a release assay, especially if whole unselected ventricle was studied. It is interesting to note that the InsP3-gated channels appear concentrated in a region of heart containing specialized conducting tissue that has many characteristics similar to neurons, some of which have the highest known density of InsP3 receptors (23). A possible role for the unexpected differences we observed in the calcium-release channels of the septum and free wall of the left ventricle was suggested over 50 yr ago by Carl Wiggers (24), who in the first edition of his classical textbook noted that during ventricular contraction, "blood is not merely pressed out by a decrease in their cavities; it is virtually wrung out. The effectiveness of contraction is further enhanced by the fact that the thick interventricular septum is excited first and becomes a point of fixation for the contracting spirals of muscles [of the free wall] which terminate there." These early observations raise the possibility that the smaller conductance and, possibly more densely packed, septal channels are adapted to provide for sustained tension in the septum, which is the first part of the ventricles activated. Sustained contraction of the septum would "anchor" contraction of the free wall and also allow the right and left ventricles to empty in their different manners: the left ventricle as a pressure pump; the right ventricle, which develops much less pressure, as a volume pump. In contrast, the larger conducting channels of the free wall may be specialized to allow these portions of the left ventricle to carry out their more rapid "squeezing" motion, which ejects blood into the aorta. The present study adds to growing evidence for a remarkable heterogeneity in myocardial cell composition that contributes to the adaptation of form to function, which is essential for efficient contraction of the heart (2). Our results also suggest that mechanical abnormalities in the hypertro-

Physiolcoggy/Phannacology: Borgatta et al. phied and failing heart (25) may be due, in part, to changes in the regional expression of calcium-release channel isoforms. We thank Dr. Karol Ondrias for his comments for improving the experiments described here and for his encouragement to forge ahead despite numerous technical snafus; Drs. R. Klonoski and F. Mkparu for assistance with the ryanodine-binding assays; and Drs. I. Bezprozvanny, W. K. Chandler, L. B. Cohen, C. Miller, and S. Garber for their critical comments on the manuscript. This work was supported by National Institutes of Health Grants HL-33026 and GMS-39029. B.E.E. is a Pew Scholar in the Biomedical Sciences. 1. Lewis, T. (1915) Lectures on the Heart (Hoeber, New York). 2. Katz, A. M. & Katz, P. B. (1989) Circulation 79, 712-717. 3. Sartore, S., Gorza, L., Pierobon Bormioli, S., Dalla Liber, L. & Schiaffino, S. (1981) J. Cell Biol. 88, 226-233. 4. Bouvagnet, P., Leger, J., Dechesne, C., Pons, F. & Leger, J. J. (1984) Circ. Res. 55, 794-804. 5. Spach, M. S., Dolber, P. C. & Anderson, P. A. W. (1989) Circ. Res. 65, 1594-1611. 6. Bugaisky, L. B., Anderson, P. G., Hall, R. S. & Bishop, S. P. (1990) Circ. Res. 66, 1127-1132. 7. Litovsky, S. H. & Antzelevich, C. (1988) Circ. Res. 62, 116126. 8. Kimura, S., Bassett, A. L., Furukawa, T., Cuevas, J. & Myerburg, R. (1990) Circ. Res. 66, 469-477. 9. Nilius, B., Hess, P., Lansman, J. B. & Tsien, R. W. (1985) Nature (London) 316, 443-446. 10. Armstrong, C. M. & Matteson, D. R. (1985) Science 227, 65-67.

Proc. Natl. Acad. Sci. USA 88 (1991)

2489

11. Watras, J. & Benevolensky, D. (1987) Biochim. Biophys. Acta 931, 354-363. 12. Smith, J., Coronado, R. & Meissner, G. (1986) J. Gen. Physiol. 88, 573-588. 13. Ehrlich, B. E. & Watras, J. (1988) Nature (London) 336, 583-586. 14. Ondrias, K., Borgatta, L., Kim, D. H. & Ehrlich, B. E. (1990) Circ. Res. 67, 1167-1174. 15. Rousseau, E., Smith, J., Henderson, J. & Meissner, G. (1986) Biophys. J. 50, 1009-1014. 16. Rousseau, E. & Meissner, G. (1989) Am. J. Physiol. 256, H328-H333. 17. Ondrias, K., Castellani, L. & Ehrlich, B. E. (1990) Biophys. J. 57, 276 (abstr.). 18. Ritov, B. R., Men'shikova, E. V. & Kozlov, Y. P. (1985) FEBS Lett. 188, 77-80. 19. Airey, J. A., Beck, C. F., Murakami, K., Tanksley, S. J., Deerinck, T. J., Ellisman, M. H. & Sutko, J. L. (1990) J. Biol. Chem. 265, 14187-14194. 20. Deerinck, T., Ellisman, M., Beck, C., Nichol, J. & Sutko, J.

(1990) Biophys. J. 57, 168 (abstr.). 21. Vites, A.-M. & Pappano, A. (1990) Am. J. Physiol. 258,

H1745-H1752. 22. Fabiato, A. (1986) Biophys. J. 49, 190 (abstr.). 23. Ross, C., Meldolesi, J., Milner, T. A., Satoh, T., Supattopone, S. & Snyder, S. H. (1989) Nature (London) 339, 468-470. 24. Wiggers, C. J. (1934) Physiology in Health and Disease (Lea & Febiger, Philadelphia), p. 561. 25. Katz, A. M. (1990) N. Engl. J. Med. 322, 100-110.

Regional differences in calcium-release channels from heart.

The heart is a heterogeneous tissue composed of several cell types tailored for specialized functions. We found that intracellular channels also exhib...
770KB Sizes 0 Downloads 0 Views