Review

In this review the molecular characteristics and reaction mechanisms of different Ca(2+) transport systems associated with various membranes in muscle cells will be summarized. The following topics will be discussed in detail: a brief history of early observations concerning maintenance and regulation of cellular Ca(2+) homeostasis, characterization of the Ca(2+) pumps residing in plasma membranes and sarco(endo)plasmic reticulum, mitochondrial Ca(2+) transport, Ca(2+)-binding proteins, coordinated expression of Ca(2+) transport systems, a general background of muscle excitation-contraction coupling with emphasis to the calcium release channels of plasma membrane and sarcoplasmic reticulum, the structure and function of dihydropyridine and ryanodine receptors of skeletal and cardiac muscles, and finally their disposition in various types of muscles.


Brief history of early observations
The recognition of the role of Ca 2+ as the regulator of a broad range of cellular processes is one of the great accomplishments of the 20th century biology (Cambell, 1983;Carafoli & Klee, 1999;Putney, 2000).It evolved from Ringer's observations that Ca 2+ is required for cardiac muscle contractility (Ringer, 1883), continued with observations of a similar role of Ca 2+ in skeletal muscle (Kamada & Kinoshita, 1943;Heilbrunn & Wiercinsky, 1947), and culminated in the discoveries of the Ca 2+ -transport ATPase of SR (Ebashi, 1961;Hasselbach & Makinose, 1961) and the regulation of actomyosin activity by Ca 2+ (Weber, 1959) through its interaction with troponin (Ebashi, 1963;Ebashi et al., 1999).In the intervening years the messenger role of Ca 2+ has been demonstrated in virtually all living cells and all cellular compartments, in processes as diverse as cell division, cellular motility, hormone secretion, intermediary metabolism, neural activity, protein trafficking, gene expression, developmental regulation and apoptosis (Berridge et al., 1998;Brini & Carafoli, 2000;Martonosi, 2000;Berridge, 2001;Mellstrom & Naranjo, 2001;Bootman et al., 2002).
The cytoplasmic free Ca 2+ concentration of muscle and other cells at rest is around 20-50 nM; this is 10 3 -10 4 times lower than the free Ca 2+ concentration in the extracellular space (usually millimolar) or in the lumen of sarco(endo)plasmic reticulum (SR/ER) (0.1-2.0 mM).The large Ca 2+ gradients across cellular boundaries are established and maintained by powerful Ca 2+ pumps located in the cell surface membranes, and in the endoplasmic reticulum (MacLennan et al., 1997;Guerini & Carafoli, 1999;Philipson & Nicoll, 2000), with contributions by the mitochondria (Rizzuto et al., 2000;Duchen, 2000).

The Ca 2+ pumps of the plasma membranes
The plasma membrane Ca 2+ -ATPase (PMCA).A (Mg 2+ + Ca 2+ )-activated ATPase of about 130 kDa is present in the plasma membranes of most eukaryotic cells, including plants (Guerini & Carafoli, 1999;Strehler & Zacharias, 2001).It catalyzes the ATP-dependent transport of Ca 2+ from the cytoplasm into the extracellular space.The Ca 2+ -dependent cleavage of ATP occurs with the transient formation of an acylphosphate enzyme intermediate, that is linked to Ca 2+ translocation.The enzyme has high affinity for Ca 2+ when complexed with calmodulin (K m < 0.5 mM), that serves as a physiological activator; other activators are acidic phospholipids and phosphorylation by several protein kinases.The predicted structure of PMCA contains 10 transmembrane domains; about 80% of its mass protrudes into the cytoplasm, and only short loops are exposed to the outside.
Mammalian PMCAs are encoded by four distinct genes, yielding the PMCA1-4 isoforms (Table 1), (Strehler & Zacharias, 2001).Additional isoform variants are generated by alter-native splicing of the primary gene transcripts at two major conserved locations, that correspond to the regulatory domains of the pump.PMCA1 and 4 are present in most adult tissues, while PMCA2 and PMCA3 are primarily expressed in the nervous system and in muscles.The distinct functional characteristics of the isoforms, their tissue-specific localization, and the dynamic regulation of their expression during development and in adult animals suggest specific roles for each isoform in Ca 2+ homeostasis.
The surface membrane Ca 2+ pumps are of primary importance in the regulation of total cell Ca 2+ content.Their cellular concentration and activity are adjusted to compensate for the influx of extracellular Ca 2+ , but very little is known about the mechanism of this regulation.In healthy red blood cells the ATP-dependent Ca 2+ pump is sufficient to balance the relatively slow influx of extracellular Ca 2+ .In excitable cells (heart, skeletal muscle, nervous system), where a large influx of Ca 2+ occurs during activity, the Na + :Ca 2+ exchange system is also required to restore the cell Ca 2+ content to normal levels.The Na + :Ca 2+ exchanger (NCX).The Na + :Ca 2+ exchangers catalyze the electrogenic exchange of 3 Na + for 1 Ca 2+ across the surface membrane (Blaustein & Lederer, 1999;Philipson & Nicoll, 2000).The exchange is driven by the transmembrane gradient of Na + , that in turn is maintained by the Na + , K + -transport ATPase (Skou, 1985;Sweadner & Donnet, 2001;Kaplan, 2002).The net direction of Na + :Ca 2+ exchange is determined by the Na + and Ca 2+ concentrations in the medium and the cytoplasm, and by the membrane potential.The primary role of the exchanger is the export of Ca 2+ from the cell when the cytoplasmic Ca 2+ concentration is elevated upon activation.Due to its relatively low Ca 2+ affinity (K m >1 mM), it is likely to operate most efficiently in its Ca 2+ efflux mode in areas of high local Ca 2+ concentration at the peaks of Ca 2+ transients.Under some special conditions the exchanger may reverse and catalyze net Ca 2+ import, but the physiological significance of this is not clearly established.
The exchanger is present in particularly high concentration in the surface membranes of excitable cells (heart and nerve cells).The Na + :Ca 2+ exchange is not coupled to the hydrolysis of ATP, but ATP serves as a physiological activator by phosphorylation and by promoting the synthesis of phosphatidylinositol 4,5-bisphosphate.In addition to serving as substrates for transport, Na + and Ca 2+ exert distinct regulatory effects on the exchange activity.Na + inactivates the exchanger by binding to the internal Na + transport sites The three known mammalian Na + :Ca 2+ exchangers (NCX1-3) are encoded by distinct genes (Philipson & Nicoll, 1999).NCX1 and 2 are ubiquitously expressed in various tissues, while NCX3 is confined to the brain.Several splice variants of NCX1 have been detected that are functionally distinct and show tissue specific distribution.
The full-length mature cardiac sarcolemmal NCX1 contains 938 amino-acid rsidues.In contrast to earlier predictions of 11 transmembrane helices, its revised topology now contains 9 transmembrane domains with a large cytoplasmic loop of about 550 aminoacid residues between transmembrane helices 5 and 6 (Philipson & Nicoll, 1999).The N-terminus is glycosylated at position 9 and presumed to be extracellular.The large cytoplasmic loop contains the regulatory Ca 2+ -binding sites and the regions of alternative splicing.The human NCX1 gene is organized into 12 exons and is located on chromosome 2.The mouse NCX1, NCX2 and NCX3 genes were mapped to chromosomes 17, 7, and 12, respectively.The Ca 2+ pump of sarco(endo)plasmic reticulum (SERCA).The ATP-dependent Ca 2+ pumps of sarco(endo)plasmic reticulum (SERCA) constitute a large family of proteins of 100-110 kDa (de Meis, 1995;Moller et al., 1996;Toyoshima et al., 2000;Toyoshima & Nomura, 2002;Lee, 2002;MacLennan et al., 2002).They are structurally distinct from the Ca 2+ pumps of the plasma membrane, but share similarities in the mechanism of Ca 2+ translocation.The intracellular location of SERCA ATPases exclusively in SR/ER membranes is maintained by the presence of specific retention/retrieval motifs in their primary sequences (Newton et al., 2003).
The SR Ca 2+ pump catalyzes the electrogenic transport of 2 Ca 2+ ions per one ATP molecule hydrolyzed from the cytoplasm into the lumen of SR.The key steps of the process are outlined in Fig. 1.The Ca 2+ transport is reversible and under favorable conditions results in the formation of 1 ATP molecule for 2 Ca 2+ ions released from the lumen of SR (Sumbilla et al., 2002).The transfer of 2 Ca 2+ ions represents four positive charges.Countertransport of H + and fluxes of ions through the anion and cation channels of SR prevent large changes in membrane potential during Ca 2+ transport (Yu et al., 1993).
The SERCA pumps have high affinity for Ca 2+ (K m about 0.1 mM), and are capable to maintain a resting cytoplasmic [Ca 2+ ] of 10-20 nM.Like in other P-type ATPases, transient phosphorylation of the enzyme by ATP on an active site aspartyl group is a key step in Ca 2+ translocation.The activity of the Ca 2+ pump is dependent on membrane phospholipids, but in distinction from the plasma membrane Ca 2+ -ATPase, it is not influenced by calmodulin.The recently obtained high resolution structure of SERCA1 Ca 2+ -ATPase in two conformations (Toyoshima et al., 2000;Toyoshima & Nomura, 2002;Lee, 2002) will reveal the structural basis of Ca 2+ translocation and the mechanism of its regulation.
The SERCA1a and SERCA1b isoforms are preferentially expressed in adult and neonatal fast-twitch skeletal muscles, respectively.SERCA2a is the principal isoform in slow-twitch skeletal, cardiac and neonatal skeletal muscles, but is also expressed in non-muscle cells.The SERCA2b isoform is primarily located in smooth muscles and in non-muscle cells.The SERCA3 gene product is broadly distributed in various muscle and non-muscle cells.The differential but overlapping expression pattern and regulation of the various SERCA isoforms is a key feature of the functional plasticity of muscles displayed in response to altered physiological requirements.
The density of Ca 2+ transport sites in the SR membrane is very high (about 30 000/mm 2 surface area) and in fast-twitch fibers the Ca 2+ transport ATPase may account for 80% of the protein content of the SR membrane (Fig. 2).Due to the extensive development of SR in most types of muscles, and the high density of Ca 2+ transport sites within the membrane, 1-2 cycles of Ca 2+ transport are usually sufficient to restore resting cytoplasmic [Ca 2+ ] during muscle relaxation.The Ca 2+ transport activity of fast-twitch muscle fibers is greater than that of slow-twitch muscle fibers, nearly in proportion to their rate of relaxation.Experimentally induced changes in fiber type composition are reflected in corresponding changes in Ca 2+ transport activities.Transcriptional control of the expression of Ca 2+ -ATPase and contractile protein isoforms (Martonosi, 2000;Pette, 2001), together with fiber type specific proteolysis (Sultan et al., 2001), are involved in this adaptation, but the detailed mechanism of the process is unknown.

Mitochondrial Ca 2+ transport
Ca 2+ influx pathways of mitochondria.Isolated mitochondria are capable to accumulate large amounts of Ca 2+ in the presence of inorganic phosphate through a Ca 2+ uniport of low apparent Ca 2+ affinity (app.K m >10 mM) located in the inner mitochondrial membrane (Bernardi, 1999;Duchen, 2000;Buntinas et al., 2001).The Ca 2+ uptake is driven by the negative membrane potential of -80 to -200 mV across the inner mitochondrial membrane generated either by the respiratory chain or by the hydrolysis of ATP.In the presence of phosphate the transported Ca 2+ is precipitated in the mitochondrial matrix as Ca 2+ -phosphate and the Ca 2+ loading capacity of mitochondria may exceed the total Ca 2+ content of the cell.Although the mitochondrial Ca 2+ uniport was discovered nearly a half century ago, it has never been isolated in pure form, and its molecular structure is unknown.Its kinetic properties were characterized largely through flux studies.
The Ca 2+ uptake rates of mitochondria are steeply dependent on cytoplasmic [Ca 2+ ], suggesting that the Ca 2+ uniport channel opens only at elevated cytoplasmic [Ca 2+ ].ATP, inorganic phosphate (P i ) and Mg 2+ modulate its opening.The Ca 2+ flux through the open uniport may be greater than 10 000 ions per second, i.e., faster than most ion pumps (Babcock & Hille, 1998).At cytosolic free Ca 2+ concentrations of 0.1-1.0 mM with a mitochondrial membrane potential of -180 mV, the equilibrium free Ca 2+ concentration in the mitochondrial matrix ([Ca 2+ ] M ) should rapidly reach 0.1-1.0 mM (Rizzuto et al., 2000); the actually measured free [Ca 2+ ] M in resting cells using mitochondrially targeted aequorin or fluorescent Ca 2+ indicators is only 100-200 nM, i.e., a tiny fraction of its expected equilibrium value.These data are consistent with observations by electron X-ray microanalysis and by electron energy loss imaging, that show surprisingly low total Ca 2+ content in the mitochondria of various cells.For example, the total Ca 2+ content of frog skeletal muscle mitochondria measured by electron X-ray microprobe was only 1.7 mmol/kg dry mass, while the corresponding value in the SR was 117 mmol/kg (Somlyo et al., 1981).The mitochondrial Ca 2+ content of muscle did not change significantly during tetanus or fatigue.Similar low values were obtained for the mitochondria of liver and other cells (Somlyo et al., 1986).These mitochondrial Ca 2+ contents represent only about 5% of the total cell Ca 2+ .The low mitochondrial Ca 2+ content of resting cells in vivo is consistent with the low apparent Ca 2+ affinity of mitochondrial Ca 2+ transporter measured in vitro (K m >10 mM) that would render it essentially inactive at the resting cytosolic [Ca 2+ ] of 20-50 nM.These observations led to the suggestion that in differentiated skeletal muscle and perhaps in most other cells, mitochondria play only a secondary role in the regulation of cytoplasmic [Ca 2+ ].
The main physiological role of mitochondrial Ca 2+ transport is the control of Ca 2+ concentration in the mitochondrial matrix ([Ca 2+ ] M ), upon which the activity of key enzymes of mitochondrial energy metabolism depends (Kunz, 2001;Jouaville et al., 1999).These include three dehydrogenases (McCormack & Denton, 1999) and the F O F 1 -ATPase (Territo et al., 2000), all of which are activated by submicromolar [Ca 2+ ] M .The activation of NAD-isocitrate dehydrogenase and a-ketoglutarate dehydrogenase by Ca 2+ is allosteric and increases their affinity for the oxidizable substrate.In the case of pyruvate dehydrogenase Ca 2+ activates a protein phosphatase that converts the inactive phosphorylated form of pyruvate dehydrogenase into an active dephosphorylated form (PDH A ) (Reed et al., 1985).
Increases in cytoplasmic [Ca 2+ ] caused by a variety of interventions (membrane depolarization, agonists, hormones, etc.) causes rapid increase in the [Ca 2+ ] M (Robert et al., 2001), that is followed by increase in the concentration of NADH due to activation of dehydrogenases (McCormack & Denton, 1999), increase in PDH A (Reed et al., 1985), increased oxygen consumption, and elevation of ATP concentration (Jouaville et al., 1999;Territo et al., 2001).The Ca 2+ activation of mitochondrial ATP synthesis coincides with the Ca 2+ activation of muscle contraction and other ATP-dependent cellular processes.
The electrophoretic Ca 2+ uniport is inhibited by ruthenium red (RR) that also blocks the activation of Ca 2+ -sensitive dehydrogenases and the Ca 2+ agonist-induced stimulation of ATP synthesis (Rizzuto et al., 2000;Territo et al., 2001).Liver and heart mitochondria contain a RR-insensitive Ca 2+ uptake mechanism, designated as rapid uptake mode (RAM) (Bernardi, 1999;Rizzuto et al., 2000;Buntinas et al., 2001).The RAM has much higher conductivity at the beginning of a pulse than the uniporter, but this conductivity rapidly declines as the [Ca 2+ ] rises, due to Ca 2+ binding to an external regulatory site (Buntinas et al., 2001).Little is known about the physiological role of RAM.
In view of the low apparent Ca 2+ affinity of the mitochondrial Ca 2+ uniport, it was suggested that the changes in [Ca 2+ ] M are more closely related to the large amplitude fluctuations of [Ca 2+ ] in microdomains of the cytoplasm adjacent to mitochondria (such as the space between the mitochondria and the SR or the plasma membrane) than to the average cytoplasmic [Ca 2+ ] derived from measurement with Ca 2+ indicators (Rizzuto et al., 2000).This would imply that mitochondrial Ca 2+ transport involves mainly the portion of mitochondrial surface that faces the Ca 2+ release channels of the SR or the Ca 2+ influx pathways of the plasma membrane (Hajnoczky et al., 2000;Seppet et al., 2001).
The mitochondrial [Ca 2+ ] M under physiological conditions probably do not exceed a few mmol per litre; at much higher [Ca 2+ ] would uncouple the mitochondrial oxidative phosphorylation.Such uncoupling was found to be negligible up to 600 nM [Ca 2+ ] M (Territo et al., 2000;2001).The [Ca 2+ ] M reflects a steady state between Ca 2+ influx and Ca 2+ efflux.The kinetic properties of these pathways effectively balance the [Ca 2+ ] M within a physiological concentration range.
Pathways of Ca 2+ efflux from mitochondria.The equilibration of mitochondrial [Ca 2+ ] M under the influence of the mitochondrial membrane potential is opposed by Na + -dependent and Na + -independent Ca 2+ exporters (Bernardi, 1999;Rizzuto et al., 2000), and by a non-selective high conductance permeability transition pore (PTP) (Rizzuto et al., 2000;Di Lisa et al., 2001) that facilitate the export of Ca 2+ from the mitochondrial matrix space back into the cytoplasm.
The Na + -dependent Ca 2+ efflux is catalyzed by the mitochondrial Na + :Ca 2+ exchanger, that is distinct from the Na + :Ca 2+ exchanger of the plasma membrane.The mitochondrial Na + :Ca 2+ exchanger operates with a probable stoichiometry of 3 Na + :1 Ca 2+ ; its K m for Na + is 8-10 mM and its V max is 18 nmol Ca 2+ /mg protein per min in the heart.The Ca 2+ efflux is inhibited by Sr 2+ , Ba 2+ , Mg 2+ , Mn 2+ , ruthenium red, verapamil, diltiazem and by the specific inhibitor CGP37157 (Co & Matlib, 1993).The Na + -dependent Ca 2+ efflux is primarily responsible for balancing the electrophoretic Ca 2+ uniport, and maintaining mitochondrial Ca 2+ loads within the physiological range of 1-4 mmoles Ca 2+ /mg protein.
The Na + -independent Ca 2+ efflux is insensitive to RR and may occur via an nH + :Ca 2+ exchange mechanism (Kunz, 2001).It saturates at a Ca 2+ load of 25 nmoles/mg protein and its V max is only 1.2 nmoles Ca 2+ /mg protein per min (Rizzuto et al., 2000).
The PTP opening is regulated by membrane potential, Ca 2+ , pH, and inhibited by cyclosporin A (Rizzuto et al., 2000;Di Lisa et al., 2001).PTP opening causes a Ca 2+ -dependent increase in the permeability of mitochondria to ions and molecules up to 1500 Da, including NAD + (Di Lisa et al., 2001).The physiological role of PTP is not clear.It contributes to cell death associated with post-ischemic reperfusion injury of the heart (Di Lisa et al., 2001), and perhaps to other forms of apoptosis (Duchen, 2000;Ferri & Kroemer, 2001).
In the resting muscle much of the cell Ca 2+ content is stored in the SR lumen and a major portion of SR Ca 2+ is bound to the low affinity Ca 2+ of the lumenal Ca 2+ -binding proteins.Upon activation of muscle, Ca 2+ is rapidly released from SR through the ryanodine-sensitive Ca 2+ channels (Meissner, 1994;Ogawa et al., 2002) with increase in the [Ca 2+ ] of the cytoplasm, mitochondria and nuclei.Depending on the amplitude and kinetics of the Ca 2+ signal, the high affinity Ca 2+ -binding proteins of these compartments rapidly bind Ca 2+ .In skeletal and cardiac muscle much of the Ca 2+ released from the SR during activation is bound to troponin causing activation of actomyosin ATP hydrolysis and muscle contraction (Holmes & Geeves, 2000;Huley, 2000;Gordon et al., 2000).Simultaneous activation of glycolysis and oxidative phosphorylation serve to maintain the ATP concentration of the muscle at nearly constant level, despite the high rate of ATP utilization (Booth & Thomason, 1991).Depending on the frequency and amplitude of Ca 2+ signals, other components of the Ca 2+ signalling networks, including those involved in the regulation of gene expression, are also activated permitting an accurate matching of both the activity and the concentration of the various enzymes to the physiological demands.
The Ca 2+ pumps of the plasma membrane and the SR have relatively slow turnover rates (<10 s -1 ) compared with the conductance of Ca 2+ release channels (10 000 s -1 ) that generate the Ca 2+ signals.Therefore, the Ca 2+ pumps have only modest effect on the rapidly rising phase of the Ca 2+ transients during muscle activation.By contrast, troponin and parvalbumin bind Ca 2+ rapidly and most of the activating Ca 2+ ends up bound to these proteins.
At the beginning of muscle relaxation, the repolarization of the plasma membrane causes the closure of the Ca 2+ release channels of SR.The cytoplasmic free Ca 2+ concentration begins a gradual decline toward resting level, due to the operation of Ca 2+ pumps and the continued binding of calcium to troponin, parvalbumin and other high affinity Ca 2+ -binding proteins.The Ca 2+ binding to troponin causes continued increase in contractile tension, while the cytoplasmic free [Ca 2+ ] is lowered to near resting level (Ashley et al., 1993).Muscle relaxation is eventually brought about by the slow release of bound Ca 2+ from troponin.The relationship between contractile tension and the kinetics of the Ca 2+ transient is consistent with the kinetics of Ca 2+ binding to troponin measured in vitro, and was realistically modeled in frog skeletal muscle assuming that the concentration of troponin, parvalbumin and the SR Ca 2+ pump are 0.24, 1.0, and 0.19-0.49mM, respectively, and the rate constants of Ca 2+ binding are 0.575 ´l0 8 M -1 s -1 for troponin, and 1.25 ĺ0 8 M -1 s -1 for parvalbumin, and the rate constants of Ca 2+ release are 115 s -1 for troponin and 0.5 s -1 for parvalbumin (Baylor et al., 1983).The simulations suggest that following a single action potential, about 300 mM Ca 2+ was released from the SR into the cytoplasm.The free cytoplasmic [Ca 2+ ] rose only to a peak value of 2.1 mM, because much of the released Ca 2+ was bound to troponin (about 110 mM) and parvalbumin (about 170 mM).Repetitive stimuli cause the additional release of only about one third as much Ca 2+ .Parvalbumin is present in resting muscle largely complexed with Mg 2+ ; therefore, the Ca 2+ binding to parvalbumin occurs with a slow displacement of Mg 2+ from its binding site.As a result, Ca 2+ binding to troponin precedes that of parvalbumin during activation and some of the Ca 2+ released from troponin during relaxation is bound by parvalbumin before it is returned to the SR lumen.The buffering action of parvalbumin requires the release of nearly twice the amount of Ca 2+ from the SR than in its absence.
There is a precise stoichiometry between actin, tropomyosin, troponin and myosin in the myofibrils, and the amount of Ca 2+ released from the SR must be sufficient to saturate a large portion of the Ca 2+ -binding sites on troponin, without reaching harmful concentrations in the cytoplasm.This delicate balance requires that the expression of each component must be coordinated to yield cellular concentrations that match the Ca 2+ requirement of the contractile apparatus.The amplitude and time course of the Ca 2+ transient was similar in frog and fast-twitch mouse skeletal muscles (Hollingworth et al., 1996).The slow release of Ca 2+ after each Ca 2+ pulse from the high affinity Ca 2+ -binding sites of the various Ca 2+ receptors is a general feature of Ca 2+ signalling networks that prolongs the physiological effects of Ca 2+ beyond the observed lifetime of the brief cytoplasmic Ca 2+ spikes.

Coordinated expression of calcium transport systems
The relative contribution of the various Ca 2+ transport systems to the intracellular Ca 2+ homeostasis varies with cell type and species as determined by their specific physiological requirements.Variations in the content and isoform composition of the Ca 2+ signalling network of muscle can be imposed by exercise, cross-innervation or chronic electric stimulation with major changes in contractility (Pette, 2001).The range of contribution of individual components of the regulatory network can be tested by their overexpression or ablation, that may also trigger compensatory changes in other components aimed at maintaining the Ca 2+ homeostasis within physiological limits.These data permit an assessment of the robustness of Ca 2+ regulation, defined as the ability of the system to function in the face of large changes in its composition.Such robustness is an important requirement of molecular networks and may serve as an evolutionary principle in their development.
Comparison of Ca 2+ regulation in skeletal and cardiac muscles.In skeletal muscles there is only small influx of extracellular Ca 2+ during activity (Bianchi & Shanes, 1959) and contractility can be maintained for a long time in virtually Ca 2+ -free medium (Armstrong et al., 1972).As over 99% of the activating Ca 2+ originates from the SR, the total Ca 2+ content ing the cytoplasmic Na + with choline, that inhibits the mitochondrial Na + :Ca 2+ exchanger (Bers et al., 1993).Similar conclusions were reached by Balke et al. (1994) who also found negligible contribution by mitochondria and the plasma membrane Ca 2+ -ATPase to Ca 2+ removal during a single Ca 2+ transient in rat ventricular myocytes following inhibition of Na + :Ca 2+ exchangers by the removal of internal and external Na + .Under these conditions the declining phase of the Ca 2+ transient could be modeled considering only the kinetics of SR Ca 2+ -ATPase (K m , V max ), the complexation of Ca 2+ by the Ca 2+ -binding proteins and the leak of Ca 2+ from the SR into the cytoplasm (F SR leak).The values of K m (0.28 mM), V max (0.21 mM s -1 ) and F SR leak (0.018 mM s -1 ) derived from these data are consistent with independent estimates based on in vitro data.
Contrasting these observations, a more recent study suggests (Choi & Eisner, 1999) that the contribution of plasma membrane Ca 2+ -ATPase to the Ca 2+ removal in rat ventricular myocytes may reach 25% of the contribution of Na + :Ca 2+ exchanger.Inhibition of PMCA by carboxyeosin increased the resting Ca 2+ content of SR by 17%, suggesting that in the resting cell the plasma membrane Ca 2+ -ATPase decreased the Ca 2+ content of SR by competition with the SR Ca 2+ -ATPase.
Overexpression and ablation of components of Ca 2+ homeostasis.The individual components of the Ca 2+ regulatory network are expressed in proper concentration and isoform composition to maintain the total Ca 2+ concentration of the cell and the free Ca 2+ concentrations of the various organelles within physiological ranges both during rest and activity.The robustness of the design was tested by manipulating the expression of the various isoforms of the SR and plasma membrane Ca 2+ -transport ATPases and Na + :Ca 2+ exchangers, followed by analysis of the resulting changes in the pattern of Ca 2+ regulation (Shull, 2000;Kiriazis & Kranias, 2000;Periasamy & Huke, 2001).The genetic manip-ulation of the individual components identified the limits of plasticity in the Ca 2+ signalling network, but also revealed major gaps in our knowledge about the integration of their function and the mechanisms that coordinate their expression.
Reduction of SERCA Ca 2+ -ATPase content in Brody's disease.Brody' disease is a rare genetically heterogeneous, autosomal recessive or dominant inherited disorder of skeletal muscles, characterized by painless contractures that develop during exercise (Brody, 1969).The slow relaxation of muscle is caused by decreased Ca 2+ transport activity of SR due to mutations in the ATP2A1 gene coding for the SERCA1 isoform of the Ca 2+ -ATPase of fast-twitch (type 2) skeletal muscle fibers (MacLennan, 2000;Odermatt et al., 2000).Some of the mutations cause virtually complete loss of both SERCA1 protein and the Ca 2+ -ATPase activity, presumably due to its defective translation and/or degradation.In other cases, the rate of Ca 2+ transport was reduced by about 50% without change in the content of Ca 2+ -ATPase protein (Taylor et al., 1988;Benders et al., 1996).There may also be cases that are not associated with defects in the ATP2A1 gene (Zhang et al., 1995), and involve other unidentified proteins that may modulate SERCA1 activity.The SERCA1a protein accounts for at least 99% of the SERCA isoforms expressed in adult Type 2 fast-twitch muscle fibers.In view of its major role, it is surprising that its severe deficiency or virtual absence in Brody's disease causes a benign clinical condition that usually does not interfere with normal life and frequently goes undiagnosed.Compensatory mechanisms for SERCA1 deficiency have not been identified so far.Refilling of the Ca 2+ stores of the SR must be a key element of compensation since the activation of contraction that depends on these Ca 2+ stores is not affected significantly in Brody's disease.The possible contributions of SERCA2, SERCA3, or the capacitative Ca 2+ entry mechanism have been considered and need to be further explored.Ca 2+ transport systems in the plasma membranes or mitochondria are not likely to play a major role in this compensation.
Null-mutation of the ATP2A2 gene.The ATP2A2 gene produces two alternatively spliced isoforms of the SERCA2 ATPase.SERCA2a is the principal isoform in cardiac and slow-twitch skeletal muscles, while SERCA2b is present in smooth muscles and non-muscle cells (Wu et al., 1995).Heterozygous transgenic mice were generated by gene targeting that eliminated the promoter and the 5¢ end of the ATP2A2 gene (Ji et al., 2000).Homozygous mutants were not observed, indicating that SERCA2 serves functions essential for life and there is no satisfactory compensation to rescue the null-mutant.The heterozygous offspring were viable and appeared healthy.In heart samples of the heterozygous animals, the SERCA2 mRNA level was reduced by 45%, and the SERCA2 protein content and Ca 2+ transport activity by 35%, without change in the Ca 2+ -concentration dependence of Ca 2+ transport.These observations indicate that both copies of the ATP2A2 gene are required for normal activity.The loss of one copy of the ATP2A2 gene and the associated decrease in Ca 2+ -transport activity have significant physiological consequences.The mean arterial blood pressure and the left ventricular systolic pressure were lower in heterozygous as compared with control animals; this difference was not observed at maximal-adrenergic stimulation (Periasamy et al., 1999) that relieves the inhibition of Ca 2+ -ATPase by phospholamban (Brittsan & Kranias, 2000;Kiriazis & Kranias, 2000).The reduction in the Ca 2+ content of SR and in the rate of Ca 2+ transport was also reflected in slower rise and decline of the left ventricular pressure wave, dP/dt (Periasamy et al., 1999), and a decrease in the rate of shortening and relaxation of heterozygous myocytes (Ji et al., 2000).These effects are similar to those observed after inhibition of the cardiac Ca 2+ pump by thapsigargin (Kirby et al., 1992).The phospholamban content was reduced by 40% in the transgenic heart, but the basal phosphorylation level of Ser16 and Thr17 of phospholamban was increased approximately 2-fold, relieving the inhibition of Ca 2+ -ATPase by phospholamban and providing some compensation for the decrease in Ca 2+ -ATPase content (Ji et al., 2000).The NCX content was 38% higher in the heterozygous than in control heart, without change in the NCX mRNA level, suggesting posttranscriptional regulation of NCX protein expression.The Ni 2+ -sensitive NCX outward current also increased by about 40% indicating increased NCX activity.
The increase in NCX activity in the forward and reverse directions may facilitate both Ca 2+ extrusion and Ca 2+ entry, providing inotropic support for the heart (Ji et al., 2000).Loss of function mutation of one copy of the ATP2A2 gene is implicated in the human autosomal dominant skin disorder, the Darier disease (Jacobsen et al., 1999;Zhao et al., 2001).This indicates a role for SERCA2 in the cell adhesion of keratinocytes and an as yet uncertain relationship to mood disorders, epilepsy and mental retardation (Jacobsen et al., 1999).Increased incidence of heart disease has not been observed in Darier patients, and Darier-like skin lesions have not been seen in mice carrying null-mutation of ATP2A2 gene, indicating species differences in the susceptibility to skin or heart disease in SERCA2a deficiency.
Ablation of the SERCA3 isoform of Ca 2+ -ATPase.The ATP2A3 gene codes for the SERCA3 Ca 2+ -ATPase that is expressed in endothelial cells, in epithelial cells of the trachea, intestine and salivary glands, in platelets, mast cells, and lymphocytes.Each of these cell types also contain SERCA2b.SERCA3 has lower Ca 2+ affinity and higher pH optimum than SERCA2b, and it is expected to participate in specialized signalling functions consistent with its limited distribution (Wu et al., 1995).Mice homozygous for null-mutation of the ATP2A3 gene were viable, fertile and did not show overt disease phenotype (Liu et al., 1997).However, after con-traction with phenylephrine, the acetylcholine-induced relaxation of their aortic rings was impaired.The acetylcholine-induced relaxation is dependent on NO synthesized by the endothelial cells, that diffuses into the smooth muscle and induces relaxation by a cGMP-dependent mechanism.As the aortas of ATP2A3-/-animals relax in response to sodium nitroprusside, an NO donor, their smooth muscle is apparently sensitive to NO (Liu et al., 1997).Therefore, the defective acetylcholine-induced relaxation in these animals may be due to depletion of the acetylcholine-responsive Ca 2+ stores in the aortic endothelial cells, that would impair the Ca 2+ signalling required for the activation of endothelial NO synthase.Ablation of SERCA3 also interfered with the epithelium-dependent relaxation of tracheal smooth muscle (Kao et al., 1999).Since SERCA3 is not expressed in tracheal smooth muscle, this effect implies a role for SERCA3 in mediating epithelial Ca 2+ signal transduction.Based on these studies, SERCA3 does not appear to play a critical role in basic housekeeping functions essential for cell viability either during development or in mature animals.Rather it has some specific role in Ca 2+ signalling in a restricted set of cells, that requires further investigation.
Overexpression of various isoforms of the Ca 2+ -ATPase in cardiac muscle of transgenic animals provides an opportunity to analyze the effect of Ca 2+ -ATPase content and isoform specificity on the contractility of the heart.Besides their basic biological interest, such studies have a potential therapeutic importance.Cardiac failure and chronic pressure overload cardiac hypertrophy are associated with decreased expression levels of the Ca 2+ -ATPase, altered Ca 2+ homeostasis, and impaired myocardial function, that may be corrected by increasing the Ca 2+ -ATPase content of the heart (Martonosi, 2000).
Overexpression of SERCA2a.The rat SERCA2a transgene was overexpressed in various tissues of mice using the chicken b-actin promoter which is not tissue specific (He et al., 1997), or selectively in the heart using the a-myosin heavy chain promoter (Baker et al., 1998).The level of SERCA2a mRNA increased 2.6-7.9-fold in the transgenic hearts, while the increase in Ca 2+ -ATPase content was only 1.2-1.5-fold.The modest increase in the SERCA2a content may indicate some regulatory mechanisms that control the translation, membrane insertion or degradation of the Ca 2+ -ATPase in the heart.The increased expression of the Ca 2+ -ATPase led to increased rates of ATP-dependent Ca 2+ uptake in cardiac homogenates without change in Ca 2+ affinity, and to increased rates of contraction and relaxation both in isolated myocytes and in work-performing heart (He et al., 1997;Baker et al., 1998).There was no significant change in the phospholamban and Na + :Ca 2+ exchanger content of the transgenic heart (He et al., 1997).The SERCA2a mRNA and protein levels are significantly reduced in hypothyroid as compared with euthyroid mice (Bluhm et al., 1999).Overexpression of SERCA2a in hypothyroid mice compensated for some of the loss of SERCA2a expression due to hypothyroidism and improved the cardiac contractile function.The results obtained with SERCA2a overexpression in transgenic animals confirm the results of earlier studies using adenovirus-mediated transient gene transfer (Miyamoto et al., 2000).Adenoviral-mediated SERCA2a gene transfer also restored contractile function in cardiomyocytes isolated from failing human hearts (Del Monte et al., 1999).Overexpression of SERCA1a.The SERCA isoform specificity of cardiac function was analyzed by overexpression of the rat SERCA1a isoform in mice (Loukianov et al., 1998).The SERCA1a is the major isoform of fast-twitch skeletal muscle fibers, and it is not expressed normally in heart.SERCA1a overexpression increased the total SERCA content of the transgenic heart 2.5-fold, with 1.7-fold increase in the maximal velocity of ATP-induced Ca 2+ uptake, indicating functional expression of the SERCA1a isoform in the cardiac SR.
This was reflected in increased rates of contraction and relaxation both in transgenic myocytes and in work-performing heart preparations (Loukianov et al., 1998).The overexpression of SERCA1a caused a 50% decrease in the SERCA2a content of the transgenic hearts, without change in the phospholamban, calsequestrin, actin and tropomyosin content.Similar observations were made using an adenovirus vector for transient expression of SERCA1a in cardiac myocytes (Cavagna et al., 2000).The high level of expression of SERCA1a in transgenic hearts with a 50% reduction in SERCA2a content suggests that SERCA1a competes effectively with SERCA2a for insertion sites in cardiac SR.With the expression levels obtained in these studies the cardiac SR may have reached physical saturation with the Ca 2+ -ATPase.
Overexpression of SERCA2b.The role of SERCA isoform specificity was further analyzed by overexpression of the rat SERCA2b isoform in transgenic mice (Greene et al., 2000).SERCA2b is found in smooth muscle and in most other cells, but it is only a minor isoform in cardiac muscle.The SERCA2b mRNA increased 20-fold and the SERCA2b protein 8-10-fold in transgenic hearts, without change in SERCA2a mRNA or protein content.The maximal rate of Ca 2+ uptake in cardiac homogenates remained unchanged, but there was an increase in the Ca 2+ affinity of the pump consistent with the high Ca 2+ affinity and low velocity of the Ca 2+ transport catalyzed by the SERCA2b Ca 2+ -ATPase.The SERCA2b transgenic hearts also had higher contraction and relaxation rates than control hearts indicating functional insertion of the non-cardiac isoform into cardiac SR.While the SERCA2a isoform is widely distributed in the SR membrane, the overexpressed SERCA2b was preferentially localized near the transverse tubules and the Z line.The distinct localization of the two isoforms implies that they do not compete for the same insertion sites.This may explain that SERCA2b overexpression did not affect the SERCA2a content of the heart.The distinct localization may be related to structural differences between SERCA2b and SERCA2a.SERCA2b has a long carboxyl terminal tail of 50 amino -acid residues (Lytton et al., 1989), that may form an extra transmembrane domain placing the C-terminus of SERCA2b on the lumenal side of the SR membrane.By contrast, SERCA1 and SERCA2a have only 10 putative transmembrane domains with C-termini on the cytoplasmic side.
Overexpression of Na + :Ca 2+ exchanger in cardiac muscle.In ventricular myocytes of transgenic mice overexpressing the canine cardiac NCX (Adachi-Akahane et al., 1997;Terracciano et al., 1998), the rates of contraction and relaxation were faster, and the Ca 2+ content of SR was 69% greater than in ventriculocytes of control litter mates.Even when the Na + :Ca 2+ exchanger was inhibited by removing the Na + and Ca 2+ from the medium, the rate of Ca 2+ removal from the cytoplasm remained faster in the transgenic than in control myocytes, indicating faster Ca 2+ uptake by the transgenic SR.There was no evidence for changes in the SERCA2, phospholamban, and calsequestrin contents determined by Western blot, suggesting increased activity rather than increased content of the SR Ca 2+ pump.Reversal of the NCX at low cytoplasmic [Ca 2+ ] may contribute to the increased Ca 2+ content of SR in transgenic animals.Inhibition of SR Ca 2+ transport by thapsigargin slows the declining phase of the Ca 2+ transient.Close to 30% inhibition of SR Ca 2+ transport by thapsigargin was required to slow the decline of Ca 2+ transient in transgenic animals to control levels (Terracciano et al., 1999).The increased Ca 2+ -transport activity of SR may serve to compensate for the increased Ca 2+ extrusion by the overexpressed Na + :Ca 2+ exchanger, but the mechanism of this compensation remains to be established.A somewhat different relationship between NCX and SERCA expression was observed during development in mouse heart.The ex-pression of NCX is 2-fold higher in fetal than in adult heart and decreases during development, while the SERCA2 and phospholamban expression increase during the same period nearly 8-fold (Reed et al., 2000).This antithetical relationship between NCX and SERCA expression levels is consistent with the greater contribution of Ca 2+ influx to the excitation/contraction coupling of fetal heart, while in adult heart Ca 2+ release from the SR plays a dominant role (Martonosi, 2000).Hypothyroidism tends to reduce the expression of SERCA2a mRNA and protein, while the expression of NCX1 increases.Opposite changes are observed in hyperthyroidism (Reed et al., 2000).More information is clearly needed about the mechanisms that regulate the expression of these two proteins before a functional connection between their expression level could be rationalized.
Overexpression of plasma membrane Ca 2+ -ATPase.The PMCA4CI isoform (corresponding to human PMCA4b isoform) was overexpressed in rats under the control of ventricle specific myosin light chain-2 promoter (Hammes et al., 1998), resulting in 2.5and 1.6-fold increase in PMCA protein level in neonatal and adult hearts, respectively.The Ca 2+ -dependent PMCA ATPase activity in adult hearts increased in proportion to the increase in PMCA content.The PMCA transgenic animals showed no differences in cardiac performance from control animals, and the expression of NCX, SERCA2a Ca 2+ -ATPase and the endogenous rat PMCA1 mRNA was also unchanged.
These observations suggest that PMCA has no major role in the beat-to-beat regulation of the contraction-relaxation cycle in adult transgenic rat hearts.However, overexpression of PMCA increased about 2-fold the rate of protein synthesis in cardiomyocyte cultures stimulated with fetal calf serum or isoproterenol, suggesting a role in growth regulation during development (Hammes et al., 1998).Several-fold overexpression of PMCA4CI in L6 myogenic cell lines (Hammes et al., 1996) sig-nificantly decreased the resting cytosolic free Ca 2+ levels and accelerated myogenic differentiation as measured by the formation of multinucleated myotubes, and by a 3-to 4-fold increase in the activity of creatine phosphokinase.The relationship of the decrease in cytoplasmic free [Ca 2+ ] to the differentiation promoting effect is unclear since activity-dependent increase in cytoplasmic Ca 2+ promotes differentiation of muscle (Martonosi, 2000) and nerve cells (Carafoli et al., 1999).PMCA4CI was also stably expressed and correctly targeted to the plasma membrane in Chinese hamster ovary (CHO) cells (Guerini et al., 1995).Based on the amount of phosphorylated intermediate, the level of expressed PMCA4CI in transgenic cells was 10-15 times greater than the PMCA level in control cells, while the concentration of endogenous SERCA pump was reduced.There was no change in Na + , K + -ATPase activity.The PMCA transfected cells effectively released Ca 2+ into the medium, and inside out plasma membrane vesicles isolated from transfected cells accumulated 8-10 times more Ca 2+ than those of control cells.Cells expressing high levels of PMCA recovered slowly from trypsinization and grew slowly after plating but after this initial growth delay behaved like non-transfected CHO cells.
Stable overexpression of PMCA1a in rat aorta endothelial cells (Liu et al., 1996) yielded 3-4-fold increase in the amount of functional PMCA1a, that was matched by similar increase in the Ca 2+ -calmodulin stimulated Ca 2+ pump activity tested in isolated vesicle preparations.The expression and activity of the SERCA pump was down-regulated.There was also an unexplained 2.6-fold increase in Ca 2+ influx into the PMCA1a expressing cells.Transient overexpression of PMCA4CI in CHO cells (Brini et al., 2000) also reduced the Ca 2+ content of SR by about 20% without significant effect on cytosolic or mitochondrial free Ca 2+ levels.The decrease in SERCA expression and/or activity associated with PMCA overexpression is difficult to view as part of a regulatory mech-anism, since it contributes to a decrease in SR Ca 2+ content.The possibility must be considered that the transfection technique itself may interfere with the expression of some endogenous proteins.
In conclusion, the data on ablation and overexpression of key components of the Ca 2+ signalling network indicate a surprising degree of robustness of Ca 2+ regulation.Only complete ablation of the SERCA2 gene in mice caused loss of viability but even in this case, heterozygotes remained essentially healthy.Nearly complete loss of the human SERCA1 protein in some forms of Brody's disease caused only mild delay of muscle relaxation, while ablation of SERCA3 had no effect either on viability or fertility.Massive overexpression of SERCA1, 2a, and 2b in heart improved cardiac performance, and there was no obvious disease phenotype associated with overexpression of Na + :Ca 2+ exchanger or the plasma membrane Ca 2+ transport ATPase.Induced change in the expression level of one component was frequently accompanied by changes in others, but at this stage, it is difficult to assign these changes to cooperative effects that may be expected to improve Ca 2+ regulation.In fact, some of the associated changes may be artifacts of the genetic manipulation technique.The story of gene regulation by Ca 2+ is only beginning to unfold.

EXCITATION-CONTRACTION (E-C) COUPLING. THE Ca 2+ RELEASE CHANNELS OF THE PLASMA MEMBRANES AND THE SARCOPLASMIC RETICULUM
The mechanism of E-C coupling was the subject of intense interest during the last decade that led to a vast expansion of literature.
Only key elements of this progress will be briefly reviewed here.For more detailed information the reader is referred to the review articles listed in the text.
The depolarization of the surface membrane of muscle cell initiated by the nerve impulse (Sperelakis & Gonzales-Serratos, 2001;Wahler, 2001) is conducted into the interior of the muscle fiber by extensions of the surface membrane called the transverse (T) tubules (Heiny, 2001).The T tubules establish specialized contacts (T-SR junctions, triads, diads) with the junctional membrane of SR, that contain the two key components of the excitation-contraction coupling apparatus, the dihydropyridine receptor (DHPR) and the ryanodine receptor (RyR) (Franzini-Armstrong, 1999).
The DHPR of the T tubules and surface membrane (Hofmann et al., 1999;Bezanilla, 2000;Catterall, 2000) serves both as voltage sensor and as a calcium channel by responding to changes in membrane potential with intramembrane charge movement and inward calcium current (Hofmann et al., 1999;Bezanilla, 2000;Catterall, 2000).The RyR Ca 2+ channels of SR briefly open during the action potential permitting the rapid release of stored Ca 2+ from the SR into the cytoplasm (Meissner, 1994;2001;Franzini-Armstrong & Protasi, 1997;Niggli, 1999;Mar & Marks, 2000;Ogawa et al., 2002).The activation of RyR is under the tight control of DHPR in both skeletal and cardiac muscles, resulting in efficient coupling of membrane depolarization to calcium release from the SR (Fig. 3).However, the mechanism appears different in skeletal and cardiac muscles due to differences in the structure and disposition of the two receptors.
Skeletal muscles are able to contract in an essentially Ca 2+ -free medium for several hours (Armstrong et al., 1972); the Ca 2+ current through DHPR is slow and apparently unrelated to E-C coupling.In skeletal muscles DHPR serves mainly as a voltage sensor that controls the opening of the SR Ca 2+ channel either by direct interaction with RyR or through contacts with other proteins (Meissner, 2001).
The contraction of cardiac muscles is dependent on extracellular Ca 2+ (Ringer, 1883).
The voltage-dependent Ca 2+ current through the activated cardiac DHPR Ca 2+ channel is fast and sufficiently large to play a fundamental role in the activation of RyR in cardiac SR by a process called Ca 2+ -induced Ca 2+ release (Fabiato, 1989;Cannell et al., 1995;Bers, 2002); nevertheless a direct control of RyR by interaction with DHPR may also contribute.

The structure and function of dihydropyridine receptor in skeletal and cardiac muscles
The DHPRs of skeletal and cardiac muscles are structurally distinct; they are coded by different genes, and their expression is differentially regulated (Hofmann et al., 1999;Catterall, 2000).In dysgenic mice the functional DHPR content of skeletal muscle is severely reduced, while in the heart and in the sensory neurons the DHPR content remains normal (Chaudhari, 1995).The skeletal muscle DHPR is a pentamer containing a1, a2-d, b and g subunits (Hofmann et al., 1999;Catterall, 2000).The a1 subunit is the central component of the complex and it occurs in two forms with relative molecular masses of 175 kDa and 212 kDa (Tanabe et al., 1987;Catterall, 2000).The 175 kDa form accounts for more than 90% of the a1 subunit content.The a1 175 form may arise by posttranslational proteolysis from the a1 212 form.The a1 subunit of the cardiac DHPR has a molecular mass of 242 771 (Mikami et al., 1989); it is larger than the a1 212 subunit of the skeletal muscle DHPR (Tanabe et al., 1987).Analysis of the amino-acid sequence of skeletal and cardiac a1 subunits reveals four internal structural repeats (I-IV), each containing five hydrophobic segments (S1, S2, S3, S5, and S6), and one positively charged segment (S4).The S1 through S6 segments are assumed to represent helical membrane spanning domains arranged around a central Ca 2+ channel, while the hydrophilic N-and C-terminal regions of the molecule are positioned on the cytoplasmic surface (Hofmann et al., 1999;Catterall, 2000).The S4 segment is assumed to serve as voltage sensor (Bezanilla, 2000).
The structural basis of the differences between E-C coupling in skeletal and cardiac muscles was analyzed by expressing the cDNAs of skeletal and cardiac DHPR in dysgenic myotubes in tissue culture, and studying the characteristics of spontaneous or electrically evoked contractions.The slow Ca 2+ current together with the voltage-dependent charge movement and E-C coupling can be fully restored to dysgenic muscle by microinjection of the expression plasmid PCAC6 for the a1 subunit of the skeletal muscle DHPR (Adams et al., 1990;Tanabe et al., 1990a).The restored E-C coupling persists in a Ca 2+ -free medium or in the presence of 0.5 mM Cd 2+ like the E-C coupling of normal skeletal muscle.
Dysgenic muscles injected with the expression plasmid of cardiac DHPR (pCARD1) expressed a Ca 2+ current that was similar to the large, fast activating L type Ca 2+ current of cardiac muscle (Tanabe et al., 1990b).Substitution of Ba 2+ for Ca 2+ caused a large increase in peak current, while the skeletal muscle Ca 2+ channel was only slightly affected by Ba 2+ (Adams et al., 1990;Tanabe et al., 1990aTanabe et al., , 1990b)).The pCARD1-injected dysgenic myotubes produced electrically induced contraction in normal Ringer solution, but not in Ca 2+ -free Ringer solution or in the presence of 0.5 mM Cd 2+ , consistent with the Ca 2+ requirement for cardiac E-C coupling (Tanabe et al., 1990b).
Chimeric DHPR cDNAs constructed by replacing various segments of the cardiac DHPR with corresponding segments from skeletal DHPR identified the loop between domains II and III as an important determinant for skeletal or cardiac type E-C coupling (Tanabe et al., 1990a); the highly conserved III-IV loop may also contribute.Peptides corresponding to the II-III loop region activate the skeletal, but not the cardiac RyR Ca 2+ release channels suggesting that this region is involved in the interaction between DHPR and RyR in the junctional complex (Leong & MacLennan, 1998;Stange et al., 2001).The effects of this interaction are bi-directional resulting in the activation of RyR1 Ca 2+ release by depolarization, in the inactivation of RyR1 during repolarization, and in the enhancement of the DHPR Ca 2+ channel activity.Phosphorylation of Ser687 in the II-III loop by cAMP dependent protein kinase abolished these effects.Repeat I of the 1 subunit determines the kinetics of Ca 2+ channel activation, that is slow in skeletal and fast in cardiac muscle (Tanabe et al., 1991).The SS1-SS2 segment between helices V and VI is assumed to line the channel pore (Catterall, 2000).

The structure of ryanodine receptor
There are three genetically distinct isoforms of RyR in mammals (Franzini- Armstrong & Protasi, 1997;Mar & Marks, 2000).RyR1 is the dominant isoform in skeletal muscle, RyR2 in cardiac muscle, while RyR3 is ubiquitously distributed in various tissues including muscle (Ogawa et al., 2000).In avian, fish and amphibian muscles, two major isoforms a and b were identified (Franzini- Armstrong & Protasi, 1997).The RyR1 isoform is present both in slow and fast-twitch skeletal muscles.The rabbit and human RyR1 contain 5037 and 5032 amino-acid residues, respectively, with Depolarization of the surface membrane of skeletal or cardiac muscle cells initiated by the nerve impulse is conducted into the interior of the muscle fiber by extensions of the plasma membrane, i.e. the T tubules.Therefore, T tubules establish specialized contacts with the SR membrane that contain the two key components of the E-C coupling apparatus, the dihydropyridine receptor (DHPR) and the ryanodine receptor (RyR).Other explanations are within the text.
molecular masses of close to 560 kDa (Zorzato et al., 1990).Much of the mass of the molecule is exposed in the cytoplasm, anchored by four to twelve transmembrane sequences into the junctional SR membrane (Zorzato et al., 1990).The transmembrane helices are assumed to form the Ca 2+ channels.The gene of human RyR1 was localized to chromosome 19q13.1.Mutation in this gene may be responsible for some forms of malignant hyperthermia (Mickelson & Louis, 1996;Mac-Lennan, 2000).The RyR2 of rabbit cardiac muscle contains 4969 amino-acid residues with a molecular weight of 564 711 (Nakai et al., 1990;Otsu et al., 1990).The gene for the cardiac ryanodine receptor (RyR2) is on chromosome 1 (Otsu et al., 1990).The RyR1 and RyR2 isoforms have 66% amino acid homology and similar topology, but differ in cellular distribution, channel activation and interaction with DHPRs that contribute to the differences in E-C coupling between skeletal and cardiac muscles.Three-dimensional reconstitution of negatively stained or frozen-hydrated electron microscope images of the skeletal muscle RyR revealed a tetrameric structure with dimensions of 27 ´27 ´14 nm, that is consistent with the dimensions of "junctional feet" seen in triad junctions of skeletal muscle (Sharma et al., 1998;Liu et al., 2001;Samso & Wagenknecht, 2002).In the large cytoplasmic domain there is a central channel which branches into four radial channels that extend to the lateral surface of the feet.These channels probably serve as the pathways for Ca 2+ release.A square shaped basal platform of 14 ´14 ´4 nm contains the membrane spanning domain.RyR2 and RyR3 have similar structures (Sharma et al., 1998;Liu et al., 2001).
The side of the "feet" facing the T-tubules is presumably attached directly or with the assistance of other junctional proteins to the dihydropyridine receptor tetramers in the T tubules.Residues 922-1112 of RyR1 bind specifically to the II-III and III-IV loops of the skeletal muscle dihydropyridine receptor or to the II-III loop of the Na + channel but there is no interaction with the same region of the cardiac DHPR (Leong & MacLennan, 1998).The corresponding peptide fragment from the cardiac RyR2 did not interact with skeletal DHPR.These observations indicate that the DHPR II-III and III-IV loops interact with a broad region of the RyR1 surface.The mechanical strength of the SR T tubule interaction is critical to prevent disruption during muscle contraction and stretch.
The RyR1 interacts with FKBP12 (Wagenknecht et al., 1996) and calmodulin (Samso & Wagenknecht, 2002) on its cytoplasmic surface, at about 10 nm distance from the transmembrane channel; therefore, the effects of these agents on Ca 2+ channel activity are mediated by long-range changes in the conformation of RyR.Conformational changes visible by electron microscopy indeed accompany transitions between the closed and the open states of the Ca 2+ channel (Orlova et al., 1996).
The Ca 2+ channel is activated by cytoplasmic Ca 2+ and ATP and inhibited by calmodulin (Meissner, 1994;2001); phosphorylation on Ser2809 by calmodulin-dependent protein kinase activated the cardiac RyR (Meissner, 1994).The activation by cytoplasmic Ca 2+ is the biochemical basis of Ca 2+ -induced Ca 2+ release.The rate of Ca 2+ release increases with increase in the Ca 2+ concentration of SR lumen (Hidalgo & Donoso, 1995).This may be due to Ca 2+ binding to lumenal Ca 2+ -binding sites on RyR or to an effect of lumenal Ca 2+ on junctional proteins that may interact with RyR (Hidalgo & Donoso, 1995).

Disposition of dihydropyridine and ryanodine receptors in skeletal and cardiac muscles
The differences in the mechanism of E-C coupling between skeletal and cardiac muscles are in part related to the distinct structural organization of DHPR and RyR in the junctional complexes of the two muscle types.
In both muscles the RyR tetramers form dense structures (feet) visible by electron microscopy, that span the junctional gap of 10 nm between the SR and the T tubules or plasma membranes (Franzini- Armstrong & Protasi, 1997).
In the toadfish swim-bladder muscle and in mammalian skeletal muscles the RyR "feet" are arranged in tetragonal arrays within the junctional SR, facing arrays of tetrads containing four DHPR molecules located in the junctional T tubules and in the plasma membranes (Franzini- Armstrong & Protasi, 1997).This spatial arrangement permits interaction of the four DHPR molecules within a tetrad with the RyR tetramers in alternate feet.The association of alternate RyR feet with the DHPR tetrads provides the structural basis for the direct mechanical activation of half of the RyR Ca 2+ channels by DHPR (Rios & Pizarro, 1991).RyR feet that are not in contact with the DHPR tetrads may be activated by Ca 2+ released from adjacent feet.Since mammalian skeletal muscles contain only one major RyR isoform (RyR1), it is not clear why DHPR tetrads interact only with alternate RyR tetramers.In several non-mammalian muscles that contain a and b isoforms of the RyR in similar amounts, only the a RyR tetramers appear to interact with DHPR tetrads, and the b isoform is presumably activated by Ca 2+ .
In cardiac muscles the DHPR molecules are generally not grouped into tetrads and their position does not correspond to the position of the RyR feet (Sun et al., 1995).Extra-junctional RyRs associated with calsequestrin aggregates are seen in the non-junctional "corbular" SR of cardiac muscles located at some distance from the T tubules or surface membranes (Jorgensen et al., 1993;Franzini-Armstrong et al., 1998).The absence of regular contacts between DHPR and RyR in cardiac muscles supports the primary role of voltage-dependent Ca 2+ influx and Ca 2+ -induced Ca 2+ release in cardiac E-C coupling (Fabiato, 1989).
Purified skeletal muscle DHPR and RyR did not interact in vitro (Brandt et al., 1990), and after coexpression of the two proteins in CHO cells there was no junction formation (Takekura et al., 1995).The RyR expressed alone was correctly inserted into the ER membrane of CHO cells and formed functional tetrameric Ca 2+ channels arranged in arrays similar to those seen in skeletal muscle (Takekura et al., 1995).By contrast, most of the skeletal muscle DHPR expressed in CHO cells was retained in the perinuclear membranes and did not assemble into tetrads, suggesting a requirement for other junctional components for correct targeting and self-assembly of DHPR.Some of these components may be specific to muscle cells since DHPR is correctly targeted in dyspedic mouse embryo muscles lacking RyR1 (Takekura & Franzini-Armstrong, 1999).Even in this case, coexpression of RyR1 was also required for the formation of DHPR tetrads and for their alignment into orthogonal arrays (Protasi et al., 2000).The other skeletal muscle isoform, RyR3, was ineffective in inducing DHPR tetrad formation (Protasi et al., 2000).
The precise role of these proteins in the structure and function of junctional complexes remains to be defined.

Figure 2 .
Figure 2. Freeze-etch replicas of SR vesicles isolated from rabbit fast-twitch skeletal muscles.Fragmentation of SR membranes, a necessary step in isolation of membranous fractions from skeletal muscles, leads to formation of membrane vesicles of various orientation.Most of the vesicles is right-side-out oriented, i.e. with cytoplasmic leaflet of the membrane facing outside (the fraction enriched with right-side oriented vesicles is shown on right panel; this fraction is characterized by the presence of 7.5-nm particles on the concave or outer fracture faces of the vesicle replicas).Hovewer, 10-15% of total number of vesicles tends to form inside-out oriented structures (left panel).Different orientation of obtained vesicles explains why 7.5-nm intramembranous particles, representing the Ca 2+ -ATPase molecules, are visible on the concave faces of right-side-out oriented vesicles (right panel, black arrows), and in opposite, on the convex or inner fracture faces of inside-out oriented vesicles (left panel, red arrows) (photographs obtained by curtessy of Dr. M. Gabriela Sarzala-Drabikowska).

Figure 3 .
Figure 3.The disposition of Ca 2+ release channels of plasma membrane and SR in skeletal (A) and cardiac (B) muscles.

Table 1 . Classification of Ca 2+ -ATPase present in mammalian tissues
Ca 2+ .The Na + -dependent inactivation is counteracted by Ca 2+ binding to internal high affinity Ca 2+ regulatory sites at elevated Ca 2+ concentration (15 mM).