The functions of two species of calcium channel in cardiac ......cardiac muscle excitation-contraction coupling A. J. Williams Cardiac Medicine, National Heart and Lung Institute,

European Heart Journal (1997) 18 {Supplement A), A27-A35

The functions of two species of calcium channel incardiac muscle excitation-contraction coupling

A. J. Williams

Cardiac Medicine, National Heart and Lung Institute, Imperial College, Dovehouse Street, London SW3 6LY,U.K.

The contractile state of cardiac muscle cell is determined bythe level of calcium in the cytosol. Each action potentialproduces a transient elevation of cytosolic calcium. Theinteraction of calcium ions with the contractile elements ofthe cell initiates a series of events which culminate in cellshortening. The process by which cell excitation is coupledto contraction involves the function of two distinct speciesof calcium-selective membrane ion channel; the L-type,dihydropyridine-sensitive calcium channel of the cell sur-face membrane or sarcolemma and the ryanodine-sensitivecalcium-release channel of the sarcoplasmic reticulum intra-

cellular membrane network. The aim of this article is toprovide information on the properties of these channels andto discuss the mechanisms involved in the coordinationof their function in the coupling of cardiac muscle cellexcitation to contraction.(Eur Heart J 1997; 18 (Suppl A): A27-A35)

Key Words: L-type calcium channel, calcium-releasechannel, ryanodine receptor, excitation-contractioncoupling.

Introduction

The coupling of the excitation of a cardiac muscle cell tocontraction cannot occur in the absence of extracellularcalcium. This was first demonstrated to be the case forfrog heart by Ringer in 1883[l] and has subsequentlybeen shown to be the case in mammalian cardiac prepa-rations'21. On excitation, calcium ions enter the cell viadihydropyridine-sensitive calcium channels. In amphib-ian cardiac muscle cells this calcium influx is sufficient toraise cytosolic calcium to the level required to initiatecontraction; however, in mammalian cardiac musclecells calcium influx is insufficient to cause contraction. Inthese systems the bulk of the calcium that initiatescontraction is released from an intracellular storageorganelle named the sarcoplasmic reticulum or SR, via aprocess of calcium-induced calcium release (CICR)'3>4].In this article I will provide information on the func-tional properties of the cardiac dihydropyridine-sensitive calcium channel and SR calcium-releasechannel and will discuss how the function of these twochannels might be integrated in the mammalian cardiacmuscle cell during excitation-contraction coupling.

Correspondence: A. J. Williams, Cardiac Medicine, National Heartand Lung Institute, Imperial College, Dovehouse Street, LondonSW3 6LY, U.K.

The pathway for calcium entry — theL-type, dihydropyridine-sensitive

calcium channel

Although the sarcolemmal membranes of cardiac musclecells contain more than one form of calcium-selectiveion channel, the bulk of calcium influx into ventricularcells occurs via L-type, dihydropyridine-sensitivechannels'4"51. L-type calcium channels are induced toopen by depolarization of the sarcolemma and calciumions flow through open channels down an electro-chemical gradient into the cell. Calcium currents (Fig. 1)are activated rapidly, with peak currents achieved within2-10 ms depending on temperature and potential. Dur-ing sustained depolarizations, calcium current declinesdue to time-dependent, voltage-dependent and calcium-dependent inactivation processes'51.

Calcium currents through L-type channels canbe modified by several classes of clinically useful drugs.The most widely studied of these are the dihydropyrid-ines, exemplified by nifedipine and nitrendipine, thephenylalkylamines such as verapamil and D 600, and thebenzothiazepines such as diltiazem. Members of thesethree classes interact with separate, high affinity,allosterically-linked binding sites on the channel protein.Globally, these compounds are often referred to ascalcium channel blockers or calcium antagonists.Studies of the interaction of these drugs with singleL-type calcium channels have revealed that they reduce

0195-668X/97/0A0027+09 S18.00/0 © 1997 The European Society of Cardiology

A28 A. J. Williams

c'flrOus C

3 1 ^ - 4 0

J3

'2•so

O

63.

J_200Time (ins)

400

Figure 1 Traces obtained from isolated guinea-pig ven-tricular myocytes superfused with a solution containing(in m,M): NaCl, 140; KC1, 6; MgCI2, 1; CaCl2, 2; glucose,10; HEPES, 10; pH 7-4 at a rate of 2-3 ml. min ~ '. Cellswere voltage clamped following impalement with micro-electrodes (20-30 MQ resistance) filled with 2 M K C I ;0 1 mM EGTA; 5 mM HEPES, pH 7-2. Cell shorteningwas recorded using a video based motion detection system.Changes in cytosolic calcium were monitored usingIndo-1. Cells were voltage-clamped at — 40 mV. Depo-larization pulses to 0 mV (duration 200 ms) were appliedevery 2 s. The middle trace shows inward calcium currentthrough L-type calcium channels and the lower traceshows cell shortening. Both traces were obtained byaveraging five consecutive depolarization pulses.

calcium current by modulating the gating of the channelrather than by blocking current flow. They reduce theprobability of channel opening rather than reducecalcium current flow through the open channel'51.

Current flow through L-type calcium channelscan also be modified through a variety of physiologicalpathways. The most clearly defined of these is the in-crease in calcium current associated with the interactionof agonists with /7-adrenergic receptors. The interactionof the agonist can enhance calcium current via twopathways with different time courses. The slower of theseinvolves the stimulation of a membrane-bound GTPbinding protein (Gs), which activates a membrane-bound adenylate cyclase to produce cyclic AMP. CyclicAMP is released into the cytosol and stimulates the

dissociation of the catalytic and regulatory subunits ofsoluble cyclic AMP-dependent protein kinase (proteinkinase A). The catalytic subunit then catalyses the phos-phorylation of the L-type calcium channel'51. The fastermechanism, which is thought to occur entirely within thesarcolemmal membrane, involves the direct interactionof Gs, released following the binding of the agonist tothe /f-adrenergic receptor, with the calcium channel pro-tein. The increased calcium current observed in the pres-ence of /?-adrenergic receptor agonists stems from anincrease in the probability of channel opening151.

The pathway for calcium release — theSR calcium-release channel or

ryanodine receptorAs outlined above, calcium enters the mammalian car-diac muscle cell during each action potential. Whilst thisinflux of calcium is essential for the initiation of contrac-tion, it is not, in itself, sufficient to raise the cytosoliccalcium concentration to the level required to bringabout contraction. The calcium signal is amplified by therelease of additional calcium from a storage site, the SR.The SR is an entirely intracellular membrane networkwhich surrounds the cell's contractile apparatus. Itforms specialized structural features at regions of the cellwhere it comes into close apposition with the cell surfacemembrane or sarcolemma. The SR membrane networkis a highly efficient calcium-handling organelle. Themembrane system contains abundant ATP-driven cal-cium pumps. These play a major role in the maintenanceof the resting cytosolic calcium concentration at a levelbelow that needed to initiate contraction and at the sametime establish a reservoir of stored calcium. The functionof the cardiac muscle SR calcium pump can be regu-lated by various systems including the phosphorylationof a closely associated SR protein, phospholamban,which results in an increase in pump efficiency'4671. Thecapacity of the SR membrane network as a calcium storeis enhanced by the presence of luminal calcium-bindingproteins such as calsequestrin'4'891. In mammalianmyocardium the SR calcium accumulating apparatusprovides the principal mechanism for the removal ofcalcium from the cytosol during relaxation.

In addition to the protein components necessaryfor calcium accumulation and storage, the SR membranenetwork contains the pathway for the regulated releaseof calcium from the store during excitation-contractioncoupling. The pathway for calcium release has beenidentified as a high-conductance cation-selective chan-nel. This is sometimes referred to as the ryanodinereceptor-channel because the protein contains a highaffinity binding site for this plant alkaloid'10"121. Thedistribution of this channel is limited to the regions ofthe SR network which come into close contact with thesarcolemma.

The function of the SR calcium-release channelcan be studied using a number of different experimentalapproaches. These include the measurement of calcium

Eur Heart J, Vol. 18. Suppl A 1997

Calcium channels in cardiac muscle excitation-contraction A29

transients in intact cardiac myocytes14'131, monitoringcalcium efflux from isolated SR membrane vesicles1'41

and at the single-channel level, under voltage-clampconditions, following the incorporation of either thenative or purified channel protein into a planar phos-pholipid bilayer'15"181. In all cases the SR calcium-releasepathway displays properties consistent with a role in ascheme for cardiac excitation-contraction coupling inwhich the calcium signal provided during the actionpotential is amplified by the release of stored calcium viaa mechanism of CICR.

To illustrate the characteristics of the system Iwill cite work involving the monitoring of single-channelactivity of reconstituted cardiac SR calcium-releasechannels (Fig. 2). The primary activating ligand for theSR calcium-release channel is calcium acting at itscytosolic face. In the absence of other activating ligands,micromolar concentrations of calcium will inducechannel opening and the probability of opening riseswith increasing concentrations of activating calcium;increased open probability is achieved via an increase inthe frequency of channel opening. However, calciumalone cannot fully activate the channel; under theseconditions the probability of channel opening (Po) doesnot exceed 0-5[17-19].

The calcium-induced opening of the cardiac SRcalcium-release can be modulated by a range of physio-logically and pharmacologically important secondaryligands acting at distinct sites on the channel protein.

Magnesium appears to compete with calcium atthe cytosolic binding site and reduces the probability ofchannel opening'171. Millimolar concentrations of ATPand non-hydrolyzable analogues of ATP act synergisti-cally with calcium to produce very high open probabili-ties (Po=l)[20]. The degree of channel opening attainedin the presence of cytosolic calcium and a secondaryligand such as ATP can be modulated by the concen-tration of calcium at the luminal side of the channel12'1.Additionally, phosphorylation, mediated by a numberof protein kinases[22~24] and direct effects of calmodulinmay regulate the function of the cardiac muscle SRcalcium-release channel1'4'251.

Of the pharmacologically important ligands,caffeine and structurally related compounds interact at aspecific site on the cytosolic side of the channel protein,and in the presence of micromolar concentrations ofcytosolic calcium, can fully activate the channel(Po= 1)[26'27'. Interestingly, therapeutic concentrations ofcardiac glycosides, such as digoxin, have recently beenshown to interact with and increase the open probabilityof the cardiac, but not the skeletal, isoform of the SRcalcium-release channel'281. This action may contributeto the positive inotropic effect of this drug in cardiacmuscle (see the recent review by Levi et al.l29] for adiscussion of the complexity of cellular actions of thedigitalis glycosides in the heart).

The SR calcium-release channel has also beenidentified as a target for potential mediators of cellinjury under pathophysiological conditions such asreactive oxygen species and proteases130'3'1.

ATP

caffeine

calmodulin

^—-U ryanodine

©'©/

\©Ca

SR lumen

Figure 2 Cartoon showing the cytosolic and luminalligand binding sites on the SR calcium-release channel.+indicates that the agent increases channel Po (prob-ability of channel opening), — indicates that the agentdecreases channel Po. Calcium, acting from the cytosolicside of the channel, is the primary activating ligand as it iseffective in the absence of all other ligands. The effective-ness of ligands acting at the ATP and caffeine sitesis influenced by both cytosolic and luminal calcium.The influence of ryanodine on channel gatingis dependent upon concentration; low concentrations(nM-u.\i) lock the channel into a reduced conductance stateof very high Po|4S|, high concentrations (high (iM-mM)close the channel. Channel gating may also be modified byprotein kinase mediated phosphorylation (see text). Theshape of the calcium-release channel reflects structuralfeatures revealed in recent studies involving cryo-electronmicroscopy and three dimensional image reconstruction ofthe skeletal muscle SR calcium-release channel1501.

Excitation-contraction coupling

In the preceding sections I have provided a very briefdescription of the properties of the sarcolemmal L-typecalcium channel and the sarcoplasmic reticulumcalcium-release channel. In this section of the article Iwill discuss how the function of these two populations ofchannel may be coordinated to link the excitation of acardiac muscle cell (the action potential) to contraction.It is clear that the properties of the L-type calciumchannel and the SR calcium-release channel equip thecardiac muscle cell with the apparatus required toamplify a calcium trigger to the extent necessary toinduce contraction via CICR from the SR. However, thedetails of the mechanisms involved in the function ofthese channels and their relationship in excitation-contraction coupling remain areas of intensive research.

In cardiac muscle the release of calcium from theSR appears to be tightly coupled to both the size andthe duration of the calcium current flowing through thesarcolemmal L-type calcium channel'32"341 (Fig. 3). Putanother way, the probability of the SR calcium-release

Eur Heart J, Vol. 18, Suppl A 1997

A30 A. J. Williams

40

-40

-30 -20-10

10 20 30 40

C o3 O

oo

5!

"a ©

C= 'S SQ ® -i

3 * -

Figure 3 Traces obtained from isolated guinea-pig ven-tricular myocytes under the experimental conditionsdescribed in Fig. 1 showing the relationship between clampvoltage and L-type calcium channel current, indo-l fluor-escence and cell shortening. The traces in the top panel ofthe figure show the voltage clamp protocol. The cell wasvoltage-clamped at — 40 niV and depolarization pulses tothe indicated voltages (duration 200 ms) were appliedevery 2 s. The current traces shows the typical bell-shapedrelationship between applied voltage and calcium current.The indo-l fluorescence and cell shortening traces havevery similar bell-shaped forms.

channel being open appears to be determined by thecalcium entering the cell via the L-type calcium channel.At first sight this observation appears to be at odds withcalcium release via CICR. Intuitively one might expectthat CICR would display positive feedback; calciumreleased from an SR calcium-release channel would itselfactivate the channel from which it was released andother channels in the SR membrane network. We mightpredict that this process would result in uncontrolledregenerative calcium release rather than the observedgraded regulation of release by calcium current throughthe sarcolemmal L-type calcium channel.

These issues have been addressed in considerabledetail by Michael Stern'35361. Here I will discuss some ofthe general concepts that have emerged from theseelegant studies. Based on mathematical modelling ofCICR Stern has argued that, as might be predicted froma simplistic interpretation, it is not possible to constructa stable model of CICR which produces the requiredamplification and is controlled in a graded fashion bycalcium influx via the L-type calcium channel if calciumis released from the SR into a common pool, i.e. a poolshared with other SR calcium-release channels.

A theoretical solution to this problem is providedby so-called 'local control' models. In the simplest ofthese models discussed by Stern each L-type calciumchannel is assumed to exist in a pair with one SRcalcium-release channel (Fig. 4). It is further assumedthat the two-channel proteins in the pair are located sothat the calcium entering the cell via the L-type calciumchannel has preferential access to its complimentary SRcalcium-release channel and that each pair is separatedfrom every other pair by enough distance to preventinteraction between the pairs, other than via the bulkcytosolic calcium. In this arrangement the SR calcium-release channel would be induced to open, not by thecalcium concentration in the bulk cytosol but by thelocal calcium concentration resulting from the openingof the L-type calcium channel. In such a system theconcentrations of calcium needed to activate the SRcalcium-release channel could be relatively high, com-pared to that present in the bulk cytosol, and this wouldreduce the likelihood of regenerative release.

As it stands, this system is still prone to positivefeedback; although release of calcium from one SRcalcium-release channel would not stimulate the openingof other release channels, it seems likely that it wouldcontinue to activate the channel from which it wasreleased. The solution that Stern has proposed to over-come this problem reflects an inherent property of theSR calcium-release channel. As is the case with allchannel protein molecules, the gating of the SR calcium-release channel is stochastic; over a period of time in thepresence of calcium (the activation ligand), records ofsingle SR calcium-release channels consist of a series ofrandom opening and closing events of differing dura-tion. In other words, the channel will close spon-taneously even in the presence of activating cytosoliccalcium. If the level of activating cytosolic calcium ismaintained, the SR calcium-release channel will beinduced to open again; however, if the level of activatingcytosolic calcium is reduced significantly, the channelwill remain closed. We can then envisage a situationin which calcium influx through the L-type calciumchannel ceases. Under these conditions, when the SRcalcium-release channel closes, calcium in the localenvironment of the channel will diffuse away in a frac-tion of a millisecond. The signal for channel openinghas been removed and the SR calcium-release channelremains closed. The SR calcium-release channel wouldthen only be stimulated to re-open by the re-introduction of calcium via the L-type calcium channel.Stern has used a mathematical model based on thisscheme. It simulates stable calcium release that varieswith the duration and degree of depolarization of thesarcolemma in a similar fashion to the variation seen inL-type calcium current with these parameters.

The local control scheme outlined above pro-vides a working description of the control of SRcalcium-release by L-type calcium channel current.However, if each L-type calcium channel controls thegating of a single SR calcium-release channel, the am-plification of the calcium signal is dependent upon the

Eur Heart J, Vol. 18, Suppl A 1997

Calcium channels in cardiac muscle excitation-contraction A31

(a)

Open •

Closed 20 pA

250 ms

Ca

Ca Ca CaCa-release channel

OPEN

Ca Ca Ca

Ca-release channelCLOSED

Ca Ca CaCa-release channel

OPEN

Ca

Ca Ca CaCa-release channel

OPEN

Ca Ca CaCa-release channel

CLOSED

Ca Ca CaCa-release channel

CLOSED

Figure 4 Local control model for CICR with an L-type calciumchannel to SR calcium-release channel stoichiometry of 1:1. (a)Single-channel current fluctuations of sheep cardiac SR calcium-release channel monitored under voltage clamp conditions'191 in thepresence of 10 JIM cytosolic calcium as activating ligand. Note thecharacteristic very brief opening events and somewhat longer closedevents, (b) A cartoon depicting an equivalent scheme assuming localcontrol of CICR during sustained influx of calcium through the L-typecalcium channel in a cardiac muscle cell. In all cases it is assumed thatcalcium will diffuse away from the vicinity of the SR calcium-releasechannel to the bulk cytosol in a fraction of a millisecond. In the leftpanel the release channel is open and calcium flows into the cytosol.The middle panel represents the situation following the spontaneousclosing of the release channel (see trace a). The local concentration ofcalcium at the cytosolic face of the release channel is maintained bycalcium entry through the L-type calcium channel and so the releasechannel is re-opened (right panel), (c) Closing of the release channeldue to stochastic attrition. In the presence of trigger calcium enteringthrough the L-type calcium channel, the SR calcium-release channelwill show normal random gating behaviour as seen in (a) and (b) (leftpanel). When the flow of trigger calcium is stopped (because theL-type calcium channel closes — middle panel) normal spontaneousclosing of the release channel will now result in the cessation ofcalcium release from the SR until trigger calcium is re-introduced. Itis important to remember that at all stages of this process calcium willdiffuse away from the vicinity of the release channel in less than onemillisecond. In the absence of trigger calcium influx through theL-type calcium channel, a random closing of the release channellasting for tens of milliseconds (see (a)) will allow sufficient time forcalcium at the cytosolic face of the release channel to fall to a levelbelow that required to re-open the channel.

Eur Heart J, Vol. 18, Suppl A 1997

A32 A. J. Williams

relative conductance of the two systems. Single-channelmeasurements of the conductance of the two species ofchannel under ionic conditions approximating to thosefound in the cell certainly indicate that the SR calcium-release channel has a considerably higher conductancethan the L-type calcium channel (~2pA compared with~0-3pA)'5'371; however, this is not sufficient to producethe signal amplifications monitored in intact cells.Stern's solution to this problem involves modifying thestoichiometry between SR calcium-release channels andL-type calcium channels in his local control model;amplification is increased if each L-type calcium channelcontrols a cluster of SR calcium-release channels.

The essential features of this 'cluster bomb'model are as follows. Calcium-release channels arearranged in clusters in the cardiac SR membrane (e.g.each cluster could contain five channels) and eachcluster is located in close apposition to an L-typecalcium channel. Calcium-release channel opening isinitiated by calcium entry via the L-type calcium chan-nel and this in turn leads to the opening of otherchannels within the cluster due to regenerative recruit-ment. Intra-cluster release will not spread to otherclusters if the clusters are far enough apart and ifrelatively high levels of activating calcium are requiredto open a release channel. Importantly, the modelpredicts that release of calcium from the SR membranenetwork can still be graded. Very brief or low ampli-tude depolarizations of the sarcolemma will not initiaterelease in all clusters. Increases in the duration or sizeof the depolarization will activate more L-type calciumchannels and hence more clusters. The final feature ofthe cluster bomb model concerns the termination ofcalcium release. Stern suggests three mechanisms whichmay contribute to the cessation of calcium release froma cluster; calcium-dependent inactivation of the releasechannel, depletion of releasable calcium and stochasticattrition. This latter mechanism is based on the intrin-sic gating properties of the release channel describedabove. In the cluster bomb model, calcium releasedfrom a channel within the cluster will stimulate theopening of other channels within the cluster. This isregenerative and release could continue in the absenceof a sustained trigger from the L-type calcium channel.However, in clusters containing a small number ofrelease channels, once the L-type calcium channel trig-ger has ceased, there is a definite probability thatsufficient channels will close simultaneously to lowerthe local calcium to a level below that required tomaintain channel opening. Under these conditionsopen release channels within the cluster will quicklyclose and the cluster will remain in this state untilactivated by the influx of external calcium.

As acknowledged by Stern'361 these models,although complex, do not provide realistic quantitativedescriptions of experimental observations; however,they have provided a very useful framework in which toset and discuss experimental data. In the final section ofthis article I will provide a brief review of structural andfunctional information, some of which has emerged

since the formulation of local control models, which hascontributed to our current image of cardiac excitation-contraction coupling.

The structural relationship between theL-type calcium channel and the SR

calcium-release channel

The cluster bomb model for CICR requires that cardiacSR calcium-release channels are arranged in groups.Each group must be functionally separate from itsneighbouring groups and must be closely associated witha sarcolemmal L-type calcium channel to form a func-tional release unit. Is there evidence to support thispicture?

Recent studies employing double labellingimmunofluorescence and laser confocal microscopyindicate that the L-type calcium channel and the SRcalcium-release channel co-distribute in ventricular cells,occurring as transversely orientated punctate bands at2 (im intervals along the entire muscle fibre'381. Imagesobtained using standard transmission electronmicroscopy have established that the sarcolemmal andSR membrane systems of mammalian ventricular musclecells form structures termed dyads where the SRmembrane network comes into apparent contact withsarcolemmal transverse-tubules. Similar structures areobserved infrequently at points of contact between theSR membrane network and the surface sarcolemma. Theinter-membrane space in both of these structures isoccupied by so called 'feet' structures which have beenidentified as the cytoplasmic domain of the SR calcium-release channel'391.

Freeze fracture studies of skeletal muscle cellshave produced estimates of a stoichiometry of two SRcalcium-release channels to one L-type calcium channel(or more precisely to one tetradic array of particleswhich has been interpreted as representing an L-typecalcium channel140411). No equivalent studies are avail-able for cardiac muscle; however, indirect estimates ofthe stoichiometry of SR calcium-release channel andL-type calcium channel are available from ligand bind-ing studies which indicate that the ratio of ryanodinebinding sites to dihydropyridine binding sites is consid-erably greater in cardiac muscle than skeletal muscle'42'.These observations would be consistent with an arrange-ment in which one L-type calcium channel was associ-ated with a group or cluster of SR calcium-releasechannels in the mammalian cardiac myocyte.

Calcium sparks

Recent developments have made possible the imaging ofcalcium indicators within cardiac muscle cells. This hasprovided important information on the spatial featuresand time course of variations in intracellular calcium(see the recent review by Niggli and Lipp'43').

Eur Heart J, Vol. 18, Suppl A 1997

Calcium channels in cardiac muscle excitation-contraction A33

The term 'calcium spark' was introduced byLederer and colleagues'441 to describe spontaneous localincreases in the concentration of intracellular calcium.These were observed in quiescent rat ventricular cardiacmyocytes using laser scanning confocal microscopy to-gether with the fluorescent calcium indicator fluo-3.Under these conditions a calcium spark did not activateadditional sparks at other locations within the cell;however, an increase in the calcium load of the SR bothincreased the probability of seeing a spark and resultedin some sparks triggering propagated calcium waveswithin the cell. These authors initially suggested thatsparks result from spontaneous openings of single SRcalcium-release channels'441. Subsequent investigationshave led to the conclusion that sparks represent releasefrom 'release units' composed of several SR calcium-release channels. Limiting depolarization-induced cal-cium entry through the L-type calcium channel, eitherby termination of calcium current after a very briefperiod by further depolarization or by exposing cells toL-type calcium channel antagonists has established that,under these conditions, sparks are triggered by localcalcium entry, that sparks occur at a limited number ofdifferent locations within the cell, that the sites of sparkgeneration vary with each depolarization and that theevoked sparks are very similar to the spontaneoussparks observed in quiescent cells'45^*81. In addition, ithas been demonstrated that the calcium release during aspark is maintained for a longer period than the dura-tion of the calcium trigger and that the amplitude ofevoked calcium sparks are independent of membranepotential. It is proposed that the large synchronousincrease in cytosolic calcium induced by the normalactivation of cardiac sarcolemmal L-type calcium chan-nels, which leads to contraction, results from the sum-mation of a large number of calcium sparks arising fromindependent release units'451.

These concepts have been incorporated into ascheme for cardiac excitation-contraction coupling'461

which shares many features with the cluster bomb modelproposed by Stern'35'361. As in the cluster bomb model,calcium entry through a single L-type calcium channel isthought to activate a group of SR calcium releasechannels. The activation of this release unit results in acalcium spark. The more L-type channels recruited bythe depolarization the greater the probability of sparkoccurrence. The amount of calcium released in a spark isnot dependent upon membrane potential. As in thecluster bomb model, the cessation of calcium releasefrom the functional unit will depend upon the intrinsicgating of the SR calcium-release channel, includingstochastic attrition.

According to Cannell et at, the overall amplifi-cation of the calcium trigger is achieved in two ways'461.Firstly, the quantity of calcium released per unit timefrom the release unit is considerably larger than thetrigger calcium; this is because the single channel con-ductance of the SR calcium-release channel is consider-ably greater than the L-type calcium channel (see above)and because the release unit contains several release

SL

L-type Cachannel

Closed

Open

Closed

Ca-releasechannel cluster

I Ca-releaseJ unit

=SPARK

Figure 5 CICR involving local control of calcium releaseunits by a single L-type calcium channel. A calciumrelease unit is assumed to be made up of a single L-typecalcium channel and a number of SR calcium-releasechannels. Calcium influx through the L-type channelactivates release and the calcium signal is amplified asdescribed in the text resulting in a 'spark'. When thesupply of trigger calcium ceases, release from the SR stopsdue to the gating characteristics of the release channels inthe cluster. Local control ensures that release from oneunit does not activate neighbouring release units. Factorswhich increase the probability of L-type calcium channelopening will increase the probability of spark occurrence.The depolarization of the sarcolemma during an actionpotential will produce an increase in bulk cytosoliccalcium and hence initiate contraction due to a highprobability of L-type calcium channel opening and thesummation of numerous sparks.

channels and only one L-type channel. This phenom-enon is termed 'analogue gain'. The second componentof signal amplification results from the proposed mech-anism for the cessation of calcium release. Followingdepolarization, the L-type channel will open, triggercalcium will enter the cell and then calcium entry willcease as the L-type channel inactivates. However, asoutlined above, calcium release will continue to occur inthe absence of trigger calcium due to positive feedbackbetween the SR calcium-release channels within the

Eur Heart J, Vol. 18, SuppI A 1997

A34 A. J. Williams

release unit, which will result in regenerative release. Theintrinsic gating mechanisms of the channels within therelease unit will quickly switch off release; however,during the period between the cessation of the calciumtrigger and the switching off of the release unit, calciumrelease will continue and will contribute to the overallamplification of the signal. Cannell et al. refer to thisphenomenon as 'digital pulse stretching'1461.

Conclusion

The roles of the sarcolemmal L-type calcium channeland the SR calcium-release channel in mammaliancardiac muscle excitation-contraction coupling havebeen recognised for many years. There is a broadagreement that this process involves calcium signalamplification by calcium-induced calcium release fromthe SR. However, our understanding of this fundamen-tal process is far from complete. In this article I havediscussed some recent experimental data together withtheoretical considerations which have contributed to ourcurrent understanding of how the function of these twospecies of channel may be integrated to produce con-trolled muscle function. These are summarised in Fig. 5.

I thank Drs C. Terracciano and K. MacLeod for providing thedata presented in Figs 1 and 3 of this article. I am also grateful toDrs R. Sitsapesan and K. MacLeod for very helpful discussions onthe mechanisms of cardiac excitation-contraction coupling and tothe British Heart Foundation, BBSRC and Wellcome Trust forfinancial support.

References

[1] Ringer S. A further contribution regarding the influence of thedifferent constituents of the blood on the contraction of theheart. J Physiol 1883; 4: 29^2.

[2] Rich TL, Langer GA, KJassen MG. Two components ofcoupling calcium in single ventricular cell of rabbits and rats.Am J Physiol 1988; 254: H937^6.

[3] Fabiato A. Appraisal of the physiological relevance of twohypotheses for the mechanism of calcium release from themammalian cardiac sarcoplasmic reticulum: calcium-inducedrelease versus charge-coupled release. Mol Cell Biochem 1989;89: 139^0.

[4] Bers DM. Excitation-contraction coupling and cardiac con-tractile force. Dordrecht: KJuwer, 1991.

[5] McDonald TF, Pelzer S, Trautwein W, Pelzer DJ. Regulationand modulation of calcium channels in cardiac, skeletal, andsmooth muscle cells. Physiol Rev 1994; 74: 365-507.

[6] Sham JSK, Jones LR, Morad M. Phospholamban mediatesthe y?-adrenergic-enhanced Ca2+ uptake in mammalian ven-tricular myocytes. Am J Physiol (Heart Circ Physiol) 1991;261: H1344-9.

[7] Davis BA, Edes I, Gupta RC et al. The role of phospho-lamban in the regulation of calcium transport by cardiacsarcoplasmic reticulum. Mol Cell Biochem 1990; 99: 83-8.

[8] Yano K, Zarain-Herzberg A. Sarcoplasmic reticulum cal-sequestrins: Structural and functional properties. Mol CellBiochem 1994; 135: 61-70.

[9] Guo W, Campbell KP. Association of triadin with the ryano-dine receptor and calsequestrin in the lumen of the sarcoplas-mic reticulum. J Biol Chem 1995; 270: 9027-30.

[10] Pessah IN, Waterhouse AL, Casida JE. The calcium-ryanodine receptor complex of skeletal and cardiac muscle.Biochem Biophys Res Comm 1985; 128: 449-6.

[11] Lai FA, Erickson HP, Rousseau E, Liu Q-Y, Meissner G.Purification and reconstitution of the Ca release channel fromskeletal muscle. Nature 1988; 331: 315-19.

[12] Lai FA, Erickson HP, Rousseau E, Liu Q-Y, Meissner G.Evidence for Ca2+ channel within the ryanodine receptorcomplex from cardiac sarcoplasmic reticulum. BiochemBiophys Res Comm 1988; 151: 441-9.

[13] Cheng H, Cannell MB, Lederer WJ. Propagation ofexcitation-contraction coupling into ventricular myocytes.Pflugers Arch Eur J Physiol 1994; 428: 415-17.

[14] Meissner G, Henderson JS. Rapid Ca release from cardiacsarcoplasmic reticulum vesicles is dependent on Ca2+ and ismodulated by Mg2+, adenine nucleotide and calmodulin.J Biol Chem 1987; 262: 3065-73.

[15] Rousseau E, Smith JS, Henderson JS, Meissner G. Singlechannel and 45Ca2+ flux measurements of the cardiacsarcoplasmic reticulum calcium channel. Biophys J 1986; 50:1009-14.

[16] Anderson K, Lai FA, Liu Q-Y, Rousseau E, Erickson HP,Meissner G. Structural and functional characterization of thepurified cardiac ryanodine receptor-Ca2+ release channel com-plex. J Biol Chem 1989; 264: 1329-35.

[17] Ashley RH, Williams AJ. Divalent cation activation andinhibition of single calcium release channels from sheepcardiac sarcoplasmic reticulum. J Gen Physiol 1990; 95:981-1005.

[18] Lindsay ARG, Williams AJ. Functional characterisation ofthe ryanodine receptor purified from sheep cardiac musclesarcoplasmic reticulum. Biochimica et Biophysica Acta 1991;1064: 89-102.

[19] Sitsapesan R, Williams AJ. Gating of the native and purifiedcardiac SR Ca2+-release channel with monovalent cations aspermeant species. Biophys J 1994; 67: 1484-94.

[20] Holmberg SRM, Williams AJ. Single channel recordings fromhuman cardiac sarcoplasmic reticulum. Circ Res 1989; 65:1445-9.

[21] Sitsapesan R, Williams AJ. Regulation of the gating of thesheep cardiac sarcoplasmic reticulum Ca2+-release channel byluminal Ca2+. J Memb Biol 1994; 137: 215-26.

[22] Witcher DR, Kovacs RJ, Schulman H, Cefali DC, Jones LR.Unique phosphorylation site on the cardiac ryanodine recep-tor regulates calcium channel activity. J Biol Chem 1991; 266:11144-52.

[23] Hain J, Onoue H, Mayrleitner M, Fleischer S, Schindler H.Phosphorylation modulates the function of the calcium releasechannel of sarcoplasmic reticulum from cardiac muscle. J BiolChem 1995; 270: 2074-81.

[24] Lokuta AJ, Rogers TB, Lederer WJ, Valdivia HH. Modula-tion of cardiac ryanodine receptors of swine and rabbit bya phosphorylation-dephosphorylation mechanism. J Physiol(Lond) 1995; 487: 609-22.

[25] Smith JS, Rousseau E, Meissner G. Calmodulin modulationof single sarcoplasmic reticulum Ca2+-release channels fromcardiac and skeletal muscle. Circ Res 1989; 64: 352-9.

[26] Sitsapesan R, Williams AJ. Mechanisms of caffeine activationof single calcium-release channels of sheep cardiac sarcoplas-mic reticulum. J Physiol 1990; 423: 425-39.

[27] McGarry SJ, Williams AJ. Activation of the sheep cardiacsarcoplasmic reticulum Ca2+-release channel by analogues ofsulmazole. Br J Pharmacol 1994; 111: 1212-20.

[28] McGarry SJ, Williams AJ. Digoxin activates sarcoplasmicreticulum Ca2+-release channels: a possible role in cardiacinotrophy. Br J Pharmacol 1993; 108: 1043-50.

[29] Levi AJ, Boyett MR, Lee CO. The cellular actions of digitalisglycosides on the heart. Prog Biophys Molec Biol 1994; 62:1-54.

[30] Holmberg SRM, Cumming D, Kusama Y et al. Reactiveoxygen species modify the structure and function of thecardiac sarcoplasmic reticulum calcium-release channel.Cardioscience 1991; 2: 19-25.

Eur Heart J, Vol. 18, Suppl A 1997

Calcium channels in cardiac muscle excitation-contraction A35

[31] Holmberg SRM, Williams AJ. The calcium-release channelfrom cardiac sarcoplasmic reticulum: Function in the failingand acutely ischaemic heart. In: Hasenfuss G, Holubarsch Ch,Just H, Alpert NR, eds. Cellular and Molecular Alterations inthe Failing Human Heart (Supplement to Basic Research inCardiology; Vol. 87,1). New York: Springer-Verlag, 1992:255-68.

[32] Beuckelmann DJ, Wier WG. Mechanism of release of calciumfrom sarcoplasmic reticulum of guinea-pig cardiac cells.J Physiol 1988; 405: 233-55.

[33] Cleemann L, Morad M. Role of Ca2+ channel in cardiacexcitation-contraction coupling in the rate: Evidence fromCa2+ transients and contraction. J Physiol 1991; 432: 283-312.

[34] Cannell MB, Berlin JR, Lederer WJ. Effect of membranepotential changes on the calcium transient in single ratecardiac muscle cells. Science 1987; 238: 1419-23.

[35] Stern MD, Lakatta EG. Excitation-contraction coupling inthe heart: The state of the question. FASEB J 1992; 6:3092-3100.

[36] Stern MD. Theory of excitation-contraction coupling in car-diac muscle. Biophys J 1992; 63: 497-517.

[37] Tinker A, Lindsay ARG, Williams AJ. Cation conduction inthe cardiac sarcoplasmic reticulum calcium-release channelunder physiological and pathophysiological conditions.Cardiovasc Res 1993; 27: 1820-5.

[38] Carl SL, Felix K, Caswell AH el al. Immunolocalizationof sarcolemmal dihydropyridine receptor and sarcoplasmicreticular triadin and ryanodine receptor in rabbit ventricle andatrium. J Cell Biol 1995; 129: 673-82.

[39] Inui M, Saito A, Fleischer S. Isolation of the ryanodinereceptor from cardiac sarcoplasmic reticulum and identitywith the effect structures. J Biol Chem 1987; 262: 15637-42.

[40] Block BA, Imagawa T, Campbell KP, Franzini-ArmstrongC. Structural evidence for direct interaction between the

molecular components of the transverse tubule/sarcoplasmicreticulum junction in skeletal muscle. J Cell Biol 1988; 107:2587-2600.

[41] Franzini-Armstrong C, Jorgensen AO. Structure and develop-ment of E-C coupling units in skeletal muscle. Annu RevPhysiol 1994; 56: 509-34.

[42] Bers DM, Stiffel VM. Ratio of ryanodine to dihydropyridinereceptors in cardiac and skeletal muscle and implicationsfor E-C coupling. Am J Physiol (Cell Physiol) 1993; 264:Cl 587-93.

[43] Niggli E, Lipp P. Subcellular features of calcium signalling inheart muscle: what do we learn? Cardiovasc Res 1995; 29:441-8.

[44] Cheng H, Lederer WJ, Cannell MB. Calcium sparks: Elemen-tary events underlying excitation-contraction coupling inheart muscle. Science 1993; 262: 740-4.

[45] Cannell MB, Cheng H, Lederer WJ. Spatial non-uniformitiesin [Ca2+]: during excitation-contraction coupling in cardiacmyocytes. Biophys J 1994; 67: 1942-56.

[46] Cannell MB, Cheng H, Lederer WJ. The control of calciumrelease in heart muscle. Science 1995; 268: 1045-9.

[47] Lopez-Lopez JR, Shacklock PS, Balke CW, Wier WG. Localcalcium transients triggered by single L-type calcium channelcurrents in cardiac cells. Science 1995; 268: 1042-5.

[48] Santana LF, Cheng H, Gomez MB, Cannell MB, Lederer WJ.Relation between the sarcolemmal Ca2+ current and Ca2+

sparks and local control theories for cardiac excitation-contraction coupling. Circ Res 1996; 78: 166-71.

[49] Lindsay ARG, Tinker A, Williams AJ. How does ryanodinemodify ion-handling in the sheep cardiac sarcoplasmic reticu-lum Ca2+-release channel? J Gen Physiol 1994; 104: 425-47.

[50] Wagenknecht T, Radermacher M. Three-dimensional archi-tecture of the skeletal muscle ryanodine receptor. FEBS Letts1995; 369: 43-6.

Eur Heart J, Vol. 18, Suppl A 1997