HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (12) Citing Articles via Google Scholar Google Scholar Articles by Cheng, H. Articles by Lederer, W. J. Search for Related Content PubMed PubMed Citation Articles by Cheng, H. Articles by Lederer, W. J.

Calcium Sparks

Institute of Molecular Medicine, National Laboratory of Biomembrane and Membrane Biotechnology, Peking University, Beijing, China; and Medical Biotechnology Center, University of Maryland Biotechnology Institute, Baltimore, Maryland

ABSTRACT

I. INTRODUCTION

II. MICROSCOPIC Ca2+ GRADIENTS IN LIVING CELLS

III. Ca2+ SPARKS

A. Spontaneous Sparks

B. Evoked Sparks

C. Sparks and EC Coupling

D. Sparks and Ca2+ Waves

E. Sparks in Different Types of Cardiac Cells

F. Sparks in Other Types of Cells

G. Terminology

IV. THEORY OF SPARK FORMATION AND DETECTION

A. Spark Formation

B. Spark Detection

C. Ca2+ Spikes

D. Inferring [Ca2+]i Gradients From Sparks

V. SPARK MORPHOLOGY

A. Amplitude

B. Width

C. Kinetics

D. Autonomy

VI. Ca2+ SPARKLETS

VII. INTERMOLECULAR SIGNALING BETWEEN LCCS AND RYRS

A. Triggering Sparks by Single LCC Openings

B. Power Law of Spark Activation

C. Kinetics of Sparklet-Spark Coupling

VIII. Ca2+ BLINKS

A. In Cardiac Myocytes

B. In Skeletal Muscle Cells

IX. SPARK MECHANISMS: ACTIVATION

A. Spontaneous Sparks

B. Triggered Sparks

X. SPARK MECHANISMS: COORDINATION IN A CRU

A. Overview

B. Key Observations

C. CICR and Coupled Gating

D. How Many RyRs Open During a Spark?

XI. SPARK MECHANISMS: TERMINATION

A. Local Ca2+ Depletion

B. Stochastic Attrition

C. Desensitization of CICR by Ca2+ Store Depletion

D. Coupled Gating as a Termination Mechanism

E. Channel Inactivation and Adaptation

1. Ca2+-dependent inactivation

2. Planar lipid bilayer experiments

3. RyR channel adaptation

F. Refractoriness of Local CICR

G. RyR Gating In Vivo Versus In Vitro

XII. Ca2+ SPARKS IN HEART DISEASE

A. Arrhythmic Disease

B. Heart Failure

C. Remodeling of the Ca2+ Signaling System

1. CRU organization

2. SERCA2a and NCX

3. Dyssynchrony of Ca2+ sparks

4. SR Ca2+ leak

5. AP duration and arrhythmias

D. Genetic and Acquired Channel Dysfunction

E. Cardiac Ca2+ Signaling in Diverse Diseases

XIII. SPARKS AND EMBERS IN SKELETAL MUSCLES

A. Overview

B. Sparks in Amphibians

1. Activation mechanisms

2. Coordination mechanisms

3. Termination mechanisms

C. Embers and Sparks in Mammals

1. Embers and ghost sparks

2. Inhibitory role of conformational coupling between DHPRs and RyRs

3. Role of the RyR3 isoform

XIV. SMOOTH MUSCLE SPARKS

A. Activation of Ca2+-Sensitive Channels by Subsurface Sparks

B. Ca2+ Sparks Relax Smooth Muscle

C. Spark Modulation of Membrane Excitability

XV. NEURAL Ca2+ SPARKS

A. Characteristics and Mechanisms

B. Properties of CICR in DRG Neurons

C. Possible Role of Neuronal Sparks

XVI. IP3R Ca2+ PUFFS AND BLIPS

XVII. SPARKLESS RELEASE

XVIII. NEW INSIGHTS INTO Ca2+ SIGNALING

A. Mechanisms Underlying Excitation-Ca2+ Release Coupling

B. Ca2+ Sparks and the CICR Paradox

C. Cellular Architecture of Ca2+ Signaling

D. The Digital-Analog Dichotomy

E. Biochemistry of Ca2+ Sparks

F. Building Complexity From Simplicity

XIX. PERSPECTIVE

GRANTS

ACKNOWLEDGMENTS

REFERENCES

	ABSTRACT

Top Next References

Ja, Kalzium das ist alles... Otto Loewi (1936 Nobel Laureate)

	I. INTRODUCTION

Top Previous Next References

 
Calcium signaling is paradoxically both simple and complex in that single-atom calcium ions (Ca²⁺) act as the most versatile biological messenger known. Rapid, transient changes in Ca²⁺ concentration directly control muscle contraction, cell locomotion, hormonal secretion, and neural transmission. Sustained elevations of Ca²⁺ signals play pivotal roles in biological processes ranging from fertilization to gene expression to apoptosis (25–27, 79, 309). Yet, being a single-atom cation, Ca²⁺ exists in only one biologically relevant form that undergoes neither catabolic degradation nor anabolic synthesis. Its biological information-coding ability derives almost entirely from its binding and unbinding from target proteins as well as the electrical currents it generates when moving across a biological membrane. This simplicity of Ca²⁺ signaling seems to be at odds with the complexity and exquisiteness of the processes that it controls. One ion connecting thousands of proteins, each harboring one or more evolutionarily conserved Ca²⁺-binding motifs (S. Wei, personal communication), has given rise to the enigma of how Ca²⁺ can choreograph these molecular players to fulfill the amazingly diverse and even opposing physiological functions in a given cell. As Ca²⁺ is the best characterized among common biological messengers [cAMP, inositol 1,4,5-trisphosphate (IP₃), and nitric oxide (NO), for example], resolving the paradox of Ca²⁺ signaling should be enlightening with respect to fundamental principles of intracellular signal transduction.

Over the past two decades, vibrant research on local Ca²⁺ signaling in muscles, neurons, and other excitable and nonexcitable cells has laid the foundation for this review. Multiple studies by the authors and by many others have led to a body of work that addresses old dilemmas and raises new questions. In particular, the discovery of discrete, elemental Ca²⁺ signaling events exemplified by "Ca²⁺ sparks" (Table 1) (69) has illuminated the building principles of the Ca²⁺ signaling system and brought teleological insights into many supramolecular Ca²⁺ signaling nanomachines in diverse types of cells. This work has started to delineate in some detail the character of intracellular Ca²⁺ signals in the dimensions of space, time, and magnitude and has begun to unify the simplicity of elementary Ca²⁺ signaling (in terms of design principles) and the versatility of physiological functions.

	II. MICROSCOPIC Ca2+ GRADIENTS IN LIVING CELLS

Top Previous Next References

 
Sustained, macroscopic gradients of free Ca²⁺ concentration ([Ca²⁺]) occur across cell surface and intracellular membranes of organelles. A 10⁴-fold [Ca²⁺] gradient exists across the plasma membrane that partitions the extracellular space from the cytosol. A similar gradient is found between the lumen of the endoplasmic and sarcoplasmic reticulum (ER/SR) and the cytosol. The cytosolic [Ca²⁺] ([Ca²⁺]_i) is actively maintained at a very low level around 100 nM by Ca²⁺ homeostatic mechanisms, including the plasmalemmal Na⁺/Ca²⁺ exchanger (NCX) and Ca²⁺-ATPase, the ER/SR Ca²⁺-ATPase (SERCA), and by a battery of Ca²⁺ buffering molecules of differing capacity and kinetics. A low resting cytosolic [Ca²⁺], while avoiding the toxicity of sustained high [Ca²⁺], bestows on the Ca²⁺ signaling system a wide dynamic range and a high signal-to-background ratio.

Many Ca²⁺ signaling cascades are initiated by rapid mobilization of Ca²⁺ among membrane-bound compartments, usually down its electrochemical gradient across the membrane. Ca²⁺ entry from the exterior of the cell is mediated by diverse Ca²⁺-permeant channels and their gating mechanisms, which include voltage, ligands, mechanical force, temperature change, pH, reactive oxygen species (ROS), and intracellular signals arising from G protein coupling receptor (GPCR) activation, and even depletion of the ER/SR store. NCX operating in the reverse mode, i.e., translocating Ca²⁺ inward and Na⁺ outward, also brings Ca²⁺ in under certain circumstances. In contrast, Ca²⁺ release from the ER/SR is mediated by only two families of Ca²⁺ channels, the ryanodine receptor (RyR) (121, 128, 255) and the IP₃ receptor (IP₃R) (21, 125), each with three major isoforms (types 1, 2, and 3). Recently, it has been suggested that presenilins form Ca²⁺ channels of tiny conductance, accounting in part for the constant ER Ca²⁺ leak (382). Restoration of Ca²⁺ homeostasis following a transient change in [Ca²⁺] is achieved by extrusion across the plasma membrane by NCX and Ca²⁺-ATPase, and resequestration into the ER/SR by SERCA, at the expense of biochemical and electrochemical energy.

At the molecular level, transmembrane Ca²⁺ translocation should be considered a discrete, stochastic process. When in action, Ca²⁺-permeant channels and transporters act as Ca²⁺ sources (in the destination compartment) and sinks (in the source compartment), creating dynamic [Ca²⁺] gradients in their immediate vicinity. Many determinants are thought to shape transient, microscopic [Ca²⁺] gradients (179, 186, 301, 344, 362). These include 1) Ca²⁺ fluxes, their magnitudes, and time courses; 2) intracellular Ca²⁺ buffers, many of which are also Ca²⁺ effectors, their abundance, affinity, and kinetics; 3) diffusion of Ca²⁺ in the cytosolic milieu; and 4) spatial restriction (212, 350). [Ca²⁺] gradients are accentuated and sustained by spatial confinement, e.g., the subspace of a couplon in muscles (Table 1), the tortuous ER/SR network, synaptic terminals, dendritic spines, fine processes in neurons, and lamellipodia and filopodia in migrating cells. When the spatial scale of interest is comparable to the Debye length (1 nm) of the negatively charged lipid biomembrane, the electrostatic effect on Ca²⁺ distribution and diffusion also becomes evident (350). As we illustrate later, such nanodomain (1–100 nm) and microdomain (0.1–10 µm) Ca²⁺ signals are of profound significance in determining signaling efficiency, specificity, and diversity.

	III. Ca2+ SPARKS

Top Previous Next References

 
Ca²⁺ signals in living cells were first "seen" in the late 1960s as intracellular [Ca²⁺]_i transients during single twitches in barnacle muscle fibers (7, 157) and first imaged in the late 1970s as Ca²⁺ waves in fertilizing Medaka fish eggs (136, 305). In these early experiments, [Ca²⁺] was measured with aequorin, the chemiluminescent indicator obtained from jellyfish (34). Shortly after, [Ca²⁺]_i transients during cardiac excitation-contraction (EC) coupling were seen as aequorin signals in frog cardiac muscle (2) and in canine Purkinje fibers (409). Measurement of [Ca²⁺]_i transients began to seriously expand with metallochromic dyes: first arsenazo III, used by Brown et al. (49) in squid axon and then in muscle, and later antipyrylazo III used by Kovacs et al. (204). The creation by Tsien and co-workers of novel fluorescence indicators like fura 2 (147) catalyzed an explosion of new measurements and [Ca²⁺]_i imaging activity in isolated smooth muscle cells (413) and in single cardiac muscle cells under normal conditions (56) and during Ca²⁺ overload when propagating waves of elevated [Ca²⁺]_i were observed (20, 410). Elegantly spiraling Ca²⁺ waves were also documented in Xenopus oocytes (217). The convergence of two technical developments, the advent of a flexible confocal microscope (the BioRad MRC500), and the introduction of fluorescein and rhodamine-based Ca²⁺ indicators (261) prepared the field for the discovery of ever subtler dynamic Ca²⁺ signaling events. With the use of fluo 3 as the Ca²⁺ indicator (Table 2) in conjunction with confocal microscopy, Ca²⁺ sparks were first visualized in quiescent cardiac myocytes (Figs. 1 and 2).

View larger version (30K): [in this window] [in a new window]   FIG. 1. Ca2+ sparks. A: two-dimensional confocal images of Ca2+ sparks in a quiescent cardiac myocyte (scan rate 1.0 s/frame). B: line scan confocal images of an action potential (AP)-elicited [Ca2+]i transient (top) and a spontaneous spark (bottom) (scan rate 2.0 ms/line). Time and space ordinates are displayed in the horizontal and vertical directions, respectively. [Modified from Cheng et al. (69) and Cheng and Wang (71).]

View larger version (35K): [in this window] [in a new window]   FIG. 2. Depolarization-evoked Ca2+ sparks. Under conditions of whole cell patch clamp and intracellular dialysis of the indicator fluo 3, a small depolarization (from –50 mV holding potential to –40 mV) evoked Ca2+ sparks randomly in space and time. Summation of these discrete, brief, and localized events determines the time course and magnitude of the global [Ca2+]i signal (bottom). Scale bar: 10 µm. [Modified from Wang et al. (390).]

In unstimulated single cardiac myocytes, a Ca²⁺ spark appears abruptly amidst a seemingly featureless background, reaches its peak of about a twofold increase in fluorescence intensity within 10 ms, and dissipates in another 20 ms. It is confined to an area of 2.0 µm in diameter or 8 fl by volume. In a high concentration of exogenous Ca²⁺ buffer EGTA (4 mM), the local Ca²⁺ event corresponding to a spark is called a "Ca²⁺ spike" (Fig. 3) which spans 0.6 µm, lasts 8 ms, and may better resolve the location and kinetics of the Ca²⁺ flux underlying the spark (391). The characteristics of Ca²⁺ spikes are consistent with the idea that a Ca²⁺ spark arises from a "point" or narrow Ca²⁺ source, a single Ca²⁺ release unit (CRU) (Table 1).

View larger version (23K): [in this window] [in a new window]   FIG. 3. Ca2+ spikes. Cardiac myocyte under whole cell voltage clamping was dialyzed with 4 mM EGTA and 1 mM Oregon Green 488 BAPTA 5N via the patch pipette, and the confocal scan line was placed along the cell length, perpendicular to the transverse tubule (TT). A: depolarization to 0 mV evoked discrete Ca2+ spikes at TTs that were roughly 1.8 µm apart. B: spatially averaged fluorescence signal, which, to the first approximation, is proportional to the SR Ca2+ release flux (Jsr). C: contour plot of Ca2+ spikes in A. D: surface plot of a line scan image from another cell, showing both spontaneous (at the holding potential –70 mV) and evoked Ca2+ spikes (at –30 mV). (Figure courtesy of L. S. Song.)

Ca²⁺ spark rate increases in a concentration-dependent manner in the presence of the methylxanthine caffeine, which sensitizes RyRs to Ca²⁺, or decreases in the presence of the RyR inhibitor tetracaine (243). The response usually involves a transient phase because of the self-correcting nature of the process: as the Ca²⁺ spark rate increases, the efflux of Ca²⁺ from the SR increases. The SR Ca²⁺ depletion produced by the increased Ca²⁺ spark rate tends to lower the Ca²⁺ spark rate back towards its original rate. A high concentration of caffeine (>5 mM) opens all RyRs and thus completely empties the SR, giving rise to a full-fledged [Ca²⁺]_i transient and then masquerades as a release inhibitor. Ryanodine itself causes complex, concentration-dependent actions. At submicromolar concentrations, it induces repetitive Ca²⁺ sparks at the same sites, or promotes sustained, low-amplitude Ca²⁺ elevations or "embers" (142) (Table 1) that trail their initial peaks (69). At supramicromolar concentrations, ryanodine abolishes sparks altogether. Other RyR modifiers such as FK506, rapamycin, and Imperatoxin A also evoke trailing embers in a subpopulation of sparks (242, 375, 420), in qualitative agreement with the planar lipid bilayer observations.

B. Evoked Sparks

During cardiac EC coupling, Ca²⁺ sparks are evoked by Ca²⁺ influx through LCCs via the Ca²⁺-induced Ca²⁺ release (CICR) mechanism (Fig. 2). The evoked sparks appear to be identical to spontaneous ones in terms of amplitude, kinetics, and spatial properties. Depolarization per se is ineffective in evoking sparks. For instance, no depolarization-triggered sparks are seen upon removal of extracellular Ca²⁺, or replacing it with Ba²⁺ as the charge carrier, or depolarization beyond the apparent reversal potential of Ca²⁺ currents (+80 mV). These findings indicate that Ca²⁺ entry through LCCs serves as the trigger of evoked Ca²⁺ sparks in heart.

Individual Ca²⁺ sparks can be resolved only when they are activated at a low density, not fusing with each other in space or time. Accordingly, evoked sparks are seen at the leading edge of a [Ca²⁺]_i transient, with near-threshold activation of LCC, during very brief (ms) depolarization, with reduced extracellular [Ca²⁺], or after pharmacologically reduced LCC availability. In the latter case, both the residual Ca²⁺ currents (I_Ca) and the probability of Ca²⁺ spark activation (P_spark) display a bell-shaped voltage dependence. However, the voltage dependence of P_spark shifts to more negative voltages than I_Ca under these conditions, indicating that the total I_Ca does not uniquely determine the number of evoked sparks (315). The finding is consistent with an earlier observation that the gain function of EC coupling (G), defined as the ratio of whole cell release Ca²⁺ flux over the whole cell I_Ca, decreases at positive potentials (407). It is thought that microscopic properties of I_Ca such as the amplitude of unitary LCC current (i_LCC) constitute additional determinants of EC coupling efficiency (see below).

C. Sparks and EC Coupling

It is now widely accepted that Ca²⁺ sparks constitute the elementary events of cardiac EC coupling. Stated in a slightly different way, depolarization-triggered [Ca²⁺]_i transients arise from the stochastic summation of discrete sparks. This view is based on two criteria: 1) evoked sparks are "atomic" or indivisible under physiological conditions (even though they are splittable under certain experimental conditions and are not of uniform size, see below) and 2) summation of Ca²⁺ sparks can explain prominent features of [Ca²⁺]_i transients. In this regard, it has been shown that evoked sparks account for the vast majority, if not the totality, of Ca²⁺ release flux in heart cells (357). Spatial nonuniformity in the rising phase of [Ca²⁺]_i transients can be reconstructed by appropriately placing Ca²⁺ sparks in space and time (58). During full-fledged cardiac EC coupling, 10⁴ sparks are evoked within a few tens of milliseconds in a single myocyte, summing into a global [Ca²⁺]_i transient of 1 µM. This estimate agrees roughly with the estimated numbers of CRUs in a typical cell (10⁴ CRUs each consisting of 10² RyRs) (58, 130), suggesting that most, if not all, CRUs are activated during cardiac EC coupling.

The view that global [Ca²⁺]_i transients consist of atomic digital units rather than being a continuum has revolutionized our understanding of cardiac functional regulation. Simple, robust, and graded control of cardiac EC coupling can be achieved by varying the number of sparks recruited. Specifically, the characteristic bell-shaped voltage dependence of global [Ca²⁺]_i transients can now be ascribed to a similar voltage dependence of spark production, while the unitary properties of evoked sparks are voltage independent. Furthermore, many physiological modulations of cardiac contractility are now appreciated to be mediated primarily by altering the magnitude and kinetics of spark activation. Examples include contractile regulation by β-adrenergic receptor (β-AR) stimulation (444), NO production during stretch (294), mitochondrially derived ROS (425), and reactive nitrogen species (RNS) (378). Intrinsic variability in unitary sparks, the properties and nature of "sparkless" release, as well as the "paradox of CICR" (363) are discussed later.

D. Sparks and Ca²⁺ Waves

Spontaneous Ca²⁺ sparks from single CRUs usually remain local and solitary despite the CICR mechanism that operates in ventricular myocytes. Within a sarcomere, Ca²⁺ sparks tend to center on the transverse tubules (TTs) at the Z-disk of a sarcomere, where discrete CRUs are spread on a plane that is perpendicular to the long axis of the cell, with a nearest-neighbor average distance of 0.7 µm in the rat (130, 352). In the roughly 8 fl volume of the Ca²⁺ spark, there are between 2 and 5 CRUs (48, 352). It appears that none of these neighboring CRUs engulfed in the volume of a spontaneous Ca²⁺ spark is activated. Under Ca²⁺ overload conditions (e.g., high extracellular Ca²⁺), however, spontaneous Ca²⁺ sparks do activate nearby CRUs. This is reflected by nearly synchronous activation of multiple sparks from neighboring CRUs, seen as a "compound spark" or "macrospark" (Table 1). Such spark-induced spark activation occurs more readily in the transverse direction, among CRUs separated by 0.5–1.5 µm on the same Z-disk, than in the longitudinal direction, among CRUs on adjacent Z-disks at 1.8 µm intervals (288). A number of factors, including greater spark amplitude and duration, and elevated [Ca²⁺]_i and [Ca²⁺]_SR that sensitize CICR, may act in a synergistic fashion to promote spark propagation. Orderly activation of sparks in space and time then evolve into a propagating wave of elevated [Ca²⁺]_i. As best seen during initiation and low-velocity propagation, Ca²⁺ waves are often triggered by compound sparks, and then hop forward from one Z-disk to the next (67). Microscopic spiral Ca²⁺ waves (231) also develop in cardiac myocytes displaying high spark activity. These observations, together with theoretical modeling (181, 344), support the notion that Ca²⁺ sparks participate as crucial events in the initiation and propagation of Ca²⁺ waves.

In neonatal cardiac myocytes, Luo et al. (244) have shown that perinuclear sparks trigger perinuclear Ca²⁺ waves that engulf the entire nucleus without traveling into the bulk cytosol, affording a mechanism for active nuclear Ca²⁺ regulation independent of the bulk cytosolic Ca²⁺. Specifically, Ca²⁺ sparks arising from both RyRs and IP₃Rs are relatively concentrated in the perinuclear region, where immunocytochemical evidence suggests the existence of a higher release channel density. When the IP₃-linked receptor signaling pathway is activated, perinuclear sparks and waves become more frequent, and intermittently generate Ca²⁺ waves that sweep over the entire cytoplasm, indicating that perinuclear sparks and waves also act as the pacemaker of global [Ca²⁺]_i oscillations in developing cells (244).

E. Sparks in Different Types of Cardiac Cells

To date, Ca²⁺ sparks are found in the ventricular myocytes of mammals, including rat (69), mouse (68), rabbit (48, 233, 317), ferret (319), guinea pig (239, 240), dog (356), human (134, 226, 241), and bird (H. Cheng, personal communication). Similar Ca²⁺ sparks have also been recorded in isolated cardiac trabeculae (408), heart slices, and intact perfused heart (H. Cheng, personal communication). In atrial myocytes, which largely lack TTs and contain both peripheral junctional and central nonjunctional CRUs, spontaneous Ca²⁺ sparks are larger and longer lasting than their ventricular counterparts (300,000 Ca²⁺ in 12 ms versus 100,000 Ca²⁺ in 7 ms), with a high prevalence at the periphery (330, 415, 416). The action potential (AP) directly evokes subsurface Ca²⁺ sparks, which then activate inwardly propagating waves of CICR. In Purkinje fiber cells that contain RyR2, RyR3, and IP₃Rs organized in a complex, layered structure, regional differences in the properties of Ca²⁺ sparks and "Ca²⁺ wavelets" (Table 1) have been documented (368). Rhythmic local Ca²⁺ release events, manifested as Ca²⁺ sparks and compound sparks, are also found in sinoatrial nodal cells and are thought to cause diastolic depolarization via local activation of the electrogenic NCX (35, 36), thereby serving as a "Ca²⁺ clock" that is coupled to the "electrophysiological clock" made up of ionic currents to regulate the pacemaker activity (208, 386, 387) (Table 3). Perinuclear Ca²⁺ sparks and waves play an important role in the autonomous regulation of nuclear Ca²⁺ signaling in neonatal rat cardiac myocytes (244) and embryonic (E18) mouse cardiac myocytes (182). Likewise, robust spark activity is detected in cardiac myocytes derived from in vitro stem cell differentiation, but is absent after genetic ablation of RyR2 (426). In the H9c2 cardiac cell line, spontaneous sparks are thought to trigger single-mitochondrial [Ca²⁺] transients ("Ca²⁺ marks") (Table 1), though direct evidence is lacking (285). Sparklike events are also found in cardiac fibroblasts (C. Wei, personal communication). Thus spark production is a property common to all types of cardiac cells. Characteristics of subcellular Ca²⁺ signaling differ in different types of cells, perhaps matching cell-specific functions. In particular, distinctive perinuclear Ca²⁺ dynamics in embryonic and neonatal cardiac cells may encode information for the transcriptional regulation of cardiac development.

Microdomains of elevated [Ca²⁺] resulting from intracellular Ca²⁺ release are universal to virtually all cell types examined. In particular, Ca²⁺ sparks are present in skeletal (16, 168, 197, 381) and smooth muscle myocytes (52, 274, 446), neuroendocrine cells (e.g., chromaffin cells) (447), and neurons containing different isoforms of RyRs (95, 202, 284). Discrete, local Ca²⁺ release events mediated by IP₃R, termed "Ca²⁺ puffs" (428) and "blips" (289), are well characterized in Xenopus oocytes. Local Ca²⁺ release events with spark characteristics have been recorded even in nonexcitable cells, including endothelial cells (172), oligodendrocyte progenitors (156), and HeLa cells (40) and are mediated by either RyRs or IP₃Rs or both. See below for discussion of sparkless release.

G. Terminology

The terminology for the ever-growing collection of local Ca²⁺ signaling events is summarized in Table 1. Traditionally, the term "spark" is used to describe Ca²⁺ release events originating from RyRs while "puff" refers to those from IP₃Rs only and "blip" for a very small puff. "Spark" has also been used to name an event mediated by a mixture of RyRs and IP₃Rs. For uniform nomenclature, we use "spark" as a generic name to describe a local Ca²⁺ release event from a single CRU regardless of the molecular origin or cell type. Other terms stemming from it include "sparklet" referring to a single-channel event and "compound spark" referring to an event consisting of multiple CRUs. For instance, "IP₃R Ca²⁺ spark" is used interchangeably with "Ca²⁺ puff" or "Ca²⁺ blip," depending on the context (Table 1).

	IV. THEORY OF SPARK FORMATION AND DETECTION

Top Previous Next References

 
A. Spark Formation

Before delving into the details of the spark formation mechanism, we wish to distinguish among the following: 1) the underlying event of Ca²⁺ release flux; 2) the local increase in Ca²⁺ concentration (a local [Ca²⁺]_i transient), in the presence of a dye; 3) the fluorescence transient (a fluorescence spark in object space); and 4) the image of the fluorescence spark (the true fluorescence spark blurred by the imaging system). The term Ca²⁺ spark usually denotes either the underlying local [Ca²⁺]_i transient or the observed fluorescence spark, in a context-sensitive manner.

Theoretically, the formation of a spark is a "diffusion-reaction" process: the released Ca²⁺ from a spatially delimited source diffuse in the cytosol, being removed by various buffers and transport mechanisms (69, 301, 344). Governed by Fick's law of diffusion, a sheer [Ca²⁺]_i gradient develops around the point source. The vast majority of Ca²⁺ (up to 99%), however, are initially bound to a variety of Ca²⁺ binding sites on proteins and lipid membranes and, in the declining phase of the spark, are liberated from these sites. Ca²⁺ is also transported into the SR by SERCA2a, extruded from the cytosol through the NCX and the sarcolemmal Ca²⁺ ATPase, and taken into the mitochondria. Among these mechanisms, diffusion and buffering dominate in shaping the local [Ca²⁺]_i transient. In a global [Ca²⁺]_i transient, however, the decline of [Ca²⁺]_i is due largely to extrusion and the SR reuptake mechanisms.

Inside the subspace of a "couplon" (Table 1), the immediate volume into which Ca²⁺ release flux is injected, two novel determinants shape the nanoscopic [Ca²⁺] gradients: the geometric constraints (200–500 nm in diameter, 12–15 nm thick, large fractional volume occupied by the densely packed RyR feet) and the surface charge effect on the electrodiffusion of Ca²⁺. Numerical simulations by Soeller and Cannell (61, 350) show that the establishment and dissipation of [Ca²⁺]_subspace gradients occur on the time scale of 100 µs upon the "on" or "off" of Ca²⁺ flux, and the magnitude of [Ca²⁺]_subspace depends nearly linearly on the magnitude of Ca²⁺ current injected.

There is evidence suggesting that diffusion of Ca²⁺ and indicator in the cytosol should be considered anisotropic (i.e., diffusion coefficient differs in different directions). In ventricular myocytes, myofilaments and other cytoskeletal element are arranged in a complex reticular network. Mitochondria, which occupy 35% of the cell volume, are assembled into two groups: longitudinally oriented intramyofibrillar mitochondria and shorter and rounder perinuclear mitochondria. These structures may enforce a "channeled" diffusion. If this is the case, it might help to explain the observation that cardiac Ca²⁺ spark width is about double numerical model predictions and is, on average, 20% greater in the longitudinal than the transverse direction (68, but see also Ref. 11). An idea of non-Fickian diffusion in spark formation has been proposed recently (370).

B. Spark Detection

As discussed above, the [Ca²⁺] gradients due to local release flux are first transformed into the [Ca²⁺ indicator] or fluorescence gradients. Wide ranges of imaging techniques have been applied to measure this fluorescence signal , including single-photon (57, 69, 197, 240, 274, 381) and two-photon excitation confocal microscopy (97, 184, 291, 351), total internal reflection fluorescence microscopy (TIRFM) (80), and wide-field microscopy coupled with a low-noise CCD camera (446). The generic problems with optical fluorescent spark detection are twofold. 1) Optical blurring is defined by the point spread function (PSF) of the imaging system. A PSF depicts the three-dimensional (3D) intensity description of the image produced by a point object of unit intensity. For a 3D object (e.g., a fluorescence spark), its image is the convolution of the 3D object with the PSF as the weighing matrix. The sharper the PSF, the smaller the blurring effect. Confocal microscopy offers a sharp PSF, but is never blur-free. A typical confocal PSF (0.4 µm by 1.0 µm in the axial and radial directions for spark imaging) is almost comparable to the size of a spark. 2) Out-of-focus sparks: the observed population of sparks represents a random sampling of sparks at uncertain locations, and there are no established criteria to discriminate in-focus from out-of-focus sparks. Owing to random sampling, the amplitude distribution of observed sparks should follow a monotonic decreasing function, regardless of the true spark amplitude distribution (70, 334).

Another confounding factor for Ca²⁺ spark event detection and measurement is related to noise in the digital fluorescence images. For spark imaging, we are dealing with an extremely low-level fluorescence signal that is limited by the indicator concentration and the amount of illumination applicable to the cell. It is recognized that indicator inclusion perturbs the physiological process under study, and illumination stimulates photochemical reactions (e.g., ROS production) and, when in excess, causes damage to the cell and photobleaches the indicator. Furthermore, resolving a fast signal from a spatially confined source is demanding in both speed (temporal resolution) and sharpness (spatial resolution). The typical sampling time and volume of a pixel is of the order of microseconds and femtoliters, respectively, in confocal spark imaging. Under stringent experimental conditions, only a few photons can be collected in a pixel. Thus the photon-shot noise intrinsic to fluorescence emission and detection imposes a physical limit on spark resolution. In practice, there are tradeoffs among the conflicting requirements of signal-to-noise ratio, spatiotemporal resolution, photostimulation, photodamage, and the indicator perturbation of the [Ca²⁺]_i gradients under study. With regard to spark image processing, automated spark detection algorithms are necessary for objective spark detection in a noisy background (70, 295) (Table 4).

A rather unique experimental setting has been developed to resolve the source location and time course of Ca²⁺ release in a spark (324, 357). This involves an admixture of a fast, low-affinity Ca²⁺ indicator (e.g., Oregon Green 488 BAPTA 5N, 1 mM, K_d = 31 µM) and an excess of high-affinity, but slow Ca²⁺ chelator (e.g., EGTA, 5 mM, K_d = 150 nM at pH 7.2). Owing to the kinetic disparity between the indicator and the buffer, newly released Ca²⁺ ions tend to bind to the indicator before they are captured by EGTA. The initial ratio of Ca²⁺ binding to the fluorescent indicator and to EGTA is determined by k_on,indicator [indicator]/k_on,EGTA[EGTA]. Since the association rate constant k_on,indicator is typically 100-fold faster than k_on,EGTA, EGTA is far less effective in Ca²⁺ binding in the close vicinity of the release site. As Ca²⁺ ions diffuse away, however, nearly all of them are eventually captured by the EGTA in excess. As a result, the fluorescent signal is spatially more restricted than in a regular spark and, in the limit of a high concentration of Ca²⁺ buffer, closely follow the waveform of Ca²⁺ release flux (357). The sharply defined local Ca²⁺ signals are called "Ca²⁺ spikes" (357) (Table 1), which track the time course and pinpoint the origin of local Ca²⁺ flux. In heart muscle cells, Ca²⁺ spikes have been used to visualize Ca²⁺ release flux during spontaneous Ca²⁺ sparks (70) and TT-SR couplon activation during full-fledged EC coupling (Fig. 3) (32, 314, 324, 357, 360). A TT-SR Ca²⁺ spike reflects a single or a compound Ca²⁺ spark, depending on the number of CRUs recruited. Similar approaches have been widely applied in two-dimensional (2D) mapping of membrane Ca²⁺ influx in neurons (436), the organization of CRUs in dorsal root ganglion (DRG) cells (283), LCC Ca²⁺ sparklets in smooth muscle cells (273), and Ca²⁺ puffs and blips (i.e., IP₃R Ca²⁺ sparks and sparklets) in Xenopus oocytes (55).

By the same reasoning that EGTA does not suppress local indicator signal, inclusion of EGTA up to 5–10 mM should not significantly affect CICR between LCCs and RyRs (324, 357, 411) nor the retrograde inactivation of LCCs by released Ca²⁺ (323), both occurring on the nanometer scale. Nevertheless, excessive Ca²⁺ buffering would hamper Ca²⁺ communication on the micrometer scale, such as CICR between adjacent CRUs. In addition, high [EGTA] would "clamp" the bulk [Ca²⁺]_i, suppressing global [Ca²⁺]_i transients and slowing the refilling of the ER/SR (324, 357, 360).

D. Inferring [Ca²⁺]_i Gradients From Sparks

When we infer local [Ca²⁺]_i gradients from the fluorescence signal, it is important to distinguish between [Ca²⁺]_i sparks per se (i.e., the aforementioned microscopic, transient [Ca²⁺]_i gradients), true fluorescence sparks (with no optical distortion), and the fluorescence sparks as captured by an optical imaging system. In the Ca²⁺-to-fluorescence transform, indicator saturation introduces nonlinearity when [Ca²⁺]_i is well above the K_d of the indicator. With regard to the kinetics, the ability of the fluorescent signal to track rapid [Ca²⁺]_i rise is primarily determined by K_on,indicator, whereas K_{off,indicator} serves as a low-pass filter to the dissipation of the fluorescence spark. The diffusion of the Ca²⁺-bound indicators also smears the spatial [Ca²⁺]_i gradients of a developing spark. Numerical simulations confirm that the fluorescence spark, even free of optical blurring, differs drastically from the underlying [Ca²⁺]_i spark, underestimating the high [Ca²⁺]_i near the origin and overestimating [Ca²⁺]_i at distant regions, whereas its time-to-peak roughly tracks the duration of Ca²⁺ flux (assuming a sharp turnoff of the release) (69, 301, 344). The observations that the protein-binding of the indicator alters Ca²⁺ association and dissociation kinetics (159) should also be taken into account for quantitative understanding of spark formation and detection.

	V. SPARK MORPHOLOGY

Top Previous Next References

 
The Ca²⁺ spark is a direct measure of the gating of arrayed RyRs and the resultant microdomain Ca²⁺ in intact cells, providing a real-time report on the behavior of the CRU and primary information on Ca²⁺ efflux from the ER/SR. Spark morphology thus bears rich information about CRU gating, release flux, microscopic Ca²⁺ diffusion-reaction, and physiological regulation of these processes. However, tapping the information concealed in spark morphology often requires judicious examination of various aspects of spark formation and detection.

A. Amplitude

Measured fluorescence spark amplitude varies widely from the lower end set by the detection threshold (F/F₀ 0.2) to the upper end set by the brightest ones (F/F₀ 4 in ventricular myocytes). Confocal sampling theory states that the apparent spark, which represents an optical section of the true fluorescence spark at random angles, displays reduced amplitude, broadened spatial spreading, and blunted kinetics, with the degree of distortion depending on the relative positioning between the true spark origin and the focal plane or scan line. Regardless of the true spark amplitude, the apparent spark amplitude always obeys a monotonically decaying distribution (181, 334), analogous to the situation that more dim than bright stars are visible to an earth-bound stargazer. Experiments aided by an automated spark detection algorithm corroborate this prediction for both cardiac and skeletal muscle sparks (70, 141). In this regard, the average amplitude of the observable population of sparks underestimates the mean of true sparks. The practice of using the brightest sparks as the representative events (141, 179), however, is also biased if sparks are not of one size. It should be noted that the F/F₀ value also depends on [Ca²⁺] at F₀, which likely varies from cell to cell.

In an effort to restore the true population statistics, Rios et al. (308) extracted the true distribution of amplitude from the measured amplitude distribution and showed that the true distribution was dispersed, and could be modal. Some investigators have exploited sparks at fixed sites, produced either by hyperactive CRUs or by trains of electrical stimuli with partial inhibition of the LCC, to analyze the variability in spark amplitude. Considerable event-to-event variability was detected but has been largely attributed to photon shot noise (46). More recently, Wang et al. (390) and Shen et al. (331) have used the so-called "loose-seal" patch-clamp technique to activate Ca²⁺ sparks from in-focus CRUs. By pressing a patch pipette (tip diameter 1 µm) against the surface membrane to form a low-resistance (20–50 M) seal, repetitive depolarization can evoke a population of Ca²⁺ sparks beneath the patch membrane, free of out-of-focus sparks. With this rigorous characterization of spark properties, the true amplitude of cardiac sparks exhibits a broad, modal distribution. In addition, polymorphism of sparks has been demonstrated even for events from the same CRUs. Similarly, Lipp and Niggli (230) demonstrated that Ca²⁺ sparks triggered by local photorelease of caged Ca²⁺ are highly variable in amplitude. Collectively, these results suggest a revision of the initial thought that Ca²⁺ sparks are stereotypical.

Computer simulations suggest that spark peak amplitude roughly reflects the amount of Ca²⁺ released in the brief rising phase (63, 179, 301, 344). For long-lasting events, the plateau amplitude is proportional to the time-averaged release flux (346). In Ca²⁺ spike measurement, however, the spike amplitude is more directly related to the instantaneous release flux (357).

B. Width

The spatial width of a spark, usually represented by the full width at half-maximum (FWHM), is a revealing parameter of microscopic Ca²⁺ diffusion and reaction. Cardiac muscle sparks are of 2.0 µm FWHM when measured at the time of the peak signal. Ca²⁺ spikes obtained in the presence of 4 mM EGTA suggest an upper limit of 0.6 µm for the Ca²⁺ source extension (391). However, nearly all numerical models of spark formation predict a FWHM that is only half of the experimentally observed value (63, 179, 186, 344, 351), and this twofold difference in FWHM would translate into an eightfold difference in volume or "signal mass" (defined as the integral of the fluorescence increase over the volume of a spark), creating the "spark width paradox" (344). This paradox has been difficult to reconcile by simply varying model parameters (source current, geometry of an extended source, Ca²⁺ and indicator diffusion, Ca²⁺ buffering, and Ca²⁺ pumping) within their physiological limits (344). For indicator saturation to render a 2.0-µm wide spark, it would require an exceptionally large release flux (20–40 pA) and predict a spark amplitude of F/F₀ 6–7 (179). Soeller and Cannell (351) show that a spatially extended Ca²⁺ source combined with a triangle-shaped waveform of release flux affords a possible solution to the spark width paradox. Sparks in intact frog skeletal fiber are of 1.0 µm FWHM that can be fully accounted for by a spark model, whereas the spark width paradox persists in permeabilized or cut-end skeletal fiber preparations from the same species (63, 167). Differences in spark width in these muscle fiber preparations might be attributable to different Ca²⁺ buffering or release flux or both.

In a more drastic view, Tan et al. (370) questioned whether the Fickian diffusion model applies to spark formation. Fick's second law of diffusion states that the diffusive flux is proportional to the first order of the derivative of the concentration gradient.

In anomalous or non-Fickian diffusion model, diffusion with reaction can be given by

where 0 < 1, 0 < β < 2, D is anomalous diffusion constant and f is the reaction function represnting Ca²⁺ sources or sinks. When = β = 1, the equation is simplified into Fick's law.

It is known as anomalous space superdiffusion or subdiffusion for 0 < β < 1 or 1 < β < 2, which corresponds to long or short jumps of the random walker, respectively. The spark data are best depicted by = β = 1.25, suggesting a subdiffusion model in the milieu of the cytoplasm. The physical basis for this non-Fickian behavior is unclear, but it is thought to reflect the complex microscopic and nanoscopic structures and viscoelastic properties of the cytoplasm.

C. Kinetics

The average time to peak of a spark is 5 and 10 ms for skeletal muscle and cardiac sparks, respectively; the half decay time is 20 ms in either case (69, 197, 206, 391). The brevity of Ca²⁺ sparks ensures the temporal locality of Ca²⁺ signaling and implicates a powerful spark-termination mechanism (see below).

Detailed analysis of the spark release duration, indexed by the rise-time of sparks or duration of spikes, uncovered a modal distribution in both skeletal muscle fibers and cardiac myocytes (308, 391). This is not expected, if the underlying channel or channel group is governed by a Markovian gating mechanism and is free from external free energy. The expected open or bursting time distribution should be the sum of decaying exponentials, as required by thermodynamic laws of microscopic reversibility. It was then suggested that the preferred release duration thus necessitates a coupling of the release channels to a free energy source (304, 391). In this regard, the rise of [Ca²⁺]_subspace as well as the fall of [Ca²⁺]_SR during a spark could link RyR gating to the chemical potential of [Ca²⁺] across the SR membrane. Notably, an inverse relationship between the release duration and the rising rate has been observed (206, 392), consistent with either a [Ca²⁺]_subspace-mediated inactivation or a local store depletion-mediated spark termination (see below).

In the presence of reagents that stabilize RyR opening at subconductance (e.g., submicromolar ryanodine) or full conductance (e.g., Imperatoxin A for type 1 RyRs), a subpopulation of Ca²⁺ sparks may last for hundreds of milliseconds or even longer (69, 141, 142, 420). Such a long-lasting spark usually displays a brief normal peak followed by a plateau at reduced amplitude, dubbed a trailing "Ca²⁺ ember" (142). Embers without the initial peak (or lone embers) have been seen in mammalian adult skeletal muscle (440), and 10-s-long "Ca²⁺ glows" of regular amplitude (Table 1) have recently been observed in superior cervical ganglion neurons (427). The nondecaying plateau suggests that the release and replenishment of local ER/SR Ca²⁺ quickly reach a balance. The replenishment mechanisms consist of local cytosolic Ca²⁺ reuptake and, more importantly, luminal Ca²⁺ transfer within the connected ER/SR network.

D. Autonomy

Several lines of evidence indicate the autonomy of the Ca²⁺ spark: unitary Ca²⁺ spark properties are independent of the trigger. As the first clue, it has been demonstrated that AP-evoked sparks are morphologically indistinguishable from spontaneous sparks (58, 65). A close examination revealed that unitary properties of evoked sparks remain constant over the physiological voltages from –50 to +30 mV, despite gross changes of single-channel LCC currents (240). Properties of evoked Ca²⁺ sparks are also invariant regardless of their latency relative to the onset of the voltage pulse (240). Furthermore, abrupt turnoff of the LCC Ca²⁺ current by hyperpolarization after a 2 ms depolarization does not curtail Ca²⁺ release in a spark. Conversely, prolonging LCC mean open time with FPL64176 increases neither spark amplitude nor duration (390). Thus, once activated, the evolution of cardiac sparks is no longer controlled by the trigger Ca²⁺ signal; rather, it appears to be highly autonomous, determined by properties intrinsic to the CRU and its connected ER/SR cistern.

The polymorphism of Ca²⁺ sparks mentioned above should not be taken as contradictory to the autonomy of spark production. Rather, it reflects stochastic gating of CRUs. With exceptions, pharmacological and physiological modulation of sparks is mainly achieved by altering the propensity of Ca²⁺ sparks. This is particularly the case for RyR activators (low or moderate concentrations of caffeine, ROS) (51, 197, 319, 378), whereas a RyR inhibitor (e.g., tetracaine) reduces the spark amplitude (390) as if it reduces the effective size of a CRU. The spark autonomy is not absolute in skeletal muscle where CRUs are physically linked to the voltage sensor and cannot run freely (see below). Taken together, the autonomy of spark genesis strongly supports the notion that sparks are the building blocks of global [Ca²⁺]_i transients.

	VI. Ca2+ SPARKLETS

Top Previous Next References

 
Ca²⁺ passing through a single Ca²⁺-permeant channel opening create a Ca²⁺ microdomain that is fundamental to the architecture of intracellular Ca²⁺ signaling. The feasibility of optical recording of single-channel Ca²⁺ microdomains has now been convincingly established for many types of Ca²⁺-permeant channels, including L- and N-type Ca²⁺ channels, stretch-activated cation channels, acetylcholine receptor channels, RyRs, and IP₃Rs. Combining confocal microscopy with the cell-attached patch-clamp technique, Wang et al. (390) first resolved "Ca²⁺ sparklets," microscopic [Ca²⁺]_i transients at the mouth of single LCCs in heart cells (Fig. 4). This was done with the aid of the channel agonist FPL64176 to prolong the channel open duration and high extracellular Ca²⁺ to increase i_LCC amplitude. Simultaneous recording of i_LCC indicates that the amount of Ca²⁺ entry correlates linearly with the total fluorescence of a Ca²⁺ sparklet seen in the line scan image, suggesting that the sparklet provides an "optical yardstick" for calibration of the local Ca²⁺ signal.

View larger version (28K): [in this window] [in a new window]   FIG. 4. Ca2+ sparklets. Simultaneous recordings of single L-type Ca2+ channel (LCC) Ca2+ current (iCa) and local Ca2+ changes (Ca2+ sparklets) at the mouth of a channel. The ER/SR Ca2+ release was disabled by thapsigargin and caffeine in a cardiac cell, while the iCa was enhanced by inclusion of high Ca2+ and the LCC agonist FPL64176 (FPL) in the filling solution of the cell-attached patch pipette. [Modified from Wang et al. (390).]

	VII. INTERMOLECULAR SIGNALING BETWEEN LCCS AND RYRS

Top Previous Next References

A. Triggering Sparks by Single LCC Openings

Voltage-dependent spark production in rat ventricular myocytes has been quantified at voltages near the threshold of LCC activation by using a voltage ramp protocol (from –70 to –40 mV) (57) or small depolarizing steps (2 mV increments from a –60 mV holding potential) (315). The threshold voltage for spark production is about –55 mV, and the spark probability increases e-fold per 7.5 mV depolarization. Both the threshold voltage and the voltage dependence of spark activation are therefore identical to those for LCC activation. Under conditions when the LCC is largely blocked, Lopez-Lopez et al. (240) first resolved individual sparks over the entire physiological voltage range (–50 mV to +40 mV), and at all voltages tested, the time course of spark production overlaps the time course of the residual I_Ca. The nearly identical voltage dependence (over the narrow threshold voltage range) and the similarities in kinetics (at a given voltage) suggest a linear relationship between LCC activation and spark production. These results indicate that spark activation can be the result of a single LCC opening (57, 240, 315). The rationale is that, if spark activation requires the simultaneous activation of n independently gated LCCs, then P_spark should be the power function of the probability of LCC activation (P_LCC) with exponent n; that is

The aforementioned results suggest n 1 under the respective experimental conditions. It should be noted that an n value greater than unity is found when P_LCC is high but i_LCC is small (174).

An unequivocal demonstration that an opening of a single LCC can indeed trigger a spark came later with combined cell-attached patch clamping and confocal microscopy. Typical sparks are activated beneath the patch membrane during an LCC opening under gigaseal patch-clamp conditions (390). However, this is an extremely challenging experiment; and for routine recording of intermolecular signaling events, a loose-seal patch-clamp technique is employed. As shown in Figure 5, LCC sparklets (modified by FPL64176) can be recorded along with RyR sparks beneath the patch membrane. The two signals are distinguishable by amplitude and by pharmacological means. Nearly every spark arises on top of a discernible sparklet, while 30% of sparklets do not trigger any sparks (Fig. 5), indicating a sparklet-spark coupling fidelity () of 0.7 in the presence of the LCC modifier.

View larger version (30K): [in this window] [in a new window]   FIG. 5. Sparklet-spark coupling. A: LCC sparklets (the low-amplitude events) and triggered sparks (the high-amplitude events) recorded under loose-seal patch-clamp conditions. Note that not every sparklet can trigger a spark. B: histogram of the latency of the sparklet-spark coupling and its monoexponential fit ( = 6.7 ms). Data were obtained in the presence of the LCC agonist FPL64176. [Modified from Wang et al. (390).]

Santana et al. (315) investigated the relationship between trigger [Ca²⁺] and spark production by controlling i_LCC and thereby [Ca²⁺]_subspace, at varying voltages. Qualitatively, the same I_Ca does not uniquely determine the number of evoked sparks: more sparks are evoked at negative rather than positive voltages for a given amount of I_Ca, supporting the idea that the microscopic properties of I_Ca are also important determinants of EC coupling (411). Quantitatively, the voltage dependence curve for the spark-triggering efficiency of I_Ca (G = P_spark/I_Ca) closely resembles that of i_LCC as depicted by the Nernst-Planck equation, that is

where N refers to the total number of LCCs in a cell. Therefore

Given the linear relationship between [Ca²⁺]_subspace and i_LCC (350), we have

That is, spark activation is a quadratic function of [Ca²⁺]_subspace, requiring the cooperative action of the trigger Ca²⁺. Similarly, a power function with the power being 2 or higher has been reported for RyRs responding to [Ca²⁺] step jumps or brief [Ca²⁺] pulses in planar lipid bilayer experiments (152, 435).

C. Kinetics of Sparklet-Spark Coupling

Simultaneous visualization of sparklets and sparks enabled us to determine the kinetics of intermolecular coupling between LCCs and RyRs. The latency from the onset of the sparklet to the ignition of the spark varies widely; its histogram is well depicted by a single exponential function with a mean value 6.7 ms, suggesting that the LCC-to-RyR coupling is a stochastic process of first-order kinetics. To determine the latency and fidelity under physiological conditions, we need to develop some theoretical considerations. If LCC of first-order kinetics displays a mean open time _LCC, the closing rate constant is 1/_LCC. When a spark is activated by a sparklet, the sparklet is interrupted at a rate constant of 1/_latency,, where _latency, is the latency of spark activation when an LCC opens indefinitely. Let _{latency,apparent} denote the sparklet duration; the sparklet termination rate constant (1/_{latency,apparent}) is therefore the sum of the two rate constants governing the closure of LCC and the triggering of a spark, respectively

For the coupling fidelity , defined as the fraction of openings that do trigger a spark, we have

With an i_LCC of 0.3 pA at 0 mV in the presence of FPL64176 and high Ca²⁺ (390), the data fit roughly with _{latency,apparent} 6 ms, _LCC 16 ms, _latency, 9.6 ms, and 0.63.

What insights can we gain about LCC-to-RyR coupling under physiological conditions? For a _LCC of 0.5 ms and i_LCC of 0.1 pA at 0 mV and 1 mM Ca²⁺ in the absence of an LCC modifier (149), we estimate that _latency, 86 ms (by applying the power law of spark activation), _{latency,apparent} 0.5 ms, and 0.006. A rough estimate of can also be inferred from the ratio between macroscopic (G, whole cell) and microscopic (, single CRU) gains of EC coupling. The latter refers to the total Ca²⁺ flux in a spark relative to that in a sparklet. For spark Ca²⁺ flux (I_spark) of 3 pA and release flux duration of 10 ms, the value is 600 when a sparklet does trigger a spark, and = 0 when it does not. Given that G 10 at 0 mV, it can be deduced that, on average, only 1 out of 60 LCC openings can trigger a spark or = 0.016 at 0 mV, which is within threefold of those derived from the sparklet-spark coupling. From these two independent approaches, we conclude that many LCC openings trigger a spark at 0 mV under physiological conditions. There is thus a teleological need for multiple LCCs to cluster over a 0.2 µm² dyad (assuming a cisternal diameter of 500 nm) and direct demonstration of this awaits future investigation.

	VIII. Ca2+ BLINKS

Top Previous Next References

 
The ER/SR membrane encloses one continuous space that occupies only a few percent of the cell volume, but is extended throughout the cytosol as interconnected cisternae and tubules in heart cells (48, 287, 418). In addition to serving as the Ca²⁺ storage and release organelle, increasing evidence indicates that the ER/SR Ca²⁺ store actively regulates many Ca²⁺ signaling processes, such as store-operated Ca²⁺ entry (23, 43, 227, 303) and ER/SR stress-mediated cell death (271, 322). In striated muscles, the SR is organized into two distinct domains: the free (fSR) and the junctional SR (jSR); the latter bears the CRUs and contains the low-affinity, high-capacity Ca²⁺ buffer calsequestrin (CSQ) in the cisternae (126, 127).

A. In Cardiac Myocytes

Visualization of Ca²⁺ dynamics in the intricate ER/SR space has been made possible by the development of an empirical method to load low-affinity Ca²⁺ indicators (e.g., fluo 5N) preferentially into the ER/SR (287, 327). Furthermore, the ER/SR and the cytosol, while overlapping under an optical microscope, can be distinguished experimentally by targeting spectrally distinctive indicators into the two conjugate spaces (48). Building on these, the local jSR Ca²⁺ change during a spark has been investigated along with cytosolic Ca²⁺ measurement. Brief darkenings of the fluo 5N signal, called "Ca²⁺ blinks" (Fig. 6), accompany the production of spontaneous or evoked Ca²⁺ sparks in rabbit (48), rat, and mouse ventricular myocytes (D. X. P. Brochet, personal communication). Similar Ca²⁺ blinks are evoked by APs in the presence of an LCC inhibitor (to allow for activation of only a few CRUs). Compared with sparks, blinks are much more confined, displaying a typical FWHM of 0.8 µm. Given the size of a pancake-shaped SR cistern (thickness 30 nm; diameter 465 nm, wrapping around a TT of 100–200 nm diameter), this observation indicates that Ca²⁺ depletion during a spark is largely confined to a single cistern, supporting the idea that a spark arises from a single CRU. The magnitude of jSR depletion is substantial, approaching 85% of that during a full-fledged global Ca²⁺ release (30–50% depletion). The existence of a jSR-fSR [Ca²⁺]_SR gradient implies some diffusion restriction connecting the fSR and the jSR. Ultrastructural study reveals that a cistern communicates with the fSR network via only one or a few 30 nm diameter tubules at the outer edge of the jSR cistern (48). The rate of Ca²⁺ refilling via these narrow connections is 35 s^–1, as inferred from the blink recover kinetics (omitting the somewhat 6 times slower reuptake by SERCA). The presence of Ca²⁺ blinks vividly demonstrates the locality of the ER/SR luminal Ca²⁺ signaling and reveals a new dimension of Ca²⁺ signaling specificity and diversity.

View larger version (15K): [in this window] [in a new window]   FIG. 6. Ca2+ blinks. A: Ca2+ blink mirrors Ca2+ spark. Data were obtained in an intact rabbit ventricular myocyte whose SR lumen and cytosolic space were dual stained with fluo 5N and rhod 2, respectively. Line scan images are presented as F/F0. Left inset: SR was stained more intensely at the TT-SR junctions (jSR) by fluo 5N, due to greater local SR volume. B and C: time courses (B) and spatial profiles (C) of the spark-blink pair. (Figure courtesy of D. X. P. Brochet.) D: a typical Ca2+ blink rendered in surface plot. [From Brochet et al. (48), copyright 2005 National Academy of Sciences, USA.]

View larger version (76K): [in this window] [in a new window]   FIG. 7. Sparklet-spark-blink trilogy of cardiac EC coupling. [From Brochet et al. (48), copyright 2005 National Academy of Sciences, USA.]

In saponin-permeabilized amphibian skeletal muscle, store depletion in the spark, dubbed "skraps" (213) (Table 1), has also been visualized, but with a number of surprises. Launikonis et al. (213) developed a shift excitation and emission ratioing (SEER) method of Mag indo 1 fluorescence imaging to measure [Ca²⁺]_SR when mitochondrial Ca²⁺ uptake is suppressed to limit the contaminating mitochondrial fluorescent signal. Their basic observation is that local [Ca²⁺]_SR depletes by 7% during a spark, comparable to the global SR depletion during a twitch (8–15%). The lower magnitude of SR depletion in skeletal versus cardiac muscle (30%) reflects a skeletal SR with larger terminal cisternae containing a high concentration of a CSQ with greater Ca²⁺ binding capacity, thus constituting a much greater Ca²⁺ reservoir. The first surprise is that, after the brief peak (10 ms) of the spark, local [Ca²⁺]_SR continues to decline for nearly 50 ms, apparently violating the law of mass conservation. More surprisingly, a positive [Ca²⁺]_SR transient develops during a cell-wide Ca²⁺ release initiated by caffeine or a low [Ca²⁺] and low [Mg²⁺] solution. The positive [Ca²⁺]_SR transient is most prominent under moderate stimuli and is often preceded by a brief decrease in [Ca²⁺]_SR. These paradoxical observations prompted a model in which CSQ depolymerizes upon local release, liberating its bound Ca²⁺ and giving rise to the positive [Ca²⁺]_SR transient; after cessation of release, repolymerization of CSQ accounts for the anomalous kinetics of the skraps; there is no violation of mass conservation if the CSQ-bound Ca²⁺ pool is also taken into account.

	IX. SPARK MECHANISMS: ACTIVATION

Top Previous Next References

 
In this and the next two sections, we discuss spark mechanisms. What activates a spark? How many RyRs open in a spark? If more than one RyR is involved, how do different RyRs coordinate in a spark? What terminates a spark? Does a spark impart any after-effect on subsequent spark activation? See Table 6 for a range of mechanisms pertinent to sparkology. It should be noted that not all of them are fully expressed in any given cell, though many may coexist. Their expression and relative dominance depends on the type of cell, its developmental stage, and the subcellular regions of concern (e.g., perinuclear versus subsurface CRUs). We begin by focusing on the mechanisms that trigger a spark (see also Table 7).

Spontaneous sparks occur in quiescent unstimulated ventricular myocytes and persist upon transient removal of extracellular Ca²⁺ (69), blockade of LCCs (58), or chemical permeabilization of the surface membrane (242). They are unlikely to be due to Ca²⁺-indpendent or "constitutive" activity (Table 6), because RyRs reconstituted in planar lipid bilayers do not display any significant channel activity at subnanomolar [Ca²⁺] (121, 152). Instead, they could simply reflect the probabilistic opening of RyRs at diastolic [Ca²⁺]_i of 100 nM, via the CICR mechanism. Other trigger signals also influence spontaneous Ca²⁺ spark production. Zhang et al. (437) reported that 50% of spontaneous sparks are sensitive to Cd²⁺ blockade and are therefore attributable to triggering by the spontaneous and infrequent openings of LCCs at resting membrane potential (V_m), in contrast to Katoh et al. (192).

In saponin-permeabilized cells, the DHPR antagonists nifedipine, nimodipine, FS-2, and calciseptine decrease spark frequency (88), while the DHPR agonists BAY K 8644 and FPL-64176 increase it (88, 192, 318), independently of Ca²⁺ entry. Furthermore, the calmodulin-binding LA motif of the _1C COOH-terminal tail increases spark frequency in the periphery of atrial cells (417), while the _1C II-III loop peptide inhibits resting Ca²⁺ sparks in ferret ventricular myocytes (224). These observations can be accounted for by physical coupling between DHPR and RyRs, a mechanism vestigial in heart but fully expressed in skeletal muscle (Table 6).

RyR activity varies as a function of [Ca²⁺]_SR (150, 152, 422), perhaps through altering local [Ca²⁺]_i at the mouth of the channel pore (422), directly acting on RyR from the luminal side, or interacting with luminal Ca²⁺ binding proteins and both junctin and triadin in the SR membrane (151, 154). This "store Ca²⁺-induced Ca²⁺ release" (SCICR) (Table 6) mechanism may be involved in smooth muscle cells under physiological conditions (52, 84, 113). In cardiac cells, however, during time- and thapsigargin concentration-dependent SR Ca²⁺ depletion, Ca²⁺ sparks become smaller but their rate of occurrence remains constant if the small-amplitude events missed from detection are accounted for (357). This suggests that lower than physiological [Ca²⁺]_SR does not affect Ca²⁺ spark production. In contrast, one striking manifestation of SCICR is the so-called "store overload-induced Ca²⁺ release" (SOICR) (185), whereby [Ca²⁺]_SR reaching a threshold "triggers" sudden and near-synchronous global SR release (67, 150, 243, 346). The SOICR mechanism may be particularly relevant to spontaneous spark production under Ca²⁺ overload conditions. Both Ca²⁺ spark frequency and amplitude increase with elevated extracellular Ca²⁺, so rogue Ca²⁺ sparks can now trigger neighboring CRUs to form compound sparks and sometimes initiate a propagating Ca²⁺ wave (67, 69, 316). Nonetheless, there is no direct evidence to support the hypothesis that spontaneous Ca²⁺ sparks arise only when a specific SR cistern becomes "overloaded," as suggested in the literature (185).

Ample evidence shows that the Ca²⁺ spark rate can increase or decrease at a constant SR Ca²⁺ load by the application of agents that sensitize or inhibit RyR activity, respectively; a spark can also progress into an ember in the presence of RyR modifiers that increase the channel open duration (see above). For instance, caffeine, by sensitizing CICR, increases spark production in a dose-dependent manner (319) and may cause apparent spark inhibition by emptying the SR store at high concentrations (10 mM or higher). A more potent and powerful spark activator is perhaps nitroxyl. The nitroxyl donor, Angeli's salt, dose-dependently increases Ca²⁺ spark frequency in cardiac myocytes within minutes of addition. At 1 mM, it causes an 18-fold increase in the rate of occurrence of Ca²⁺ sparks without affecting unitary spark properties (378). Step increase of [Ca²⁺]_i produced by photorelease of caged Ca²⁺ evokes Ca²⁺ sparks of variable sizes, independent of membrane depolarization (230).

Bidirectional regulation of Ca²⁺ spark activity by mitochondrially derived ROS has been observed in intact cardiac myocytes (425). In both cardiac and skeletal muscles, chemical permeabilization of the cell greatly enhances spontaneous Ca²⁺ spark production (175, 176, 242, 244), a phenomenon thought to be related to altered mitochondrial ROS metabolism (175, 176). Finally, in the case of skeletal muscle, removal of DHPR or Mg²⁺ inhibition of RyR1 permits the de novo appearance of spontaneous Ca²⁺ sparks in mammals and augments spark production in amphibians (see below).

B. Triggered Sparks

In cardiac myocytes, sparks evoked during EC coupling are triggered predominantly, if not exclusively, by Ca²⁺ entry through LCCs via the CICR mechanism. Depolarization in the absence of extracellular Ca²⁺, replacement of Ca²⁺ with Ba²⁺, or depolarization to high voltages beyond the apparent reversal potential of LCC Ca²⁺ currents are unable to evoke any sparks (57, 240, 390). Blockade of LCC gating (verapamil, nifedipine) or Ca²⁺ permeation (Cd²⁺) also inhibits spark production (58, 65, 239). In contrast, spark activation in skeletal muscle is controlled by dual mechanisms, primarily depolarization-induced Ca²⁺ release (DICR) and CICR. The former refers to the conformational coupling between the DHPR (voltage sensor) and the RyR (release channel) (197). Both DICR and CICR trigger subsurface sparks in neurons (95).

It has long been known that stretching cardiac muscle causes enhanced [Ca²⁺]_i transients and contraction, a cellular basis for Starling's law of cardiac function (greater contraction is generated by greater preload of the heart). Stretch-induced Ca²⁺ release (SICR) (Table 6) has recently been demonstrated at the level of elementary releases. Ca²⁺ spark production is proportionately enhanced by cell stretch up to 20%, and this effect is completely abolished by the nitric oxide synthase (NOS) inhibitor N^G-nitro-L-arginine methyl ester (L-NAME) and is absent in cardiac myocytes from endothelial nitric oxide synthase (eNOS)-null mice (294). In urinary bladder smooth muscle myocytes, SICR in the form of Ca²⁺ sparks and propagating Ca²⁺ waves can be demonstrated in Ca²⁺-free extracellular solution (183). Similarly, Ca²⁺ sparks independent of Ca²⁺ entry are observed during integrin-mediated mechanotransduction in vascular smooth muscle cells (8).

	X. SPARK MECHANISMS: COORDINATION IN A CRU

Top Previous Next References

 
A. Overview

Delineating the number of RyRs activated in a spark is a matter of fundamental importance to sparkology. Depending on species and muscle type, a CRU in striated muscle contains a variable number of homotetrameric RyRs, ranging from a few tens up to a few hundred of these SR Ca²⁺ release channels. They are assembled into a 2D paracrystalline array of round or elongated geometry (Fig. 8). The vast majority of RyRs are found in such CRUs, the exquisite supramolecular assembly is common to all three RyR isoforms, and this is evolutionarily conserved from invertebrates to mammals (117, 130, 237). Teleologically, such RyR arrays must exist for a purpose, performing biological functions that are impossible or less than optimal if RyRs exist as rogue channels. Reporting a CRU in action, Ca²⁺ sparks can best serve as an investigative tool to reveal Nature's secrets hidden in this supramolecular Ca²⁺ signaling nanomachine.

View larger version (50K): [in this window] [in a new window]   FIG. 8. Organization of cardiac and skeletal ryanodine receptor (RyR) arrays. A: front and side views (inset) of RyR arrays in an invertebrate skeletal muscle of cardiac-type EC coupling (i.e., via the Ca2+-induced Ca2+ release mechanism). About a hundred RyRs (marked by arrows) are assembled into a 2-dimensional lattice. Such assembly is thought to be representative of RyR arrays in cardiac myocytes. [From Loesser et al. (237). Courtsey of C. Franzini-Armstrong.] B: conceptual model showing a cardiac RyR array relative to LCCs in the plasma membrane. C and D: organization of RyR arrays in mammalian versus frog skeletal muscle. A couplon in frog skeletal muscle (C) consists of a central double row of RyR1 (blue squares), which are alternately coupled to DHPR tetrads (red circles, TT not shown), and a significant presence of RyR3/β (gray squares) in a parajunctional position (117). D shows the lack of a parajunctional RyR3/β array in a mammalian skeletal muscle couplon (129). [Modified from Ward and Lederer (394).]

B. Key Observations

Seven key observations in cardiac cells guide the current investigations and provide a context for past findings. 1) The I_spark is 3 pA and lasts for 10 ms (69, 351, 390, 392). 2) The fractional Ca²⁺ current through a single RyR2 channel reconstituted in planar lipid bilayer has been measured to be smaller than initially estimated and appears to be within the range of 0.5–0.7 pA under physiological ionic conditions (194). 3) Polymorphism of Ca²⁺ sparks can be demonstrated for events from the same CRUs (331, 390). Nevertheless, the amount of Ca²⁺ released and the duration of release are somewhat stereotypical. 4) Special conditions can lead to very long Ca²⁺ embers at about half of the peak spark amplitude (69, 346, 420). 5) Substantial depletion of cisternal Ca²⁺ develops during a spark (48). 6) Many (50–300) RyRs are clustered to form each CRU as noted above (130, 352) 7) Recent results suggest that there may be substructure (quanta) in I_spark and the number of quanta in a spark varies from event to event (390, 392). In these studies, Wang et al. used sparklets produced by a known LCC current as an optical "yardstick" and scaled the rising phase of a spark to infer the I_spark. The histogram of this calibrated I_spark suggests peaks every 1.2 pA, as if sparks consist of Ca²⁺ quanta of 1.2 pA Ca²⁺ flux (Table 1). However, little is known as to whether a quantum corresponds to a single RyR or a group of tightly coupled RyRs. If such a spark substructure is supported by future experiments, it will certainly be an important basis for developing mechanistic insights into the behavior of RyR2 CRUs.

C. CICR and Coupled Gating

Once an LCC or RyR opens, Ca²⁺ floods into the subspace and can thereby activate other RyRs. This CICR is central to all views of Ca²⁺ spark triggering and activation, and clearly involves a high-gain positive feedback regulation. Based on the discussion above, that _latency, is 10 ms at 0.3 pA Ca²⁺ flux, and the power law of spark activation, we predict that _{latency,apparent} is 0.9 or 0.06 ms at 1 or 4 pA Ca²⁺ flux, respectively. Thus we might expect that CICR alone is sufficient to activate all RyRs in a fraction of a millisecond. Thus, if not all RyRs in a CRU are activated in a spark, there must be an equally powerful, yet-to-be-found inhibitory mechanism that inactivates the channels even before they can be activated.

The second mechanism invokes the notion that RyRs interact via conformational coupling. In this manner, gating of the RyRs in the CRU involves cooperativity among the RyRs, which is also known as "coupled gating" (249, 251), since it was first identified in planar lipid bilayer experiments with two RyRs that appeared to open and close synchronously. How much does cooperative gating contribute to RyR activation in a spark? To a first approximation, not much. Since subspace CICR is already expected to be powerful, an additional activation mechanism would be redundant. Nonetheless, one important benefit of cooperative RyR gating is to confer a steeper Ca²⁺ dependence for CRU activation compared with activation of rogue RyRs (347) (Fig. 9). CRUs can thus be made stable at resting or low [Ca²⁺] but highly responsive to subspace trigger [Ca²⁺]. Indeed, Stern and colleagues (367) showed that no published gating schemes of individual RyRs incorporated into EC coupling models can explain the systolic and diastolic features of cardiac EC coupling simultaneously; introducing cooperative gating among arrayed RyRs effectively cures the poor behavior of the models. Such cooperative gating would provide a compelling rationale for packing RyRs into large arrays.

View larger version (15K): [in this window] [in a new window]   FIG. 9. Rogue RyRs underlie sparkless release: a hypothesis. Plots show model predictions of steady-state Po as a function of cytosolic [Ca2+] for isolated (i.e., "rogue") RyRs (solid line) and strongly coupled RyRs (solid circles and dashed line). In this model, the energetic coupling between RyRs represses the channel activity at low [Ca2+], causes a steep transition at a critical [Ca2+] level, and enhances the channel activity at higher [Ca2+]. One implication of this is that rogue RyRs tend to be more easily activated by small increases in intracellular [Ca2+] above the resting level than clustered RyRs that gate cooperatively. The release from rogue RyRs in response to low-level trigger Ca2+ thus might manifest as sparkless release. [Modified from Sobie et al. (347).]

D. How Many RyRs Open During a Spark?

There is no simple answer. Reflecting on the range of spark models proposed, we examine two schools of thought, the "one and a few channels" model (16, 69, 141, 321) and the "whole-CRU" model (30, 179, 346).

The first model suggests that only a tiny fraction of RyRs in a CRU are activated to release Ca²⁺ in a spark. The supporting evidence includes the average 4 pA flux, which is not far from that of a few RyRs; the plateau of modified cardiac sparks at 50% of the peak level, presumably reflecting the sustained opening of a single RyR; and the quantal substructure in I_spark (if a quantum represents a single or a small number of RyRs). The problems are twofold, however. The first is a sort of teleological consideration: Why on earth would only a small percentage of RyRs be activated in EC coupling? Could it be a solution to the pressing need of maintaining stability at diastole while enabling high-gain amplification at systole? The second is: What prevents RyRs from activating when exposed to high [Ca²⁺]_subspace? This model has to invoke either a Ca²⁺-dependent inactivation that is much stronger than previously thought (392), or negative cooperativity among RyRs (an open RyR represses neighboring RyRs from activation), which is yet to be experimentally demonstrated. In this model, all RyRs in a CRU participate in the "to activate or to inactivate" computation, but only one or a few end up open, and the others are inactivated.

The whole-CRU model, however, states that nearly all RyRs in a CRU are activated by CICR and coupled gating during a spark. In this model, spark termination does not invoke channel inactivation, but rather, a powerful luminal Ca²⁺ desensitization of CICR upon cisternal Ca²⁺ depletion. The supporting evidence includes the lack of rapid and powerful channel inactivation in vitro, and on the demonstration of luminal Ca²⁺ regulation of CICR sensitivity (150, 151, 153, 328, 422; see also Ref. 359). This model is also consistent with the somewhat stereotypical amount of Ca²⁺ and release duration in a spark (392). Furthermore, it predicts that, as long as a CRU contains many RyRs, spark generation should be robust and independent of the number of RyRs in a CRU. For long-lasting sparks, the peak and plateau ratio now depends on the time-dependent cisternal Ca²⁺ depletion. Some degree of polymorphism is expected from stochastic variation in CRU open duration. The problem of the tiny unitary current per RyR (4 pA for 100 RyRs or 0.04 pA/RyR) could possibly be accounted for by a diminishing electrochemical potential of Ca²⁺ {due to the rise of [Ca²⁺]_subspace (346, 350) and the fall of cisternal [Ca²⁺]_SR (48)}. How such a model can accommodate the I_spark substructure (392) remains to be determined.

	XI. SPARK MECHANISMS: TERMINATION

Top Previous Next References

 
CICR, in its simplest form, is expected to produce everlasting local release of Ca²⁺. That I_spark lasts only 5–10 ms is indicative of powerful termination mechanism(s) that turn off the elemental Ca²⁺ signals, sharply and promptly. So, what terminates a spark? To date, no single mechanism can adequately account for spark termination, and more questions have been raised than answered. In this section, we discuss current concepts and experimental evidence pertinent to spark termination.

A. Local Ca²⁺ Depletion

Emptying the cisternal Ca²⁺ store could, in principle, extinguish the spark. Significant cisternal depletion is also a requisite for deactivating or desensitizing RyRs via the SICR mechanism operating in reverse mode (Table 6) (150, 151, 153, 185, 276, 328). As discussed above, brief and substantial jSR Ca²⁺ depletion does develop during a cardiac spark, and evidence for this comes from the Ca²⁺ blink that mirrors the spark (48). However, store depletion cannot be the sole mechanism of spark termination because of a simple "counterexample": the rate of cisternal refilling, measured as the recovery time of a blink, is about six times faster than the rate of restitution of Ca²⁺ spark amplitude (48). Similar results are obtained when the restitution is measured by other means including the probability of spark reactivation (349). Inferred from their data in skeletal muscle, Launikonis et al. (213) proposed that CSQ, a high-capacity Ca²⁺ buffering protein in the cisternal lumen, depolymerizes and reduces its Ca²⁺-binding capacity upon store release. If this is applicable to cardiac myocytes, the slow recovery of spark amplitude could reflect a slow restoration of the CSQ polymer-bound Ca²⁺ pool. Taken together, these data suggest that local store depletion may work in synergy with additional inhibitory mechanisms for spark termination and restitution.

B. Stochastic Attrition

"Stochastic attrition" refers to the inherent random closing of individual channels that can turn off an activated channel cluster (363) (Table 6). Albeit appealing for parsimony, stochastic attrition works only if a single or very few RyR channels (or units of coupled gating) are activated in a spark (363). For stochastic attrition alone to terminate CICR in a CRU, the predicted release duration would increase astronomically with increasing number and P_o of independently gated channels in the spark. Furthermore, stochastic attrition does not impose any after-effect, in contrast to the local refractoriness of CICR that prevents a CRU from immediate reactivation (48, 324).

C. Desensitization of CICR by Ca²⁺ Store Depletion

The sensitivity of the RyRs to be activated by cytosolic Ca²⁺ is augmented by increasing luminal Ca²⁺ (67, 150, 243, 346). When the local store depletes, not only do the I_spark and [Ca²⁺]_subspace diminish, but also the RyRs deactivate or desensitize. That is, the SCICR can operate in reverse mode. Recent advances in Ca²⁺ blinks support a substantial local store depletion in a spark, a requisite for this mechanism to work in intact cells. In a theoretical model, Sobie et al. (346) have shown how luminal dependence of cooperative gating among RyRs can underlie robust Ca²⁺ spark termination without Ca²⁺-dependent inactivation.

D. Coupled Gating as a Termination Mechanism

In contrast to the prevalent view on coupled gating as a CRU activation mechanism, we emphasize here that it is more properly considered as a release stabilization and termination mechanism. In other words, the RyR-RyR interaction in a CRU may impart a powerful basal repression of spontaneous release in resting cells and terminate the release during a spark (347) or a global [Ca²⁺]_i transient (367). Putative RyR inhibition of neighboring RyRs (in a CRU) is analogous to the DHPR inhibition of RyR, which has been experimentally demonstrated in skeletal muscle couplons (206, 443).

E. Channel Inactivation and Adaptation

RyR inactivation can be Ca²⁺ dependent or "fateful" (296). Fateful inactivation refers to spontaneous transition of an open channel to an inactivated state via a Ca²⁺-independent mechanism. In a broad sense, the Ca²⁺ dependence of inactivation may involve binding of Ca²⁺ to inactivating sites, Ca²⁺-dependent phosphorylation or dephosphorylation of the channel per se, or of a molecular partner in the macromolecular complex. The net effect is the loss of channel activity in the presence of sustained or new stimuli.

1. Ca²⁺-dependent inactivation

Early demonstration of CICR by rapid application of solutions of defined [Ca²⁺] to mechanically skinned cardiac Purkinje fibers revealed a bell-shaped Ca²⁺ dependence of force generation, with the peak at low micromolar [Ca²⁺]. At a given [Ca²⁺], the faster the application rate, the greater the CICR effect (114–116). From these experiments, Fabiato (116) postulated that Ca²⁺-dependent inactivation of high affinity but slow kinetics occurs concurrently with CICR. Näbauer and Morad (270) tested Ca²⁺-dependent release inactivation in intact cardiac myocytes. They showed that a preconditioning elevation of cytosolic [Ca²⁺] (from 90 to 110 nM for 500 ms) does not prevent I_Ca from triggering substantial release. This observation has been widely interpreted as evidence against a Ca²⁺-dependent inactivation in situ. This argument, however, may be flawed. Their observation can be readily explained if the K_d of inactivation is well above the conditioning [Ca²⁺] involved. Moreover, the conditioning protocol would upload Ca²⁺ into the SR due to the elevation of steady-state [Ca²⁺]_i and therefore mask the inactivating effect, if any. A similar argument was made by Pizarro et al. (296) against similar proposals of inactivation by relatively low Ca²⁺ in skeletal muscle (340). In another experiment, they showed that photolytic release of Ca²⁺ delivered 50–100 ms after the onset of depolarization during the rise of contraction always potentiates cell contraction. This observation is not inconsistent with Ca²⁺-dependent inactivation either, because the photolytic Ca²⁺ not only can directly contribute to cell contraction, but also can elicit CICR on its own. Insights gleaned from mathematical modeling (350) show that [Ca²⁺]_subspace rises to the 100 µM range in a spark, suggesting that local Ca²⁺-dependent inactivation remains an appealing possibility, perhaps it is such local inactivation that matters in terminating a spark (but see discussion elsewhere).

2. Planar lipid bilayer experiments

A reductionist approach to the spark mechanism is to investigate RyR single-channel behavior in planar lipid bilayers. However, it should be cautioned at the outset that many regulatory features, including interactions between RyRs and their molecular partners (252), are stripped away in such experimental settings (Table 8). Direct single-channel measurement suggests that neither activation nor inactivation features of RyRs are simple. Inactivation, when it is observed, is too slow at physiologically relevant [Ca²⁺] levels. Is inactivation of RyR invoked? Yes. However, the inactivation and its characterization are complicated by varying experimental conditions and methods. There are multiple "kinds" of inactivation reported in the literature, and in this focused presentation, we divide the reports into two groups for the sake of simplicity. There is a complex Ca²⁺-dependent slow inactivation called "adaptation" (66, 122, 152, 383) and a slow complex inactivation that we will call "other." The "other" group is unified by the denial of adaptation as a valid inactivation mechanism (209, 210, 343). None of the "other" inactivation schemes is fast enough to enable RyRs to inactivate in a few milliseconds (165, 215), the approximate rate needed to underlie spark termination. The rate of adaptation inactivation ranges from 150 ms to seconds, while the rate of the other forms of inactivation range from more than a second to longer than 30 s (165, 215). Clearly, as they stand, neither adaptation nor "other" inactivation is fast enough.

A complex kind of inactivation called "adaptation" (152, 383) (Table 6) is observed following a rapid increase of cytosolic Ca²⁺ by photolysis that produces an increased P_o. Adaptation describes the decline in P_o of RyRs, and it differs from inactivation because an "adapted" channel retains responsiveness to subsequent higher [Ca²⁺] steps (152) (Table 6). Conceptually, the adaptive behavior at the single-channel level can be explained on the basis of complex time-dependent modal gating schemes (66). It should be cautioned that not every group can demonstrate the reactivation of the "adapted" channel by a second, higher [Ca²⁺] stimulus, particularly when purified RyRs are involved (see Refs. 122, 210 for reviews). There has also been discussion in the literature about the possibility that there is a complex time course of [Ca²⁺]_i following photolytic uncaging of Ca²⁺, and this complex time course may contribute to adaptation (211, 384). Adaptation, when found, is relatively slow (100 ms to seconds) and thus kinetically inadequate to account for Ca²⁺ spark termination (ms to tens of ms), but could contribute to changes in RyR responsiveness over a longer time scale.

In crayfish skeletal muscle, where EC coupling is mediated by the CICR mechanism, Yasui et al. (431) demonstrated that depolarization to +30 mV in the presence of FPL64176 elicits a transient Ca²⁺ release that terminates despite continued I_Ca. Yet, additional Ca²⁺ release can be triggered by tail I_Ca upon repolarization. Using Ca²⁺ spike measurement, Sham et al. (324) found that the tail current produced on repolarization elicits Ca²⁺ spikes that most likely originate from RyRs unfired during depolarization, rather than from those in the adapted state. After a maximal activation, a multifold increase in I_Ca (by hyperpolarization to –120 mV) at 50 ms interval fails to evoke any additional release, indicating absolute refractoriness of CICR. These results argue in favor of inactivation, rather than adaptation, as the specific mechanism of release refractoriness (and spark termination). However, the high degree of intracellular buffering that significantly slows jSR refilling may complicate the interpretation.

F. Refractoriness of Local CICR

Following a global Ca²⁺ transient or wave, triggered Ca²⁺ release becomes refractory. Refractoriness of global CICR is exemplified by annihilation of two Ca²⁺ waves in head-on collision and suppression of depolarization-activated [Ca²⁺]_i transients in the wake of a Ca²⁺ wave. It has been thought that the mechanisms of refractoriness may have a lot in common with those that terminate the release in the first place.

Refractoriness of local CICR has been convincingly demonstrated by different groups. An early approach by Cheng et al. (67) examined restitution of AP-elicited Ca²⁺ release in the wake of a Ca²⁺ wave and suggested that restitution may be as fast as 150 ms. Tanaka et al. (371) showed that, immediately following a spontaneous spark, there is oftentimes a void at the sparking site when a global [Ca²⁺]_i transient is elicited by an AP. This approach suggested that restitution of Ca²⁺ release takes at least 25 ms. DelPrincipe et al. (96) showed that there is little change in the triggered release of Ca²⁺ produced by long pulses of two-photon uncaging of Ca²⁺, but restitution over periods less than the minimum time between pulses (200 ms) would have been missed. Sham et al. (324) used the Ca²⁺ spike method to investigate Ca²⁺ release triggered by depolarization and by repolarizing I_Ca tail currents. Use-dependent inhibition of triggered Ca²⁺ release was observed, but the kinetics of the findings are difficult to interpret because of the significant level of cytosolic Ca²⁺ buffer.

Terentyev et al. (375) and Sobie et al. (349) used Imperatoxin A in permeabilized cells and low levels of ryanodine in intact myocytes to investigate Ca²⁺ spark restitution. These agents induce repetitive Ca²⁺ sparks and can thus be used to examine spark restitution. Sobie et al. (349) reported that the amplitude of Ca²⁺ sparks recovers with a time constant of 91 ms, while the triggering probability recovers with a time constant of 80 ms. Using a loose patch-clamp method, Brochet et al. (48) found that the sparklet-spark coupling fidelity and the amplitude of the succeeding spark decrease substantially following spark activation. The recovery of local refractoriness appears to be much faster (spark amplitude restoration with a half-time of 150 ms) than the restitution of global CICR (half-time 600 ms); the latter is rate-limited by the refilling of the SR Ca²⁺ store. Taken together, these findings suggest that local refractoriness does develop and the Ca²⁺ blink measurement suggests that the important region of depletion is the cistern. However, cisternal refilling alone cannot fully explain the kinetics of restitution, necessitating the involvement of other mechanisms with slower kinetics.

G. RyR Gating In Vivo Versus In Vitro

At the heart of the uncertainty and controversy regarding the mechanisms of spark activation, coordination, and termination is the paucity of information on RyR gating in situ. It has been increasingly appreciated that caution must be exercised in translating results from the planar lipid bilayer to intact cells, and vice versa. Generic differences between RyRs in a spark and those in planar lipid bilayers are summarized in Table 8. In the first place, gating of RyRs in a spark is constrained by the 2D array assembly to which they belong. The array formation is a property intrinsic to the RyR, as purified RyRs can form large-scale arrays in solution or on the surface of positively charged lipid membrane (432, 433). In a skeletal muscle couplon, DHPRs also physically constrain the gating of RyRs in the junctional CRU. Second, the channel behavior in a spark may be modified by a large number of molecular partners in the macromolecular complex on the cytosolic side (e.g., calmodulin, calstabin, sorcin, kinases, phosphatases) (47, 222, 250, 252, 259), in the lipid membrane (e.g., triadin, junctin), and inside the lumen (e.g., CSQ). Most of these are dislodged during vesicle preparation and particularly by channel purification prior to reconstitution. Third, the steady-state behavior studied in vitro does not serve as a good predictor of transient, nonequilibrium responses of the channel, for there is no unique or simple relationship between kinetic and steady-state responses. Last but not least, many positive- and negative-feedback loops (e.g., local CICR, desensitization by luminal Ca²⁺ depletion, Ca²⁺-dependent phosphorylation and dephosphorylation) could be involved in shaping the sparks. Not all these gating and regulatory mechanisms can be readily reproduced in reductionist approaches in vitro. With these differences kept in perspective, we can judiciously integrate the findings in cell-free systems with the spark measurement in intact cells. Both types of approaches are indispensable in defining how RyR behavior is linked to Ca²⁺ sparks.

	XII. Ca2+ SPARKS IN HEART DISEASE

Top Previous Next References

 
Diseases of the heart may affect Ca²⁺ sparks and their properties. These changes reflect the alterations in many regulatory features of the RyR2 macromolecular complex, such as altered RyR2 itself and its associated proteins, spatial organization of the RyR2 clusters, intercluster distance, sensitivity of RyR2 to [Ca²⁺]_i, the SR Ca²⁺ content, and Ca²⁺ "leak" from the SR, as well as AP duration and shape. Recent findings suggest that diverse diseases (e.g., diabetes, muscular dystrophy) with cardiovascular consequences may also contribute to cardiac Ca²⁺ signaling dysfunction. With respect to these new findings and current exploration of treatments involving stem cells and novel small molecular therapeutic agents, much work remains to be done to characterize the links between diseases and altered behavior of Ca²⁺ sparks and SR Ca²⁺ release. Importantly, inventive mathematical modeling may provide new understanding of Ca²⁺ sparks and Ca²⁺ signaling in heart and lay the foundation for future experiments and analysis.

A. Arrhythmic Disease

An early clue that Ca²⁺-activated current contributes to arrhythmogenesis came from studies of current oscillations in heart that were attributed to SR Ca²⁺ overload (191, 219). Follow-up investigations revealed small current fluctuations that were linked to mechanical and [Ca²⁺]_i "noise" (60, 189, 365, 412). The arrhythmogenic membrane current, identified as the "transient inward current" (I_TI) appeared to arise from a synchronization of Ca²⁺-dependent current fluctuations and what would turn out to be a propagating wave of elevated [Ca²⁺]_i (20). I_TI can activate sufficient inward current to produce arrhythmias in two ways. The first is to augment early afterdepolarizations (EADs) due to the interactions of I_Ca and repolarizing K⁺ currents (219); the second is to produce de novo delayed afterdepolarizations (DADs) (118, 119, 219, 388). Both EADs and DADs can produce aberrant electrical activity and extrasystoles (arrhythmias).

This classic view of Ca²⁺-dependent arrhythmogenesis was initially clarified by the discovery of Ca²⁺ sparks when it was shown that Ca²⁺ waves are initiated by a "triggering" Ca²⁺ spark in the condition of Ca²⁺ overload (69). Ca²⁺ overload is the condition that occurs when there is an elevated amount of Ca²⁺ in the SR. The high SR Ca²⁺ content makes the SR unstable (190). That the Ca²⁺ wave is linked to I_TI at the cellular level became clear in multiple studies (20, 87, 193, 242, 355, 408). This did not address the question, however, whether or not these events are linked to arrhythmogenesis in the intact heart. Early work did suggest that the underlying cellular events can be "scaled up" to account for arrhythmogenicity in intact cardiac tissue (408). Some of the complexities of these connections are discussed below.

How the waves of elevated Ca²⁺ are propagated in heart cells is central to the understanding of the role of Ca²⁺ sparks in arrhythmogenesis. Early examination showed three relevant features of Ca²⁺ waves. 1) As noted above, they appear to be initiated by Ca²⁺ sparks. 2) The propagation depends on sequential propagating activation of Ca²⁺ sparks by the locally elevated Ca²⁺ due to the earlier Ca²⁺ sparks (67). 3) A common feature identified is the increased sensitivity of the CICR process that enables Ca²⁺ sparks that would not normally activate neighboring Ca²⁺ sparks to do so (67, 178, 180, 408).

Ca²⁺ overload itself as an RyR2 "sensitizer" plays an important role in producing Ca²⁺ instability in heart under many conditions. While it was first described following blockade of the Na⁺ pump by cardiotonic steroids (82, 83, 109, 191, 219), it also plays a key role in the development of Ca²⁺ instability in a variety of conditions. For example, increased SR Ca²⁺ load has been put forward as a central factor in increased RyR2 sensitivity following β-adrenergic stimulation (225), although phosphorylation of RyR2 may also influence the process (see below). Ca²⁺ overload is an important contributor to the generation of Ca²⁺ alternans (103, 108, 138) and in many channelopathies (see below).

B. Heart Failure

There are many distinct insults that presage heart failure (HF) and for much of this discussion, we mean left-sided HF. In this late-stage condition, the left ventricle is unable to pump enough blood to meet the demands of the tissue. HF is produced in reaction to acute pump failure or sustained excess afterload or both, and pulmonary edema often develops during HF. The acute injury that leads to HF is severe but not lethal as in an acute myocardial infarction (MI). Following that insult, the unaffected tissue must compensate for the infarcted tissues, and this leads to "remodeling" of the healthy cells. In a similar manner, the stress of hypertension is not necessarily acutely associated with HF, but the HF develops with sustained hypertension and consequent cardiac remodeling. In these contexts, the remodeling is the response of healthy tissue to sustained excess work and is associated with the onset of Ca²⁺ signaling dysfunctions. While it may provide some adaptive benefit initially, the remodeling becomes maladaptive when it continues in the face of sustained demand. There are six clear changes that appear to occur with this maladaptive remodeling: 1) reduction in the SR Ca²⁺ pump, SERCA2a (93); 2) increase in NCX level (342, 421); 3) prolonged APs (see below); 4) altered TTs (361); 5) the development of dyssynchronous Ca²⁺ release (233); and 6) increased SR Ca²⁺ leak (see below).

C. Remodeling of the Ca²⁺ Signaling System

1. CRU organization

The spatial organization of the CRUs may change in disease and affect the likelihood that a Ca²⁺ spark at one CRU will trigger a Ca²⁺ spark at a neighboring CRU. As noted above, there is 1 CRU/fl (µm³) in rat ventricular myocytes separated by 0.65 µm in the plane of the Z-disk and by 0.16 µm from the centroid of the Z-disk (352) (Fig. 10). The couplons (made up of between 120 and 260 RyRs) in one Z-disk are separated from the couplons in the next Z-disk by the sarcomere spacing (1.8 µm in human and rat). Despite the asymmetry of couplon spacing, Ca²⁺ waves, when observed, are near-spherical. Cannell and colleagues (352) suggest that this may come about because of the asymmetry of longitudinal (faster) versus transverse (slower) diffusion (68). In this regard, anisotropy of Ca²⁺ diffusion appears to dictate the disparity of inter-CRU distance in differenct directions. Such possibilities are being investigated by several groups.

View larger version (35K): [in this window] [in a new window]   FIG. 10. Constellation of Ca2+ release units (CRUs) in cardiac myocytes. A: CRU distribution in a transverse section of a human ventricular myocyte. Each circle represents a CRU detected after deblurring of confocal RyR-immunofluorescence images. B: color-coded map of the areas served by each CRU in a Z-disk of a human ventricular myocyte. Scale bar: 2 µm. [See Soeller et al. (352).]

The reduction in the expression of SERCA2a leads to a reduction in the amount of Ca²⁺ in the SR, and this alone would reduce the [Ca²⁺]_i transient. The increase in NCX expression would tend to reduce the SR Ca²⁺ still further but would predispose to the generation of a larger I_TI for a given increase in [Ca²⁺]_i during a Ca²⁺ wave (298). The depolarization-produced I_TI is enhanced further by the lower level of repolarizing K⁺ current that is also found in HF (298, 312). Exactly how changes in SERCA2a and NCX are produced and the molecular signaling involved remain unknown, but there are some clues about Ca²⁺ and nuclear factor of activated T-cell (NFAT) signaling pathway contributing to the changes in K⁺ current (312). The major question is how does the elevation of [Ca²⁺]_i come about? What roles are played by Ca²⁺ sparks? How does this come about in a cellular environment that has, in principle, reduced total Ca²⁺?

3. Dyssynchrony of Ca²⁺ sparks

Song et al. (360) first proposed that synchrony of Ca²⁺ sparks is an important determinant for physiological modulation of cardiac EC coupling. A key observation was made by Litwin, Zhang and Bridge in the context of EC coupling in HF (233). They observed that in cells from an MI model of HF, Ca²⁺ sparks do not behave normally during the AP (233). Although depolarization of the HF myocyte activates a somewhat diminished I_Ca, the surprise comes from the timing of the Ca²⁺ sparks. The Ca²⁺ sparks are not synchronized with the depolarization, and "late" Ca²⁺ sparks are present so these Ca²⁺ sparks are "dyssynchronized" with respect to the depolarization. Isoproterenol treatment of the afflicted HF ventricular myocytes leads to reduction of the dyssynchronization (233). The many actions of isoproterenol on the heart cell make this observation useful, but it does not point to a more specific cause.

Kamp and co-workers (9), using tachycardia-induced HF, suggested that Ca²⁺ signaling dysfunction arises from depletion of the TTs in HF heart cells. While this seems a less subtle cause of dysfunction than dyssynchronization of Ca²⁺ sparks may have required, the concept of TT "disease" is appealing in principle. Sipido and colleagues (241) examined pig ventricular myocytes in culture that had lost TTs and compared their findings to human ventricular myocytes from individuals in HF, supporting the hypothesis that TT loss is linked to dyssynchronous Ca²⁺ sparks. Using spontaneously hypertensive rats (SHR) in HF, Song et al. (358) showed that the TTs (that have a preponderance of LCCs) appear to dedifferentiate and become increasingly disorganized in SHR ventricular myocytes in HF while the CRUs and the RyR2 clusters associated with the SR remain largely in place (Fig. 11). There is thus increasing separation between the trigger (I_Ca) for Ca²⁺ sparks and the RyR2 release channels. Recent work of Kamp and co-workers (254) supports these conclusions by simultaneously imaging TTs and Ca²⁺ sparks in HF.

View larger version (44K): [in this window] [in a new window]   FIG. 11. Intermolecular relationships between LCCs and RyRs in normal and diseased heart cells. A: TT disorganization in ventricular myocytes from failing SHR heart (SHR/HF) and age-matched control (WKY). B: immunostaining of RyRs (top) and LCCs (bottom) in SHR/HF and WKY ventricular myocytes. Note the chaotic distribution of LCCs in the diseased heart cell. [A and B from Song et al. (358), copyright 2006 National Academy of Sciences, USA.] C: diagram of TT-SR structure.

RyR2-mediated Ca²⁺ efflux from the SR is thought to depend on two components: triggered Ca²⁺ release and "leak" Ca²⁺ loss. While triggered SR Ca²⁺ release underlies the normal [Ca²⁺]_i transient and involves Ca²⁺ sparks, the leak loss of Ca²⁺ from the SR occurs during diastole. In addition to spontaneous Ca²⁺ spark contribution to the leak, there is also the probability of RyR2s opening as rogue RyR2s (1 to about 5 RyRs in a cluster) so that there is no visible Ca²⁺ spark (346, 414). If RyR2s behave cooperatively when they are in a cluster (249), then the [Ca²⁺]_i dependence of P_o is much steeper (Fig. 9). This means that with large clusters, as are seen in the CRU, cooperativity may be significant, and with that high degree of cooperativity, there is a reduced likelihood that the cluster is triggered at low [Ca²⁺]_i. The rogue RyR2s with no or small degrees of cooperativity may be more readily activated than CRUs by low [Ca²⁺]_i. In the simplest circumstance, when there is an increase in SR Ca²⁺ leak without other changes, there is a decrease in SR Ca²⁺ content and hence a decrease in the tendency to produce Ca²⁺ sparks and Ca²⁺ waves (29). Thus such an increase in Ca²⁺ leak tends to be antiarrhythmic. Recent work, however, has suggested a "paradox of Ca²⁺ signaling" in the diseased heart: there is an increased leak, and this SR Ca²⁺ leak is associated with arrhythmias (18, 220–222, 377, 385, 398, 399).

It is, however, a challenge to square the relationships between and among the signals with the requirements of SR pump-leak balance, the tendency of one Ca²⁺ spark to trigger a wave and of Ca²⁺ sparks to sustain the propagation of the wave. This is a clear example where mathematical modeling of Ca²⁺ sparks and Ca²⁺ waves within a single cell is needed to inform our thinking on how the arrhythmias arise, what role may be played by Ca²⁺ sparks during the initiation, how Ca²⁺ sparks may contribute to sustained activation, and how leak and "invisible" SR Ca²⁺ release may contribute to Ca²⁺ wave propagation. Some of these considerations provide a context for future modeling and experimentation.

5. AP duration and arrhythmias

Changes in AP duration and shape occur in diverse acquired and genetic diseases (78, 275). These changes can affect Ca²⁺ sparks and Ca²⁺ waves by changing SR Ca²⁺ load and diastolic [Ca²⁺]_i through voltage-dependent action on Ca²⁺ entry and the NCX Ca²⁺ efflux across the surface membranes. In addition, the changes in AP profile alter the triggering of Ca²⁺ sparks (46, 59, 146, 313, 345). Thus the tendency for the ventricular AP to be longer in HF and other diseases, often due to less repolarizing K⁺ current (188, 207, 216, 260, 269, 406), tends to promote both Ca²⁺ loading and Ca²⁺ sparks.

When EADs develop due to prolonged APs, they are attributed to interactions between the repolarizing K⁺ currents and the depolarizing Ca²⁺ currents (92, 161, 281, 282) or possibly larger than normal inward Na⁺ currents (1, 78, 123). The EADs can, however, be significantly affected by Ca²⁺ sparks and Ca²⁺ waves via I_TI (20, 73, 89, 219), the latter making them more readily reach threshold to trigger an extrasystole. DADs, in contrast, depend primarily on Ca²⁺ sparks and Ca²⁺ waves to activate I_TI. Importantly, however, the amount of NCX that is expressed and is functional in a cell, as well as the amount of K⁺ current present, appear to play an important role in determining how large the voltage excursion is when a Ca²⁺ wave actives I_TI (298).

D. Genetic and Acquired Channel Dysfunction

Ion channel dysfunction may have a profound effect on Ca²⁺ sparks and Ca²⁺ waves through gain- or loss-of-function mutations or by other means. For channels that pass Ca²⁺ or Na⁺, the ion flux may lead to Ca²⁺ overload directly or indirectly (via NCX). Loss of function mutations or drugs that block K⁺ channels lead to prolonged APs. Channel trafficking problems can lead to mistargeting of ion channels so that they end up in the wrong membrane microdomain or get stuck at some point in the processing. For example, K⁺ channel dysfunction may be associated with ion channel trafficking errors (10). When the abundance of functional K⁺ channels is decreased (due to any cause including trafficking errors), APs lengthen. Prolonged APs increase SR Ca²⁺ load by favoring Ca²⁺ entry and decreasing Ca²⁺ extrusion and may lead to EADs and Ca²⁺-dependent arrhythmogenesis as noted above. Excess Na⁺ entry through the expression of late Na⁺ channels or poorly inactivating I_Na (132) can affect EADs but can also increase the intracellular Na⁺ concentration and hence [Ca²⁺]. It was appreciated over 30 years ago that the intracellular Na⁺ level has a significant affect on Ca²⁺ in the SR and hence on the cellular vulnerability to Ca²⁺ instability (110, 190, 268), and this has been seen in diseased rodent heart cells (3, 101). The elevation of intracellular Na⁺ leads to an increase in Ca²⁺ sparks, Ca²⁺ waves, and I_TI and thus acts to produce arrhythmias through a Ca²⁺ overload mechanism.

RyR2 dysfunction can also develop, and this produces the "leaky" RyR2 that has been widely presented in literature. Specifically mutations in RyR2 (248) increase its P_o at constant [Ca²⁺]_i and make ventricular myocytes unusually vulnerable to triggered or spontaneous activity. The molecular mechanisms by which this may happen remain uncertain because of the dual-dependency of RyR2 P_o on Ca²⁺; it is sensitive to both [Ca²⁺]_i and [Ca²⁺]_SR as noted above. Since the rapid triggering of a CRU, is affected by the opening of an apposing DHPR, the more parsimonious explanation is that local [Ca²⁺]_i "triggers" the CRU, while [Ca²⁺]_SR "tunes" the sensitivity of triggering (Table 7).

Just as RyR2 mutations may increase the P_o of RyR2s by increasing their sensitivity to [Ca²⁺]_i, mutations in proteins associated with the RyR2 can do the same. Mutations in the cardiac isoform of CSQ, CSQ2, have also been recently reported to increase RyR2 P_o and sensitivity to [Ca²⁺]_i and cause Ca²⁺ instability, Ca²⁺ sparks and waves, and arrhythmias (76, 105, 151, 154, 155, 201, 234, 326, 353, 374). The other associated proteins (junctin and triadin), when they are mutated or expressed at abnormal levels, may also bring about altered sensitivity of RyR2 to [Ca²⁺]_i and contribute to arrhythmogenesis (17, 131, 135, 143, 151, 155, 196, 373, 376).

E. Cardiac Ca²⁺ Signaling in Diverse Diseases

Complex diseases may affect heart function and, in many cases, the links between the primary disease and the myocyte remain obscure. Diabetes is associated with Ca²⁺ signaling dysfunction including abnormal Ca²⁺ sparks and Ca²⁺ release (75, 94, 329, 341, 429, 430). Dystrophin mutations account for Becker and Duchenne muscular dystrophy, the two most common muscular dystrophies. While both are primarily diseases of skeletal muscle, each has severe cardiac consequences; among other things in mouse models there appear to be abnormal Ca²⁺ signaling in heart (187) and altered mitochondrial function. In each of these cases, the molecular causes of the cardiac signaling dysfunction are being investigated.

	XIII. SPARKS AND EMBERS IN SKELETAL MUSCLES

Top Previous Next References

 
A. Overview

In a vertebrate skeletal muscle couplon, type 1 RyRs in a CRU form a double row assembly; tetrads of DHPRs in the plasma membrane overlay and form conformational couplings with RyR homotetramers, but only at alternate RyRs (129) (Fig. 8). In this "double checkerboard" arrangement, the RyRs unconnected to DHPRs are apparently in direct contact with those that are engaged with DHPRs. In the skeletal muscle of frog and other nonmammalian species, there are also parajunctional type 3 RyR stripes along the jSR that run parallel on either side of the double RyR1 swath (Fig. 8) (117). The presence of such RyR3 tetramers or the lack of them confers distinctive phenotypes on the respective muscles: sparks in amphibians and embers in mammals (197, 300, 381) (Fig. 12). Investigation of the lack of sparks, the presence of embers, and the reappearance of "ghost sparks" (394) (Table 1) in mammalian skeletal muscle as well as the spark mechanisms in amphibians has led to the appreciation of two new mechanisms controlling the skeletal muscle Ca²⁺ release from store. The first is related to an inhibitory role of DHPR on RyR1 and the second, the requirement of parajunctional RyR3 in supporting depolarization-evoked sparks. Next, we first discuss the activation, coordination, and termination of skeletal muscle sparks with reference to their counterparts in heart and then sparkless release, embers, and the conditions for reappearance of sparks in mammalian skeletal muscles.

View larger version (77K): [in this window] [in a new window]   FIG. 12. Sparks and embers in amphibian and mammalian skeletal muscle fibers, respectively. A, left: x-y scan image from a frog semitendinosus muscle at rest, voltage-clamped in a 2-Vaseline gap chamber, and equilibrated with an internal solution containing fluo 3. Right: line scan (along the line in blue on the left); a depolarizing pulse was applied as indicated. The image is pseudo-colored for visibility. B: line scan image of a rat extensor digitorum longus (EDL) fiber, voltage-clamped in a 2-Vaseline gap clamp, and similarly equilibrated with a fluo 3-containing internal solution. The voltage protocol starts from a depolarized situation and includes an interval of 2.8 s at –90 mV. This allows for partial recovery of the voltage-inactivated voltage sensors, which results in a reduced density of active release channels and improved resolution of individual embers. [Modified from published and unpublished work of E. Rios, J. Zhou, N. Shirokova, and L. Csernoch.]

1. Activation mechanisms

First suggested on the basis of noise analysis, spontaneous Ca²⁺ sparks and those induced by depolarization or RyR ligands have been extensively characterized in amphibian skeletal fibers. In permeabilized or cut-end preparations, skeletal muscle sparks are wider (2–4 µm) but briefer (3–5 ms rise-time, 15 ms FDHM) than cardiac sparks (197). They are localized to regions where couplons or TT-CRU junctions are expected (197, 381). Reducing [Mg²⁺] (from 2 to 0.1 mM) profoundly enhances the rate of spark production (197, 205, 442), suggesting an inhibitory role of intracellular Mg²⁺ at physiological concentrations. Interestingly, caffeine not only increases the spark frequency, but also makes sparks wider, which was interpreted as recruiting more channels into the spark (16, 142). Baylor et al. (16) carefully examined important morphometric differences between sparks from the intact fiber preparation (with indicator loaded by microinjection) and the cut-open preparations: sparks in intact fibers are briefer (FDHM = 6 vs. 9–15 ms), narrower (FWHM = 1.0 vs. 2.0 µm), and 3- to 10-fold smaller by signal mass. These differences suggest that diffusible cellular constituents and exact ionic conditions are important in determining the CRU gating, the I_spark, and the appearance of sparks.

Dual mechanisms, DICR and CICR, control depolarization-evoked release, which does not require any Ca²⁺ entry and is largely made up of sparks (141, 197, 199) (Fig. 12). Spark activation at around threshold voltages displays an e-fold increase of rate occurrence every 4.1 mV of depolarization, similar to the voltage-dependent activation of DHPR, the voltage sensor. By brief repolarization to reprime inactivated voltage sensors in depolarized frog fibers, Klein et al. (199) resolved sparks over the full range of skeletal muscle fiber depolarization (from –60 to +40 mV) and demonstrated autonomy of Ca²⁺ spark formation irrespective of voltage and latency.

2. Coordination mechanisms

As is the case for cardiac sparks, the estimates for the number of RyRs activated in a skeletal muscle spark cover a wide range. In one school of thought, the "one or a few channel model" explains spark production, because features of intact muscle sparks, including spark width, can be reproduced by model simulation with a point source I_spark of 1–3 pA (63). The time course of Ca²⁺ sparks in amphibian skeletal muscle has been meticulously examined with video-rate (63 µs/line) confocal microscopy (206). The fluorescence time course is systematically well fit by an empirical equation consisting of an exponential rising phase followed by an exponential falling phase. This feature is consistent with an immediate and full activation of the release flux at the start of a spark, constant throughout the rising phase, and sudden and complete cessation of the release at the peak of the spark. Since spark formation involves low-pass filtering components (Ca²⁺ buffering, indicator kinetics, diffusion), the optical measurement cannot resolve brief closures during an ongoing spark. Hence, such a square wavelike flux can be generated by sustained or bursting opening of either a single channel or a channel group (321).

In the competing school of thought, amphibian skeletal muscle sparks arise from up to a 30 pA release flux (based on model extraction from the brightest sparks), requiring the activation of as many as 60 RyRs (306). The estimated release flux underlying an ember induced by Imperatoxin A is merely 1/8 to 1/20 of the estimated peak I_spark. The center of an ember may be off-set from the center of its precedent spark by a fraction of a micron, indicative of an extended source of I_spark. Moreover, fast 2D imaging visualizes a wavelike evolution of a spark: the wavefront propagates over 1–2 µm at a speed of 100 µm/s, as if the spark is generated by CICR along the direction of an elongated skeletal CRU (366, 441). Caffeine increases the propensity and propagation speed of the wavelike sparks. Importantly, the appearance of wavelike sparks also suggests that RyRs in a skeletal muscle CRU do not gate in unison, arguing against a tight CICR or coupled gating under the experimental conditions.

An amphibian couplon consists of the DHPR voltage sensor, the junctional RyR1 array, and parajunctional RyR3 arrays on the same cistern (Fig. 8). During a depolarization-evoked spark, junctional RyR1 may serve as the master to control the activation as well as termination of the slave RyR3 arrays (206, 300). The mechanism by which these RyR1 and RyR3 arrays communicate with each other presumably involves CICR on the cytosolic side and, perhaps, Ca²⁺-dependent regulation in the lumen.

3. Termination mechanisms

Skeletal muscle sparks are more abruptly terminated than their cardiac counterparts and exhibit an average rise time of 5 ms. As in cardiac cells, use-dependent refractoriness of spark production has been demonstrated by several groups (197, 198, 214). Only 7% cisternal depletion (213) or 10% global store depletion (286) is associated with a spark or a muscle twitch, respectively, in amphibian skeletal muscle, suggesting a limited contribution of cisternal depletion to spark termination. Compared with RyR2, RyR1 is readily inactivated by a Ca²⁺-dependent mechanism, and its P_o displays a bell-shaped Ca²⁺ dependence with profound inactivation at millimolar [Ca²⁺] (256, 423). However, RyR3 in the parajunctional arrays is known to be resistant to inactivation even in the presence of 10 mM [Ca²⁺] (238).

A novel termination mechanism, "termination by voltage sensor deactivation," has been demonstrated for depolarization-evoked sparks (Table 6). Lacampagne et al. (206) have shown that DHPR deactivation on repolarization curtails an ongoing spark, violating spark autonomy. They suggested that continued DHPR activation is essential to sustain the evolution of a spark in the first few milliseconds of spark generation. Beyond this time period, the release inactivation mechanism would terminate the spark independently of DHPR. This finding is of great importance for several reasons. First, it shows that, in contrast to the cardiac counterpart, a skeletal spark is, to some extent, still under the control of the DHPR during its evolution; the autonomy of skeletal muscle sparks is not absolute. Second, DHPR deactivation terminates CICR, if any CICR occurs between junctional (RyR1) and parajunctional channels (RyR3) and within the junctional array. If we envisage the immediate effect of DHPR deactivation is the cessation of Ca²⁺ flux from the DHPR-coupled RyR1, it can be inferred that CICR in a spark operates in a low-gain regime (G < 1) and is unable to support regenerative release after removal of the trigger Ca²⁺. Third, DHPR may repress RyR1 activity at rest, serving to inhibit SR Ca²⁺ leakage at rest. This idea is now supported by more direct evidence from mammalian muscle (443). Hence, DHPR in skeletal muscle functions not only as the voltage sensor and the Ca²⁺ entry pathway, but also the release terminator and the SR leak repressor.

C. Embers and Sparks in Mammals

1. Embers and ghost sparks

The quest for elementary release events in adult mammalian species first met with a surprise: the almost absolute absence of spontaneous sparks in intact quiescent adult fibers (332) and, during depolarization, the lack of evoked Ca²⁺ sparks (335). Subsequent studies under certain experimental conditions (e.g., inclusion of in the internal solution) (91, 440) revealed that depolarization-evoked release consists of lone embers: long-duration, low-amplitude (F/F₀ = 0.2), low flux (0.5 pA) release events (Fig. 12). These lone embers are narrower (1.3 µm) than sparks and lack the initial peaks seen in spark-trailing embers in amphibian skeletal muscle in the presence of RyR modulators. A true irony, however, is that mammalian skeletal muscle does generate spark-featured releases under certain circumstances: robust spark production after saponin permeabilization of the surface membrane or in the presence of in the internal solution in cut-end preparations (440), or locally in membrane-permeabilized ends (91). Moreover, typical sparks and long-lasting Ca²⁺ "bursts" (401) (Table 1) appear in intact rat skeletal muscle fibers upon exposure to hypertonic solution, or returning from hypotonic swelling, or after strenuous excerise of the animal (393). A higher propensity of Ca²⁺ spark activity has also been associated with aging skeletal muscle (258, 400, 402). Thus the same Ca²⁺ release machinery is capable of producing no sparks (in quiescent fibers), sparks, or another type of elemental release event, in stark contrast to the amphibian and cardiac counterparts. These results suggest that mammalian skeletal muscle does contain the spark-generating machinery, but it is normally repressed to manifest sparkless release; experimental and pathophysiological conditions can unmask these "ghost sparks" (394).

2. Inhibitory role of conformational coupling between DHPRs and RyRs

While the traditional view is that DHPRs serve solely as the voltage sensor to their RyR1 partners in physical contact, the current understanding is that DHPRs, when not activated, prevent their RyR1 partners from activation. The first clue was from the observation that voltage sensor deactivation abbreviates an amphibian spark as discussed above. More recently, it has been shown that repolarization truncates ongoing embers in rat skeletal muscle (440). These findings indicate that DHPR activation and deactivation tightly control the commencement and termination of CRU release. The observation of lone embers in mammalian skeletal fibers provides a hint. If the embers each arise from a single RyR, their mere presence implies little or no CICR (or coupled gating) among RyR1 channels in the same junctional array. The reappearance of sparks after chemical permeabilization, exercise, or osmotic shock from otherwise sparkless release is generally consistent with this idea, as if subtle membrane deformation disrupts DHPR inhibition of CICR. In immature (85, 86, 336), differentiating (443), or dedifferentiating (50) mouse skeletal muscle fibers, depolarization evokes a sparkless type of release at spatially segregated sites, whereas discrete sparks occur at locations devoid of direct EC coupling (and hence DHPR inhibition). Additional evidence comes from investigation of Ca²⁺ release in myotubes from dysgenic mice (mdg) that lack DHPRs but contain both RyR1 and RyR3 isoforms. By comparing mdg with the wild type, Zhou et al. (443) have shown that DHPRs reduce the rate of spontaneous opening of the release channels and their susceptibility to activation by high Ca²⁺ and 1 mM caffeine. Thus DICR, DHPR-dependent spark repression, and spark termination by voltage sensor deactivation appear to be different manifestations of the same underlying mechanism (see Ref. 394 for a review). Physiologically, direct voltage control of both release activation and termination affords tight control of Ca²⁺ signaling by brief APs (of millisecond duration) and, in quiescence, strong inhibition of background SR Ca²⁺ leak in skeletal muscles. A disruption of such an inhibitory mechanism or reappearance of ghost sparks is associated with skeletal muscle aging and dystrophy (393, 400).

3. Role of the RyR3 isoform

The striking species difference between amphibians and mammals in spark production has been linked to their respective CRU organization (300, 335). Only in the amphibians are there parajunctional sets of RyRs (presumably RyR3) not directly controlled by the DHPRs (117); CICR activation of accessory RyR3 is thought to account for the appearance of spontaneous sparks when junctional RyR1 channels are normally inhibited by DHPRs. This hypothesis is supported by the observation that tetracaine eliminates the depolarization-evoked sparks but spares the smaller, diffuse, and sparkless release in amphibian skeletal muscle fibers (335). In an elegant study, Pouvreau et al. (300) used electrotransfection of cDNA to express either RyR3 or RyR1 in mouse skeletal muscle and showed that exogenous expression of RyR3, but not RyR1, enables spontaneous sparks and macrosparks at rest and produces abundant sparks upon depolarization. It has been concluded that depolarization-evoked sparks in skeletal muscle require a voltage sensor, a master junctional RyR1 channel that provides the trigger Ca²⁺, and a slave parajunctional RyR3 cohort (300). Conversely, the lack of embers in amphibians and at RyR3-expressing mammalian couplons suggests that the parajunctional RyR3 cohort may enslave the junctional RyR1 for release termination, via perhaps depletion of the common cisternal Ca²⁺ pool.

	XIV. SMOOTH MUSCLE SPARKS

Top Previous Next References

 
As in striated muscle, Ca²⁺ sparks are present in a wide variety of smooth muscle myocytes including those from artery (133, 262, 274), portal vein (144, 263), urinary bladder (84, 163, 164, 183, 266), ureter (42, 52), airway (203, 235, 339), and gastrointestinal tract (145, 405). They occur spontaneously and can be evoked by depolarization, either directly by CICR or indirectly by SCICR due to SR loading by Ca²⁺ influx. Instead of an extensive review of spark mechanisms, here we focus on physiological functions consequential to Ca²⁺ spark activation in smooth muscles. In particular, we illustrate how the high local [Ca²⁺] produced by sparks activates plasma membrane Ca²⁺-sensitive channels or electrogenic transporters to impart their physiological actions.

A. Activation of Ca²⁺-Sensitive Channels by Subsurface Sparks

SR elements in close apposition (<20 nm) to the inner side of the plasma membrane have been observed in different types of smooth muscle myocytes (102, 223) with RyRs within 100 nm of the cell membrane (139, 223). A subsurface Ca²⁺ spark, which measures 2–3 µm in diameter and occupies 1% of the cell volume, can affect 1% of the cell membrane (293), on which reside the Ca²⁺-sensitive channels, including big-conductance Ca²⁺-activated K⁺ (BK_Ca) and Ca²⁺-activated Cl^– channels (Cl_Ca), and electrogenic Ca²⁺ transporters such as the NCX.

Spontaneous outward currents (STOCs) mediated by BK_Ca channels were first described by Benham and Bolton (19) and were thought to arise from sudden discharges of the subsurface SR Ca²⁺ store. In a classic work, Nelson et al. (274) showed that subsurface ryanodine-sensitive sparks activate iberiotoxin- and TEA-sensitive STOCs in arterial smooth muscle. Simultaneous recording of STOCs and sparks suggests that virtually every spark activates a STOC (293), indicating a near unity fidelity for spark-to-STOC coupling ( 1). In rat cerebral arterial myocytes, electrophysiological measurement of single-channel and whole cell BK_Ca channel activity yields an estimate of 3,000 BK_Ca channels/myocyte or 2 channels/µm², and an average STOC amplitude of 148 pA and unitary BK_Ca current of –8.5 pA, at –40 pA and symmetric 140 mM K⁺ concentration (292). From this, it is estimated that 18 BK_Ca channels open simultaneously at the peak of a STOC. Furthermore, by determining the Ca²⁺ dependence of BK_Ca channels in excised patches (K_d = 19 µM, Hill coefficient = 2.9 at –40 mV), the lowest estimate for local [Ca²⁺] to trigger BK_Ca channels would be 4 µM (assuming 100% channels clustered over the spark) and the highest [Ca²⁺]_i estimate 30 µM (assuming uniform distribution or 1% of channels affected by the spark) (292), similar to the [Ca²⁺] expected to be found in the subspace of a cardiac couplon (350). Spark-to-STOC coupling appears to be extremely robust. Loading the myocytes with a fast Ca²⁺ chelator, BAPTA, causes the disappearance of Ca²⁺ sparks, but the STOCs remain unchanged (292), suggesting that the fast Ca²⁺ chelator is unable to break the coupling, and the underlying Ca²⁺ signals that mediate the coupling must be too local to be detected by current optical methods. In addition to the LCC-to-RyR coupling, RyR-to-BK_Ca coupling provides another good example of high, local [Ca²⁺] being sensed by a local target.

Such stringent local control of Ca²⁺ signaling would predict that BK_Ca activity can discriminate Ca²⁺ signals from different sources and, conversely, sparks can differentially affect different Ca²⁺-sensitive channels, depending on relative localization between the source and the target. Indeed, it has been shown that, in urinary bladder myocytes, blockade of RyR release inhibits iberiotoxin- and TEA-sensitive BK_Ca currents without affecting the Ca²⁺-activated small-conductance K⁺ channel (SK_Ca) currents that are apamin sensitive. This suggests that SK_Ca channels are somewhat excluded from the microdomains of RyR CRUs. In contrast, Ca²⁺ entry through voltage-dependent Ca²⁺ channels (VDCCs) activates both BK_Ca and SK_Ca channels, but with distinctive kinetics that suggest VDCCs are more distant from BK_Ca than SK_Ca channels (164).

In addition to STOCs, sparks can activate Cl_Ca channels to produce transient inward currents (STICs) (Fig. 13) (166). When BK_Ca and Cl_Ca channels coexist, spontaneous transient outward then inward currents (STOICs) are recorded (Fig. 12) (446). These STOCs and STICs are expected to modulate V_m and excitability, with STOCs being hyperpolarizing and inhibitory and STICs being depolarizing and excitatory. In addition, the increase in membrane conductance (G_m) negatively influences membrane excitability (see below). The spark-STOC and spark-STIC couplings, akin to sparklet-spark coupling, provide a real-time demonstration of intermolecular signaling events in intact cells.

View larger version (30K): [in this window] [in a new window]   FIG. 13. Spark modulation of membrane currents. Depending on membrane potential, three sparks from the same site (A: surface plots; B: time courses) in a tracheal smooth muscle myocyte evoked a BKCa current (STOC), a combined BKCa and ClCa current (STOIC), and a ClCa current (STIC) (C) at 0, –50, and –80 mV, respectively (446).

An increase in [Ca²⁺]_i in arterial smooth muscle myocytes is expected to initiate muscle contraction, as is the case in striated muscles. Counterintuitively, Ca²⁺ sparks relax smooth muscle. Nelson et al. (274) have shown that the underlying mechanism involves, sequentially, subsurface Ca²⁺ sparks, STOCs, membrane hyperpolarization, resulting shutoff of LCC Ca²⁺ entry, and a loss of intracellular Ca²⁺ (Fig. 14). Specifically, the direct contribution of sparks to global [Ca²⁺]_i is almost inconsequential: a spark would raise the global [Ca²⁺] by <2 nM if its Ca²⁺ flux is uniformly distributed in the cytosol. The rate of occurrence of spontaneous sparks is only 1 Hz next to the cell membrane, and thus the time-averaged change in global [Ca²⁺]_i is even smaller. On the other hand, the spark-activated STOCs would cause a time-averaged hyperpolarizing current of the order of 10 pA and a >10 mV hyperpolarization assuming a few G input resistance of the cell, leading to reduction of VDCC-mediated Ca²⁺ entry. Myocyte relaxation and vasodilatation ensue as a result of the net reduction in [Ca²⁺]_i (Fig. 14). That sparks cause arterial myocyte relaxation vividly exemplifies how the same Ca²⁺ messenger fulfils opposing physiological functions in a given cell.

View larger version (21K): [in this window] [in a new window]   FIG. 14. How do sparks relax arterial smooth muscle and regulate vascular tone? A: subsurface Ca2+ sparks promote membrane hyperpolarization through the activation of BKCa and STOCs in arterial smooth muscle myocytes. Hyperpolarization tends to shut off Ca2+ influx via the voltage-operated channels (VOC), resulting in a net decrease of intracellular Ca2+ and relaxation of the arterial smooth muscle. B: in intact vessels, increases in intravascular pressure lead to membrane depolarization, activation of voltage-dependent Ca2+ channels, and increased arterial tone. The spark-STOC coupling thus acts as a negative-feedback mechanism to limit pressure-induced arterial tone. [Modified from Nelson et al. (274).]

C. Spark Modulation of Membrane Excitability

While arterial smooth muscle tone is regulated by varying V_m, APs are generated in smooth muscles that undergo phasic contraction. In this type of smooth muscle, oscillatory spark production may serve as a "Ca²⁺ clock" (208, 391) (Table 3) that is coupled to the electrophysiological clock made up of pacemaker ionic currents, and collectively, they set the rhythm of the firing of APs. For instance, phasic spark production has been shown to regulate the AP refractory period for the peristalsis of ureter smooth muscle, and is associated with an exceptionally long refractory period (10–100 s) (42, 52). Ca²⁺ entry via LCC during an AP loads up the SR and increases the spark activity, perhaps via the SCICR mechanism. Spark activity gradually fades away in 70 s as the SR gradually depletes. Blockade of spark activation or BK_Ca channel activity causes premature firing of the next AP. Intriguingly, the spark modulation of membrane excitability occurs with no persistent hyperpolarization (52), in contrast to arterial smooth muscle tone regulation. Possible explanations for the lack of significant hyperpolarization include 1) the resting V_m is closer to the reversal potential of K⁺ currents in phasic than tonic smooth muscles, limiting the amplitude of STOCs; 2) the membrane input resistance might be lower (higher conductance) such that STOCs are less effective in hyperpolarizing the cell; and 3) there might be Ca²⁺-activated inward currents, such as STICs, that nullify the outward currents.

Despite the lack of a direct V_m-hyperpolarizing effect, BK_Ca channel activation nonetheless increases in G_m and thus makes the membrane difficult to depolarize (i.e., smaller voltage excursion in response to a given current), conferring the refractoriness. If this interpretation is correct, the spark modulation of ureter smooth muscle refractoriness employs a novel "V_m-clamping" rather than a V_m-hyperpolarizing mechanism. In summary, the firing of an AP sets the Ca²⁺ clock consisting of the SR, the RyR, and the BK_Ca channel modules; the Ca²⁺ clock then slows the ticking of the electrophysiological clock until it itself times out; the electrophysiological clock then catches up with the pace and sets off the next AP. Then, the next cycle begins (Table 3).

	XV. NEURAL Ca2+ SPARKS

Top Previous Next References

 
A. Characteristics and Mechanisms

An elaborate ER system is found in the cell body of many types of neurons, extending into the shanks of dendrites and axons and perhaps to dendritic spines and presynaptic terminals (23). The possibility of the existence of Ca²⁺ sparks as the elemental release events in neurons was first suggested on the basis of excessive "noise" or "outlier" bright pixels in neuronal Ca²⁺ images in the absence of discernible local events (257). Meanwhile, Koizumi et al. (202) demonstrated elementary Ca²⁺ release events arising from both RyRs and IP₃Rs in nerve growth factor-differentiated PC12 cells and cultured hippocampal neurons. These neuronal sparks exhibit 2-fold greater spatial width and at least 20-fold longer duration than a typical cardiac spark. Later, Llano et al. (236) visualized RyR-mediated spontaneous Ca²⁺ release at presynaptic terminals, where individual events last for a few seconds and spread 5–10 µm along the processes. It is unknown whether these are neuronal sparks from single CRUs or complex supraspark events involving multiple CRUs. More recently, neuronal sparks termed "syntillas" (95) (Table 1) have been characterized in hypothalamic neural terminals: these can be activated via the DICR mechanism. Another unequivocal demonstration of neuronal Ca²⁺ sparks is in mammalian DRG sensory neurons (Fig. 15), where sparks from RyR3 on the subsurface cisternae are activated by physiological Ca²⁺ influxes or the RyR sensitizers caffeine and DMPX (283, 284). Highly localized (0.66 µm width) 10-s-long ryanodine-sensitive "Ca²⁺ glows" (427) (Table 1), as well as typical Ca²⁺ sparks, have also been observed in superior cervical ganglion neurons in primary culture.

View larger version (36K): [in this window] [in a new window]   FIG. 15. Subsurface sparks and their possible roles in neurons. A: immunostaining shows subsurface localization of RyR3 puncta in a DRG sensory neuron from rat. "N" marks the nucleus. B: subsurface localization of sparks. C: possible functions of subsurface sparks in neurons. RyR/IP3R sparks from the ER cisterns, which can be activated by Ca2+ entry through the voltage-operated Ca2+ channels (VOCs), could modulate exocytosis and membrane excitability by activating BCa, ClCa channels, and NCX currents in the plasma membrane (PM). [Data from Ouyang et al. (284).]

B. Properties of CICR in DRG Neurons

Virtually all CRUs in DRG cells manifest repetitive spark activity (283), indicating a lack of refractoriness. Though unexpected, this is consistent with the planar lipid bilayer observation that RyR3 from DRGs displays no Ca²⁺-dependent inactivation even with 10 mM [Ca²⁺] on the cytosolic side. Yet, it was noted that the DRG CICR system is unable to support spontaneous propagating Ca²⁺ waves under various experimental conditions. Quantification of cellular [Ca²⁺]_i transients revealed that CICR in DRG neurons (283) appears to operate in a low-gain, linear regime (G = 0.54). In this regard, Stern (363) showed mathematically that a CICR system with G <1 is intrinsically stable in spite of positive feedback. These results illustrate that CICR in DRG neurons operates in a regime of unconditional stability. Through a low-gain CICR system, DRG neurons can still generate large intracellular [Ca²⁺]_i transients because the surface Ca²⁺ current density (156 pA/pF at –10 mV) is exceptionally high compared with that in cardiac cells (Table 9). The striking differences as well as the similarities between neural and cardiac CICR (Table 9) are instructive as to how the same signaling mechanism can be adaptive and plastic in different physiological contexts.

The role of sparks is less well understood in neurons than muscles. The presence of extensive subsurface cisternae (SSC) implies that, among other possibilities, neural sparks are instrumental in mediating bidirectional communications between the plasma membrane and the ER. The forward mode of communication, i.e., spark activation by both CICR and DICR mechanisms has been demonstrated by independent groups. In the retrograde signaling mode, spark modulation of membrane excitability, analogous to that in smooth muscle cells and cardiac pacemaker cells, represents an appealing possibility that awaits future investigation (Fig. 15).

Ca²⁺-dependent vesicle secretion of transmitters and peptides is widely involved in neural functions. Most neurons release their neurotransmitters from terminals at the synapse in a Ca²⁺-dependent manner. However, small DRG neurons (15–25 µm in diameter, C type) are capable of Ca²⁺-dependent somatic exocytosis of, among others, the pain-related peptides, calcitonin gene-related peptide and substance P (12, 170, 438). Emerging evidence suggests that somatic vesicle release might be common to central nervous system neurons (169). In this regard, Ca²⁺ sparks from SSCs could elevate local [Ca²⁺] to trigger exocytosis. Since neural secretion is a superlinear function of [Ca²⁺] (77), sparks can also modulate secretion through synergism with depolarization-induced Ca²⁺ influx. Both possibilities are supported by data from simultaneous measurement of subsurface Ca²⁺ sparks and membrane capacitance (C_m) increases (as an index of vesicular secretion). Ouyang et al. (283) found that caffeine- or DMPX-induced Ca²⁺ sparks increase C_m, independently of membrane depolarization and external Ca²⁺. Inhibition of spark production by ryanodine suppresses C_m changes and abolishes endotoxin-induced secretion of pain-related neuropeptides. By signal averaging, a tiny C_m change associated with individual sparks has been discerned, and the spark-secretion coupling fidelity in DRG neurons is 1 vesicle/10 sparks ( 0.1). At present, it remains an open question whether spark-secretion coupling is generally involved in mediating synaptic release (see Ref. 140 for a review of the possible role of local Ca²⁺ release in the regulation of axon pathfinding).

	XVI. IP3R Ca2+ PUFFS AND BLIPS

Top Previous Next References

 
The cell logic is more complex for IP₃R than for RyR-mediated Ca²⁺ signaling, because the former requires both IP₃ and Ca²⁺ as coagonists (21, 125). A useful view point is that IP₃ serves as a physiological ligand to tune the Ca²⁺-dependent inhibition of the IP₃R release channel (246). IP₃Rs are ubiquitously distributed and are the common effector of many GPCRs and tyrosine kinase receptors that are linked to the activation of phospholipases and the production of the second messengers IP₃ and DAG from PIP₂. It is generally believed that IP₃Rs also form clusters or CRUs (289, 338), but direct ultrastructural evidence is yet to be seen.

Ca²⁺ puffs and mixed IP₃R-RyR sparks have been reported in many types of cells, including Xenopus oocytes (287a, 289), HeLa cells (40, 156), oligodendrocytes (156), neonatal cardiac myocytes (244, 299), and adult atrial cells (448), where they fulfill cell-specific functions. The Xenopus oocyte, which contains exclusively type 1 IP₃R, has been a model system for investigating IP₃R signaling. Immunocytochemical data visualize a dense distribution of IP₃R to a subsurface band 5 µm thick, with the outermost layer separated by 6 µm from the surface membrane (54). A distinctive feature of IP₃R CRU operation is that it can generate both puffs and smaller release events ("blips") (Table 1) in response to varying IP₃-uncaging light intensity (53, 310), i.e., gradation of Ca²⁺ release at the level of single CRUs. We speculate that the concentration of IP₃ determines the effective size of a CRU, i.e., the number of IP₃-occupied channels for CICR. A fully developed puff exhibits a mean intensity of 1.5 (F/F₀ of Oregon Green 488 BAPTA-1) and 3.5 the brightest, a long duration (100–600 ms), and a large spatial spread (up to 6 µm). The properties of puffs can be explained by the synchronous opening of 25–35 IP₃Rs, each carrying a Ca²⁺ current of 0.4 pA, with the channels distributed through a cluster 300–800 nm in diameter (310). As for the puff termination mechanism, it is well-documented that Ca²⁺ exerts a biphasic action on single IP₃R P_o. In type 1 IP₃Rs from a Xenopus oocyte, Ca²⁺-dependent channel activation occurs in an IP₃-independent manner, with a half-activation concentration of 210 nM. Ca²⁺-dependent inhibition, however, is tuned by IP₃, with the half inhibition [Ca²⁺] varying from 160 nM to 42 µM as IP₃ concentration decreases from 180 µM to 10 nM (246). Annihilation of colliding Ca²⁺ waves indicates that the cytoplasm as an excitable medium becomes refractory in the wake of recent excitation (218). Because of the refractoriness and other nonlinear behavior of CICR, complex spatial and temporal patterns emerge, elegantly manifested as spiral Ca²⁺ waves (218). In Xenopus oocytes, the Cl_Ca current displays a steep IP₃ concentration dependence. Due to the 10 µm spatial separation, the surface membrane Cl_Ca channels can hardly be affected by rogue puffs but can be effectively activated by coordinated puff activation in the form of local or global Ca²⁺ waves traversing over the subsurface IP₃R band (289).

An intriguing feature of IP₃R signaling, discovered in the early 1990s, is the so-called "quantal release" (120, 264, 380): cells respond to an IP₃ signal with a [Ca²⁺]_i transient that decays almost completely to the resting Ca²⁺ level; however, a subsequent greater IP₃ signal can still elicit yet another [Ca²⁺]_i transient of similar amplitude. By spike measurements, quantal release has recently been demonstrated at the single CRU level (55), indicating that it is a property intrinsic to the CRU. At present, it is unknown whether the quantal release feature requires channel clustering or is a feature of single-channel adaptation (see discussion of RyR adaptation). Luminal Ca²⁺ regulation of IP₃R sensitivity to cytosolic Ca²⁺ and IP₃ has also been suggested as a candidate mechanism underlying the quantal Ca²⁺ release.

All three types of IP₃Rs are coexpressed, with type 2 IP₃R being predominant in developing heart (14, 229, 267). In mammalian adult ventricular myocytes, IP₃Rs are expressed at a low level and are thought to be concentrated at the perinuclear region and are found at the Z-line. These IP₃Rs may be inconsequential to cardiac EC coupling but may play an important role in transcriptional regulation (13, 419), evidenced by IP₃R-mediated perinuclear sparks and waves (244). However, there is no convincing demonstration of intranuclear IP₃R-mediated Ca²⁺ signals. In atrial cells, IP₃Rs comingle with RyRs to mediate subsurface Ca²⁺ sparks and are thus implicated in cardiac arrhythmogenesis (448). The expression of IP₃Rs is generally elevated in various heart conditions (137, 424), suggesting possible roles in cardiac functional and structural remodeling.

	XVII. SPARKLESS RELEASE

Top Previous Next References

 
Despite the fact that sparks are ubiquitous in different tissues and cells, a sparkless type of release is associated with ectopic expression of cardiac or skeletal RyRs in Chinese hamster ovary cells (32, 33), whereas expression of RyR1 or RyR3 in myotubes null for native RyRs produces sparks (395, 396). Even in heart cells where Ca²⁺ sparks predominate, sparkless release has also been documented. NCX in the reverse mode (extruding intracellular Na⁺ in exchange for extracellular Ca²⁺) triggers release that is far more uniform than expected if sparks were the elementary release units. Furthermore, Niggli and Lipp initially reported sparkless releases upon Ca²⁺ steps produced by ultraviolet photolysis in cardiac myocytes (232), but later revised their conclusion by showing variable local release events triggered by two-photon focal photorelease of caged Ca²⁺ (230). In oocytes, a small, spatially uniform Ca²⁺ signal immediately upon photolytic production of IP₃, termed "pacemaker Ca²⁺," has been shown to precede the occurrence of IP₃R blips and puffs (247). Sparkless and subspark release events have also been systematically investigated in adult mammalian skeletal muscles (see above).

Sparkless release indicates that the underlying elementary release events involve quantities of Ca²⁺ that are beyond the detection limit in these cases, suggesting that Ca²⁺ sparks do not account for the totality of ER/SR release under all circumstances. The coexistence of sparks and sparkless release in heart cells is difficult to reconcile if both are based on the same CRUs. To this end, Sobie et al. (347) speculated that there are solitary, yet-to-be-visualized RyRs scattered over the ER/SR, or "rogue RyRs," either in transit toward the junctions or by biological design. The rogue RyRs might be more sensitive to low levels of Ca²⁺ but less sensitive to high levels of Ca²⁺ than arrayed RyRs, the latter displaying a steeper, highly cooperative Ca²⁺-activation curve (Fig. 9). Besides gating kinetics, the connected store capacity and molecular partners on both cytosolic and luminal sides may also differ for rogue and CRU RyRs. Release via rogue channels, rather than CRUs, might also account for the sparkless release in the ectopic expression system and for the pacemaker Ca²⁺ signal of IP₃-induced sparks.

	XVIII. NEW INSIGHTS INTO Ca2+ SIGNALING

Top Previous Next References

 
A. Mechanisms Underlying Excitation-Ca²⁺ Release Coupling

Distinctive mechanisms are used in different types of cells to trigger store Ca²⁺ release upon membrane depolarization. As discussed above, direct coupling in skeletal muscle as well as in some types of neurons involves conformational interaction between the voltage sensor and the release channel, independent of Ca²⁺ entry. Tight CICR coupling in cardiac as well as invertebrate skeletal muscles operates in the nanoscopic subspace of couplons. In contrast, many types of smooth muscle (e.g., urinary bladder and arterial smooth muscles) display an "indirect or loose coupling" between membrane excitation and Ca²⁺ release (84, 113, 164). Features of the indirect or loose intermolecular communication between LCCs and RyRs include the following: 1) triggered Ca²⁺ sparks are often observed long after LCC closure; 2) CICR is a function of total LCC Ca²⁺ influx, rather than i_LCC; and 3) CICR is sensitive to cytosolic exogenous Ca²⁺ buffers. It may also involve an SR Ca²⁺ reuptake and SCICR component. In this regard, Essin et al. (113) recently demonstrated that disruption of LCC expression in arterial smooth muscle cells leads to fewer STOCs of reduced amplitude, accompanied by reduced cytosolic and SR Ca²⁺. This effect is fully reversed after restoration of the global and SR Ca²⁺ levels.

Both direct and tight couplings can provide the speed and stability for the mobilization of stored Ca²⁺. One particular advantage of direct coupling is perhaps related to its ability to inhibit CICR when the muscle is not activated (443), which is of obvious advantage in bioenergetics, noise reduction, and prevention of false firing. Furthermore, direct coupling may also be a requisite for sharp termination of the release flux when an AP lasts merely a few milliseconds (206). In heart, however, the AP is of much longer duration (hundreds of milliseconds), so tight CICR coupling appears to suit the purpose better to release Ca²⁺ in systole and then completely terminate it in diastole, the latter being crucial to cardiac relaxation for refilling. In the situation of smooth muscles with tonic depolarization and very long AP duration, signal sustainability might dictate the choice of possible mechanisms that couple Ca²⁺ influx to store release. CICR via global, rather than, local [Ca²⁺]_i transients would greatly prolong the signal duration. If this is coupled with store Ca²⁺ uptake and SCICR, the Ca²⁺ release signal would last even longer. In this way, distinctive coupling mechanisms may fulfill cell type-specific functions.

B. Ca²⁺ Sparks and the CICR Paradox

CICR as the ubiquitous control mechanism of ER/SR Ca²⁺ release was once thought to be paradoxical. With Ca²⁺ as both the input and the output, CICR represents a positive feedback, ensuring the needed sensitivity, speed, and amplification of Ca²⁺ signals. However, uncontrolled CICR is expected to operate in an all-or-none fashion, unless the gain of CICR amplification is sufficiently low (G < 1) (279, 363). In contrast, cardiac EC coupling as a high-gain CICR system is characterized by a smoothly graded response to the trigger I_Ca in terms of amplitude and duration.

To solve this CICR paradox, early experimental and theoretical work invoked the "local control theory" of CICR (Fig. 16) (279, 363). This theory consists of four postulates, each of which has now been independently validated by findings stemming from sparkology.

View larger version (29K): [in this window] [in a new window]   FIG. 16. Sparks and the CICR paradox. A: in a common pool model, high-gain CICR is expected to be explosive, operating in an all-or-none fashion. B: the spark model (or local control model) is able to support high-gain CICR within a CRU, while CICR between and among adjacent CRUs is abrogated. C: when the Ca2+ signaling system is perturbed and made unstable (e.g., with increased SR Ca2+ load), spatially regenerative activation of CRUs manifests as propagating Ca2+ waves.

In a nutshell, the discrete architecture of couplon, the extremely high-gain amplification within a couplon, the extremely low-gain amplification between CRUs, and the prompt termination of couplon activation provide the basis for local control of CICR and account for the paradoxical Ca²⁺ phenotypes in the heart. These insights have general implications in understanding the control of Ca²⁺ signaling in different types of cells.

C. Cellular Architecture of Ca²⁺ Signaling

The discovery of Ca²⁺ sparks and other elementary Ca²⁺ signals has revealed new insights into the organizational principles and cellular logic of the Ca²⁺ signaling system. In the prespark era, it was thought that Ca²⁺ rises and falls uniformly in the cytosol, a continuous and homogeneous excitable common pool. Coding and decoding the biological information are achieved by Ca²⁺ concentration and time as two analog variables. It is now understood that nearly all cells contain discrete elemental Ca²⁺ signaling units, which build up an exquisite hierarchical and discontinuous architecture in space, time, and intensity, over scales spanning many orders of magnitude (Fig. 17).

View larger version (23K): [in this window] [in a new window]   FIG. 17. Ca2+ signaling multiscaled in space, time, and amplitude.

The cellular organization of the Ca²⁺ signaling system in skeletal muscle follows the same general design but with important differences. Skeletal CRUs consist of elongated double rows of RyRs (instead of the round or oval geometry of cardiac CRUs), with some parajunctional RyRs in amphibian skeletal muscle fibers, and the sarcomere length is considerably greater (3.0 µm). In DRG sensory neurons, CRUs of RyR3 on subsurface cisterns are particularly enriched on a single-layered shell at lateral intervals of 1.7 µm; CRUs are also sprinkled over the entire cytoplasm at a reduced density, and void in the nucleus (284). Another distinct and remarkable cellular architecture of CRUs is found in Xenopus oocytes, where IP₃R CRUs are almost exclusively packaged in a subsurface band about 6 µm thick, and the outer boundary of the zone is separated from the surface membrane by 6 µm (54).

The temporal scales of cardiac Ca²⁺ dynamics range from 100 µs (the establishment and dissipation of subspace Ca²⁺ gradients) to submillisecond (LCC open time) to 5–50 ms (RyR open time, sparks, blips, blinks) to 0.1–1 s (Ca²⁺ puffs, global [Ca²⁺]_i transients or waves) and to 10 s (Ca²⁺ bursts and glows). Modulation of the Ca²⁺ signaling by various physiological and pathophysiological stimuli occurs over a time scale of seconds to minutes (e.g., fight-or-flight response, exercise). Remodeling of the Ca²⁺ signaling system, as those in development, cardiac hypertrophy, and heart failure, takes effect over much longer time scales, i.e., days and months.

In the dimension of concentration, we now have physiologically relevant cytosolic [Ca²⁺]_i levels in three different zones, spanning at least three orders of magnitude. [Ca²⁺]_subspace during Ca²⁺ sparklets and sparks is expected to be as high as a few hundred micromolar, be confined to a volume of 0.002 fl, and to last 1–10 ms. [Ca²⁺]_i in the close atmosphere of a spark is on the order of 1–10 µM and to spread over a sphere with a radius of 0.1–0.5 µm in the immediate vicinity of the subspace. Bulk cytosolic [Ca²⁺]_i is 0.1 µM at rest and varies by 10-fold during a [Ca²⁺]_i transient. The microdomain [Ca²⁺] in a spark thus extends the dynamic range of Ca²⁺ signaling by about an order of magnitude; formation of junctional subspace extends it by yet another order of magnitude.

D. The Digital-Analog Dichotomy

Cumulative advances over the last 15 years or so have led to the suggestion that there is a "digital-analog" duality of Ca²⁺ signaling. In this scenario, discrete Ca²⁺ microdomains (e.g., those of CRUs) form a digital Ca²⁺ signaling subsystem that is intertwined with an analog global Ca²⁺ signaling subsystem. The digital subsystem is mainly built for high-threshold Ca²⁺-dependent processes (such as activation of BK_Ca and Cl_Ca, arrayed RyR, activation of CaMKII and calcineurin, and vesicle secretion). These processes are discontinuous in space and operate intermittently in coincidence with gating of the Ca²⁺ signaling units. The autonomy of elemental signaling events further indicates that they operate essentially in the frequency-dependent modulation (FM) mode, though stochastic variations are intrinsic to the location, timing, and duration of CRU activation. Hence, low-threshold Ca²⁺ signaling processes are turned on by bulk Ca²⁺. Ca²⁺ signaling should be considered as an analog function, S([Ca²⁺], t), and high-threshold Ca²⁺ signaling is thus more appropriately depicted by a digital function, S(n_CRU, t_event), where n_CRU and t_event denote the number and time of CRU activation.

E. Biochemistry of Ca²⁺ Sparks

At the molecular level, the basic logic of Ca²⁺ signaling is the association and dissociation of a Ca²⁺ to and from an effector protein. The Ca²⁺ binding either activates or inhibits a physiological function of the effector, and the Ca²⁺ unbinding either deactivates or disinhibits the function. If an effector protein harbors two or more functionally opposing Ca²⁺-binding motifs, interplay between these motifs can render bell-shaped [Ca²⁺] dependence and biphasic kinetics, such that maximal biological activity is attained at an intermediate [Ca²⁺] over an optimal time window. At the level of the Ca²⁺ signaling proteome, the K_d values of Ca²⁺-binding proteins vary over 7 logarithmic units, ranging from nanomolar to 10 mM with a broad mode around 10 µM (Fig. 18), indicating the ability of signal transduction at high (>10 µM), intermediate (1 µM), and low Ca²⁺ concentrations (<0.1 µM). At the cellular level, the rise and fall of [Ca²⁺] in different compartments are decoded by the vast panel of Ca²⁺ signaling proteins.

View larger version (9K): [in this window] [in a new window]   FIG. 18. Diversity of Ca2+ binding affinity of Ca2+-sensing proteins. The dissociation constants (Kd) of 68 Ca2+ binding proteins were obtained by text search and manual collection from the literature. [Modified from Cheng et al. (72).]

From a single-atom messenger connecting thousands of signaling proteins, the Ca²⁺ signaling system is arguably the most fascinating biological signal transduction network known. Characterization of the family of elementary Ca²⁺ signaling events has unveiled a "digital-analog" dichotomy and an amazing richness in Ca²⁺ signaling modalities. Key features of the digital subsystem include discreteness of CRU organization, local control of CRU signaling, autonomous CRU operation (with stochastic variability), nonlinear responsiveness, and FM coding and decoding of biological information. Building on these, hierarchical structures of Ca²⁺ dynamics emerge, ranging from sparklets, sparks, and compound sparks, to abortive wavelets and full-fledged propagating Ca²⁺ waves, to spiral Ca²⁺ waves that encompass multiscaled spatiotemporal substructures (106, 107, 177, 218, 231). Furthermore, local control of Ca²⁺ signaling applies to subcellular compartments. Concomitant with sparks in the cytosol, the presence of blinks in the jSR cisternae and marks in single mitochondria suggests an ability of sparks to coordinate distinct yet closely related biological functions in different cellular organelles. These experimental and conceptual advances put us in a good position to unlock the secrets underlying Ca²⁺ signaling complexity.