The sarcoplasmic reticulum (SR) and apposed regions of the sarcolemma passively trap Ca2+ entering the cell to limit the rise in cytoplasmic Ca2+ concentration without SR pump involvement. When “leaky” the SR facilitates Ca2+ entry to the cytoplasm. SR Ca2+ release via inositol 1,4,5-trisphosphate receptors (IP3Rs) propagates as calcium waves; IP3Rs alone account for wave propagation.
The ability to change the cytosolic free Ca2+ concentration ([Ca2+]c) is an inherent characteristic of eukaryotic cells. The amplitude, duration, frequency, and location of each of these [Ca2+]c changes are important determinants of the Ca2+ signal; they combine to target specific cellular activities such as gene expression and metabolism and so control the nature of the overall cellular response. The cell’s ability to modulate each of these aspects of Ca2+ change provides the flexibility required to enable a single messenger to generate and manage many different and often opposing cellular functions.
Two main sources of [Ca2+]c are recognized: the extracellular fluid and the intracellular stores of the sarcoplasmic reticulum (SR). One of the main Ca2+ entry pathways from the extracellular fluid is the voltage-dependent Ca2+ channel in the sarcolemma. Membrane depolarization increases the activity of these channels to enhance Ca2+ entry and raise [Ca2+]c. The other main [Ca2+]c source is the internal SR store from which release proceeds via two main receptor-controlled channels: the inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) and the ryanodine receptor (RyR). The SR may liberate Ca2+ in response to activation of IP3R as a result of agonist activity at the sarcolemma that increases IP3 production. Ca2+ influx from outside may activate RyR in the process of Ca2+-induced Ca2+ release (CICR). The SR may also release Ca2+, via RyR, when the store’s Ca2+ content exceeds normal physiological values, i.e., in “store overload.” Indeed, store overload may facilitate the occurrence of CICR in cells that do not normally display this feature.
In our studies, depolarization activates voltage-dependent Ca2+ channels to increase [Ca2+]c uniformly throughout the bulk of the cytoplasm (Fig. 1A⇓). The increase can be explained solely by Ca2+ influx and does not involve the SR or CICR (3). Thus depolarization does not increase [Ca2+]c in the presence of voltage-dependent Ca2+ channel blockers, in the absence of extracellular Ca2+, or when the membrane potential is held at the equilibrium potential for Ca2+. Each of these findings demonstrates that, in response to sarcolemma depolarization, Ca2+ influx is essential for a rise in [Ca2+]c and that channel activity or membrane potential changes themselves failed to alter [Ca2+]c in the absence of influx. Moreover, Ca2+ influx produced by depolarization of the sarcolemma is quantitatively adequate to account for the entire [Ca2+]c increase, without SR involvement (3). At least 100 times more Ca2+ enters the cell than appears as free Ca2+ in the cytosol (3). Not surprisingly, therefore, depletion of the SR Ca2+ stores using the Ca2+ pump inhibitors thapsigargin or cyclopiazonic acid or by ryanodine and caffeine to open RyR on the SR (Fig. 1B⇓) does not reduce depolarization-evoked [Ca2+]c increases (3). These results suggest that the SR does not contribute to the rise in [Ca2+]c evoked by Ca2+ influx via CICR, although controversy persists as to the precise contribution of CICR in smooth muscle. Significantly, in equine tracheal myocytes, Ca2+ influx may activate, albeit after a considerable delay, Ca2+ release from the internal SR store in a process referred to as “loosely coupled CICR” (10). The delayed Ca2+ release from the SR was attributed to the time required for subsarcolemmal [Ca2+] to build up in the gap between the voltage-dependent Ca2+ channels in the sarcolemma and the RyR in the SR to the value necessary to activate CICR. However, the delay between Ca2+ influx and Ca2+ release from the SR may also reflect the time required for the SR store to accumulate and subsequently discharge Ca2+ via RyR when store overload conditions had developed. In those cells in which “loosely coupled CICR” has been described, experiments are required to distinguish between store overload-evoked Ca2+ release and the proposed loosely coupled CICR.
Rather than contributing to an increase in [Ca2+]c, the SR may in fact limit the rise that occurs after influx. SR pump activity reportedly contributes to this limiting effect of the SR in the “superficial buffer barrier” hypothesis (see Ref. 2 for references). Our own observations, however, suggest that pump activity does not substantially reduce the bulk average [Ca2+]c rise that occurs during depolarization (2). Although unable to modify the rise in [Ca2+]c occurring during influx, the SR accumulates Ca2+ and contributes to the decline in [Ca2+]c when influx ends by pump activity throughout the SR structure (3). The increased SR Ca2+ load after sarcolemma depolarization (as a result of its accumulation of the ion) is exported slowly to the outside of the cell by Na+/Ca2+ exchanger activity at the sarcolemma to restore the steady-state SR Ca2+ content (3).
Although SR Ca2+ pump activity does not limit the rise in [Ca2+]c, the SR itself nonetheless restricts the increase in bulk [Ca2+]c that follows influx by forming a passive Ca2+ trap. Inducing leakiness of the SR store, using ryanodine together with caffeine (and fully depleting the SR of Ca2+), significantly increases the [Ca2+]c rise that occurs after depolarization and Ca2+ influx (Fig. 1B⇑). Yet fully depleting the SR by using either thapsigargin (Fig. 1C⇑) or cyclopiazonic acid (each of which inhibits the SR Ca2+ pump) does not increase the amplitude of the depolarization-evoked Ca2+ transient (Fig. 1C⇑). These results suggest that, under physiological conditions, the SR store limits the rise in [Ca2+]c that occurs following influx. However, pump activity is not required for this limiting action of the SR store; neither thapsigargin nor cyclopiazonic acid increased the amplitude of the depolarization-evoked [Ca2+]c rise. Rather, the store forms a physical barrier with the sarcolemma to passively limit the access of Ca2+ that has entered by influx to the bulk of the cytoplasm, i.e., a passive “Ca2+ trap” (2) (Fig. 1D⇑). As much as 40% of the Ca2+ entering the cell is held in the trap, generating high subsarcolemma [Ca2+] in the tens of micromolar range (2).
One possible role for the Ca2+ trap may be to activate low-affinity Ca2+-sensitive ion channels in the sarcolemma to modulate Ca2+ entry via membrane potential changes. Another role may be to maintain the SR Ca2+ content itself. Ca2+ influx is essential to maintain the IP3-sensitive store content of Ca2+. Two influx pathways, one voltage insensitive (i.e., not voltage gated) and one voltage sensitive (i.e., voltage gated), allow store refilling to be maintained over a wide membrane voltage range (12). Refilling appears to take place substantially from a subsarcolemma Ca2+ pool in which Ca2+ concentration exceeds bulk average cytoplasmic values. The Ca2+ trap may contribute to the generation and maintenance of the high subsarcolemma Ca2+ concentration to permit store refilling (12).
In addition to the extracellular fluid source, the SR also plays an integral role in the generation of those Ca2+ signals that give rise to the biological response. These signals emanate from the SR and appear as localized events: Ca2+ “puffs” and “sparks” that are restricted to small areas of the cell. In turn, these localized events may activate sarcolemma ion channels such as Ca2+-activated K+ channels and generate spontaneous transient outward currents (STOCs). Under certain, as yet unspecified conditions the cytoplasm becomes excitable, and Ca2+ puffs and/or sparks may coalesce to give rise to repetitive Ca2+ transients called “oscillations,” which may in turn propagate as Ca2+ waves (Fig. 2A⇓). Significantly, removal of external Ca2+ inhibits both waves and oscillations. Yet external Ca2+ may serve primarily to fill the SR store or to maintain [Ca2+]c at a level that can support waves rather than provide the underlying mechanisms for wave initiation (8). Nonetheless, it is the ordered release of Ca2+ from the SR and the characteristics of this release that influence the location, frequency, and amplitude of the Ca2+ signals that in turn generate the biological response.
Ca2+ release from the SR involves both the IP3R and RyR. Release from one receptor is not independent of release from the other; the organization of receptors may contribute to the level of interaction between them. Our studies suggest that two SR stores exist in myocytes, one possessing both IP3R and RyR (each of which has common access to Ca2+ in that store) and another, separate store with RyR alone (5). Release from IP3R reduces the Ca2+ available for release from RyR by depleting the common store and contributes to the excitatory effects of IP3-generating agonists by reducing negative feedback systems that operate to limit Ca2+ influx (11). In smooth muscle, the number of IP3R exceeds that of RyR by ~10-fold (18). Not surprisingly therefore, sarcolemma agonists (e.g., muscarinic agonists) that generate IP3 are perhaps the most common means of evoking Ca2+ waves in smooth muscle. Other means of evoking waves, such as occur at RyR in conditions of store overload (e.g., Ref. 4), may require an increase in the (normally low) sensitivity of RyR to Ca2+. The physiological significance of these conditions is not clear.
The wave components
Waves are induced from the IP3R on the SR by IP3-generating agonists and proceed, at nearly constant amplitude, by sequential Ca2+ release from one receptor to the next throughout the store (Fig. 2A⇑). Each wave comprises both temporally and spatially related elements, and both a rising and a declining phase have been identified. The rising phase consists of a localized “initiation” component derived from the release of Ca2+ from the IP3R followed by an “amplification” component during which this release is augmented by CICR by positive feedback at either the IP3R or RyR. The positive feedback process thus amplifies the initiation (local) response to produce the rising phase of the wave.
One interesting but disputed aspect of wave initiation is whether or not IP3 concentration oscillations are required, i.e., are IP3 concentration changes coupled to wave production? In the “cross-coupling hypothesis,” IP3R activity, from an as yet ill-defined initiating stimulus, releases Ca2+, which results in more IP3 production in a positive feedback loop until the store presumably is depleted of Ca2+, at which time IP3 production falls to basal levels. However, in other cases changes in IP3 levels, which occur for example following agonist activity, are not mirrored by corresponding changes in Ca2+ levels. Indeed, Ca2+ oscillations and waves can be induced by application of constant concentrations of IP3 (17). Implicit in these findings is the view that IP3R open and close in a coordinated fashion in the presence of constant concentrations of IP3, i.e., oscillating IP3 levels are not necessary for rhythmic activity. It may be that the IP3R responds to released Ca2+ by increasing its open probability, i.e., synergy exists between Ca2+ and IP3 in gating the receptor channel (7) to account for the rising phase of the wave.
Another area of particular interest and controversy concerns the precise nature of the receptors involved in the amplification phase of wave production, especially as to whether IP3R and RyR each contribute to the process. In one proposal waves initiated by agonists propagate exclusively as a result of IP3R activity. Here waves originate from properties inherent in the IP3R themselves that allow them to open and close (deactivate) in the continued presence of a constant concentration of IP3. This proposal requires that 1) synergism exists between Ca2+ and IP3 in which the IP3R may act as CICR site, 2) a high [Ca2+]c inhibits IP3R opening, and 3) a refractory period for the IP3-gated channel follows channel opening; this period may be induced by IP3 itself or by cytoplasmic and/or luminal Ca2+ concentration.
In another proposal, the IP3-mediated Ca2+ release serves only to initiate the wave and thereafter the role of the IP3R ends (6). The initial IP3-mediated Ca2+ release then triggers a more substantial Ca2+ release from the RyR via CICR. This in turn activates more RyR so that a positive feedback cycle of Ca2+ release ensues exclusively at RyR. In support of RyR involvement, drugs that alter RyR activity (ryanodine, ruthenium red, or tetracaine) sometimes abolish [Ca2+]c oscillations (e.g., Ref. 9). Yet different interpretations of these findings are possible. For example, the experimental protocol may have created abnormal store overload conditions in which the sensitivity of RyR to Ca2+ may have increased, thus predisposing the cell to wave production at RyR (e.g., Ref. 4), an event that may not occur under physiological conditions.
Alternatively, the pharmacological agents used to assess the contribution of RyR may lack specificity and also block IP3R. Ryanodine, tetracaine, and ruthenium red may each inhibit IP3-mediated Ca2+ signals independently of RyR involvement (e.g., Refs. 5, 8, 11, 15, 16, and 20). The actions of ryanodine are particularly complex. The drug binds to the open state of the RyR and may prolong receptor open time, albeit at a lower conductance (e.g., Ref. 19; Fig. 2B⇑). Persistently open RyR would presumably increase Ca2+ leak from the SR and lower its Ca2+ content. Under such conditions, depletion of these stores to which IP3R and RyR have common access would indirectly inhibit IP3-mediated responses. This seems a more likely explanation for the inhibition than the proposed amplification of Ca2+ release by CICR at RyR. Several of our own observations support this view. First, ryanodine by itself does not inhibit IP3-mediated Ca2+ transients (at a membrane potential of −70 mV), suggesting that Ca2+ released via the IP3R does not activate RyR (Fig. 2C⇑). On the other hand, after transient activation of RyR by caffeine, IP3-mediated Ca2+ release is inhibited (Fig. 2D⇑). Caffeine opens RyR, allowing ryanodine to persistently activate the channel and deplete the store of Ca2+. Thus the IP3-mediated Ca2+ transient arises from IP3R activity alone without RyR involvement. Block of IP3-mediated Ca2+ transients by ryanodine occurs indirectly as a result of a reduction in the Ca2+ content of the SR store to which IP3R and RyR have common access rather than from RyR involvement in the IP3 response. Ryanodine may also inhibit IP3 responses without prior activation of RyR by caffeine. For example, under experimental conditions that increase the SR Ca2+ content (e.g., depolarization to −20 mV) and activate RyR (as evidenced by the occurrence of STOCs; Fig. 2E⇑), ryanodine reduces the IP3-evoked Ca2+ transient (Fig. 2F⇑). These results can again be explained by ryanodine’s maintaining the RyR in an open configuration to reduce the store content and attenuate the IP3-evoked Ca2+ transients.
Tetracaine also lacks specificity and blocks IP3-mediated Ca2+ release (e.g., Ref. 15). In our experiments, tetracaine blocked submaximal (unpublished results) but not maximal (11) responses to IP3. Ruthenium red also lacks specificity and exerts complex actions on IP3-mediated Ca2+ responses. For example, in avian atria ruthenium red inhibits the response to IP3 by an action unrelated to RyR since the response to caffeine was potentiated (16). Clearly there is a problem of lack of specificity of those drugs used to distinguish between the contributions of RyR and IP3R. Control experiments demonstrating that IP3R are not inhibited by “RyR blockers” are required to permit the correct interpretation of drug effects on Ca2+ signals.
In our studies of Ca2+ waves (13), the potential difficulties introduced by poor drug specificity were avoided as far as possible and the conditions proposed to generate waves were employed, i.e., a localized increase in IP3 and Ca2+ that subsequently, it is proposed, activates RyR. Even though IP3-generating agonists evoked waves (Fig. 2A⇑), the localized release of IP3 (by localized photolysis of the caged phosphoinositide) did not (Fig. 3A⇓). If activation of RyR had been involved in propagation, it would be expected that the resulting localized increase in Ca2+ following photolysis would propagate throughout the cell as a wave. Clearly, our results showed that a localized increase in Ca2+ originating from IP3R activity was, by itself, insufficient to trigger a Ca2+ wave, i.e., RyR were not recruited by localized increases in Ca2+ from the IP3R even under conditions where RyR were active and generating STOCs (i.e., at −20 mV) (13) (Fig. 3A⇓). On the other hand, when IP3 concentrations throughout the cell were elevated, e.g., by subthreshold concentrations of IP3-generating agonists (Fig. 3B⇓) or following dialysis of the cell with IP3 to sensitize and prime the IP3R to release Ca2+ (13), a local increase in Ca2+ from the activated IP3R following a localized photolysis of the caged compound successfully evoked a Ca2+ wave. Thus a global elevation in IP3, i.e., a general increase throughout the cell, is required for wave production. One proposal to accommodate these findings is that a global elevation in IP3 sensitizes the IP3R to Ca2+. When Ca2+ release occurs, propagation throughout the cell takes place by the activity of the Ca2+-sensitized IP3R. This accounts for the rising phase (the leading edge) of the propagated Ca2+ wave. Wave propagation as implied by the present findings extends only to those cellular sites already sensitized to Ca2+ by an increased IP3 concentration. This permits a spatial control of the Ca2+ signal by the cell.
The decline of the wave
After peaking, [Ca2+]c declines (the back of the wave), although propagation continues (Fig. 2A⇑), i.e., release of Ca2+ proceeds (the front of the wave) at the same time as [Ca2+]c declines at the back (Fig. 2A⇑). The basis for the decline may lie in the construction of the SR itself. The SR store may be visualized either as one element luminally continuous throughout or as a series of self-contained discontinuous elements. Depletion of a luminally continuous store is unlikely to account for the decline, and it is likely that sufficient Ca2+ is retained in the store to support further release at the back of the wave (Fig. 2A⇑). If, on the other hand, the store was arranged as a series of compartmentalized and discontinuous elements (Fig. 3C⇑) rather than as a single entity in luminal continuity, then both wave progression and simultaneous decline could be explained by each element releasing and being successively depleted of Ca2+ in turn (Fig. 3C⇑). To test whether or not the store was compartmentalized, successive Ca2+ release events were evoked from small regions of the IP3-sensitive store again by the localized photolysis of caged IP3 in the same cell (Fig. 4A⇓). When release was activated on two separate occasions at the same site within the same cell, the second was unsuccessful in evoking a Ca2+ increase. On the other hand, both the first and second activations at different (separate) sites within the same cell evoked comparable Ca2+ release. These results, at first sight, could imply the existence of a store arranged as discontinuous elements, each event depleting only that region of the store of Ca2+ at which release had occurred. This explanation is unlikely. To determine if the IP3-sensitive store was depleted after releasing Ca2+, the same region at which IP3 was photolyzed was subsequently activated by using caffeine [the IP3-sensitive store also contains RyR (5)]. RyR activation, with caffeine, evoked Ca2+ release from the region that had prior IP3 activation (Fig. 4B⇓). This result suggests that the region of the store refractory to IP3 retained Ca2+ and had not been depleted. Indeed, the content decreases relatively little, as assessed by the magnitude of a caffeine-evoked Ca2+ rise (Fig. 4B⇓). Yet the second of two successive IP3 increases at the same site was unsuccessful in evoking a Ca2+ increase, suggesting that the SR was compartmentalized. The compartmentalization, therefore, appears to be functional rather than structural and does not arise because the store is composed of a series of separate subunits each capable of being independently depleted of Ca2+. Rather, the results suggest that the store retained sufficient Ca2+ for release and by implication that release failed on the second occasion because IP3R had become refractory to IP3.
There may be several mechanisms responsible for the development of the refractoriness to IP3, e.g., a lowered SR luminal Ca2+ concentration, an increased IP3 concentration, or an increased [Ca2+]c. Of these, a lowered SR content seems an unlikely candidate because the SR Ca2+ content decreased relatively little after local photolysis of caged IP3 (as assessed by the magnitude of the caffeine-evoked Ca2+ transient; Fig. 4B⇑). Additionally, IP3-mediated release was normal in regions not previously exposed to the phosphoinositide. These regions had been subject presumably to the same lowered Ca2+ content as that at the refractory site, assuming that Ca2+ is in free diffusional equilibrium within the SR lumen.
Since the IP3 response persisted after sustained global elevations in IP3 (after dialysis with the phosphoinositide or activation with sarcolemma agonists), this suggests (by elimination) that an increased [Ca2+]c may be sufficient to account for the refractoriness under the present experimental conditions. Indeed, the refractoriness of IP3-mediated Ca2+ release was mimicked by an increased [Ca2+]c in the absence of either an increased cytoplasmic IP3 concentration or a reduced SR luminal [Ca2+] (Fig. 4C⇑). The IP3R type 1 (as exists in the present preparation; unpublished result) is activated only over a narrow range of Ca2+ concentrations (up to ~300 nM); higher concentrations inhibit the receptor. Previous studies in a variety of tissues and in isolated IP3R in bilayers under steady-state conditions found that Ca2+ deactivation of IP3R occurred at concentrations >300 nM (e.g., Ref. 7). The deactivation was rapid in onset with, for example, a t1/2 of 50 ms in rat permeabilized hepatocytes (1). The Ca2+-dependent deactivation reported in our studies (11) was much slower in onset and required pulse generations >1 s to be significant (Fig. 4D⇑). After onset, the deactivation persisted in our study for tens of seconds, whereas the recovery (Fig. 4E⇑) from other steady-state studies took considerably less (cf. t1/2 = 400 ms; Ref. 1). Others have, however, also reported a slow onset and recovery from deactivation. For example, in rat basophilic leukemia cells, a Ca2+-dependent suppression of the IP3-mediated Ca2+ transient took 30–60 s for recovery (14). It seems possible that separate, different, Ca2+-dependent inhibitory mechanisms of IP3-mediated Ca2+ release may exist, one rapid in onset and recovering quickly on Ca2+ removal, a second slow in onset and persistent. Indeed, deactivation of IP3R, in the present study, once initiated by an increased [Ca2+]c, persisted long after restoration of [Ca2+]c even to resting values, i.e., the deactivation became, at least partially, Ca2+ independent (see also Ref. 14). Since some signaling pathways that are activated by Ca2+ subsequently become Ca2+ independent, e.g., Ca2+-dependent protein kinase II, a study of the ability of kinase and phosphatase inhibitors to modulate the deactivation of IP3-mediated Ca2+ release seemed justified. In the event, no support for the participation of either of these enzymes emerged. Perhaps in response to substantial (~1 μM) transient elevations in Ca2+, a persistent inhibitory effect of Ca2+ on the receptor ensues.
In summary, depolarization evokes uniform increases in [Ca2+]c throughout the bulk of the cytoplasm. However, a significant fraction of the Ca2+ entering the cell is retained in a Ca2+ trap created by the close apposition of the sarcolemma and SR structure to limit the [Ca2+]c increase in the bulk of the cytoplasm and to create a high subsarcolemma Ca2+ concentration. IP3-generating agonists evoke [Ca2+]c waves, which progress through the cytoplasm by CICR acting on IP3R without RyR involvement. A delayed onset and thereafter a persistent negative feedback accounts for the decline of the wave. Properties of the IP3R are themselves sufficient to account for wave progression (Fig. 4F⇑).
The Wellcome Trust (060094/Z/00/Z) and British Heart Foundation (PG/2001079; PG/02/161) funded this work.
- © 2004 Int. Union Physiol. Sci./Am.Physiol. Soc.