Cytoplasmic calcium oscillations and store-operated calcium influx

Abstract

Intracellular calcium oscillations have fascinated scientists for decades. They provide an important cellular signal which, unlike most signalling mechanisms, is digitally encoded. While it is generally agreed that oscillations most frequently arise from cyclical release and re-uptake of intracellularly stored calcium, it is becoming increasingly clear that influx of calcium across the plasma membrane also plays a critical role in their maintenance and even in delivering their signal to the correct cellular locus. In this review we will discuss the role played by Ca²⁺ entry mechanisms in Ca²⁺ oscillations, and approaches to understanding the molecular nature of this Ca²⁺ entry pathway.

Introduction

Activation of a wide variety of plasma membrane receptors results in activation of phospholipase C and generation of the Ca²⁺ mobilizing messenger, inositol 1,4,5-trisphosphate (IP₃) (Berridge, 1993). IP₃ induces the release of Ca²⁺ from endoplasmic reticulum stores, which in turn leads to activation of store-operated Ca²⁺ channels in the plasma membrane (Parekh & Putney, 2005). These two processes are readily demonstrable with maximal or near maximal concentrations of phospholipase C-linked agonists, which induce a biphasic Ca²⁺ signal comprising an initial release of Ca²⁺ followed by a sustained phase of steady Ca²⁺ entry (Putney et al. 1981). However, intracellular Ca²⁺ ([Ca²⁺]_i) signals resulting from lower, more physiological concentrations of agonists generally result in repetitive spikes of [Ca²⁺]_i, sometimes referred to as Ca²⁺ oscillations (Thomas et al. 1996). These oscillations are ideally suited for signalling a host of cellular responses because their digital nature provides a high degree of signal-to-noise discrimination (Berridge & Galione, 1988). However, the analysis of the sources and extents of Ca²⁺ release and the mechanisms of associated Ca²⁺ entry poses problems because of the very low upstream signalling strengths. The nature and regulation of associated Ca²⁺ influx mechanisms are of special interest because plasma membrane channels have promise as pharmacological targets. In addition, there is evidence that Ca²⁺ entering through plasma membrane channels is often tightly coupled to downstream effectors, even when larger global changes in Ca²⁺ arise from intracellular release (Chang et al. 2006, 2007).

Mechanisms of [Ca²⁺]_i oscillations

The most commonly encountered form of [Ca²⁺]_i oscillations involves baseline transients or spikes (Berridge, 1990; Meyer & Stryer, 1991). Baseline spikes were first described by Woods et al. (1986) and are characterized by rapidly rising transient increases in [Ca²⁺]_i rising from a baseline of [Ca²⁺]_i which is generally quite close to the resting level. In most instances in which [Ca²⁺]_i transients arise periodically from a stable baseline, increasing the agonist concentration increases the frequency of the spikes with little effect on the amplitude of the spikes.

Considerable effort and speculation has been focused on the mechanism underlying the baseline spike type of calcium oscillation. Usually baseline spiking will continue for at least a few cycles in the absence of extracellular Ca²⁺, and thus it is generally agreed that it represents cycles of discharge and reuptake of Ca²⁺ by intracellular stores. The regenerative, frequency regulated nature of Ca²⁺ spiking suggests minimally a negative feedback as well as an amplifier or feed forward mechanism. The discoverers of this phenomenon advocated a model according to which it is the cellular level of IP₃ which rises in spikes, and the [Ca²⁺]_i changes reflect these oscillations in IP₃ (Woods et al. 1987). This model suggests regulation of Ca²⁺ signalling at an early step in the signalling pathway, at or prior to the formation of IP₃. In a conceptually similar model for baseline spikes, also featuring spikes in the level of IP₃, Meyer & Stryer (1988) proposed feed forward activation of phospholipase C by [Ca²⁺]_i together with highly cooperative opening of intracellular Ca²⁺ channels by IP₃, but have suggested that the inactivation of the [Ca²⁺]_i signal might result from reuptake of Ca²⁺ into IP₃-insensitive pools, or by a negative feed back of [Ca²⁺]_i causing desensitization of the IP₃ receptor.

The second general proposal for the mechanism of baseline [Ca²⁺]_i spiking involves fluctuations in [Ca²⁺]_i while cellular levels of IP₃ are constant. Berridge has perhaps been the major proponent of models of Ca²⁺ oscillations involving regulation at the IP₃ receptor level (Berridge, 1990). Such mechanisms require that the IP₃ receptor has the capacity to be regulated both positively and negatively by downstream signals, presumably by Ca²⁺ itself. There is considerable evidence for such regulation. The IP₃ receptor is known to be biphasically regulated by Ca²⁺; in the low cytoplasmic range of [Ca²⁺], as [Ca²⁺] is increased, the ability of IP₃ to release Ca²⁺ is increased (Iino, 1990; Finch et al. 1991; Marshall & Taylor, 1993), while at higher [Ca²⁺], and with more slowly developing kinetics (see below; Finch et al. 1991) the action of IP₃ is inhibited (Jean & Klee, 1986; Marshall & Taylor, 1993; for a review see Foskett et al. 2007). In addition, there may be positive regulation of the receptor by Ca²⁺ within the lumen of the endoplasmic reticulum, which could also serve as a limiting factor for the extent of Ca²⁺ release during a single spike (Parys et al. 1993; but see Missiaen et al. 1999).

Recent work from Politi et al. (2006) provided strong evidence in support of an oscillating mechanism upstream of the IP₃ receptor. These investigators expressed a soluble polypeptide from the IP₃ binding domain of the IP₃ receptor that is known to act as an IP₃ ‘buffer’ in the cytoplasm (Gaspers & Thomas, 2005). As predicted from mathematical models, buffering changes in cytoplasmic IP₃ prevented Ca²⁺ oscillations; this is not expected if oscillations occur at constant IP₃ level.

Role of Ca²⁺ entry in [Ca²⁺]_i oscillations

In the absence of extracellular Ca²⁺, oscillations are not maintained, indicating that some mechanism of Ca²⁺ entry across the plasma membrane is required for their maintenance (Kawanishi et al. 1989; Lewis & Cahalan, 1989). The discharge of Ca²⁺ stores by IP₃ is known to activate store-operated Ca²⁺ channels in the plasma membrane (Putney, 1986; Putney et al. 2001). However, activation of phospholipase C-linked receptors is sometimes linked to other types of Ca²⁺ permeable channels (Barritt, 1999), and it has been argued that with low, physiological levels of agonist, such as those that produce [Ca²⁺]_i oscillations, these non-store-operated channels might be more important (Shuttleworth, 1999). A technical difficulty in assessing the nature of the Ca²⁺ entry mechanism with low levels of activation is that the predicted entry of Ca²⁺ will be very small. Indeed given that baseline Ca²⁺ is usually unchanged, the prediction is that a very small amount of entry occurs that is adequately buffered by intracellular stores. The amount of Ca²⁺ lost from the endoplasmic reticulum during a single oscillation appears to be a very small fraction of stored Ca²⁺ (Bird & Putney, 2005). Thus, regardless of the mechanism of entry that replenishes this deficit, it is likely to be very small. In an earlier review, Shuttleworth argued that in many instances, Ca²⁺ entry may be a primary signal that secondarily activates intracellular release, resulting in [Ca²⁺]_i oscillations (Shuttleworth, 1999). However, as mentioned above in most instances oscillations continue for at least a few cycles in the absence of extracellular Ca²⁺. Also, it is sometimes possible to block both Ca²⁺ entry and extrusion with high concentrations of lanthanides; this then results in [Ca²⁺]_i oscillations that persist even in the absence of extracellular Ca²⁺ (Sneyd et al. 2004; Bird & Putney, 2005).

We will consider the argument that the entry that supports oscillations is store-operated entry, and discuss what kinds of experimental evidence can support or refute this idea. First of all, the best evidence would be the demonstration in an oscillating cell of a current with properties similar to the well-characterized calcium-release-activated calcium current (I_crac). However, in haematopoetic cells, I_crac is of the order of −2 to −3 pA pF⁻¹, and then only under conditions whereby it is hugely amplified by preventing Ca²⁺ feedback inhibition (Parekh & Putney, 2005). Even with maximal store depletion, I_crac has never been observed (or at least reported) under physiological conditions, that is with weak intracellular Ca²⁺ buffering and at physiological membrane potentials. In other cell types, for example in one often used model the HEK293 cell, the maximal I_crac is of the order of −0.5 pA pF⁻¹. So measurement of I_crac, or for that matter any agonist-activated Ca²⁺ current (Parekh, 2007), in a cell that is oscillating (which requires minimal intracellular buffering) is highly unlikely, and thus more indirect evidence is needed to test this idea.

The first approach is to compare the pharmacological sensitivity of [Ca²⁺]_i oscillations to that known to apply to I_crac. To this end, Bird & Putney (2005) utilized two agents known to block store-operated entry and I_crac, Gd³⁺ and 2-aminoethyldiphenylborate (2APB). The latter compound is known to affect a number of other channel types, but notably only blocks members of the TRPC family partially (Lievremont et al. 2005), and is reported not to block channels responsible for arachidonic acid-activated currents (Mignen et al. 2003), two candidates for non-store-operated Ca²⁺ entry in a number of cell types (Vazquez et al. 2004; Shuttleworth et al. 2004). Gd³⁺ is believed to be highly specific for store-operated channels when employed in a concentration of 1 μm or less (Broad et al. 1999). Thus, the pharmacological profile of the [Ca²⁺]_i oscillations in HEK293 cells is that which is expected of store-operated channels, likely involving I_crac. No other known Ca²⁺ entry pathway shares this pharmacological profile.

The second approach is to examine the roles of specific gene products known to be involved in store-operated entry, specifically, members of the Stim and Orai gene families. Stim1 serves as the major Ca²⁺ sensor in the endoplasmic reticulum when Ca²⁺ levels fall below a critical level due to active release (Roos et al. 2005; Liou et al. 2005; Brandman et al. 2007). Stim2 appears to respond to fluctuations in endoplasmic reticulum Ca²⁺ near the steady-state resting level (Brandman et al. 2007). The Orai (also known as CRACM) proteins function as pore-forming subunits of the store operated CRAC channels (Feske et al. 2006; Vig et al. 2006b; Zhang et al. 2006; Prakriya et al. 2006; Yeromin et al. 2006; Vig et al. 2006a). In HEK293 cells, knockdown by RNAi of either Stim1 or Orai1 results in almost complete abrogation of muscarinic receptor-linked [Ca²⁺]_i oscillations (Wedel et al. 2007). This result is consistent with the pharmacological data and with the conclusion that in this cell line, the Ca²⁺ entry that supports [Ca²⁺]_i oscillations comes through store-operated CRAC channels. One other channel has been shown to depend upon both Stim1 and Orai proteins, the arachidonic acid gated channels underlying the arachidonate-regulated calcium current, I_arc (Mignen et al. 2007, 2008). However, the function of Stim1 in I_arc is clearly not related to store depletion because it appears to be specifically Stim1 in the plasma membrane that is required (Mignen et al. 2007). Thus, it is significant that in the earlier study of the roles of these proteins in HEK293 cell [Ca²⁺]_i oscillations, it was demonstrated that the effects of knockdown of Stim1 by RNAi could be rescued by an N-terminal fluorescent protein-tagged Stim1 (Wedel et al. 2007), which distributes only to endoplasmic reticulum and does not reach the plasma membrane (Mercer et al. 2006; Hauser & Tsien, 2007). Thus it is clear that the role of Stim1 in supporting [Ca²⁺]_i oscillations in HEK293 cells involves its action in the endoplasmic reticulum where it functions as the calcium sensing initiator of store-operated Ca²⁺ entry.

The major point to be taken from this short review, as well as from previous studies from the authors' laboratory, is that a combination of carefully controlled pharmacological and molecular biological studies can provide clear and incontrovertible evidence on the role of store-operated Ca²⁺ entry in the Ca²⁺ signalling that occurs during modest and physiological strengths of cell activation. However strong this conclusion may be for the HEK293 cell line, this of course does not assure, or even suggest, that other Ca²⁺ mechanisms will not have major roles in other cell types. But hopefully these studies establish strategies and criteria for establishing whether, or to what degree store-operated entry contributes to the process of Ca²⁺ oscillations, or for that matter to any important physiological end point.