Spatial Ca²⁺ profiling: decrypting the universal cytosolic Ca²⁺ oscillation

Abstract

Stimulation of cell‐surface receptors that couple to phospholipase C to generate the second messenger inositol trisphosphate often evokes repetitive oscillations in cytosolic Ca²⁺. Signalling information is encoded in both the amplitude and frequency of the Ca²⁺ spikes. Recent studies have revealed that the spatial profile of the oscillation also imparts signalling power; Ca²⁺ microdomains near store‐operated CRAC channels in the plasma membrane and inositol trisphosphate‐gated channels in the endoplasmic reticulum both signal to distinct downstream targets. Spatial profiling therefore increases the transduction power of the universal oscillatory cytosolic Ca²⁺ signal.

Introduction

A rise in cytosolic Ca²⁺ concentration is universally used as a signalling mechanism in eukaryotic cells (Clapham, 2008). The Ca²⁺ rise activates responses over a broad temporal bandwidth. These range from exocytosis of neurotransmitter‐containing vesicles that operates on a submillisecond time scale, through stimulation of mitochondrial and cytoplasmic enzymes that occur over a period of seconds, up to regulation of Ca²⁺‐dependent gene expression, cell growth and differentiation and even cell death, which can develop hours or even days after the initial Ca²⁺ signal has terminated (Berridge et al. 2003). Many of these responses are not necessarily independent and need to be co‐ordinated in order for an organism to adapt to changes in its environment. For example, an increase in cytosolic Ca²⁺ that occurs following a burst of action potentials in neurons not only triggers neurotransmitter release but also needs to stimulate mitochondrial metabolism to provide ATP to replenish and prime the mobilizable pool of secretory vesicles at the active zone. In addition, the Ca²⁺ rise often signals to the nucleus, typically through Ca²⁺‐dependent transcription factors such as CREB, to increase expression of genes associated with forms of memory such as long‐term potentiation (Wheeler et al. 2008). On the other hand, activation of Ca²⁺‐dependent apoptotic programmes needs to be tightly regulated. This remarkable versatility of cytosolic Ca²⁺ in activating a multitude of, and sometimes opposing, responses in a cell raises the old chestnut of specificity: how can some responses be activated by Ca²⁺ and not others in the same cell?

One mechanism that will lead to activation of distinct Ca²⁺‐dependent response is an affinity‐based one, where different pathways have different affinities for cytosolic Ca²⁺. In this scheme, higher affinity processes will be recruited first and then, as the amplitude of the Ca²⁺ signal increases, those with lower affinity will be subsequently engaged. This can be refined further by having multiple identical Ca²⁺ binding sites on the Ca²⁺ sensor. Calculations have shown that, for a given cytosolic Ca²⁺ concentration, fractional Ca²⁺ occupancy of the sensor, and therefore its signalling strength, becomes considerably less as the number of identical Ca²⁺‐binding sites increases (Parekh, 2011). Hence varying the number of Ca²⁺ binding sites in a sensor could provide a simple yet effective means for maintaining amplitude‐encoded selectivity. However, the main limitation with an amplitude‐encoded system is that different Ca²⁺‐dependent processes will be recruited sequentially, in a manner dictated by their relative affinities for Ca²⁺ (Parekh, 2011). To activate a low affinity process, the large rise in cytosolic Ca²⁺ that is needed would indiscriminately stimulate sensors with a higher affinity for Ca²⁺, compromising specificity.

Additional mechanisms are therefore needed to confer specificity to the cytosolic Ca²⁺ signal. One way is through temporal filtering, whereby the kinetic or temporal properties of the Ca²⁺ signal are detected by downstream Ca²⁺ sensors and then transduced into distinct cellular responses. Different Ca²⁺ sensors exhibit different on‐ and off‐rates for Ca²⁺. A sensor with a slow off‐rate would retain activity for longer after the Ca²⁺ signal has decayed. A sensor with a faster on‐rate is more likely to respond to transient Ca²⁺ signals. Different cytosolic Ca²⁺ kinetics can therefore bring different Ca²⁺ sensors into play. Ca²⁺ sensors that have been found to respond to the frequency of Ca²⁺ signals include protein kinase Cγ (Oancea & Meyer, 1998; Violin et al. 2003) and Ca²⁺/calmodulin‐dependent protein kinase II (Meyer et al. 1992; De Koninck & Schulman, 1998).

A protein that can discriminate between different kinetics of a cytosolic Ca²⁺ signal is the ubiquitous Ca²⁺‐binding protein calmodulin (Hoeflich & Ikura, 2002). Calmodulin has an amino acid sequence that is invariant amongst vertebrates. The protein has globular amino‐ and carboxy‐terminal lobes, each containing 2 EF‐hand Ca²⁺‐binding motifs. The lobes are connected by a flexible hinge region. The on‐rate and the off‐rate for Ca²⁺ binding to the EF‐hands are ∼70‐fold and ∼170‐fold faster in the amino‐terminal lobe than that in the carboxy‐terminal lobe (Johnson et al. 1996), and this imparts lobe‐specific regulation to P/Q‐type Ca²⁺ channels (DeMaria et al. 2001). Lobe‐specific regulation of proteins increases the range of Ca²⁺ signals that calmodulin can decode. However, calmodulin activity is essentially limited to signalling through the N‐lobe, C‐lobe or both lobes. Yet calmodulin can regulate dozens of proteins including various types of Ca²⁺ channel, Ca²⁺‐activated K⁺ channel and TRP ion channels as well as transporters such as the plasma membrane Ca²⁺‐ATPAse pump along with enzymes including calcineurin, Ca²⁺–calmodulin‐dependent protein kinases and myosin light chain kinase. Therefore the ability of a Ca²⁺ sensor like calmodulin to detect only amplitude and frequency components of the cytosolic Ca² signal is not sufficient to guarantee specificity. As will be elaborated upon below, a fundamental mechanism that ensures specificity to the Ca²⁺ signal is its spatial profile.

Cytosolic Ca²⁺ oscillations: the AM and FM catechism

In non‐excitable cells, stimulation of G protein‐coupled cell‐surface receptors that activate phospholipase C often generates repetitive cytosolic Ca²⁺ oscillations through regenerative Ca²⁺ release from the endoplasmic reticulum by InsP₃ (Thomas et al. 1996). In a series of seminal papers, agonist‐elicited cytosolic Ca²⁺ oscillations were predicted from analysis of oscillations in membrane potential and membrane conductance, following the repetitive opening and closing of Cl⁻ channels (Prince & Berridge, 1973; Berridge & Rapp, 1979). Cytosolic Ca²⁺ oscillations were first observed in hepatocytes stimulated with vasopressin or phenylephrine (Woods et al. 1986), and have now been found in many different cells types and in response to activation of numerous cell‐surface receptors (Thomas et al. 1996). Oscillations in cytosolic Ca²⁺ concentration in response to receptor stimulation are thought to confer numerous signalling advantages over a sustained rise in Ca²⁺ (Parekh, 2011). First, because high amplitude Ca²⁺ signals are attained only transiently during each oscillation, the toxic effects that arise from a sustained elevation in Ca²⁺ are circumvented. Second, oscillatory Ca²⁺ signals support prolonged stimulation of downstream responses such as gene expression because they minimize Ca²⁺‐dependent inactivation of Ca²⁺ channels. An example of this is seen following activation of cysteinyl leukotriene type I receptors, which are Gq coupled and generate InsP₃ through an increase in phospholipase C activity. Stimulation of these receptors in rat basophilic leukaemia (RBL) mast cells with the agonist leukotriene C₄ (LTC₄) evokes a series of cytosolic Ca²⁺ oscillations that lead to increased transcription and translation of the immediate early gene c‐fos, a component of the AP‐1 transcription factor complex (Ng et al. 2012). Cysteinyl leukotriene receptors exhibit homologous desensitization, mediated by protein kinase C‐dependent phosphorylation of three serine residues in the carboxy terminus, followed by receptor internalization (Naik et al. 2005). Pharmacological inhibition, down‐regulation or knockdown of protein kinase Cα all converted the oscillatory cytosolic Ca²⁺ response into a more sustained Ca²⁺ rise but now c‐fos expression was inhibited (Ng et al. 2012). The elevated Ca²⁺ signal following loss of receptor desensitization led to Ca²⁺‐dependent inactivation of store‐operated Ca²⁺ release‐activated Ca²⁺ (CRAC) channels, the latter providing the trigger Ca²⁺ for gene expression (Ng et al. 2009). Overexpression of the scaffolding protein caveolin‐1 also accelerated loss of cytosolic Ca²⁺ oscillations through enhanced desensitization of cysteinyl leukotriene receptors (Yeh et al. 2014). This effect arose from increased coupling between receptor and phosholipase C, resulting in larger Ca²⁺ release transients. These Ca²⁺ signals activated Ca²⁺‐dependent protein kinase C, leading to more rapid receptor desensitization.

A third signalling advantage that is thought to arise from an oscillatory Ca²⁺ signal is that it increases the signal‐to‐noise ratio by enhancing the ability of low levels of stimulation to activate responses. A few large amplitude Ca²⁺ spikes will be more reliably detected by a Ca²⁺ sensor than a smaller sustained increase over the same length of time that produces the same overall integrated Ca²⁺ signal.

Finally, stemming from the observation that different concentrations of an agonist evoke Ca²⁺ oscillations with different frequencies, it is thought that oscillatory Ca²⁺ signals impart information in both the amplitude and frequency of the Ca²⁺ oscillations, thereby enabling selective recruitment of downstream responses. There are several examples that nicely document how different frequencies of Ca²⁺ oscillation activate distinct targets. In Jurkat T cells treated with the SERCA pump blocker thapsigargin in Ca²⁺‐free external solution to deplete stores and open store‐operated CRAC channels in the plasma membrane, repetitive pulses of external Ca²⁺ led to regular cytosolic Ca²⁺ oscillations (Dolmetsch et al. 1998). Varying the time interval between the external Ca²⁺ pulses resulted in Ca²⁺ oscillations with different frequencies. The pro‐inflammatory Ca²⁺‐dependent transcription factors NFAT, NFκB and Oct/OAP were found to respond to different frequencies of cytosolic Ca²⁺ oscillation: all three responded to high frequencies but only NFκB was activated by low frequency Ca²⁺ pulses (Dolmetsch et al. 1998). In RBL mast cells, Ca²⁺ oscillations evoked following photolytic release of InsP₃ were found to control gene expression, in a frequency‐dependent manner (Li et al. 1998). Frequency‐dependent coding of Ca²⁺ signals from voltage‐gated Ca²⁺ channels has also been reported for neurite extension (Gu & Spitzer, 1995).

Cytosolic Ca²⁺ oscillations arising from InsP₃‐dependent Ca²⁺ release from the endoplasmic reticulum (ER) have also been found to regulate mitochondrial metabolism (Hajnoczky et al. 1995). In hepatocytes, stimulation with the G protein‐coupled receptor agonist vasopressin triggered a series of cytosolic Ca²⁺ oscillations that arose from regenerative Ca²⁺ release from the ER. These oscillations invaded the mitochondria and resulted in oscillations in matrix Ca²⁺ that were largely in phase with the cytosolic ones. Each mitochondrial Ca²⁺ spike led to robust activation of Ca²⁺‐sensitive mitochondrial dehydrogenases, as judged through measurements of NAD(P)H or FAD autofluorescence. Ca²⁺ oscillations at > 0.5 min⁻¹ led to a sustained increase in mitochondrial metabolism. By contrast, a sustained cytosolic Ca²⁺ increase enhanced mitochondrial metabolism only transiently (Hajnoczky et al. 1995). Collectively, these findings suggested that increasing the frequency of cytosolic and thereby mitochondrial matrix Ca²⁺ oscillations maintained a high level of mitochondrial metabolism. Hence mitochondria were tuned to oscillating cytosolic Ca²⁺ signals, transducing the latter into robust changes in mitochondrial respiration.

Local Ca²⁺ signals: ménage a trois

If all the signalling information contained within Ca²⁺ oscillations is restricted to amplitude and frequency components, then a simple prediction would be that Ca²⁺ oscillations of the same amplitude and frequency to the same agonist in the same cell type should be equally effective in activating downstream responses. We tested this by taking advantage of a neat method described by Bird & Putney (2005). During a cytosolic Ca²⁺ oscillation, a sizeable fraction of the Ca²⁺ released from the ER is taken back into the store but a small portion is transported out of the cell by the plasma membrane Ca²⁺‐ATPase pump. In the presence of external Ca²⁺, store‐operated Ca²⁺ influx compensates for the export of Ca²⁺ from the cell and thus refills the ER with Ca²⁺ in readiness for the oscillatory cycle. In the absence of external Ca²⁺, however, Ca²⁺ export out of the cell is no longer matched by Ca²⁺ influx, resulting in a decrease in ER Ca²⁺ content below the level needed to maintain Ca²⁺ oscillations. Bird and Putney showed that inhibition of Ca²⁺ transport across the plasma membrane by exposure to Gd³⁺ or La³⁺, rare earth metal trivalent cations that block Ca²⁺ entry through CRAC channels as well as Ca²⁺ extrusion by the plasma membrane Ca²⁺‐ATPase pump, rendered the cell tight to Ca²⁺. Ca²⁺ released from the ER by muscarinic receptor activation was no longer exported from the cell and so was re‐sequestrated in the store. Repetitive cytosolic Ca²⁺ oscillations to receptor stimulation were therefore evoked in the absence of external Ca²⁺. We applied this approach to RBL mast cells, using cysteinyl leukotriene receptor activation to generate oscillatory cytosolic Ca²⁺ signals (Di Capite et al. 2009). Stimulation with a range of LTC₄ concentrations in the presence of external Ca²⁺ evoked Ca²⁺ oscillations that were identical, for a given agonist dose, in amplitude and frequency to those obtained in Ca²⁺‐free external solution supplemented with La³⁺. However, only those cytosolic Ca²⁺ oscillations that occurred in the presence of open CRAC channels activated expression of c‐fos (Di Capite et al. 2009). Similar findings were made for NFAT1 activation (Kar et al. 2011). These findings demonstrated a major role for the spatial profile of the Ca²⁺ oscillation, specifically Ca²⁺ signals near open CRAC channels, rather than amplitude or frequency in driving excitation–transcription coupling.

CRAC channel Ca²⁺ microdomains and transcription factor activation

The diffusion of Ca²⁺ through open Ca²⁺ channels, such as CRAC channels in the plasma membrane or InsP₃‐gated channels in the ER, leads to the rapid build‐up of a microdomain of Ca²⁺ near the mouth of the channel. The microdomain decays steeply with distance from the channel. Within a few nanometres of the channel, endogenous Ca²⁺ buffers are too slow to capture the incoming Ca²⁺ and this, coupled with the tiny volume near the channel mouth, results in a high local Ca²⁺ concentration close to the channel pore (Neher, 1998; Parekh, 2008). Depending on the unitary Ca²⁺ flux through the channel, the Ca²⁺ microdomain can be several times larger than the global Ca²⁺ rise. Very close to the channel pore, intracellular Ca²⁺ buffers have no effect on the local Ca²⁺, but can reduce the lateral spread of the Ca²⁺ signal in a manner dictated by how quickly they bind Ca²⁺. Slow chelators like EGTA (on‐rate of ∼10⁶ m ⁻¹ s⁻¹) are unable to capture incoming Ca²⁺ and therefore have no effect on the lateral expanse of the Ca²⁺ microdomain. By contrast, the fast Ca²⁺ chelator BAPTA (on‐rate of ∼5 × 10⁸ m ⁻¹ s⁻¹) reduces the lateral spread and therefore sharpens the Ca²⁺ microdomain (Neher, 1998).

c‐fos gene expression or activation of NFAT by Ca²⁺ entry through CRAC channels was unaffected by loading the cytosol with EGTA but was impaired by BAPTA, consistent with a role for Ca²⁺ microdomains in coupling receptor activation to nuclear gene expression (Di Capite et al. 2009). Moreover, NFAT activation (Kar et al. 2011) or c‐fos expression (Ng et al. 2009) was tightly linked to the unitary flux through CRAC channels and independent of the bulk Ca²⁺ rise. Finally, despite raising bulk Ca²⁺ to a similar level to that following CRAC channel opening, Ca²⁺ flux through plasmalemmal TRPC3 channels was unable to activate NFAT1 (Kar et al. 2011). Collectively, these results suggest that local Ca²⁺ entry through CRAC channels is important for signalling to the nucleus.

For Ca²⁺ microdomains at the plasma membrane to signal to the nucleus, a transduction pathway is needed. In RBL mast cells, local Ca²⁺ entry through CRAC channels activates the non‐receptor tyrosine kinase Syk, which then phosphorylates and activates the STAT5 transcription factor (Ng et al. 2009), mainly STAT5a (Yeh & Parekh, 2015). Both immuno‐cytochemical studies and co‐immunoprecipitation experiments have revealed that Syk associates with Orai1 (Ng et al. 2009; Samanta et al. 2015), the pore‐forming subunit of the CRAC channel (Prakriya et al. 2006; Vig et al. 2006; Yeromin et al. 2006). However, it is not clear whether this is a direct or indirect association. Regardless, these findings suggest that Syk will be held close to the CRAC channel Ca²⁺ microdomain, facilitating c‐fos activation. A slightly different mechanism links these Ca²⁺ microdomains to the NFAT pathway. At rest, NFAT proteins are extensively phosphorylated and are trapped within the cytosol (Hogan et al. 2003). Following a rise in cytosolic Ca²⁺, Ca²⁺–calmodulin binds to and activates the protein phosphatase calcineurin, the target for the immunosuppressants cyclosporine A and tacrolimus. Active calcineurin dephosphorylates NFAT and thereby leads to the exposure of a nuclear localization sequence, which enables NFAT to migrate into the nucleus. As in some other cell types (Li et al. 2012), a fraction of the cellular pools of calcineurin and NFAT are bound to the anchoring protein AKAP79 in RBL cells (Kar et al. 2014). Under non‐stimulated conditions, Orai1 and AKAP79 do not co‐immunoprecipitate but do so after store depletion (Kar et al. 2014). In this way, calcineurin is brought close to the trigger Ca²⁺ and, once active, has immediate access to its target NFAT.

Spatial profiling at the ER–mitochondrial junction

The spatial profile of a cytosolic Ca²⁺ oscillation, particularly the Ca²⁺ microdomain near open CRAC channels, plays an important role in disseminating the information contained within the oscillatory Ca²⁺ signal (Di Capite et al. 2009; Kar et al. 2011, 2012). Is ‘spatial profiling’ of an oscillation restricted to responses activated by Ca²⁺ entry or is it also a significant factor in the activation of responses evoked by Ca²⁺ release from the stores? To address this, we re‐investigated the link between cytosolic Ca²⁺ oscillations, oscillations in mitochondrial matrix Ca²⁺ and enhanced mitochondrial metabolism. Pioneering work in hepatocytes demonstrated a correlation between the frequency of global cytosolic Ca²⁺ oscillations and those detected within the matrix (Hajnoczky et al. 1995). One possibility is that the matrix Ca²⁺ oscillations mirror global cytosolic Ca²⁺ oscillations, with the latter propagating into the mitochondria. In this scheme, it is the global Ca²⁺ oscillations that are the drivers of mitochondrial Ca²⁺ oscillations and metabolism (Fig. 1 A). An alternative view stems from morphological and functional studies that have demonstrated close physical coupling between ER and mitochondria (Rizzuto et al. 1993, 1998; Csordas et al. 2006), at specialized sites called mitochondrial‐associated membranes (MAMs). Tethering of ER to mitochondria brings the two organelles within 20 nm of one another and enables Ca²⁺ microdomains near open InsP₃ receptors to be detected by juxtaposed mitochondria, leading to rapid shuttling of Ca²⁺ from the ER into the matrix (Rizzuto et al. 1998; Moreau et al. 2006). In this scheme, it is the pulsatile Ca²⁺ release from InsP₃ receptors that causes mitochondrial Ca²⁺ oscillations, with the latter being driven by Ca²⁺ microdomains near the open InsP₃‐gated Ca²⁺ channels rather than the global cytosolic Ca²⁺ signal (Fig. 1 B).

Figure 1. Two possible mechanisms that link Ca²⁺ release to mitochondrial Ca²⁺ oscillations.

In model 1, Ca²⁺ release by InsP₃ from the endoplasmic reticulum (ER) leads to oscillations in cytosolic Ca²⁺. These then propagate into the mitochondria to elicit oscillations in matrix Ca²⁺. In model 2, the ER is tethered to mitochondria so pulsatile Ca²⁺ release from InsP₃ receptors rapidly enter the matrix.

One way of distinguishing between these two possibilities is to increase the buffering capacity of the cytosol, dampening cytosolic Ca²⁺ oscillations but leaving Ca²⁺ microdomains near open InsP₃ receptors intact. Stimulation of cysteinyl leukotriene type I receptors in RBL mast cells with a low concentration of LTC₄ evoked a series of cytosolic Ca²⁺ oscillations (Fig. 2 A), which led to oscillations in matrix Ca²⁺ (Fig. 2 E). Oscillatory cytosolic Ca²⁺ signals in RBL cells enter the matrix by permeating through the mitochondrial uniporter channel (Samanta et al. 2014), a protein that is in the inner mitochondrial membrane (Baughman et al. 2011; De Stefani et al. 2011). Flux through the MCU is determined by the electrochemical gradient, a major component of which arises from the hyperpolarized potential (∼ –180 mv) across the inner mitochondrial membrane. Dissipation of this potential with the protonophore FCCP inhibited mitochondrial Ca²⁺ signalling in response to LTC₄ (Samanta et al. 2014).

Figure 2. Matrix Ca²⁺ oscillations are retained despite increasing cytosolic Ca²⁺ buffering with parvalbumin.

A, oscillations in cytosolic Ca²⁺, measured with fura 2, are shown for a mock‐transfected control RBL cell challenged with 160 nm LTC₄. B, cytosolic Ca²⁺ oscillations are shown for a cell expressing parvalbumin with a nuclear export sequence (PV‐NES‐RFP). C, aggregate data comparing the number of oscillations per 200 s recording bin (taken from when the oscillations started) are shown. Each point represents data from > 20 cells from two independent experiments. D, as in panel C, but the amplitude of each oscillation is shown. E, matrix Ca²⁺ signals, measured with the mitochondria targeted pericam, are shown following challenge with LTC₄. F, matrix Ca²⁺ is shown following stimulation with LTC₄ in a cell expressing PV‐NES. G and H, aggregate data for the number of matrix Ca²⁺ oscillations (G; mean of 12 cells from three experiments) and amplitude of each oscillation (H) are compared. Fura 2 and pericam measurements were carried out as described previously (Samanta et al. 2014).

To increase cytosolic Ca²⁺ buffering, we expressed the calcium binding protein parvalbumin that had been engineered to express a nuclear export sequence (PV‐NES) and that is located exclusively in the cytoplasm (Pusl et al. 2002; Kar & Parekh, 2015). Stimulation with LTC₄ now failed to evoke the typical series of cytosolic Ca²⁺ oscillations that were characteristic of control cells (Fig. 2 B). Instead, after an initial Ca²⁺ spike, only one or two small oscillations followed. Aggregate data comparing the number of oscillations in 200 sec bins after stimulation and the amplitude of each oscillation between control cells and those expressing PV‐NES are shown in Fig. 2 C and D.

By contrast, despite the loss of sustained oscillatory Ca²⁺ signals in the presence of PV‐NES (Fig. 2 B), oscillations in matrix Ca²⁺ still occurred following leukotriene receptor stimulation (Fig. 2 F–2 H). Although the frequency of the matrix Ca²⁺ oscillations was similar between control and PV‐NES expressing cells (Fig. 2 G), the amplitudes were reduced by PV‐NES (Fig. 2 H). Nevertheless, these results show that mitochondrial Ca²⁺ oscillations occur in the absence of global cytosolic Ca²⁺ oscillations. The data argue against the scheme in Fig. 1 A and are instead consistent with Fig. 1 B.

Stimulation with LTC₄ in control cells increased mitochondrial metabolism, measured as a decrease in FAD autofluorescence (Fig. 3 A and B). Although the reduction in FAD induced by LTC₄ was small (∼5%), it is similar in extent to that seen in hepatocytes in response to vasopressin (Hajnoczky et al. 1995). Pre‐treatment with FCCP inhibited the changes in FAD in response to cysteinyl leukotriene receptor activation by ∼70% (Fig. 3 B), consistent with the signal emanating from mitochondria. In the presence of PV‐NES, FAD levels still fell upon challenge with LTC₄ but to a level ∼40% less than control cells (Fig. 3 A and B).

Figure 3. Stimulated mitochondrial metabolism by LTC₄ is only partially reduced following PV‐NES expression.

A, FAD fluorescence was measured as an indicator of mitochondrial metabolism. Stimulation with LTC₄ decreased flavin fluorescence, indicating reduction of FAD. Mitochondrial depolarization with the combination of FCCP (1 μm) and oligomycin (0.2 μm) (labelled as F/O) suppressed the increase in metabolism induced by LTC₄. By contrast, expression of PV‐NES partially reduced the effects of LTC₄. B, aggregate data are compared for the conditions shown. Each bar is the mean of > 14 cells. To monitor FAD fluorescence, cells were excited at 470 nm and emission was collected at > 535 nm.

Collectively, these results suggest pulsatile formation of Ca²⁺ microdomains near open InsP₃ receptor channels can stimulate mitochondrial metabolism, independent of global oscillations in cytosolic Ca²⁺ in response to cysteinyl leukotriene receptor activation. However, these Ca²⁺ microdomains are not as effective as the global cytosolic Ca²⁺ oscillations in stimulating the Krebs cycle.

The cytosolic Ca²⁺ oscillation: a veritable source of information

Cytosolic Ca²⁺ oscillations are generally believed to contain information in both the amplitude and frequency of the spikes. Our recent studies have revealed that the subcellular spatial profile of the Ca²⁺ oscillation, particularly Ca²⁺ microdomains near CRAC channels at the plasma membrane, also contributes to the deciphering of the Ca²⁺ signal. Following stimulation of cysteinyl leukotriene type I receptors, typical members of the G protein‐coupled receptor superfamily, it is local Ca²⁺ signals near CRAC channels and not the global Ca²⁺ oscillations themselves that activate gene expression (Di Capite et al. 2009; Kar et al. 2011, 2012). Ca²⁺ oscillations in non‐excitable cells are generally triggered by Ca²⁺ release from InsP₃‐gated channels (Wakui et al. 1990). We now show that Ca²⁺ microdomains near these channels can propagate into the mitochondrial matrix, leading to oscillations in matrix Ca²⁺ and increased mitochondrial metabolism. However, oscillations in cytosolic Ca²⁺ were ∼2‐fold more effective than Ca²⁺ microdomains near open InsP₃‐gated channels in stimulating metabolism. This probably reflects the relatively low affinity of the mitochondrial dehydrogenases for matrix Ca²⁺, requiring an elevation of > 1.5 μm for strong stimulation (Jouaville et al. 1999). Local and global cytosolic Ca²⁺ signals may therefore interact to ensure robust activation of mitochondrial metabolism, with the global rise amplifying the rise in matrix Ca²⁺ induced by the local signals. It is tempting to speculate that the Ca²⁺ microdomains near open InsP₃‐gated channels serve to prime or kick‐start mitochondrial metabolism, in anticipation of a global cytosolic Ca²⁺ rise which then increases mitochondrial activity considerably.

InsP₃‐gated Ca²⁺ channels show a bell‐shaped dependence on cytosolic Ca²⁺; modest elevation of local Ca²⁺ increases channel activity whereas the channels inactivate as cytosolic Ca²⁺ rises further (Bezprozvanny et al. 1991; Finch et al. 1991; Ivorra & Parker, 1992). As the Ca²⁺ microdomain collapses around an inactivated non‐conducting channel, the channel will re‐open in the presence of InsP₃, leading to pulsatile Ca²⁺ release. It is likely that these pulsatile Ca²⁺ release events from clusters of InsP₃ receptors close to MAMs account for the oscillatory Ca²⁺ signals in the mitochondrial matrix, although further work is needed to address this directly.

Collectively, these results reinforce the view that the spatial profile of a Ca²⁺ oscillation is an important source of information that is decoded by sensors within the realm of the Ca²⁺ microdomain to activate downstream responses.

Ca²⁺ release from the ER occurs in response to opening of large conductance InsP₃‐gated Ca²⁺ channels (Vais et al. 2010), which generate microdomains of high Ca²⁺ with a wide lateral expanse (Bakowski et al. 2012). The reduction in Ca²⁺ content that develops following Ca²⁺ release leads to the opening of plasmalemmal CRAC channels, which cluster to produce microdomains of Ca²⁺ at the cell periphery (Hogan et al. 2010; Parekh, 2010). Both sets of Ca²⁺ microdomain activate distinct cellular responses: mitochondrial metabolism and gene expression, respectively. Therefore in addition to amplitude and frequency coding, the two spatially separate Ca²⁺ microdomains generated during an oscillatory Ca²⁺ signal provide effective mechanisms to activate specific responses (see Abstract figure). Since Ca²⁺ release from InsP₃ receptors can trigger Ca²⁺‐induced Ca²⁺ release (Petersen, 2014) and G protein‐coupled receptors can recruit additional messengers like nicotinic acid adenine dinucleotide phosphate (NAADP) (Lee, 2012), a rich tapestry of spatially segregated Ca²⁺ microdomains can be evoked, in an agonist‐ and concentration‐specific context, which will greatly enhance the signalling power of cytosolic Ca²⁺ oscillations above simple amplitude‐ and frequency‐dependent coding.

Additional information

Competing interests

The authors state that there are no competing interests.

Author contribution

The work was performed in the Department of Physiology, Anatomy and Genetics at the University of Oxford. K.S. collected the data. K.S. and A.B.P. analysed the data and were responsible for the design of experiments. A.B.P. wrote the manuscript. Both authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

Work in the authors’ laboratory is supported by MRC Programme grant L01047X/1 to A.B.P.

Biographies

Krishna Samanta studied Biochemistry at Kalyani University in West Bengal, India. Her PhD, in the Department of Biochemistry and Biophysics also at Kalyani University, purified and characterized the intracellular protease m‐calpain from the endoplasmic reticulum. She then joined Anant Parekh's group as a postdoctoral fellow, where she has focused mainly on the role of mitochondria in controlling CRAC channel activity. Following his medical studies at Oxford University.

Anant Parekh obtained his DPhil from the Pharmacology Department at Oxford. He carried out postdoctoral work in Erwin Neher's department at the Max Planck Institute for biophysical Chemistry in Goettingen, Germany working with Walter Stuehmer and then Reinhold Penner. He then joined the Physiology Department at Oxford and became a tenured Professor. His principal research interests centre on Ca²⁺ channels, how intracellular Ca²⁺ signals are decoded and how this is hijacked in diseases such as allergy and asthma.