Calcium-based signalling systems in guard cells

Authors


Abstract

Summary

Calcium is a ubiquitous intracellular signal responsible for controlling numerous cellular processes in both plants and animals. As an example, Ca2+ has been shown to be a second messenger in the signal transduction pathways by which stomatal guard cells respond to external stimuli. Regulated increases in the cytosolic concentration of free calcium ions ([Ca2+]cyt) in guard cells have been observed to be a common intermediate in many of the pathways leading to either opening or closing of the stomatal pore. This observation has prompted investigations into how specificity is encoded in the Ca2+ signal. It has been suggested that the key to generating stimulus-specific calcium signatures lies in the ability to access differentially the cellular machinery controlling calcium influx and release from intracellular stores. Several important components of the calcium-based signalling pathways have been identified in guard cells including cADPR, phospholipase C–InsP3, InsP6 and H2O2. These data suggest that the pathways for intracellular mobilization of Ca2+ are evolutionarily conserved between plants and animals.

Abbreviations

ABA, abscisic acid; [Ca2+]cyt, cytosolic free calcium concentration; [Ca2+]ext, external calcium concentration; IK,in; inward-rectifying K+ currents; InsP3, inositol-1,4,5-trisphosphate; InsP6, inositol hexakisphosphate; PLC, phospholipase C; PLD, phospholipase D; PA, phosphatidic acid; H2O2, hydrogen peroxide; AAPK, ABA-activated serine-threonine protein kinase; cADPR, cyclic adenosine 5′-diphosphoribose; U73122, 1-(6-{[17â-3-methoxyestra-1,3,5(10)-trien-17-yl]amino}hexyl)-1H-pyrrole-2, 5-dione; RyR; ryanodine receptor; CICR; calcium-induced calcium-release; ICa, inward calcium current.

Introduction

Stomata form pores on leaf surfaces that regulate the uptake of CO2 for photosynthesis and the loss of water vapour during transpiration. Many stimuli (Table 1) regulate the aperture of the stomata and they achieve this by controlling the turgor of the two guard cells that surround the stomatal pore (Willmer & Fricker, 1996). Turgor changes are driven by fluxes of K+ and anions through ion channels located in the plasma and vacuolar membranes (MacRobbie, 1998). Guard cells are a good model for studying signal transduction in plants because it is easy to quantify changes in turgor of the guard cell pairs in response to stimuli. Additionally, the activities of ion channels, which represent final effectors in stimulus-response coupling, are well characterized (MacRobbie, 1998; Assmann & Armstrong, 1999).

Table 1. Stimulus-induced increases in guard cell [Ca2+]cyt. Citation based on initial report
 Reference
Closure
ABA McAinsh et al. (1990)
Elevated [CO2] Webb et al. (1996)
Oxidative Stress McAinsh et al. (1996)
High external Ca2+ Gilroy et al. (1991)
Low external K+ Gilroy et al. (1991)
Opening
Auxin Irving et al. (1992)
Fusicoccin Irving et al. (1992)

Identifying the individual components that make up the signal transduction pathways linking signal perception to alterations in guard cell turgor has been the subject of considerable attention (MacRobbie, 1998; Assmann & Armstrong, 1999; Blatt, 2000; McAinsh et al., 2000). Calcium is a ubiquitous second messenger in plants and animals (Sanders et al., 1999; Berridge et al., 2000). An increase in the cytosolic concentration of free calcium ions ([Ca2+]cyt), above a resting level of 0.10 µM, is an important element in the regulation of numerous cellular responses (Sanders et al., 1999; Berridge et al., 2000). An increase in [Ca2+]cyt has also been shown to be a common intermediate in the signalling pathways leading to either opening or closing of the stomatal pore (Blatt, 2000; McAinsh et al., 2000). This observation has prompted investigations into how specificity is controlled in plant calcium-based signalling systems. One possible explanation is that each stimulus generates a unique increase in [Ca2+]cyt. The spatial and temporal components of this increase in [Ca2+]cyt, or ‘calcium signature’ as it is sometimes called, then dictate the outcome of the final response (Hetherington et al., 1998; McAinsh & Hetherington, 1998; Blatt, 2000; McAinsh et al., 2000). It has been suggested that the key to generating calcium signatures lies in the ability to access differentially the cellular machinery controlling calcium influx and release from internal stores (Sanders et al., 1999; Blatt, 2000; McAinsh et al., 2000).

Ca2+ as a signal in guard cells

Although guard cells are competent to respond to numerous external stimuli (Table 1) by generating an increase in [Ca2+]cyt, it is the signal transduction cascade by which abscisic acid (ABA) brings about stomatal closure that has been most inten sively investigated. ABA builds up in leaves during conditions of decreased water availability leading to a reduction in the aperture of the stomatal pore. Decreasing transpirational water loss by this mechanism enables the plant to conserve water and hence tolerate periods of drought. The site of ABA perception has been a subject of intense investigation. There is evidence to suggest that there are two sites of ABA perception, extracellular (Hartung, 1983; Curvetto & Delmastro, 1986; Anderson et al., 1994) and intracellular (Schwartz et al., 1994; Allan et al., 1994; Hamilton et al., 2000). However, the failure to identify a guard cell ABA receptor has led to much contention about the site of ABA perception. MacRobbie (1995) demonstrated that both external and internal ABA are required for the full stomatal response, and that there is a threshold level of ABA that must first be sensed before full vacuolar efflux of K+(Rb+) is achieved.

De Silva et al. (1985) first showed a highly significant interaction between the effects of Ca2+ and ABA on stomatal aperture. Definitive evidence for the involvement of Ca2+ in the response of guard cells to ABA was provided by McAinsh et al. (1990) who showed that ABA induced an elevation in [Ca2+]cyt. This elevation in [Ca2+]cyt preceded any change in stomatal aperture in about 80% of the cells measured, indicating that Ca2+ is involved in transducing the ABA signal into a decrease in guard cell turgor. However, there was no observable increase in [Ca2+]cyt in the remaining 20% of guard cells although these cells responded to ABA by a decrease in turgor. Other studies have also observed ABA-induced increases in [Ca2+]cyt (Gilroy et al., 1991; Irving et al., 1992; Allan et al., 1994). Recently, Allen et al. (1999a, 2000) showed that 85% to 100% of Arabidopsis thaliana guard cells expressing the Ca2+-sensitive yellow cameleon proteins respond to ABA by generating an increase in [Ca2+]cyt. These data are in good agreement with those of Schroeder & Hagiwara (1989), who showed that elevation of [Ca2+]cyt inhibited inward-rectifying K+ currents (IK,in) by blocking the inward-rectifying K+ channels. In addition, Schroeder & Hagiwara (1989) showed that elevation of [Ca2+]cyt activated a voltage-dependent anion-permeable channel that would lead to membrane depolarization and further activate K+ efflux through outward-rectifying K+ channels. The ability of ABA to induce an elevation in [Ca2+]cyt was further confirmed by Schroeder & Hagiwara (1990) who showed that ABA caused repetitive increases in [Ca2+]cyt as a result of passive influx of Ca2+ into the cytosol through nonselective Ca2+-permeable channels. More recently, Hamilton et al. (2000) demonstrated the presence of Ca2+-selective channels on the plasma membrane of V. faba guard cells that are activated by hyperpolarization and ABA, leading the authors to suggest that ABA-induced influx of Ca2+ through these channels may be responsible in part for elevations in [Ca2+]cyt.

The available evidence suggests a second messenger role for Ca2+ in ABA-mediated turgor signalling in stomatal guard cells. The activity of ion channels located in both the plasma and vacuolar membranes, which represent the final effectors in ABA stimulus-response coupling, have been shown to be modulated by [Ca2+]cyt (Assmann & Armstrong, 1999; Blatt, 2000; MacRobbie, 2000). It has been demonstrated that the activity of the inward-rectifying K+ channels is modulated by increases in [Ca2+]cyt (Schroeder & Hagiwara, 1990; Lemtiri-Chlieh & MacRobbie, 1994). Elevations in [Ca2+]cyt have also been shown to activate slow anion channels in the plasma membrane responsible for anion efflux during stomatal closure (Allen et al., 1999b).

In addition to inhibiting the inward-rectifying K+ channels, and activating slow anion channels, elevations in [Ca2+]cyt can also modulate the activity of the H+-ATPase. Increases in [Ca2+]cyt from 0.3 to 1.0 µM have been shown to inhibit the plasma membrane H+-ATPase in guard cells of V. faba (Kinoshita et al., 1995). This inhibitory effect can be reversed by BAPTA (Kinoshita et al., 1995). The inhibition of H+-ATPase by elevations in [Ca2+]cyt is thought to occur via intracellular Ca2+ mobilization rather than Ca2+ influx (Shimazaki et al., 1999). The activity of the H+-ATPase, which can be inhibited by ABA (Goh et al., 1992), has been shown to be important in electrogenic H+ pumping required for the membrane hyperpolarization that drives IK,in. In addition, the inhibition of H+ pumping by ABA-induced elevations in [Ca2+]cyt prevents membrane hyperpolarization thereby indirectly contributing to membrane depolarization caused by Ca2+ influx through nonselective Ca2+-permeable (Schroeder & Hagiwara, 1990) and voltage-sensitive Ca2+-selective (Hamilton et al., 2000) channels.

Elevations in [Ca2+]cyt have also been shown to affect the activities of vacuolar ion channels: the VK- and SV-channels. The VK channel is a highly selective K+ channel that gates the efflux of K+ from the vacuole to the cytosol and is activated by increases in [Ca2+]cyt above 0.1 µM (Ward & Schroeder, 1994; Allen & Sanders, 1996). The SV channel is permeable to both K+ and Ca2+ and is activated by increases in [Ca2+]cyt above 0.6 µM (Ward & Schroeder, 1994; Allen & Sanders, 1996). It has been proposed that ABA-induced elevations in [Ca2+]cyt activate the VK channel. The efflux of K+ through the VK channel in turn makes the cytosol sufficiently electro-positive to activate the efflux of K+ through the SV channels. Although vacuolar efflux of K+ contributes substantially to stomatal closure, consideration must be given to ABA-induced loss of vacuolar anions as this will affect the vacuolar membrane potential. This charge-balancing effect of the anionic efflux will affect VK channel gating. The characterization and identification of ion channels that gate vacuolar anion efflux is of considerable importance. Their identification will advance our understanding of guard cell signal transduction and ionic relations, allowing the study of how elevations in [Ca2+]cyt regulate vacuolar anion efflux.

Although ABA-induced elevations in [Ca2+]cyt correlate with inward Ca2+ currents, intracellular Ca2+ release can also contribute to the observed increases in [Ca2+]cyt. Evidence for Ca2+ influx across the plasma membrane has been provided by pharmacological and tracer flux studies (De Silva et al., 1985; MacRobbie, 1989; McAinsh et al., 1991), Ca2+ imaging studies (McAinsh et al., 1990, 1992; Schroeder & Hagiwara, 1990; Gilroy et al., 1991) and patch-clamp analysis (Schroeder & Hagiwara, 1990; Hamilton et al., 2000). It is important to note that these patch-clamp studies showed that inward Ca2+ currents could arise from the activity of different Ca2+ channels. Schroeder & Hagiwara (1990) showed the presence of nonselective Ca2+-permeable channels whereas Hamilton et al. (2000) reported the presence of Ca2+-selective channels that are activated by hyperpolarization and ABA. The activities of these Ca2+-selective channels are potentiated by ABA. ABA increases the probability that these Ca2+-selective channels are open and alters their voltage sensitivity, as evidenced by greater channel activation at increasingly depolarized membrane potentials induced by ABA (Hamilton et al., 2000). Greater activation of these Ca2+-selective channels by ABA and ABA-induced shifts to more positive voltage thresholds, together with membrane depolarization resulting from inward Ca2+ currents induced by ABA contribute to the driving force for K+ efflux from the outward-rectifying K+ channels.

Pharmacological and Ca2+ imaging studies have also provided evidence for the release of Ca2+ from intracellular stores (McAinsh et al., 1992; McAinsh et al., 1995). In addition, photolysis of caged-inositol trisphosphate (InsP3), ABA and Ca2+ (Gilroy et al., 1990; Allan et al., 1994; McAinsh et al., 1995) and microinjection of cADPR directly into the cytosol (Leckie et al., 1998) have implicated the release of Ca2+ from intracellular stores. Recent work has highlighted the importance of several Ca2+ mobilizing pathways in the generation of stimulus-induced increases in guard cell [Ca2+]cyt. These are the subjects of this review.

Phospholipase C, InsP3- and InsP6-mediated intracellular Ca2+-mobilizing pathways

The signal transduction pathway leading from perception of ABA to elevations in [Ca2+]cyt has been the subject of intense investigation. There is increasing evidence for the role of phosphoinositide-specific phospholipase C (PI-PLC)-based signalling in plants. Gilroy et al. (1990) demonstrated that UV-photolysis of caged-InsP3 microinjected into the cytosol of guard cells elevated [Ca2+]cyt and stimulated stomatal closure. This artificial elevation of cytosolic InsP3 could also reversibly inactivate IK,in, whilst at the same time activating an inward current that depolarizes the plasma membrane (Blatt et al., 1990). The inactivation of IK,in by InsP3 is indirect and likely to be mediated by Ca2+ through the action of a calcineurin-related protein phosphatase (Luan et al., 1993). In addition, many elements of an InsP3-mediated signalling pathway have been identified in stomatal guard cells (Parmer & Brearley, 1993, 1995) and Lee et al. (1996) showed that ABA induced the rapid turnover of phosphoinositide in guard cell protoplasts. These results suggest that InsP3-mediated Ca2+ release from intracellular stores may be important in regulating changes in guard cell turgor.

Recently, Staxén et al. (1999) used a pharmacological approach to demonstrate the involvement of a PI-PLC-InsP3 Ca2+-mobilizing pathway in the regulation of guard cell turgor by ABA. U73122, an aminosteroid inhibitor of PI-PLC-dependent processes in animals, inhibited in a dose-dependent manner, the activity of a recombinant delta-like PLC isolated from a guard cell-enriched cDNA library. Additionally, U73122 partially attenuated the effects of ABA on guard cell turgor and inhibited ABA-induced oscillations in guard cell [Ca2+]cyt (Fig. 1). The inhibition of ABA-induced oscillations by U73122 did not result from compromised viability as demonstrated by the ability of cells to respond to [Ca2+]ext with a regulated increase in [Ca2+]cyt (Fig. 1b). The effect of U73122 on guard cell turgor was corroborated by the recent observation of MacRobbie (2000) that U73122 interfered with vacuolar efflux of K+(Rb+). Further evidence for the involvement of InsP3-gated Ca2+ release in the regulation of guard cell turgor by ABA was shown by microinjection of heparin, an antagonist of InsP3-gated Ca2+ release, into the cytosol of guard cells. Heparin partially attenuated the effects of ABA on guard cell turgor (C. P. Leckie, C. Ng, M. R. McAinsh, D. Sanders & A. M. Hetherington, unpublished observations). Taken together, these data suggest the involvement of the PI-PLC-InsP3 Ca2+-mobilizing pathway in ABA-mediated turgor signalling in stomatal guard cells.

Figure 1.

The PLC inhibitor, U73122, abolishes ABA-induced oscillations in guard cell [Ca2+]cyt and inhibits ABA-induced stomatal closure in Commelina communis. The effect of a 15-min pulse of 1 µM U73122 on (a) stomatal closure and (b) guard cell [Ca2+]cyt in response to ABA in isolated epidermis of C. communis: control, open bars; 1 µM ABA, solid bars; 1 mM [Ca2+]ext, hatched bars; U73122, cross-hatched bars. Inset: percentage inhibition of 1 µM ABA or 1 mM [Ca2+]ext-induced stomatal closure by a 15-min pulse of 1 µM U73122 or its inactive analogue 1-(6-{[7β-3-methoxyestra-1,3,5(10)-trien-17-yl]amino}hexyl)-2,5-pyrrolidinedione (U73343). Reproduced from Staxén et al. (1999) with permission. © 1999 National Academy of Sciences, USA.

Another inositol phosphate that has been implicated in ABA-mediated turgor changes in guard cells is inositol hexakisphosphate (InsP6). When intact guard cells that have been preloaded with myo-[2–3H]inositol were treated with ABA, a 4–15-fold increase in InsP6 occurred within 5 min (Lemtiri-Chlieh et al., 2000). Furthermore, loading of InsP6 into the cytosol of guard cells inhibited IK,in in a Ca2+-dependent manner. The potency of InsP6, relative to that of InsP3, in inhibiting IK,in suggests a dominant role of InsP6 in ABA regulation of guard cell turgor. It will be of interest to determine the exact nature of the Ca2+-dependence of InsP6-mediated inhibition of IK,in and the modulation of cytosolic Ca2+ dynamics by InsP6.

Cyclic ADP-ribose-mediated Ca2+-mobilizing pathway

There is evidence for the role of cyclic adenosine 5′-diphosphoribose (cADPR), a metabolite of NAD+, in ABA signal transduction in plants. cADPR was first demonstrated to possess Ca2+-mobilizing activity using a sea urchin egg microsomal system (Clapper et al., 1987; Lee et al., 1989). There is also an increasing body of evidence for the Ca2+-mobilizing activity of cADPR in many other cell types (Galione, 1993, 1994, 2000; Lee, 2000). cADPR is synthesized from NAD+ by ADP-ribosyl cyclase or by CD38, a lymphocyte protein, and is degraded to ADP-ribose by cADPR hydrolase (Galione, 1994; Lee et al., 1995). While cADPR has been implicated as a signalling molecule in plants (Wu et al., 1997; Leckie et al., 1998), there is as yet no direct evidence for the presence of ADP-ribosyl cyclase in plants. It has been postulated that the receptor for cADPR is the ryanodine receptor (RyR) (Meszaros et al., 1993). Evidence for the involvement of RyR in cADPR-mediated Ca2+-release stems from studies showing that Ca2+ mobilization in sea urchin egg homogenates can be de-sensitized by prior addition of ryanodine and high levels of caffeine (Galione et al., 1991).

In plants, patch clamp studies on beet tap-root vacuoles and 45Ca2+ flux measurements with beet microsomes demonstrated that cADPR can elicit the release of Ca2+ from endomembrane stores (Allen et al., 1995; Muir & Sanders, 1996; Muir et al., 1997). The pharmacology and sensitivity to cADPR of this Ca2+-release strongly suggest the presence of RyR homologues in plants. Evidence that cADPR is a bona fide signalling molecule in plants has been reported by Wu et al. (1997) and Leckie et al. (1998) who showed the involvement of cADPR in ABA-mediated nuclear and turgor signalling, respectively.

Leckie et al. (1998) showed that pressure microinjection of cADPR into the cytosol of guard cells of C. communis resulted in an increase in [Ca2+]cyt and a decrease in turgor. The increase in [Ca2+]cyt elicited by cADPR ranged from 0.05 to 0.4 µM while injection of ADPR or water resulted in only a transient increase in [Ca2+]cyt that decreased rapidly to resting levels. In most of the cells injected with cADPR, a sustained increase in [Ca2+]cyt, lasting in excess of 10 min, was observed. Occasionally, injection of cADPR resulted in oscillations in [Ca2+]cyt. Furthermore, patch clamp techniques showed that cADPR can elicit Ca2+-release from guard cell vacuoles (Leckie et al., 1998). It is noteworthy that cADPR-induced Ca2+ release was down-regulated by [Ca2+]cyt above 0.6 µM, indicating that cADPR-activated channels cannot be responsible for Ca2+-induced Ca2+-release (CICR). This is in contrast to studies using animal cells, where cADPR-activated channels have been shown to contribute in part to CICR (Galione et al., 1991).

Leckie et al. (1998) used pharmacological agents to demonstrate the involvement of cADPR in ABA turgor signalling in guard cells. Microinjection of 8-NH2-cADPR, a competitive inhibitor of cADPR, into the cytosol of guard cells modulated the effects of ABA-induced decreases in turgor (Fig. 2a). By contrast, ADPR did not interfere with ABA-induced decreases in guard cell turgor (Fig. 2a, inset). In addition, nicotinamide, an inhibitor of cADPR synthesis, partially attenuated ABA-induced decreases in guard cell turgor in a dose-dependent manner (Fig. 2b,c). This observation was reiterated by MacRobbie (2000) who demonstrated the modulation of ABA-induced vacuolar K+(Rb+) efflux using concentrations of nicotinamide similar to those reported by Leckie et al. (1998).

Figure 2.

Inhibitors of the cADP-ribose signalling pathway inhibit ABA-induced stomatal closure in isolated epidermis of Commelina communis. (a) Half stomatal apertures, in response to 1 µM ABA, of a guard cell with 8-NH2-cADP-ribose or ADP-ribose (inset) microinjected into the cytosol (open circles) and of the uninjected guard cell of the same stoma (closed circles). (b) Inhibition of ABA-induced stomatal closure by nicotinamide (open squares); the effects of nicotinamide alone (open triangles). (c) Inhibition of stomatal opening: a, control; b, 50 mM nicotinamide; c, 1 µM ABA; d, 1 µM ABA and 50 mM nicotinamide. Reproduced from Leckie et al. (1998) with permission. © 1998 National Academy of Sciences, USA.

Hydrogen peroxide-mediated Ca2+-mobilizing pathway

Pei et al. (2000) have recently shown that production of hydrogen peroxide (H2O2) is part of the signal transduction pathway mediating ABA-induced changes in guard cell turgor via an increase in [Ca2+]cyt. Treatment of guard cells of A. thaliana with ABA elicited H2O2 production within 1 min (Fig. 3a). In addition, H2O2 activated inward Ca2+ currents (ICa), which could contribute to increases in [Ca2+]cyt as visualized by photometry using the fluorescent Ca2+-sensitive indicator, Fura-2 (Fig. 3b). This is in agreement with previous observations of McAinsh et al. (1996) that H2O2 could elicit increases in [Ca2+]cyt. The data of Pei et al. (2000) suggest that ABA-induced cytosolic H2O2 production is responsible for activating ICa contributing to increases in [Ca2+]cyt. Using the ABA insensitive mutant, gca2, Pei et al. (2000) showed the failure of H2O2 to elicit ICa and decreases in guard cell turgor (Fig. 3c), suggesting that gca2 is downstream of H2O2 production and upstream of or at the level of ICa. The identity of GCA2 remains to be determined but it is envisaged that its identification will contribute to our understanding of the signal transduction pathway linking perception of ABA to changes in guard cell turgor. It would be of interest to investigate this new signalling pathway in the context of existing characterized mutants such as the abi mutants.

Figure 3.

ABA-induction of H2O2 production in guard cells and H2O2 activated Ca2+ influx in guard cells of Arabidopsis thaliana. (a) Time course of ABA-induction of H2O2 production in guard cells of A. thaliana as visualized by the fluorescent indicator, 2,7-dichlorofluorescein diacetate. (b) 5 mM H2O2-activated Ca2+ influx as visualized by Fura-2 ratios ([Ca2+]cyt) and whole cell currents recorded simultaneously in a guard cell. Holding potential –116 mV (c) 50 µM ABA-induced stomatal closure in wild type (WT) and gca2 guard cells. (d) H2O2-induced stomatal closure in wild type (WT) and gca2 guard cells. Reproduced from Pei et al. (2000) with permission. © 2000 Macmillan Magazines Ltd, UK.

Ca2+-independent signalling pathways

Although changes in [Ca2+]cyt have been demonstrated to be an important downstream signalling element in guard cell ABA signal transduction, there are indications that the ABA signal can be transduced in a Ca2+-independent manner. Allan et al. (1994) showed that growth temperature could profoundly affect the ability of ABA to induce a change in [Ca2+]cyt without affecting ABA-induced turgor loss. This suggests the presence of Ca2+-independent signalling pathways mediating the effects of ABA on guard cell turgor. Jacob et al. (1999) demonstrated the involvement of phospholipase D (PLD) in the response of guard cells to ABA. PLD activity was stimulated 2.5–7.5 min and 20–25 min after application of ABA and this stimulation was associated with an increase in the concentration of phosphatidic acid (PA). In addition, application of phosphatidic acid to epidermal peels resulted in loss of guard cell turgor and 1-butanol, an inhibitor of PA production by PLD, partially attenuated the effects of ABA on guard cell turgor. Using ratio imaging of guard cells microinjected with the Ca2+-sensitive fluorescent indicator, Indo-1, Jacob et al. (1999) showed that the PA-induced turgor loss was not associated with changes in [Ca2+]cyt. In addition, PA inhibited the activity of the inward-rectifying K+ channel, suggesting that the regulation of the inward-rectifying K+ channel is Ca2+-independent. This was corroborated by Romano et al. (2000) who used simultaneous patch-clamping and confocal ratiometric Ca2+ imaging to show that increases in [Ca2+]cyt were not required for inhibition of IK,in by ABA.

The Ca2+-independent regulation of the inward-rectifying K+ channel, and hence IK,in, is in contrast with earlier observations of Lemtiri-Chlieh & MacRobbie (1994) who showed that IK,in is regulated by changes in [Ca2+]cyt. The data of Lemtiri-Chlieh & MacRobbie (1994) is consistent with the observations of Li et al. (1998) and Berkowitz et al. (2000) who showed that the activity of the inward-rectifying K+ channel can be modulated by the activity of a Ca2+-dependent protein kinase. Recently, Li et al. (2000) showed the involvement of an ABA-activated serine-threonine protein kinase (AAPK) in ABA regulation of guard cell anion channels. The activity of AAPK is ABA-dependent and Ca2+-independent (Li & Assmann, 1996; Mori & Muto, 1997). Expression of a mutant AAPK harbouring a lysine 43 to alanine 43 substitution, eliminated ABA activation of guard cell anion channels, suggesting that regulation of anion channels by AAPK is Ca2+-independent. The activity of guard cell anion channels is also regulated by post-translational modification involving a farnesyltransferase encoded by the ERA1 gene (Pei et al., 1998). Anion channels of guard cells of an ERA1 deletion mutant of A. thaliana are hypersensitive to ABA (Pei et al., 1998). This is in good agreement with the observation that hypersensitivity of guard cell anion channels to ABA can be induced through the use of farnesyltransferase inhibitors, α-hydroxyfarnesylphosphonic acid and manumycin (Pei et al., 1998). Overall, these data suggest that the activities of ion channels, which represent the final effectors in guard cell signal transduction, can be regulated by both Ca2+-dependent and Ca2+-independent pathways.

Elevations in guard cell [Ca2+]cyt– stimulus-specific ‘calcium signature’

Elevations in [Ca2+]cyt in stomatal guard cells have been shown to be the result of a shift in the dynamic balance between influx to and efflux from the cytosol. Increases in [Ca2+]cyt can result from influx of Ca2+ to the cytosol from the apoplast and internal stores (vacuole and ER) whereas decreases from elevated levels are due to efflux of Ca2+ from the cytosol to the apoplast and internal stores (vacuole and ER). In the stomatal guard cell, Ca2+-permeable channels and pumps located on the plasma and vacuolar membranes mediate Ca2+ fluxes into and out of the cytosol. The coordinated activities of these Ca2+-permeable channels and pumps will contribute ultimately to fluctuations in [Ca2+]cyt.

It is well established that increases in [Ca2+]cyt mediate changes in guard cell turgor induced by ABA and auxin (McAinsh et al., 1990; Irving et al., 1992). This is interesting, as ABA is a closing stimulus whereas auxin is an opening stimulus. The observation that an elevation in [Ca2+]cyt is a component of the signal transduction pathway by which these two diametrically opposed stimuli induce a physiological response immediately raises questions of how specificity is encoded in the Ca2+ signal. McAinsh et al. (1992) observed marked spatial heterogeneities in ABA-induced elevations in [Ca2+]cyt using Indo-1 ratio imaging. A relatively uniform level of [Ca2+]cyt was observed throughout the guard cell before ABA addition, whereas following the addition of ABA, increases in [Ca2+]cyt were observed. These increases were unevenly distributed and appeared as ‘hot spots’ and Ca2+-quiescent regions. Spatial heterogeneity in guard cell [Ca2+]cyt elevations have also been reported by Gilroy et al. (1991) and McAinsh et al. (1995) in response to increased [Ca2+]ext and decreases in [K+]ext (Gilroy et al., 1991). The spatial heterogeneities in [Ca2+]cyt elevations could result from (1) localized increases in [Ca2+]cyt due to the nonuniform distribution of the intracellular signalling machinery, or (2) accessibility of the ABA signal to only a subset of the signalling machinery (McAinsh et al., 1997; Hetherington et al., 1998; McAinsh & Hetherington, 1998). These observations suggest the potential for encoding specificity in the form of localized increases in [Ca2+]cyt. The observation of localized increases in [Ca2+]cyt immediately raises the question of how the Ca2+ signal is propagated. Evidence from animal systems suggests that the signal can be propagated in the form of Ca2+‘puffs’ or ‘sparks’, culminating in the formation of a Ca2+ wave (Berridge et al., 2000). It is tempting to suggest that the localized Ca2+‘hot-spots’ observed in guard cells represent such puffs or sparks. However, due to the spatial and temporal resolution used in these studies, it is likely that these localized elevations in [Ca2+]cyt represent longer transients in [Ca2+]cyt. The underlying mechanism contributing to such localized elevations in [Ca2+]cyt in guard cell remains to be established.

In addition to spatial changes in [Ca2+]cyt in the form of ‘hot spots’ and Ca2+-quiescent regions, temporal heterogeneities in stimulus-induced elevations in [Ca2+]cyt, in the form of oscillations in [Ca2+]cyt may also encode stimulus-specific information. Oscillations in [Ca2+]cyt allow for information to be encoded in both the amplitude and frequency (Berridge et al., 1988; Fewtrell, 1993; McAinsh et al., 1997; McAinsh & Hetherington, 1998). Perhaps the best examples of stimulus-specific Ca2+ signatures in guard cells are produced in response to [Ca2+]ext. McAinsh et al. (1995) showed that the pattern of [Ca2+]cyt oscillation varied with the strength of the stimulus and that this was correlated with the degree of turgor loss, indicating the potential for encoding information on the strength of the stimulus through differences in the oscillatory pattern (McAinsh et al., 1995). Oscillations in [Ca2+]cyt can also be modified by the addition of a second stimulus. For example, as shown in Fig. 4, guard cells exhibiting a particular Ca2+ signature in response to [Ca2+]ext generate a completely new signature when simultaneously challenged with a second stimulus (Hetherington et al., 1998). Staxén et al. (1999) reported ABA-induced oscillations in [Ca2+]cyt and showed a correlation between the pattern of [Ca2+]cyt oscillation and the degree of stomatal closure. These data demonstrate that the generation of unique Ca2+ signatures in the form of oscillations in [Ca2+]cyt have the potential to integrate information from a number of stimuli that are being perceived simultaneously by the guard cell.

Figure 4.

Modulation of [Ca2+]cyt oscillations in guard cells of Commelina communis by the presence of an additional stimulus. The effect of (a) [K+]ext (b) ABA and (c) osmotic shock induced by mannitol on oscillations in [Ca2+]cyt induced by 1 mM or 0.1 mM [Ca2+]ext. Reproduced from Hetherington et al. (1998) with permission. © 1998 The Royal Society, UK.

The potential for encoding stimulus-specific Ca2+ signatures in the form of oscillations in [Ca2+]cyt was recently reiterated by Allen et al. (2000). The authors used the det3 (de-etiolated 3) mutant of A. thaliana (which exhibits decreased endomembrane energization due to a 60% decline in the expression of the C-subunit of the V-ATPase) in order to show that a correlation exists between stomatal closure and [Ca2+]ext-induced [Ca2+]cyt oscillations. Guard cells of det3 mutants do not close in response to [Ca2+]ext and do not exhibit oscillations in [Ca2+]cyt compared with wild-type guard cells. However, guard cells of det3 mutants could be induced to close when [Ca2+]cyt oscillations were generated by imposing alternating hyperpolarization and depolarization steps, suggesting that oscillations in [Ca2+]cyt are required for physiological responses in guard cells (Allen et al., 2000).

Ca2+-induced Ca2+-release: a mechanism for signal amplification and propagation

Based on studies involving animal cells and sea urchin eggs, it has been proposed that oscillations in [Ca2+]cyt could result from Ca2+-induced Ca2+-release (CICR) by InsP3- and cADPR-gated Ca2+ release channels (Berridge et al., 2000). CICR provides a mechanism for integrating localized increases in [Ca2+]cyt into a complex Ca2+ signal (oscillations and waves). However, there is a paucity of experimental evidence supporting the operation of CICR in plants and what little information available suggests that the mechanism(s) of CICR in plants differ from animal systems. The results of Grabov & Blatt (1999) showed that influx of Ca2+ acts as a trigger for elevating [Ca2+]cyt, suggesting the operation of CICR in stomatal guard cells. Interestingly, the authors also showed that this intracellular Ca2+ release could be blocked by ryanodine, indicating that the cADPR-gated Ca2+-release channel may be involved in CICR. The endomembrane channels responsible for CICR in plants remain a topic of contention (Pottosin et al., 1997; Bewell et al., 1999; Pei et al., 1999). There is no evidence of a role for InsP3-gated Ca2+-release channels in CICR in plants. In addition, Leckie et al. (1998) demonstrated that the cADPR-gated Ca2+-release channel on the vacuolar membrane cannot support CICR as a [Ca2+]cyt greater than 0.6 µM inhibited its activity. This led the authors to propose that Ca2+-release induced by cADPR acts as the initial trigger for subsequent Ca2+ release, possibly through the SV channel whose activity is activated when [Ca2+]cyt exceeds 0.6 µM (Allen & Sanders, 1996). While the SV channel remains a likely conduit for CICR (Bewell et al., 1999; Pei et al., 1999), the role of InsP3- and cADPR-gated Ca2+-release channels in nonvacuolar membranes remains to be assessed. The possibility remains that InsP3- and cADPR-gated Ca2+ release channels on other endomembranes can support CICR in plants (MacRobbie, 2000). MacRobbie (2000) has suggested that the fast-vacuolar (FV) channel (Allen & Sanders, 1997) may function as a cADPR-sensitive channel and participate in CICR in plants by acting as the initial trigger for Ca2+-release since its activity is inhibited by [Ca2+]cyt above 100 nM (Allen & Sanders, 1997). However, direct experimental evidence for this is lacking.

Oscillations in [Ca2+]cyt can also be governed by oscillations in the cytosolic concentrations of the ligand gating the release of Ca2+. Hirose et al. (1999) used Madin-Darby canine kidney epithelial cells to show that ATP-induced elevations in [Ca2+]cyt oscillated synchronously with fluctuations in the concentration of InsP3 as visualized by green fluorescent protein-tagged pleckstrin homology domain. This suggests that changes in cytosolic concentrations of InsP3 ([InsP3]cyt) may be fundamental to the generation of oscillations in [Ca2+]cyt. Although InsP3 and delta-like PLC have been implicated in the generation of elevated [Ca2+]cyt in guard cells (Gilroy et al., 1990; Staxén et al., 1999), it remains to be determined if changes in [InsP3]cyt occur in guard cells and whether such changes can contribute to the complex temporal changes in [Ca2+]cyt mediated by PI-PLC/InsP3.

Conclusion

The identification of InsP3, InsP6, cADPR and H2O2 as components of the ABA signal transduction pathway mediating turgor changes in guard cells, and the recently identified NAADP-mediated Ca2+-release from ER (Navazio et al., 2000), suggest that Ca2+-based signalling pathways in animals and plants are evolutionarily conserved. While conservation is evident, there appear to be differences in the way the Ca2+ signal is amplified. In animals, the main feature of Ca2+ signal amplification is CICR through InsP3-, cADPR- and NAADP-gated Ca2+ release pathways. While the same ligand-gated Ca2+ release channels have been identified in plants, the mechanism of CICR in plants remains contentious.

The potential for encoding specificity in the Ca2+ signal has been demonstrated by the ability of guard cells to generate stimulus-specific temporal changes in [Ca2+]cyt in the form of oscillations. Additional information can be encoded in the form of complex patterns of [Ca2+]cyt generated in response to two or more stimuli. The ability to modulate the pattern of temporal changes in [Ca2+]cyt and the generation of novel [Ca2+]cyt oscillatory patterns suggest the presence of a complex cellular machinery capable of integrating stimuli perception into a meaningful physiological response; in this case, changes in stomatal aperture.

The existence of multiple pathways for ABA-induced Ca2+ mobilization in guard cells suggests an inherent redundancy in the ABA signalling cascade regulating guard cell turgor. More importantly, the presence of multiple Ca2+-mobilizing pathways suggest the possibility of accessing differentially the cellular machinery controlling Ca2+ influx and release from intracellular stores. It has been suggested that in animal cells the form of increase in [Ca2+]cyt in response to stimuli is determined by the pathway in which Ca2+ is mobilized (Berridge et al., 2000). Although it is plausible that the ability for encoding specificity lies in the ability to access differentially the diversity of Ca2+ mobilizing pathways, this still remains to be proved in plants.

The ability for encryption of stimulus-specific information immediately raises the question of how the cell is able to decode the information into a meaningful physiological response. We are just beginning to gain an insight into the cellular machinery used for decoding the encrypted information. Regulation of the activity of the inward-rectifying K+ channel by the phosphorylating activity of a Ca2+-dependent protein kinase (Berkowitz et al., 2000) provides a way for decoding the Ca2+ signal. Additionally, the regulation of IK,in by InsP3-mediated increases in [Ca2+]cyt is likely to involve the action of a calcineurin-related protein phosphatase (Luan et al., 1993). These data suggest the involvement of kinases, phosphatases and calmodulin as components of the cellular machinery in decoding the information encrypted in the Ca2+ signal.

Webb & Hetherington (1997) proposed the concept of a signalling cassette as a point for signal convergence. The presence of such signalling cassettes may be fundamental for the organization of the cellular machinery responsible for signal encryption and decoding the signal. It is becoming increasingly clear from animal cells that signal transduction is conducted within a complex cellular network derived from the interaction of cellular components of which lipid rafts within the plasma membrane, scaffolding molecules, anchoring proteins and the cytoskeleton feature predominantly (Jordan et al., 2000; Simons & Toomre, 2000; Teruel & Meyer, 2000). It is well established that the cytoskeleton is an integral component of guard cell function. The spatial elevations in [Ca2+]cyt as visualized by ‘hot spots’ and Ca2+-quiescent regions may be indicative of such signalling cassettes where the signalling elements responsible for information encryption and decoding are organized. In the 10 yr since the definitive demonstration of Ca2+ as a signal in guard cells (McAinsh et al., 1990), we have begun to understand the complexities involved in generating the Ca2+ signal responsible for controlling guard cell turgor. It is becoming increasingly apparent that the ‘physiological address’ or ‘poise’ of the cell (McAinsh et al., 1997), which is determined by the spatial distribution of cellular components like calmodulin, PLC, Ca2+-dependent phosphatases and kinases, influences the way in which the signal is transduced within the cell. It is envisaged that the systematic identification of genes, their products and the interactions between the various cellular components will provide a wealth of information pertaining to signal encryption and the mechanism(s) underlying the decoding of the Ca2+ signal into meaningful physiological responses.

Acknowledgements

We thank the Biotechnology and Biological Sciences Research Council, UK (MRM, JEG, LH, CPL, LM, AMH), The Royal Society, UK (MRM) and Westlakes Charitable Trust, UK (CKYN, AMH) for research funding. CKYN is in receipt of a studentship from the Westlakes Charitable Trust and MRM is grateful to The Royal Society for the award of a University Research Fellowship.

Ancillary