INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSIONS MATERIALS AND METHODS LITERATURE CITED

Calcium acts as an intracellular second messenger, coupling extracellular stimuli to intracellular and whole-plant responses (Hepler and Wayne, 1985; Sanders et al., 1999). Guard cells have been developed as a model system for dissecting early signal transduction processes in plant cells. Guard cells respond to a great variety of external stimuli, including abscisic acid (ABA; McAinsh et al., 1990; Schroeder and Hagiwara, 1990), auxin (Gehring et al., 1990, 1998), ozone (Clayton et al., 1999), and reactive oxygen species (ROS; McAinsh et al., 1996; Pei et al., 2000) with an increase in the cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_cyt) and subsequent stomatal movements (for review, see Blatt, 2000; Schroeder et al., 2001a). Cytosolic Ca²⁺ increases down-regulate inward-rectifying K⁺ channels and activate anion channels, providing mechanisms for Ca²⁺-dependent stomatal closure (Schroeder and Hagiwara, 1989). Particularly well analyzed is the Ca²⁺ response of guard cells to the phytohormone ABA (McAinsh et al., 1990; Schroeder and Hagiwara, 1990; Blatt and Armstrong, 1993; Schmidt et al., 1995; Leckie et al., 1998; Allen et al., 1999a; Staxén et al., 1999; MacRobbie, 2000; Hugouvieux et al., 2001; for review, see Blatt, 2000; Schroeder et al., 2001b).

ABA has been shown to activate a hyperpolarization-dependent Ca²⁺-permeable current in the plasma membrane of guard cells, leading to Ca²⁺ influx and an increase in the cytoplasmic free Ca²⁺ concentration (Hamilton et al., 2000; Pei et al., 2000). Furthermore, it has been demonstrated that ABA elevates levels of ROS, and that elevated ROS levels stimulate Ca²⁺-permeable cation currents in the plasma membrane termed "I_Ca" (Pei et al., 2000). I_Ca channels have been shown to be permeable to several cations including Mg²⁺ (Pei et al., 2000). The ABA-insensitive mutants gca2, abi1-1, and abi2-1 disrupt I_Ca channel activation at distinct points, providing genetic evidence for this newly recognized branch in ABA signaling (Pei et al., 2000; Murata et al., 2001).

ROS production in guard cells is induced not only by ABA, but also by the elicitors of plant defense reactions chitosan and oligo-GalUA (Lee et al., 1999). These elicitors also promote stomatal closing (Lee et al., 1999). In plant cells other than guard cells, it is known that one of the first responses to elicitors is an elevation in cytosolic Ca²⁺, which lies upstream of NADPH-oxidase activation (Knight et al., 1991; Zimmermann et al., 1997; Mithöfer et al., 1999; Blume et al., 2000). Pathogen-induced Ca²⁺ influx has been reported to occur both before (Schwacke and Hager, 1992; Blume et al., 2000) and after (Price et al., 1994; Kawano and Muto, 2000) ROS production, indicating that two distinct plasma membrane Ca²⁺ channels may function during different phases of the response. The similarities between the elicitor-activated and hyperpolarization-induced Ca²⁺ channels in tomato (Lycopersicon esculentum) cells (Gelli et al., 1997) and ABA-activated Ca²⁺ channels in guard cells (Hamilton et al., 2000; Pei et al., 2000) suggest that these two stimuli may activate related influx currents.

Ca²⁺ oscillations have been shown to be critical for induction of stomatal closure (Allen et al., 2000), and are mediated from two general sources, proposed to work in parallel: influx of Ca²⁺ across the plasma membrane (Schroeder and Hagiwara, 1990; Hamilton et al., 2000; MacRobbie, 2000; Pei et al., 2000) and release of Ca²⁺ from internal stores (Leckie et al., 1998; Staxén et al., 1999; MacRobbie, 2000). The concentration of ABA favors either induction of Ca²⁺ influx or Ca²⁺ release mechanisms in Commelina communis guard cells (MacRobbie, 2000). At high ABA concentrations (> 1 µM), Ca²⁺ influx was reported to predominantly contribute to ABA-induced [Ca²⁺]_cyt increases, whereas at low ABA, Ca²⁺ release from internal stores is predominant in Commelina communis. To understand Ca²⁺-based signal transduction pathways in guard cells, therefore, it is necessary to closely analyze the conditions under which Ca²⁺ influx or Ca²⁺ release mechanisms occur. Manganese quenching experiments show that external Ca²⁺-induced oscillations in cytosolic Ca²⁺ include plasma membrane Ca²⁺ influx (McAinsh et al., 1995). Although extracellular Ca²⁺ is required for ABA-induced changes in stomatal aperture (De Silva et al., 1985; Schwartz, 1985; MacRobbie, 2000; Webb et al., 2001), and for ABA induction of a transient [Ca²⁺]_cyt increase in Vicia faba guard cells (Romano et al., 2000), to date, ABA-induced [Ca²⁺]_cyt oscillations in Arabidopsis guard cells have not been analyzed as a function of extracellular Ca²⁺ removal. Here, we have analyzed whether intracellular Ca²⁺ release pathways that contribute to Ca²⁺ oscillations are dependent on rapid extracellular Ca²⁺ removal.

In the present study, we analyze the degree of convergence of stomatal closure pathways induced by different stimuli. We test whether I_Ca Ca²⁺ channels represent a shared branch of ABA and elicitor signaling by testing whether elicitors activate hyperpolarization-induced I_Ca Ca²⁺ influx currents. We investigated further the effects of external Ca²⁺ and external [K⁺] on [Ca²⁺]_cyt oscillations in multiple stomatal closure signaling pathways. We also reveal a new mode of ABA signaling in which ABA is shown to down-regulate spontaneous [Ca²⁺]_cyt oscillations that occur in guard cells (Allen et al., 1999b; Staxén et al., 1999).

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSIONS MATERIALS AND METHODS LITERATURE CITED

Elicitors Activate NADPH-Dependent Ca²⁺ Channel Currents

It has been shown that chitosan, an elicitor of phytoalexin production in pea (Pisum sativum) pods (Hadwiger and Beckman, 1980), induces the production of ROS and a reduction of stomatal aperture in guard cells of tomato and C. communis (Lee et al., 1999). We analyzed whether chitosan and yeast (Saccharomyces cerevisiae) elicitor, a well-studied elicitor of defense reactions in cell cultures of Eschcholtzia californica (Schumacher et al., 1987), could activate plasma membrane Ca²⁺ currents in guard cells similar to the ROS-induced Ca²⁺ currents during ABA signaling. We applied voltage ramps from 18 to 198 mV (after correction for liquid junction potentials). Both chitosan and yeast elicitor activated a hyperpolarization-dependent current in guard cell protoplasts (Fig. 1). In the absence of elicitors, only a small background current with a mean amplitude of 5.7 pA (n = 17 protoplasts) at 198 mV was observed. Upon addition of 10 µg mL¹ yeast elicitor to the bath solution, a hyperpolarization-activated inward current was observed. This current had a mean peak current amplitude of 59.9 pA at 198 mV (n = 10 protoplasts). Chitosan (10 µg mL¹) induced a hyperpolarization-activated current with a mean peak amplitude of 119.6 pA (n = 8 protoplasts). Average current/voltage relationships from untreated and elicitor treated cells are shown in Figure 1B. ROS- and ABA-activated I_Ca currents have been shown previously to have a "spiky" behavior (Pei et al., 2000; Murata et al., 2001), which was also observed for elicitor-activated currents (Fig. 1A) that are activated by hyperpolarization. Figure 1C shows currents produced by a voltage pulse protocol in the presence of H₂O₂ (n = 4). Note that I_Ca currents were not observed in voltage pulse protocols in the absence of H₂O₂ (Fig. 1C, H₂O₂; n = 4).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Elicitor activation of hyperpolarization-dependent currents in Arabidopsis guard cells. A, Whole-cell currents without or with 10 µg mL1 yeast elicitor or chitosan present in the bath solution. Elicitor-activated currents were measured approximately 5 min after elicitor exposure. Voltage ramps (1-s duration) were from 18 to 198 mV (lower). The liquid junction potential of 18 mV was corrected for. Arrows on the right show zero current levels. Bath solution: 100 mM BaCl2, 0.1 mM dithiothreitol (DTT), and 10 mM MES-Tris (pH 5.6). Pipette solution: 10 mM BaCl2, 0.1 mM DTT, 4 mM EGTA, 5 mM NADPH, and 10 mM HEPES-Tris (pH 7.1). B, Current/voltage relationship from control and elicitor-treated cells. Experimental conditions are the same as in A. Black circles, Untreated cells (n = 17); white circles, yeast elicitor-treated cells (n = 10); white squares, chitosan-treated cells (n = 8). C, Currents produced by voltage pulses (top) in the absence (middle) or presence (bottom) of 5 mM hydrogen peroxide (H2O2). The liquid junction potential of 18 mV was corrected for.

Interestingly, addition of NADPH to the pipette solution was necessary to activate elicitor- and hyperpolarization-dependent currents. Without NADPH present in the pipette solution, yeast elicitor could not induce hyperpolarization-activated currents (Fig. 2, A and C; n = 7). With NADPH in the pipette, yeast elicitor activated the hyperpolarization-activated current (Fig. 2, B and D; n = 10).

View larger version (13K): [in this window] [in a new window]   Figure 2.   NADPH dependence of elicitor-induced hyperpolarization-activated currents. A, Pipette solution, no NADPH; bath solution, no yeast elicitor (n = 4). B, Pipette solution, 5 mM NADPH; bath solution, no yeast elicitor (n = 17). C, Pipette solution, no NADPH; bath solution, 10 µg mL1 yeast elicitor (n = 7). D, Pipette solution, 5 mM NADPH; bath solution, 10 µg mL1 yeast elicitor (n = 10). In all experiments, 10 mM BaCl2 (+0.1 mM DTT, 4 mM EGTA, and 10 mM HEPES-Tris [pH 7.1]) were used as pipette solution and 100 mM BaCl2 (+0.1 mM DTT and 10 mM MES-Tris [pH 5.6]) as bath solution. The applied voltage ramp protocol is shown in the lower left. The liquid junction potential of 18 mV was corrected for. Arrows on the right show zero current levels.

Elicitor-Induced [Ca²⁺]_cyt Elevations Require External Ca²⁺

Both chitosan and yeast elicitor induced repetitive [Ca²⁺]_cyt elevations in guard cells of Arabidopsis (Fig. 3, A and B). Elicitor concentrations as low as 10 µg mL¹ were sufficient to induce repetitive [Ca²⁺]_cyt transients in guard cells treated with yeast elicitor (n = 12 of 15 cells) or chitosan (n = 9 of 9 cells). The relative amplitude and mean duration of yeast elicitor and chitosan-induced [Ca²⁺]_cyt transients are summarized in Table I. When yeast elicitor was applied at higher concentrations (50 µg mL¹), only a single, slowly declining [Ca²⁺]_cyt transient with a relative amplitude of ratio_585/480 = 0.41 ± 0.04 and a mean duration of 22.28 ± 2.67 min (duration at one-half amplitude of 10.9 ± 1.42 min) was observed (Fig. 3C; n = 14 cells).

View larger version (23K): [in this window] [in a new window]   Figure 3.   Elicitor-induced [Ca2+]cyt transients and stomatal closing. A, Repetitive [Ca2+]cyt transients induced by 10 µg mL1 yeast elicitor. B, Repetitive [Ca2+]cyt transients induced by 10 µg mL1 chitosan. C, Example of a [Ca2+]cyt transient, induced by 50 µg mL1 yeast elicitor. Note that under the imposed conditions, only a single [Ca2+]cyt transient with a slow decay time was observed. D, Yeast elicitor-induced [Ca2+]cyt transients require external Ca2+. Epidermal peels were incubated in the standard bath solution (5 mM KCl, 50 µM CaCl2, and 10 mM MES-Tris [pH 6.15]). At the indicated time point (first arrow), the bath was perfused with a solution containing 25 µg mL1 yeast elicitor (5 mM KCl, 50 µM CaCl2, 25 µg mL1 yeast elicitor, and 10 mM MES-Tris [pH 6.15]). At the second time point (second arrow), the bath solution was exchanged with a Ca2+-free solution (5 mM KCl, 250 µM EGTA, and 10 mM MES-Tris [pH 6.15]). Immediately after the perfusion with zero Ca2+ started, the yeast elicitor-induced [Ca2+]cyt transients ceased. E, Yeast elicitor reduces stomatal aperture. Each data point represents the mean stomatal aperture of 100 analyzed stomata from n = 4 replicates. Error bars = SD. Stomatal aperture was measured 2 h after elicitor application.

Yeast elicitor-induced transient increases in [Ca²⁺]_cyt required the presence of extracellular Ca²⁺ (Fig. 3D). Elicitor-induced cytoplasmic [Ca²⁺]_cyt transients immediately ceased after the bath solution was perfused with a Ca²⁺-free bath solution (5 mM KCl, 250 µM EGTA, and 10 mM MES/Tris [pH 6.15]; n = 12 cells). Although long-term external Ca²⁺ removal treatment may deplete intracellular calcium stores, it appears unlikely that this depletion is rapid enough to account for the immediate cessation of Ca²⁺ oscillations upon external Ca²⁺ removal. Note also that store-operated calcium currents, which are activated by a depletion of intracellular calcium stores, are not activated by extracellular application of EGTA alone (Kwan et al., 1990; Patterson et al., 1999). Therefore, the cessation of yeast elicitor-induced [Ca²⁺]_cyt elevations by external EGTA (Fig. 3D), together with elicitor-induced I_Ca activation (Figs. 1 and 2), demonstrate a requirement of Ca²⁺ influx for this response. We also tested whether yeast elicitor has an effect on stomatal movements in Arabidopsis and found a concentration-dependent reduction of stomatal aperture using the same concentrations that elicited [Ca²⁺]_cyt transients (Fig. 3E).

ABA-Regulated [Ca²⁺]_cyt Elevations

Having shown a requirement of extracellular calcium for elicitor-induced [Ca²⁺]_cyt oscillations, we next tested whether ABA-induced [Ca²⁺]_cyt oscillations share this requirement. As previously shown, ABA (10 µM) induces repetitive [Ca²⁺]_cyt elevations in cells (n = 18 of 40 cells) incubated in a bath solution containing calcium and 10 mM KCl (Fig. 4A). However, when extracellular Ca²⁺ was buffered to sub-micromolar concentrations by adding 250 µM EGTA (n = 12) or 250 µM BAPTA (n = 31) 10 min before ABA treatment, ABA-induced [Ca²⁺]_cyt elevations could not be observed (Fig. 4B). Re-addition of extracellular Ca²⁺ at the end of the experiments led to external Ca²⁺-induced [Ca²⁺]_cyt elevations, showing that these cells were competent to report [Ca²⁺]_cyt elevations (Fig. 4B).

View larger version (20K): [in this window] [in a new window]   Figure 4.   ABA-induced [Ca2+]cyt oscillations in Arabidopsis guard cells require extracellular Ca2+. A, Repetitive [Ca2+]cyt transients induced by 10 µM ABA. Extracellular solution: 5 mM KCl, 50 µM CaCl2, and 10 mM MES/Tris (pH 6.15). B, Guard cells that were incubated in Ca2+-free solutions show no response to 10 µM ABA. Addition of 10 mM external Ca2+ at end of experiment caused [Ca2+]cyt increases.

With evidence that external calcium is required for ABA-induced [Ca²⁺]_cyt oscillations in guard cells, we next examined the role of intracellular Ca²⁺ release pathways during ABA-induced [Ca²⁺]_cyt oscillations. We first tested the pharmacological phospholipase C (PLC) inhibitor U-73122 and its physiologically inactive analog U-73343 (Staxén et al., 1999) in Arabidopsis guard cells. Ten micromolar ABA was added to the solution bathing epidermal peels. Immediately after ABA-induced [Ca²⁺]_cyt transients became visible, the cells were perfused either with 1 µM U-73122 or 1 µM U-73343. In the case of U-73122, a partial inhibition of ABA-induced [Ca²⁺]_cyt transients was observed (Fig. 5A; n = 6 cells). Perfusion of guard cells with the inactive analog U-73343 (1 µM) did not inhibit ABA-induced transients (Fig. 5B; n = 10 cells). In further experiments, epidermal peels were pre-incubated 30 min in 1 µM U-73122 or U-73343 before ABA (10 µM) was added to the bath solution. In the pre-incubation experiments with U-73122, no ABA-induced [Ca²⁺]_cyt increases were elicited (Fig. 5C; n = 6 cells). Guard cells that were pre-incubated with U-73343 still responded to ABA (Fig. 5D; n = 15 cells).

View larger version (24K): [in this window] [in a new window]   Figure 5.   The PLC inhibitor U-73122 has a negative effect on ABA-induced [Ca2+]cyt oscillations. A, Perfusion with 1 µM U-73122 partially inhibited ABA (10 µM) induced [Ca2+]cyt elevations in guard cells. Sequence of ABA and U-73122 perfusion: time point 1 (first arrow), 10 µM ABA (+standard bath solution); time point 2 (second arrow), 1 µM U-73122 (+standard bath solution and 10 µM ABA); and time point 3 (third arrow), 10 µM ABA (+standard bath solution, without U-73122). B, Perfusion with 1 µM U-73343, an inactive analog of U-73122, has no effect on ABA-induced Ca2+ transients. C, After a 30-min pre-incubation of guard cells in 1 µM U-73122, ABA (10 µM) did not activate [Ca2+]cyt transients in Arabidopsis guard cells. D, Thirty-min pre-incubation of guard cells in 1 µM U-73343 did not inhibit ABA (10 µM)-induced Ca2+ transients in guard cells.

Studies have suggested that cADP Rib (cADPR) is a Ca²⁺-releasing second messenger in ABA signal transduction (Allen et al., 1995; Wu et al., 1997; Leckie et al., 1998). Therefore, we analyzed the effects of nicotinamide on cameleon-expressing guard cells. Nicotinamide blocks cADPR synthesis and has been used to analyze putative roles of cADPR in guard cells and tomato subepidermal cells (Wu et al., 1997; Leckie et al., 1998; Jacob et al., 1999; MacRobbie, 2000). Interestingly, nicotinamide consistently caused a rapid reduction in the cameleon fluorescence ratio of guard cells (Fig. 6; n = 44 of 50 cells), suggesting that nicotinamide has a relatively drastic effect on [Ca²⁺]_cyt in Arabidopsis guard cells. Levels of basal [Ca²⁺]_cyt dropped upon nicotinamide application in both cells that were not treated with ABA (Fig. 6A; n = 36) and ABA-treated cells (Fig. 6B; n = 8). ABA-induced oscillations were inhibited by nicotinamide. This effect of nicotinamide was not because of any influence of nicotinamide on the cameleon protein because in vitro experiments with recombinant yellow cameleon 2.1 protein showed that nicotinamide did not itself alter the fluorescence of cameleon, nor did it alter cameleon fluorescence ratio changes induced by calcium (data not shown). Nicotinamide-treated cells were still able to respond to 10 mM external Ca²⁺ with an increase in [Ca²⁺]_cyt, indicating that these cells were still responsive to external stimuli (Fig. 6C). Our studies with PLC and cADPR inhibitors show that blocking these two calcium release pathways leads to distinct alterations in calcium homeostasis and signaling.

View larger version (20K): [in this window] [in a new window]   Figure 6.   Nicotinamide causes a drop in [Ca2+]cyt in Arabidopsis guard cells. A, Fifty millimolar nicotinamide reduced the baseline [Ca2+]cyt level in untreated guard cells. B, Nicotinamide reduced the baseline [Ca2+]cyt level and inhibited ABA-induced [Ca2+]cyt increases. C, Nicotinamide (50 mM)-treated cells are still responsive to the addition of 10 mM external calcium (n = 7 of 10 cells).

ROS-Induced [Ca²⁺]_cyt Elevations

Recent work suggests that ROS participate in the induction of [Ca²⁺]_cyt elevations by ABA in Arabidopsis guard cells, and ROS have been shown to trigger [Ca²⁺]_cyt elevations. To further investigate whether the [Ca²⁺]_cyt elevations induced by these two stimuli are part of a shared pathway, we investigated whether ROS-induced [Ca²⁺]_cyt elevations are more prevalent at hyperpolarized membrane potentials, as has been shown previously for ABA-induced [Ca²⁺]_cyt elevations (Grabov and Blatt, 1998). We tested this by recording ROS-induced [Ca²⁺]_cyt elevations in bath solutions with different concentrations of K⁺. A lower extracellular K⁺ concentration ([K⁺]_ext) leads to a more hyperpolarized membrane (Saftner and Raschke, 1981; Clint and Blatt, 1989; Grabov and Blatt, 1998). In 5 mM KCl, extracellular application of 100 µM H₂O₂ induced a [Ca²⁺]_cyt transient in all 12 cells tested (Fig. 7A), either consisting of one (n = 9 of 12 cells) or two (n = 3 of 12 cells) transients. These transients had a mean duration of 3.18 ± 0.31 min, similar to that reported previously (Pei et al., 2000), and a mean relative peak [Ca²⁺]_cyt increase of ratio_535/480 (=0.49 ± 0.06; n = 12). In solutions containing 100 mM KCl, H₂O₂ induced [Ca²⁺]_cyt transients in only 12 of 29 (41%) cells (Fig. 7B); in all cases, only a single [Ca²⁺]_cyt transient was observed, and these transients showed smaller [Ca²⁺]_cyt increases (mean ratio_535/480 of 0.18 ± 0.02) than those induced in 5 mM KCl. As reported previously (McAinsh et al., 1996; Pei et al., 2000), guard cells that were incubated in Ca²⁺-free solutions showed no [Ca²⁺]_cyt elevation in response to H₂O₂. Thus, ROS-induced [Ca²⁺]_cyt elevations and ABA-induced [Ca²⁺]_cyt elevations showed enhanced activity at lower external K⁺ concentrations, and both required external Ca²⁺ (Fig. 4B).

View larger version (21K): [in this window] [in a new window]   Figure 7.   H2O2-induced [Ca2+]cyt elevations in guard cells of Arabidopsis are dependent on the extracellular K+ concentration and are not blocked by external Mg2+. A, [Ca2+]cyt transient measured in a guard cell in response to 100 µM H2O2. Extracellular solution: 5 mM KCl, 50 µM CaCl2, and 10 mM MES/Tris (pH 6.15). B, One hundred micromolar H2O2-induced [Ca2+]cyt elevations in 12 of 29 cells in high extracellular potassium (100 mM KCl, 50 µM CaCl2, and 10 mM MES/Tris [pH 6.15]). C, Five hundred micromolar Mg2+ did not block H2O2-induced Ca2+ elevations with 50 µM CaCl2 in the bath. Extracellular solution: 500 µM MgCl2, 50 µM CaCl2, 5 mM KCl, and 10 mM MES/Tris (pH 6.15). D, H2O2-induced stomatal closure is not affected by magnesium. Each data point represents the mean stomatal aperture of 50 analyzed stomata from n = 6 replicates. Error bars show SE of mean (relative to n = 6). Stomatal aperture was measured 2 h after application of: (a) 0.2 mM CaCl2 (Control), (b) 0.2 mM CaCl2 and 0.2 mM H2O2, or (c) 0.2 mM CaCl2, 2 mM MgCl2, and 0.2 mM H2O2. In all experiments, 0.1 mM EGTA was added to the bath solution to buffer external [Ca2+] close to 0.1 mM (Pei et al., 2000). H2O2 did not induce stomatal closure in the absence of this buffering, perhaps because external Ca2+, including locally released cell wall Ca2+, was too high, which can in itself cause partial stomatal closing (McAinsh et al., 1995; Allen et al., 1999a).

Previous experiments showed that ROS-activated I_Ca channels in guard cells are nonselective cation channels with a permeability to Mg²⁺, suggesting that Mg²⁺ might compete with Ca²⁺ for passage through the channel (Pei et al., 2000). To test whether external Mg²⁺ can compete with Ca²⁺ signaling, we monitored ROS-induced [Ca²⁺]_cyt transients in the presence of an extracellular Mg²⁺ concentration (500 µM) 10-fold higher than that of Ca²⁺ (50 µM). These conditions had no influence on the induction of [Ca²⁺]_cyt transients by H₂O₂ (Fig. 7C; n = 13 cells). We also performed stomatal closing assays to analyze whether H₂O₂-induced stomatal closure is affected by magnesium. Replacement of 0.2 mM CaCl₂ in the bath solution by 2 mM MgCl₂ + 0.2 mM CaCl₂ had no influence on H₂O₂-induced stomatal closure (Fig. 7D). Exposing guard cells to 10 mM MgCl₂ did not cause [Ca²⁺]_cyt oscillations (n = 4, data not shown), in contrast to the external Ca²⁺ response (Fig. 4B). These data show that external Mg²⁺ cannot replace Ca²⁺ in inducing Ca²⁺ oscillations. Similarly, in the absence of external Ca²⁺, shifts in the extracellular KCl concentration from 0.1 to 100 mM caused no changes in cameleon fluorescence ratios (Allen et al., 2001; http://www.nature.com/nature/journal/v411/n6841/extref/4111053aa.html). These data support the premise that oscillations in cameleon ratios were not because of oscillations in cytosolic Cl concentrations.

Spontaneous [Ca²⁺]_cyt Oscillations Require External Calcium

There is now quite compelling evidence that [Ca²⁺]_cyt elevations are an important component of guard cell signal transduction for multiple stimuli, such as ABA, external Ca²⁺, and ROS. Interestingly, guard cells also frequently show spontaneously arising [Ca²⁺]_cyt oscillations, i.e. oscillations that are not induced by the extracellular application of such stimuli (Fig. 8; Allen et al., 1999b, 2001; Staxén et al., 1999). We examined whether these spontaneous oscillations share properties with induced [Ca²⁺]_cyt oscillations by testing their dependence on extracellular K⁺ and external Ca²⁺. As occurred with ABA- and ROS-induced oscillations, lower extracellular potassium concentrations led to increases in the occurrence of spontaneous [Ca²⁺]_cyt oscillations (Table II). Guard cells that were incubated in 5 mM KCl showed spontaneous [Ca²⁺]_cyt transients in approximately 45% of the 33 cells analyzed in this study. Note that the percentage of cells showing spontaneous [Ca²⁺]_cyt oscillations varies in different preparations, but they are commonly observed in guard cells (Grabov and Blatt, 1998; Allen et al., 1999b; Staxén et al., 1999). Cells that showed spontaneous [Ca²⁺]_cyt transients (n = 15 of 33 cells) exhibited an average of 2.07 ± 0.29 transients in 5 mM KCl (30-min recording interval; Table II; Fig. 8A). In 0.1 mM KCl, the percentage of guard cells exhibiting spontaneous [Ca²⁺]_cyt transients was increased to 88% (n = 42 of 48 cells), and the average number of transients per cell was increased to 3.71 ± 0.25 (Table II; Fig. 8B). Spontaneous [Ca²⁺]_cyt oscillation periods were shorter than those reported for ABA-induced [Ca²⁺]_cyt oscillation periods (Table II; Allen et al., 2001). In solutions containing 100 mM KCl, spontaneous [Ca²⁺]_cyt oscillations were not observed (Fig. 8C). Interestingly, however, these same cells still produced transient [Ca²⁺]_cyt elevations when exposed to high extracellular Ca²⁺ (10 mM; Fig. 8C).

View larger version (26K): [in this window] [in a new window]   Figure 8.   Spontaneous [Ca2+]cyt oscillations in guard cells of Arabidopsis are dependent on the Ca2+ and K+ concentration in the bath solution. A, Spontaneous [Ca2+]cyt oscillations in a 5 mM KCl-solution (+50 µM CaCl2 and 10 mM MES/Tris [pH 6.15]). B, The frequency of spontaneous [Ca2+]cyt transients is increased in solutions containing low K+ concentrations (0.1 mM KCl, 50 µM CaCl2, and 10 mM MES/Tris [pH 6.15]; see also Table II). C, Guard cells showed no spontaneous [Ca2+]cyt transients in high potassium solutions (100 mM KCl and 50 µM CaCl2). Note that 10 mM external Ca2+ triggered Ca2+ transients in guard cells, incubated in the same high-potassium solutions. D, Perfusion with a Ca2+-free solution (perfusion start at time point 1) inhibited spontaneous [Ca2+]cyt transients. When cells were reperfused with a 50 µM Ca2+-containing standard bath solution (time point 2), a recovery of spontaneous [Ca2+]cyt transients was observed (n = 19 cells). E, Spontaneous [Ca2+]cyt oscillations were suppressed in 62 of 101 of guard cells (61%) by the application of 5 µM ABA. ABA was applied at the time indicated by the arrow.

Upon removal of extracellular Ca²⁺, as with ABA-induced [Ca²⁺]_cyt oscillations, spontaneous [Ca²⁺]_cyt oscillations rapidly ceased to occur (Fig. 8D; n = 19 cells). A guard cell displaying spontaneous [Ca²⁺]_cyt transients under continuous perfusion with the standard bath solution (containing 5 mM KCl) was perfused with an EGTA-containing solution at time point 1. The [Ca²⁺]_cyt oscillations immediately ceased after the perfusion with zero Ca²⁺ (Fig. 8D). At time point 2, the cell was perfused again with the standard bath solution containing 50 µM CaCl₂, which induced a rapid recovery of [Ca²⁺]_cyt oscillations.

In Ca²⁺ imaging experiments to date, ABA was added to cells that showed no spontaneous Ca²⁺ oscillations (e.g. McAinsh et al., 1990; Schroeder and Hagiwara, 1990; Gilroy et al., 1991; Allan et al., 1994; Staxén et al., 1999; Allen et al., 1999b; 2001). Here, we analyzed the effect of ABA on spontaneously oscillating guard cells. The spontaneous Ca²⁺ oscillation period (Table II) was shorter than that of ABA-induced Ca²⁺ oscillations (Allen et al., 2001). Interestingly, in cells that were showing spontaneous [Ca²⁺]_cyt oscillations at the beginning of the experiment, ABA, in a substantial number of experiments, led to a rapid cessation of [Ca²⁺]_cyt oscillation activity. Complete spontaneous oscillation suppression occurred within 10 to 15 min after ABA application via the continuous perfusion stream and occurred in n = 62 of 101 spontaneously oscillating cells treated with ABA (61%; Fig. 8E). In the remaining cells, ABA did not cause cessation of spontaneous [Ca²⁺]_cyt oscillations. In a few of the cells (n = 14) where ABA suppressed spontaneous [Ca²⁺]_cyt oscillations, oscillations reactivated after a 30- to 40-min quiescent period. The ability of ABA to abolish spontaneous [Ca²⁺]_cyt oscillations is a novel finding and may be associated with a retuning of signal transduction components for ABA signaling and/or ABA-induced depolarization.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSIONS MATERIALS AND METHODS LITERATURE CITED

Elicitor-Induced [Ca²⁺]_cyt Oscillations

In tomato (Gelli et al., 1997) and parsley (Petroselinum crispum; Zimmermann et al., 1997), two distinct types of plant defense elicitor-activated Ca²⁺ influx currents have been described. The elicitor-induced currents in tomato cells activate at hyperpolarizing membrane potentials similar to those found here (Gelli et al., 1997), and show similarities to the ABA, ROS, and hyperpolarization-activated Ca²⁺ currents (I_Ca) of Arabidopsis guard cells (Pei et al., 2000), whereas in parsley cells, elicitors activate very large conductance channels (>250 pS) at more depolarized potentials (Zimmermann et al., 1997). We attempted to establish whether I_Ca can be activated by elicitors in Arabidopsis guard cells. We used two different elicitors of plant defense reactions in our studies. Chitosan, which has been shown previously to cause the production of ROS and stomatal closing in tomato and C. communis guard cells (Lee et al., 1999) and yeast elicitor, an elicitor with broad activity (induction of benzophenanthridine alkaloids, generation of ROS, and induction of intracellular pH shifts) in many diverse plant species (Blechert et al., 1995; Roos et al., 1998). We demonstrate that both chitosan and yeast elicitor activate a hyperpolarization-dependent current in guard cells, which resembles I_Ca in its voltage dependence. Furthermore, cytosolic NAD(P)H was necessary to activate this current (Fig. 2), which correlates with recent findings of the NAD(P)H requirement for ABA activation of I_Ca (Murata et al., 2001). Therefore, the chitosan- and yeast elicitor-induced production of ROS in guard cells (Lee et al., 1999) may occur via modulation of NAD(P)H-dependent mechanisms, as has been demonstrated in plant defense responses (Keller et al., 1998). Thus, the ROS produced could activate I_Ca, leading to the observed elicitor-induced increases in [Ca²⁺]_cyt.

The presented findings support the previously proposed model that ROS activation of I_Ca channels is part of a shared branch or "cassette" of stress signaling pathways (Pei et al., 2000; Schroeder et al., 2001b) and suggests that NAD(P) H-dependent activation of I_Ca channels represents a cross talk mechanism between ABA and defense signaling. For both elicitor signaling and ABA signaling, the specific reactive oxygen intermediate(s) that mediates signal transduction remains to be determined and could include H₂O₂, O₂, ¹O₂, and OH, among others. In addition, production of the reactive oxygen molecule NO has been shown recently to be induced by ABA in guard cells (Neill et al., 2002). The abi1-1 and abi2-1 protein phosphatase 2Cs (PP2Cs) have been shown to impair ABA signaling in guard cells upstream of I_Ca activation (Murata et al., 2001). Recent findings show that H₂O₂ inhibits the PP2C activities of both ABI1 (Meinhard and Grill, 2001) and ABI2 (Meinhard et al., 2002), which in turn could contribute to I_Ca activation.

Yeast elicitor evoked stomatal closing in guard cells in a concentration-dependent manner. Ca²⁺ imaging experiments demonstrated that both elicitors induce transient elevations in the cytoplasmic Ca²⁺ concentration of guard cells. The elicitor concentrations that induced [Ca²⁺]_cyt transients (10 µg mL¹) were sufficient to trigger stomatal closing responses. Yeast elicitor-induced [Ca²⁺]_cyt increases required the availability of extracellular Ca²⁺ ions. This indicates that an initial Ca²⁺ influx is necessary for the observed [Ca²⁺]_cyt increases. This observation is in accordance with studies showing that extracellular Ca²⁺ is necessary for the induction of plant defense responses against pathogens (Yang et al., 1997; Scheel, 1998; Blume et al., 2000).

Influx and Internal Release of Ca²⁺ in Guard Cell Signaling

Although other studies have recently addressed the requirement of extracellular Ca²⁺ for guard cell Ca²⁺ increases (MacRobbie, 2000; Romano et al., 2000), the contribution of external Ca²⁺ to long-term [Ca²⁺]_cyt oscillations, which have been shown to be an important part of Ca²⁺ signaling in guard cells (Allen et al., 2000), has not yet been determined. The abundance of Ca²⁺ release mechanisms in guard cells could lead to the hypothesis that ABA-induced Ca²⁺ elevations in Arabidopsis guard cells can occur after external Ca²⁺ removal. Here, we show that [Ca²⁺]_cyt oscillations do not occur in the absence of extracellular Ca²⁺. This was shown for oscillations induced by ABA and plant defense elicitors, as well as for spontaneously arising [Ca²⁺]_cyt oscillations. This is consistent with previous work that showed that the earliest increases in [Ca²⁺]_cyt in response to ABA [Ca²⁺]_cyt elevations can occur near the plasma membrane (McAinsh et al., 1992) as well as with a recent tracer flux study showing that extracellular Ca²⁺ contributes to ABA-induced K⁺ (Rb⁺) efflux at >1 µM ABA in C. communis guard cells (MacRobbie, 2000).

Although the presence of external Ca²⁺ was a prerequisite for the induction of cytoplasmic Ca²⁺ transients under our experimental conditions, these data do not contradict the importance of internal Ca²⁺ release mechanisms. Experiments in C. communis with U-73122, an inhibitor of plant phospholipase C (Staxén et al., 1999), showed that ABA-induced cytoplasmic Ca²⁺ transients were suppressed in guard cells that were pre-incubated in 1 µM U-73122. Experiments on Arabidopsis guard cells correlate with these results (Fig. 5, A and C). These findings suggest that inositol 1,4,5 trisphosphate triggered Ca²⁺ release mechanisms may contribute to ABA-induced [Ca²⁺]_cyt increases in Arabidopsis. Present simplified models consider parallel functioning of Ca²⁺ influx and PLC-dependent Ca²⁺ release mechanisms. Interestingly, however, our results, as well as the requirement of calcium for plant PLC activity (Kopka et al., 1998; Hernández-Sotomayor et al., 1999) suggest that PLC activation and Ca²⁺ influx may be interdependent. Genetic analyses will be important to test this proposed linkage in activation of Ca²⁺ influx and PLC-dependent pathways.

Experiments using the cADPR blocker nicotinamide showed a rapid reduction in basal calcium levels in 88% of guard cells (Fig. 6, A-C). Detailed microinjection studies strongly support a role for cADPR in ABA signaling (Wu et al., 1997; Leckie et al., 1998). The results presented here show a difference in the actions of cADPR and PLC inhibitors. The reduction in basal [Ca²⁺]_cyt levels by nicotinamide may contribute to the additive effects of nicotinamide and U-73122 on inhibition of ABA-induced stomatal closing and Rb⁺ efflux (Jacob et al., 1999; MacRobbie, 2000). The present findings are consistent with models suggesting that cADPR functions parallel to other Ca²⁺-dependent pathways (Jacob et al., 1999; MacRobbie, 2000). In the present study, nicotinamide also reduced ABA-induced [Ca²⁺]_cyt elevations, which correlates with models including cADPR as a second messenger in ABA signaling (Wu et al., 1997; Leckie et al., 1998). It is also possible that nicotinamide reduces overall [Ca²⁺]_cyt levels, thus increasing the threshold required for ABA-induced [Ca²⁺]_cyt elevations to proceed. Genetic alteration and inducible silencing of individual mechanisms and Ca²⁺ imaging will be important for dissecting the underlying differential contributions of Ca²⁺ release mechanisms to Ca²⁺ signaling. In addition, the effect of nicotinamide suggests that cADPR may play a role in cytosolic calcium homeostasis.

ABA Inhibition of Spontaneous Oscillations

Surprisingly, we show that ABA causes cessation of spontaneously occurring [Ca²⁺]_cyt oscillations in the majority of guard cells (n = 62 of 101). These data reveal a new, not previously investigated mode of ABA action. A recent study has shown that a "window" of [Ca²⁺]_cyt oscillation parameters encodes steady-state stomatal closing in Arabidopsis (Allen et al., 2001). Interestingly, the periods of spontaneous [Ca²⁺]_cyt oscillations found here (6 min; Table II), based on the data of Allen et al. (2001), would correspond to little or no stomatal closure. (Note, however, that we expect that the range of [Ca²⁺]_cyt oscillation parameters that mediate stomatal movements are dynamic and exhibit a dependence on environmental and cellular conditions.) The data presented here indicate that ABA may repress [Ca²⁺]_cyt oscillations not associated with ABA signaling for ABA responses to proceed. ABA-induced depolarization is predicted to cause cessation of spontaneous [Ca²⁺]_cyt elevations (Grabov and Blatt, 1998). This hypothesis is supported by our data indicating that high external K⁺, which causes depolarization (Grabov and Blatt, 1998), also mitigate ROS-induced [Ca²⁺]_cyt elevations (Fig. 8C); thus, depolarization may contribute to the ABA inhibition of spontaneous [Ca²⁺]_cyt elevations revealed here.

Effect of [K⁺]_ext on [Ca²⁺]_cyt Oscillations

We show that the percentage of cells that show either spontaneous or ROS-induced [Ca²⁺]_cyt elevations is reduced at elevated extracellular potassium concentrations. Earlier work in V. faba demonstrated that the membrane potential of guard cells exhibits a near-Nernstian dependence on [K⁺]_ext, i.e. the lower the [K⁺]_ext, the more hyperpolarized the plasma membrane (Saftner and Raschke, 1981; Clint and Blatt, 1989). Furthermore, the nonselective nature of an ABA-activated Ca²⁺ influx current in V. faba guard cells results in reversal potentials that are more negative (e.g. 10 mV) than a Ca²⁺-selective channel would show (e.g. > +60 mV; Schroeder and Hagiwara, 1990). Consistent with these findings, 50 mM KCl produced a low probability of ABA-induced Ca²⁺ elevations in previous studies (Gilroy et al., 1991; Allan et al., 1994). Thus, these data suggest that [Ca²⁺]_cyt elevations are favored by hyperpolarized membranes. This is in agreement with other studies in V. faba guard cells that show that [Ca²⁺]_cyt elevations accompany membrane hyperpolarization (Grabov and Blatt, 1998) and low external K⁺ concentrations (Gilroy et al., 1991; Allen et al., 2000).

Conversely, we found that [Ca²⁺]_cyt elevations induced by high extracellular calcium did not appear affected by [K⁺]_ext (end of trace in Fig. 8C). Previous research has shown that Ca²⁺ influx mediates the initial phase of [Ca²⁺]_cyt transients induced by external Ca²⁺ (McAinsh et al., 1995). Even in 100 mM KCl, a large rise in [Ca²⁺]_cyt was always seen immediately after extracellular application of 10 mM CaCl₂ (Fig. 8C). This suggests that a hyperpolarization-independent mechanism for generation of [Ca²⁺]_cyt elevations is acting under these conditions or that 10 mM Ca²⁺ shifts the reversal potential of plasma membrane Ca²⁺ channels sufficiently positive to allow Ca²⁺ influx as would be expected for nonselective Ca²⁺-permeable channels found in V. faba guard cells (Schroeder and Hagiwara, 1990). Note that more than one plasma membrane Ca²⁺ channel is likely to contribute to cytosolic Ca²⁺ elevations in guard cells (Hamilton et al., 2000), assuming that the open probability of hyperpolarization-activated Ca²⁺ channels is not altered by physiological factors and ion gradients.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSIONS MATERIALS AND METHODS LITERATURE CITED

The present study demonstrates that elicitors activate plasma membrane I_Ca channels in an NADPH-dependent manner providing strong evidence for a shared signaling cassette of early ABA and elicitor signaling in guard cells. Furthermore, removal of extracellular Ca²⁺ causes rapid cessation of elicitor- and ABA-induced as well as spontaneous [Ca²⁺]_cyt oscillations in guard cells, suggesting an absolute requirement of Ca²⁺ influx for operation of the guard cell Ca²⁺ signaling system. Detailed analyses of spontaneous [Ca²⁺]_cyt oscillations in guard cells show that these differ from typical ABA-induced [Ca²⁺]_cyt elevations and that ABA can inhibit spontaneous oscillations. In addition, data suggest that Ca²⁺ influx and phospholipase C are interdependent rather than simply acting in parallel in mediating ABA-induced cytosolic Ca²⁺ elevations, whereas the pharmacological cADPR inhibitor nicotinamide has a unique effect of lowering baseline cytosolic Ca²⁺ levels in Arabidopsis guard cells.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION CONCLUSIONS MATERIALS AND METHODS LITERATURE CITED

Plant Material and Growth

Plants of Arabidopsis (ecotype Landsberg erecta) stably expressing yellow cameleon 2.1 under the control of the constitutive 35 S promoter (Allen et al., 1999b) were used in our experiments to measure [Ca²⁺]_cyt levels in guard cells. Arabidopsis seedlings were grown at 20°C in a controlled environment growth chamber (Conviron model E 15, Controlled Environments, Asheville, NC) under a 16-h-light/8-h-dark cycle with a photon fluency rate of 75 µmol m² s¹. Pots were watered every 2 to 3 d with deionized water and plants were misted with deionized water daily to keep the humidity close to 70%.

Stomatal Movement Analyses

Stomatal movement analyses were performed as described previously (Pei et al., 1997, Allen et al., 1999a). Rosette leaves from 4- to 6-week-old plants were detached and floated for 2 h in opening solution in the light (photon fluency rate of 75 µmol m² s¹). Depending on the type of experiment, we used two different opening solutions. Opening solution I (5 mM KCl, 50 µM CaCl₂, and 10 mM MES-Tris [pH 6.15]) was used to study the effect of elicitors on stomatal aperture, and opening solution II (10 mM KCl, 0.1 mM EGTA, and 10 mM MES-KOH [pH 6.15]) was used to analyze the effect of H₂O₂. After 2 h, yeast (Saccharomyces cerevisiae) elicitor or chitosan was added to opening solution I. In the case of the H₂O₂ experiments, the following solutions were added to opening solution II as indicated: (a) control, 0.2 mM CaCl₂; (b) 0.2 mM CaCl₂ and 0.2 mM H₂O₂; or (c) 0.2 mM CaCl₂, 2 mM MgCl₂, and 0.2 mM H₂O₂ (free external Ca²⁺ was about 0.1 mM). After a further incubation period, the leaves were blended in a blender (Waring, Torrington, CT), the resulting epidermal fragments were filtered out with a 30-µm nylon mesh, and guard cell aperture ratios were measured as described (Pei et al., 1997).

Elicitor Preparation

Chitosan was purchased from Calbiochem (San Diego), prepared as previously described (Walker-Simmons et al., 1984), then dissolved in stomatal opening solution I. Yeast elicitor was prepared according to the method of Schumacher et al. (1987). In brief, 1 kg of commercial baker's yeast was dissolved in 1.5 L of sodium citrate buffer (20 mM, pH 7.0) and autoclaved at 121°C and 11 Nm² bar for 60 min. The autoclaved suspension was centrifuged at 10,000g for 20 min. The resulting supernatant was mixed with 1 volume of ethanol and stirred gently overnight. The resulting precipitate was then centrifuged at 10,000g for 20 min. The supernatant of this centrifugation step was subjected to another ethanol precipitation overnight. The elicitor precipitate was lyophilized and stored at 20°C until use.

[Ca²⁺]_cyt Imaging Experiments

Epidermal strips of Arabidopsis leaves were mounted on coverslips with medical adhesive (Hollister Inc., Libertyville, IL) and incubated in a solution containing 5 mM KCl, 50 µM CaCl₂, and 10 mM MES/Tris (pH 6.15). To promote stomatal opening, the strip was illuminated for 2 h (photon fluency rate of 125 µmol m² s¹) at 22°C before measurements started. YC 2.1 [Ca²⁺]_cyt imaging experiments were performed as described previously (Allen et al., 1999b, 2000). The present imaging system differed from the one described by Allen et al. (1999) and was outfitted with a 440- ± 10-nm filter with a 455 DCLP dichroic mirror (Chroma, Brattleboro, VT) for excitation and interchanging 485- ± 20-nm and 535- ± 15-nm filters for emission. Note that in the present and also recent studies (Allen et al., 2000, 2001; Hugouvieux et al., 2001), Arabidopsis lines were used that show higher cameleon expression levels than the initially reported studies (Allen et al., 1999b) and that these lines allowed imaging at a reduced excitation intensity that did not excite measurable chloroplast fluorescence. A slow baseline drift because of yellow fluorescent protein bleaching was linearly subtracted. The lowest ratio value in each individual experiment was defined as ratio 1. Transient duration was defined as the time interval between the start and end of a transient unless otherwise noted; period was defined as the average time between two peaks of consecutive transients. Results are reported as average ± SE of the mean.

In Vitro Cameleon Assay

Fluorescence emission profiles of purified YC 2.1 protein were measured using a fluorimeter. Measurement was made in buffer (100 mM KCl and 10 mM MOPS [pH 7.2]) with or without 50 mM nicotinamide and with either nominally zero calcium (100 µM EDTA) or 4 mM CaCl₂.

Electrophysiology

Whole-cell patch-clamp experiments on Arabidopsis guard cells were performed by using an Axopatch 200 amplifier (Axon Instruments, Union City, CA) as described (Pei et al., 1997). Liquid junction potentials were corrected. For data analysis, AXOGRAPH 3.5 (Axon Instruments, Union City, CA) was used. Standard solutions contained 100 mM BaCl₂, 0.1 mM DTT, and 10 mM MES-Tris (pH 5.6) in the bath and 10 mM BaCl₂, 0.1 mM DTT, 4 mM EGTA, 0 or 5 mM NADPH, and 10 mM HEPES-Tris (pH 7.1) in the pipette. Exchange of initial bath solution with bath solution containing 10 µg mL¹ elicitor or chitosan was achieved by pipetting. Bath and pipette osmolalities were adjusted to 485 and 500 mmol kg¹, respectively, using D-sorbitol.