Calcium elevation-dependent and attenuated resting calcium-dependent abscisic acid induction of stomatal closure and abscisic acid-induced enhancement of calcium sensitivities of S-type anion and inward-rectifying K+ channels in Arabidopsis guard cells
Robert S. Siegel,
Division of Biological Sciences, Cell and Developmental Biology Section, and Center for Molecular Genetics, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0116, USA
Stomatal closure in response to abscisic acid depends on mechanisms that are mediated by intracellular [Ca2+] ([Ca2+]i), and also on mechanisms that are independent of [Ca2+]i in guard cells. In this study, we addressed three important questions with respect to these two predicted pathways in Arabidopsis thaliana. (i) How large is the relative abscisic acid (ABA)-induced stomatal closure response in the [Ca2+]i-elevation-independent pathway? (ii) How do ABA-insensitive mutants affect the [Ca2+]i-elevation-independent pathway? (iii) Does ABA enhance (prime) the Ca2+ sensitivity of anion and inward-rectifying K+ channel regulation? We monitored stomatal responses to ABA while experimentally inhibiting [Ca2+]i elevations and clamping [Ca2+]i to resting levels. The absence of [Ca2+]i elevations was confirmed by ratiometric [Ca2+]i imaging experiments. ABA-induced stomatal closure in the absence of [Ca2+]i elevations above the physiological resting [Ca2+]i showed only approximately 30% of the normal stomatal closure response, and was greatly slowed compared to the response in the presence of [Ca2+]i elevations. The ABA-insensitive mutants ost1-2, abi2-1 and gca2 showed partial stomatal closure responses that correlate with [Ca2+]i-dependent ABA signaling. Interestingly, patch-clamp experiments showed that exposure of guard cells to ABA greatly enhances the ability of cytosolic Ca2+ to activate S-type anion channels and down-regulate inward-rectifying K+ channels, providing strong evidence for a Ca2+ sensitivity priming hypothesis. The present study demonstrates and quantifies an attenuated and slowed ABA response when [Ca2+]i elevations are directly inhibited in guard cells. A minimal model is discussed, in which ABA enhances (primes) the [Ca2+]i sensitivity of stomatal closure mechanisms.
Studies from many laboratories have also reported experiments that failed to show [Ca2+]i increases in significant subsets of ABA-treated guard cells (Allan et al., 1994; Allen et al., 1999a; Allen et al., 1999b; Gilroy et al., 1991; Hugouvieux et al., 2001; Kwak et al., 2003; Levchenko et al., 2005; McAinsh et al., 1992; Romano et al., 2000; Schroeder and Hagiwara, 1990). The natural fluctuation of guard cell membrane potential between depolarized and hyperpolarized states (Gradmann et al., 1993; Thiel et al., 1992), coupled with the hyperpolarization dependence of inward Ca2+ channels, may contribute to the various cellular [Ca2+]i responses to ABA (Grabov and Blatt, 1998; Klüsener et al., 2002). It has been shown that ABA shifts the activation of inward Ca2+ channels to more positive voltages (Hamilton et al., 2000). However, imaging of [Ca2+]i in guard cells showed cells in which ABA did not elicit measurable [Ca2+]i increases. Interpretation of these results has led to hypotheses that are not strictly mutually exclusive. One hypothesis postulates that [Ca2+]i is necessary for stomatal closure in response to ABA, but that [Ca2+]i elevations are not always observed due to technical limitations of the measurements. A second hypothesis is that both [Ca2+]i-dependent and [Ca2+]i-independent pathways function in ABA-induced stomatal closure (Allan et al., 1994), a model that is presently of considerable interest. However, research has not yet determined the relative contributions of the proposed [Ca2+]i-dependent and [Ca2+]i-independent pathways to ABA-induced stomatal closure. A third hypothesis derived from long-term Ca2+ imaging studies during guard cell CO2 signaling, the ‘Ca2+ sensitivity priming hypothesis’, postulates that physiological stimuli of stomatal closure enhance the Ca2+ sensitivity of the appropriate stomatal closure mechanisms, thus allowing specific downstream Ca2+ responses to proceed, possibly even when [Ca2+]i elevations do not occur (Young et al., 2006).
Micro-injection of the Ca2+ chelator BAPTA (1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid) at elevated levels into guard cells of Commelina communis and Vicia faba abolished ABA responses in guard cells, providing evidence for a Ca2+ requirement in ABA signaling (Levchenko et al., 2005; Webb et al., 2001). Previous studies demonstrated that removal of external Ca2+ by chelation with EGTA or BAPTA inhibited spontaneous [Ca2+]i elevations, as well as [Ca2+]i elevations induced by ABA or pathogenic elicitors (Klüsener et al., 2002). In parsley (Petroselinum crispum), extracellular Ca2+ did not need to be completely removed to prevent [Ca2+]i elevations caused by pathogen elicitors (Blume et al., 2000). Buffering extracellular Ca2+ to 200 nm was sufficient to abolish pathogen-induced [Ca2+]i elevations (Blume et al., 2000). Elevated extracellular K+ concentrations depolarize guard cells and also dampen [Ca2+]i elevations (Allen et al., 2000; Grabov and Blatt, 1998; Klüsener et al., 2002). Non-invasive loading of guard cells with the membrane-permeable chelator BAPTA-AM (1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis (acetoxymethyl ester)), together with extracellular Ca2+ removal, inhibited spontaneous [Ca2+]i elevations and long-term stomatal movements induced by [CO2] (Young et al., 2006). A transient Ca2+-independent CO2 response was also found under these conditions (Young et al., 2006).
The present study reveals an attenuated partial ABA-induced stomatal closure response that occurs at resting [Ca2+]i. Furthermore, patch-clamp experiments show that ABA enhances (primes) the cytosolic Ca2+ sensitivity of both anion and K+ channels, providing mechanistic support for the Ca2+ sensitivity priming hypothesis (Young et al., 2006).
Ca2+ transients in guard cells
Yellow cameleon YC2.1 is a protein Ca2+ sensor based on a fluorescence resonance energy transfer (FRET) signal between two GFP moieties separated by Ca2+-binding calmodulin (Miyawaki et al., 1999; Miyawaki et al., 1997). Yellow cameleon allows direct monitoring of [Ca2+]i signals in Arabidopsis thaliana guard cells (Allen et al., 1999b). In the present study, we used the yellow cameleon construct 3.6 (YC3.6), which has an improved dynamic range (Mori et al., 2006; Nagai et al., 2004), to examine the role of guard cell [Ca2+]i elevations in ABA-induced stomatal closure.
Imposition of transient [Ca2+]i increases by switching the perfusion buffer between a depolarizing solution with low nanomolar Ca2+ (Ca2+ buffered by BAPTA) and a hyperpolarizing buffer containing 1 mm Ca2+ (Allen et al., 2001) resulted in greater than threefold changes in the ratiometric fluorescence reporting the [Ca2+]i, confirming the strong FRET response of this sensor (Figure 1a,b). Accurate in vivo calibration of a fluorescent [Ca2+]i sensor is difficult for various technical and biological reasons, and therefore imaging studies usually report fluorescence ratio changes rather than calibrated [Ca2+]i values. However, we have explored estimation of [Ca2+]i values from our fluorescence ratio data using in vitro calibration parameters for YC3.6 (dynamic range in the fluorescence ratio of 560%, apparent KD of 250 nm, and Hill coefficient of 1.7; Nagai et al., 2004). The large and relatively uniform peak heights in Figure 1(a) suggest near saturation of the sensor, corresponding to ≥1.4 μm Ca2+ (95% of the maximal ratio). A calibration using the above coefficients, with a recorded Rmax of 6.2 (Figure 1a) and the observed basal R of 2.0 (Figures 1 and 2), indicates a resting [Ca2+]i of approximately 120–150 nm. This approximated basal [Ca2+]i is in agreement with resting [Ca2+]i ranges reported using fura-2, indo-1 and the YC2.1 cameleon (Allen et al., 2000; Allen et al., 1999a; Grabov and Blatt, 1998; McAinsh et al., 1990; Schroeder and Hagiwara, 1990).
The guard cell [Ca2+]i in freshly prepared intact Arabidopsis leaf epidermi continuously perfused with extracellular buffer containing 50 μm added Ca2+ (Ca2+ buffer C, see Experimental procedures) showed frequent spontaneous [Ca2+]i transients (Figure 1c), similar to previous reports of spontaneous [Ca2+]i transients in Arabidopsis guard cells (Allen et al., 2001; Allen et al., 1999b; Klüsener et al., 2002; Yang et al., 2008; Young et al., 2006). The form and period of these transients were sometimes regular (Figure 1d, gray trace). However, there was no apparent consistent coordination of the spontaneous transient rises even in a guard cell pair from the same stomate (Figure 1d), which agrees with findings from previous research using YC2.1 (Allen et al., 1999b). The baseline ratios from which these transient [Ca2+]i elevations increase (Figure 1c,d) was relatively stable among experiments, with a range of approximately 1.7–2.0 (n > 15 experiments).
Spontaneous and induced [Ca2+]i transients are abolished by buffering external free [Ca2+] with BAPTA
Previous reports have observed that [Ca2+]i transients in guard cells could be inhibited by removal of extracellular Ca2+ with EGTA or BAPTA (Klüsener et al., 2002) and by incubating the cells with the membrane-permeable chelator BAPTA-AM in the absence of added Ca2+ in the external buffer (Young et al., 2006). In this study, spontaneous [Ca2+]i transients were observed during perfusion with 50 μm Ca2+ added to the extracellular buffer (Ca2+ buffers B and C, see Experimental procedures). Subsequent perfusion with a buffer that reduced the extracellular free [Ca2+] to 200 nm (BAPTA buffer, see Experimental procedures) abolished these spontaneous [Ca2+]i transients (Figure 2a; n > 20 experiments). The baseline fluorescence ratio remained between 1.7 and 2 when [Ca2+]i transients were abolished, which was equivalent to the [Ca2+]i baseline in guard cells showing repetitive [Ca2+]i transients (Figure 1).
Extracellular perfusion with hyperpolarizing buffer containing 1 mm Ca2+ rapidly restarted the [Ca2+]i transients (Figure 2a). After leaf epidermes had been held in BAPTA buffer with 200 nm free Ca2+ for 3 h or more, perfusion with 50 μm Ca2+ was sufficient to cause transient elevations in [Ca2+]i (Figure 2b). When guard cells were loaded with BAPTA using the membrane-permeable fura-2 analog BAPTA-AM, spontaneous Ca2+ elevations were inhibited (Figure 2c,d). An extracellular pulse of 50 μm Ca2+ following BAPTA-AM incubation caused increases in guard cell [Ca2+]i (Figure 2c; n = 10 treatments). When guard cells were loaded using BAPTA-AM, the baseline [Ca2+]i ratio of guard cells was similar to that of controls (Figures 1 and 2c,d; n > 10 experiments).
Further experiments were performed to reduce [Ca2+]i in guard cells to below typical resting levels by extracellular perfusion with millimolar concentrations of EGTA or BAPTA without added Ca2+. However, in Arabidopsis guard cells, EGTA did not reduce [Ca2+]i ratios in guard cells, and extracellular BAPTA (2 mm) only caused intracellular [Ca2+]i reductions in 12 of 43 guard cells. These findings confirm that YC3.6 can report lower [Ca2+]i levels than resting levels (Nagai et al., 2004), and also suggest relatively robust physiological mechanisms that maintain resting [Ca2+]i in Arabidopsis guard cells.
As incubation in BAPTA buffer with 200 nm free Ca2+ and loading guard cells with BAPTA-AM abolished spontaneous transient increases in [Ca2+]i, we examined the effect of BAPTA buffer on ABA-induced [Ca2+]i increases. Leaf epidermes were incubated in the light in BAPTA buffer for 3 h to open stomata, and loaded with BAPTA-AM for 10 min. The epidermes were then transferred to a microscope stage to monitor [Ca2+]i. Guard cells in these preparations showed no clear increases in [Ca2+]i during 20 min after application of 2 μm ABA (Figure 2d) (n = 12 experiments). As a control to test whether guard cells could report [Ca2+]i elevations, cells were exposed after this treatment to extracellular buffer containing 50 μm Ca2+. Guard cell [Ca2+]i transient elevations were triggered in response to the extracellular 50 μm Ca2+ buffer (Figure 2c,d). These Ca2+-imaging experiments showed that buffering extracellular Ca2+ to a low 200 nm level with 54 mm extracellular K+ in conjunction with BAPTA-AM loading maintained typical baseline resting [Ca2+]i ratios, while providing a robust experimental approach to abolish spontaneous [Ca2+]i transient elevations and ABA-induced [Ca2+]i increases in guard cells, thus enabling controlled analyses of the Ca2+ dependence of ABA responses.
Using BAPTA buffer (200 nm free Ca2+) in conjunction with BAPTA-AM loading, we examined the effect of blocking ABA-regulated [Ca2+]i increases on ABA-induced stomatal closure using the same solutions used for [Ca2+]i imaging. Intact epidermes from wild-type plants were placed in a high extracellular Ca2+ solution buffered to 0.2 mm free Ca2+ or a low extracellular Ca2+ solution buffered to 200 nm free Ca2+ (‘BAPTA buffer’) and incubated in the light for 3 h to open stomatal pores. Epidermes treated with BAPTA buffer (200 nm Ca2+) were also incubated with 10 μm BAPTA-AM for 10 min in the light prior to ABA application to further prevent [Ca2+]i increases (Figure 2). After incubation in the light, photomicrographs of the epidermes were obtained for measurements of stomatal apertures. The epidermes were then exposed to ABA and incubated in the light for 2 h. Photomicrographs of the same leaf epidermis regions were taken for later measurement of the same stomatal pore apertures after ABA exposures. The pre- and post-treatment apertures of the same stomata were compared pairwise to determine the degree of stomatal closure.
Guard cell viability remained intact after 5 h in the 200 nm free Ca2+ external BAPTA buffer solution. The viability dye fluorescein diacetate was added after completion of imaging experiments to monitor whether the analyzed guard cells remained viable (Young et al., 2006) (see Experimental procedures). Only a few guard cells failed to show fluorescence following fluorescein diacetate staining after several hours of experiments. In addition, the fact that stomata remained open also demonstrates viability, as maintenance of turgor requires an intact plasma membrane and active guard cell signaling and metabolism.
Stomatal apertures were individually mapped and tracked to determine the response of each stomatal aperture (Allen et al., 2001). In 0.2 mm free [Ca2+] buffer, incubation in the light for 2 h without ABA following the initial 3 h opening treatment caused the stomatal apertures to increase by an additional 5% (Figure 3); this small change was not significant at the 95% confidence level. In the 0.2 mm free [Ca2+] buffer, ABA-induced stomatal closure showed a sigmoidal response as a function of ABA concentration (Figure 3, open triangles).
When BAPTA buffer and BAPTA-AM were used to prevent [Ca2+]i increases, incubation in the light without ABA following the initial 3 h stomatal opening treatment caused stomatal apertures to increase an additional 12%; this change was significant (P < 0.01) (Figure 3). However, in this low-Ca2+ BAPTA buffer, ABA caused only a partial attenuated stomatal closure response (Figure 3, open squares). At all concentrations of ABA, stomatal closure under conditions that inhibited [Ca2+]i elevation was reduced. At the highest ABA concentration (20 μm), stomatal closure under conditions that inhibited [Ca2+]i elevations was approximately 30% of the stomatal closure under high [Ca2+] conditions (Figure 3). Similar results were obtained when final absolute stomatal apertures rather than the percentage of initial aperture are plotted (Figure 3, inset).
Although we measured stomatal closure after 2 h to determine the steady-state response of individually mapped stomata (Figure 3), much of the stomatal closure response occurs more rapidly after ABA exposure. We examined whether the reduced response to ABA in the BAPTA treatment was detectable in the early ABA response. The early time course of stomatal closure in response to ABA was analyzed in individually mapped and imaged stomata as shown in Figure 4. In both Ca2+ and [Ca2+]i-elevation-inhibited treatments, the first visible and significant reduction in stomatal aperture occurred at the 4–10 min time range (BAPTA treatment: P < 0.05 at 10 min). The response of the stomata in the low-Ca2+ BAPTA buffer was substantially reduced compared to those in physiological Ca2+ buffer (Figure 4a).
Note that when all stomatal apertures for each experiment are averaged together (Figure 4b), the error range of the responses was larger than when the relative (%) response was analyzed (Figure 4a). This increase in error (Figure 4b) is attributed to the fact that, in these experiments, each stomatal aperture was individually mapped and the time course of each stomatal aperture was determined. When the degree of the response of each individual stomatal aperture was plotted, we found that less variability occurred in the relative response of each individual stomate (Figure 4a). This is further illustrated in Figure 4(c,d), which show the relative responses of each individual stomate within an imaged field (Figure 4c), illustrating that more open stomata show a larger stomatal aperture change (Figure 4d). These data also illustrate the range of stomatal apertures that are found in Arabidopsis favoring the use of blind stomatal response assays when multiple stomatal apertures are measured in parallel-treated samples (e.g. Allen et al., 1999a; Hugouvieux et al., 2001; Murata et al., 2001), or the use of independent data confirmations by more than one experimenter (Ichida et al., 1997; Pei et al., 1997). Thus, when [Ca2+]i increases were prevented by BAPTA treatment, a greatly attenuated and slowed ‘[Ca2+]i-elevation-independent’ ABA-induced stomatal closure is revealed by analyzing individually mapped stomata (Figures 3 and 4).
ABA-induced stomatal closure in ABA-insensitive mutants
Responses of ABA-insensitive signaling mutants have not yet been analyzed under [Ca2+]i-elevation-inhibited conditions. To examine whether the partial ABA-induced stomatal closure under Ca2+-elevation-inhibited conditions bypasses ABA signaling mutants, we examined the ABA response in the following ABA-insensitive mutant lines ost1-2, abi1-1, abi2-1, gca2 and pldα1 (Himmelbach et al., 1998; Koornneef et al., 1984; Mishra et al., 2006; Mustilli et al., 2002; Yoshida et al., 2002). Stomatal closure induced by ABA was examined in these mutants and compared to wild-type responses. ABA-induced stomatal closure responses in the presence of 0.2 mm extracellular [Ca2+] are shown in Figure 5. ABA caused small mean stomatal closure responses in the Landsberg erecta mutants ost1-2, abi2-1 and gca2 in the presence of 0.2 mm extracellular Ca2+ (ost1-2, P < 0.05; abi2-1, P < 0.05; gca2, P < 0.01, paired t test). The degree of ABA-induced stomatal closure was significantly less in the Ler mutants than that in the wild-type (e.g. P < 0.02 for gca2, n = 3 experiments, one-tailed t test).
Results of ABA application to the same mutants under conditions that inhibited [Ca2+]i elevations are shown in Figure 5 (BAPTA). ABA application to the ost1-2, abi1-1 and abi2-1 mutants caused no average reduction in stomatal apertures. ABA application to the gca2 mutant caused a weak average stomatal closure response; the amount of stomatal closure in the gca2 mutant was significantly less than the stomatal closure in the presence of 0.2 mm Ca2+ (P < 0.05, n = 3 experiments). Thus the above ABA signaling mutants showed ABA insensitivity in the Ca2+-treated samples, and a further average attenuation of slight ABA responses in BAPTA-treated samples.
The phospholipase D mutant, pldα1, was selected for analyses here based on reported abolishment of ABA-induced stomatal closure (Mishra et al., 2006). The pldα1 mutant is in the Columbia wild-type background and was therefore analyzed in separate experiments. Homozygous T-DNA insertions and the lack of full-length PLDα1 mRNA were confirmed by PCR and RT-PCR (data not shown) (Mishra et al., 2006). The previously described pldα1 T-DNA knockout allele (Mishra et al., 2006) showed an ABA-induced stomatal closure response under the imposed conditions (Figure 6). We examined whether this response showed a statistical partial ABA insensitivity compared to wild-type (Columbia), and found that the difference in stomatal apertures between wild-type and pldα1 with no added ABA was not significant (Figure 6, ‘Calcium’; one-tailed t test, P > 0.27, n = 5 experiments for pldα1). The difference in stomatal apertures between wild-type and pldα1 in the presence of ABA was also not significant (Figure 6, ‘Calcium + ABA’; one-tailed t test, P > 0.28). However, when ABA responses in pldα1 without and with ABA were subtracted from one another and compared to the same change in wild-type, pldα1 stomata appeared to show a significant partial ABA hyposensitivity in this subtractive analysis (Figure 6, ‘Calcium’ − ‘Calcium + ABA’; one-tailed t test, P < 0.02). In BAPTA-treated samples, the ABA response was strongly attenuated in both wild-type and pldα1 (Figure 6, ‘BAPTA’ versus ‘BAPTA + ABA’).
The above analyses show that although guard cells show spontaneous [Ca2+]i elevations (Figures 1 and 2), ABA-induced stomatal closure exhibits a clear Ca2+ dependence (Figures 3 and 4). To further investigate this apparent paradox, we analyzed [Ca2+]i activation of S-type anion channels with and without ABA pre-incubation, as this response is an important mechanism in ABA- and Ca2+-induced stomatal closure (Allen et al., 2002; Mori et al., 2006; Negi et al., 2008; Pei et al., 1997; Schroeder and Hagiwara, 1989; Vahisalu et~al., 2008). We found that [Ca2+]i activation of S-type anion channels in patch-clamped Arabidopsis guard cell protoplasts required that protoplasts were pre-exposed to ABA, as illustrated in Figure 7. When guard cell protoplasts were not exposed to ABA, only small whole-cell currents were measured, both when [Ca2+]i was buffered to 150 nm (Figure 7a,b) and when it was buffered to 2 μm (Figure 7c top,d) under the imposed conditions (see also Allen et al., 2002). Interestingly, when protoplasts were pre-exposed to ABA, S-type anion currents were activated when [Ca2+]i was buffered to 2 μm (Figure 7c bottom,d). No significant ABA activation of S-type anion currents was observed when [Ca2+]i was buffered to 150 nm (Figure 7a,b). These data show that ABA enhances the [Ca2+]i sensitivity of S-type anion channel activation (Figure 7). These findings also suggest a mechanism that explains the apparent paradox between spontaneous [Ca2+]i elevations and the [Ca2+]i dependence of stomatal closure (Figures 1–6).
Cytosolic Ca2+ is known to cause down-regulation of inward-rectifying K+ (K+in) channels (Schroeder et al., 1987) in Vicia faba guard cells (Grabov and Blatt, 1999; Kelly et al., 1995; Schroeder and Hagiwara, 1989). Interestingly, however, this [Ca2+]i down-regulation of K+in channels has not yet been reported in Arabidopsis guard cells, and previous preliminary experiments indicated a lack of this response in Arabidopsis guard cells (Z.-M. Pei, Biology Department, Duke University, unpublished results). Further experiments confirmed that, in contrast to Vicia faba, elevation of [Ca2+]i to 2 μm did not significantly down-regulate K+in channel currents in Arabidopsis guard cell protoplasts, as illustrated in Figure 8(a,d top). The mean K+in channel current magnitudes were not significantly decreased when [Ca2+]i was buffered to 2 μm using EGTA in the patch-clamp pipette solution compared to when [Ca2+]i was buffered to 150 nm in the absence of ABA (filled circles in Figure 8b,e; P = 0.24 at −189 mV and P = 0.60 at −169 mV for experiments without ABA). Previous research in the larger Vicia faba guard cells demonstrated that using BAPTA rather than ABA in the pipette solution allows more effective [Ca2+]i buffering and K+in channel down-regulation (Kelly et al., 1995). In Arabidopsis guard cells, buffering [Ca2+]i to 2 μm using 4 mm BAPTA did not down-regulate K+in channel currents either, illustrating a robust [Ca2+]i insensitivity under the imposed conditions in the absence of ABA (filled circles in Figure 8c,f; P = 0.38 at −169 mV for experiments without ABA).
We next analyzed whether ABA exposure mediates [Ca2+]i down-regulation of K+in channels in Arabidopsis. ABA pre-incubation without elevating [Ca2+]i in guard cells (150 nm [Ca2+]i) caused no detectable ABA regulation of K+in currents using both EGTA (Figure 8a,b) and BAPTA (Figure 8c). Note that ABA was also added to the pipette solution that dialyzes the cytoplasm in these experiments. These data are consistent with reports that ABA does not always down-regulate K+in channel currents (Lemtiri-Chlieh and MacRobbie, 1994; Pei et al., 1997; Wolf et al., 2006).
Interestingly, with ABA application, [Ca2+]i elevation to 2 μm clearly down-regulated K+in channel currents when EGTA (Figure 8d,e) or BAPTA (Figure 8f) were used in the pipette solution. Taken together, the new findings in Figures 7 and 8 provide strong evidence that ABA enhances (primes) the [Ca2+]i sensitivity of Arabidopsis guard cells towards activation of S-type anion channels and [Ca2+]i down-regulation of K+in channels.
Studies have demonstrated that removal of external Ca2+ by chelation inhibits spontaneous [Ca2+]i elevations, as well as [Ca2+]i elevations induced by ABA, CO2 and pathogenic elicitors (Klüsener et al., 2002; Webb et al., 1996). It has been reported that extracellular Ca2+ need not be entirely removed to completely inhibit the [Ca2+]i response induced by pathogen elicitors in parsley (Blume et al., 2000). This [Ca2+]i response, which consisted of a rapid peak followed by a lower plateau, was completely blocked at 200 nm free extracellular Ca2+. In a recent study of [Ca2+]i responses to cold shock in Arabidopsis using an aequorin reporter, a similar temporal [Ca2+]i pattern of a sharp peak followed by a broad plateau was modelled as a consequence of an initial [Ca2+]i rise in many cells, followed by out-of-phase [Ca2+]i oscillations over multiple cells (Dodd et al., 2006). We hypothesized that buffering extracellular free [Ca2+] to 200 nm could provide ‘Ca2+-neutral’ conditions. Under these ‘Ca2+-neutral’ conditions, the long-term viability of guard cells was not adversely affected. Time-resolved imaging of individual guard cells showed that these conditions inhibited spontaneous [Ca2+]i elevations, but did not reduce the baseline resting [Ca2+]i.
YC3.6 is sensitive to changes in the 50 nm to 1.4 μm range of Ca2+ concentrations (Nagai et al., 2004). Therefore, it is unlikely that even small cell-wide changes in [Ca2+]i would be missed (Nagai et al., 2004). Our methodology could have missed rapid [Ca2+]i spikes of <1 sec duration, or localized microdomain [Ca2+]i elevations; however, the [Ca2+]i elevations previously reported in guard cells using the rapid BAPTA-based fura-2 and indo-1 reporters have durations of many seconds (Allan et al., 1994; Allen et al., 1999a; Grabov and Blatt, 1998; Marten et al., 2007; McAinsh et al., 1990; Schroeder and Hagiwara, 1990). Therefore, the observed lack of [Ca2+]i transients in the BAPTA treatment and the observed stable baseline [Ca2+]i fluorescence ratios indicate that [Ca2+]i was effectively clamped close to resting levels.
Here we demonstrate that ABA induces an attenuated and slowed stomatal closure response in Arabidopsis epidermes under conditions that prevented [Ca2+]i increases in guard cells. The substantially larger stomatal closure when [Ca2+]i increases occur shows an important function of ABA-induced Ca2+ signaling.
ABA primes the [Ca2+]i sensitivity of S-type anion and K+in channel regulation
The occurrence of spontaneous repetitive [Ca2+]i transients (Allen et al., 1999b; Klüsener et al., 2002; Staxen et al., 1999; Yang et al., 2008; Young et al., 2006) appears to contradict the conclusion that stomatal closure is [Ca2+]i-dependent. Previous research led to the hypothesis that the physiological stimulus of stomatal closure, CO2, enhances the Ca2+ sensitivity of stomatal closure mechanisms, thus allowing Ca2+ signaling to proceed (the Ca2+ sensitivity priming hypothesis) (Young et al., 2006). Interestingly, here we found that pre-exposure of Arabidopsis guard cell protoplasts to ABA enhances the ability of [Ca2+]i to activate S-type anion channels (Figure 7). Furthermore, ABA exposure also enabled [Ca2+]i down-regulation of K+in channels (Figure 8).
Note that no significant ABA activation of S-type anion currents was resolved when [Ca2+]i was buffered to 150 nm (Figure 5a,b), indicating that additional components, such as pH changes (Blatt and Armstrong, 1993; Wang et al., 2001), probably function in this stomatal closure mechanism. In Vicia faba guard cells, ion channels appear to be primed for [Ca2+]i regulation, even without exogenous ABA application (Grabov and Blatt, 1999; Schroeder and Hagiwara, 1989). However, the question of whether ABA may further enhance the [Ca2+]i sensitivity of these ion channels may be of interest for future studies. Furthermore, even in Arabidopsis guard cells, the ability of [Ca2+]i to activate S-type anion channels requires high external Ca2+ pre-treatment (in the absence of ABA pre-treatment), showing conditional priming/de-priming of [Ca2+]i-sensitive signaling (see Figure 3 in Allen et al., 2002).
Residual ABA signaling dependence on resting intracellular Ca2+ levels
Previous studies have demonstrated that injecting high concentrations of the rapid Ca2+ chelator BAPTA into Commelina communis and Vicia faba guard cells using microelectrodes effectively abolished ABA responses, including ABA-induced transcriptional responses, and anion channel activation (Levchenko et al., 2005; Webb et al., 2001). Micro-injection of the small Arabidopsis guard cells with BAPTA is not practical for analyzing many stomatal responses, and in the present study extracellular application of millimolar concentrations of BAPTA only reduced the [Ca2+]i level below resting levels in a minority of Arabidopsis guard cells. Therefore, our data cannot distinguish whether the residual ABA response at resting [Ca2+]i (Figures 3 and 4) is dependent on physiological resting [Ca2+]i (Figures 1 and 2). The only condition we have found so far that consistently reduced the [Ca2+]i level of Arabidopsis guard cells is treatment with the ADP ribosyl cyclase inhibitor nicotinamide (Klüsener et al., 2002). These findings are consistent with a role for cyclic ADP ribose (cADPR) in ABA responses (Leckie et al., 1998; Wu et al., 1997), and also with a recent study showing that nicotinamide represses the baseline circadian Ca2+ oscillation in Arabidopsis leaves (Dodd et al., 2007). ABA-induced stomatal closure was completely abolished by 50 mm nicotinamide, but not 50 mm sorbitol (Stephen Lee Biology Division, University of California, San Diego, and J.I.S., unpublished results), but this concentration of pharmacological blocker for lowering baseline [Ca2+]i is fairly high, and may inhibit additional ABA signaling components. The present results do not exclude the presence of an additional Ca2+-independent pathway (Allan et al., 1994), although a model suggesting that [Ca2+]i elevation is not required for the ABA response (Levchenko et al., 2005) is not supported by the present findings. Previous BAPTA injection studies in Commelina and Vicia faba (Levchenko et al., 2005; Webb et al., 2001) support a hypothesis in which Ca2+ sensitivity priming may enable [Ca2+]i to mediate the observed attenuated ABA response at resting [Ca2+]i (Figures 3 and 4).
[Ca2+]i-elevation-independent ABA-induced stomatal closure in ABA-insensitive mutants
In summary, the present study quantifies an attenuated and greatly slowed abscisic acid-induced stomatal closure response by experimentally clamping [Ca2+]i to resting levels. Interestingly, ABA is shown to enable [Ca2+]i to activate S-type anion channels and to down-regulate K+in channels. These ion-channel regulation findings provide strong mechanistic support for a hypothesis in which ABA enhances (primes) the [Ca2+]i sensitivity of stomatal closure mechanisms, thus enabling specificity of Ca2+ signaling (Young et al., 2006). The ABA insensitive mutants abi1-1, abi2-1, ost1 and gca2 also show Ca2+ dependence, consistent with Ca2+-dependent ABA signaling.
Stomatal movements of mapped stomata
To avoid effects of leaf-to-leaf variability, intact epidermal preparations were prepared from a single leaf for two treatments by attaching the abaxial surface of leaf halves to 25 mm round glass cover slips using medical adhesive (Hollister Inc., http://www.hollister.com), and then gently removing the upper leaf cell layers using a razor blade. Cover slips were pre-marked on the bottom with a colored grid to allow rapid re-examination of previously photographed stomatal locations. The time course of each stomatal aperture was individually tracked (Allen et al., 2001), allowing accurate analyses of the response of each individual stomatal aperture. Intact epidermis were immediately submerged in 3 ml of test buffer in 35 mm covered petri dishes, and placed under 300 μE red/blue light (230 μE red + 70 μE blue) from an array of light-emitting diodes (LUXEON III; Philips Lumileds Lighting Company, http://www.philipslumileds.com) at room temperature (21–23°C) for 3 h to open stomatal apertures. The following test buffers were used: BAPTA buffer (with 200 nm free Ca2+): 25 mm dipotassium (K2) iminodiacetic acid, 2 mm K2 BAPTA, 0.7 mm CaCl2, 10 mm MES/Tris KOH pH 6.1; Ca2+ buffer A (with 0.2 mm free Ca2+): 25 mm K2-iminodiacetic acid, 2 mm K2-EGTA, 1 mm CaCl2, 10 mm MES/Tris pH 5.6; Ca2+ buffer B (with 50 μm CaCl2 added): 28 mm K2-iminodiacetic acid, 50 μm CaCl2, 1.3 mm KCl, 10 mm MES/Tris pH 6.1, Ca2+ buffer C (with 50 μm CaCl2 added): 5 mm KCl, 50 μm CaCl2, 10 mm MES/Tris pH 5.6. Nicotinamide was added at a final concentration of 50 mm, and 50 mm sorbitol was added to controls (Dodd et al., 2007; Leckie et al., 1998). Epidermes treated with BAPTA buffer were also loaded with BAPTA-AM for 10 min where indicated during the third hour of light exposure by incubation in the same BAPTA buffer to which were added 0.3 mm eserine, 0.025% Pluronic F-127 (CAS No. 11104975) and 10–25 μm BAPTA-AM. The methods used were similar to previously reported methods that allow cytoplasmic loading of the BAPTA analog fura-2 (Kuchitsu et al., 2002) and BAPTA (Young et al., 2006) into guard cells without the need for micro-injection.
Each intact epidermal preparation was set into a viewing chamber (a 12 mm hole pre-drilled in a glass slide) by sealing the cover slip edges with silicon vacuum grease (Dow Corning Corp., http://www.dowcorning.com), and covered with 0.2 ml buffer from the incubation dish. The epidermal strip was examined using a 40× objective on an inverted microscope (Nikon TS 100; Nikon Instruments Inc., http://www.nikon.com) focused on the inner edge of stomata (Ichida et al., 1997), and an image of open stomata that contained at least five stomata was acquired using a digital color camera (CFW-1310C, Scion Corp., http://www.scioncorp.com). After image acquisition, the epidermal chamber was removed from the viewing chamber, exposed to the indicated ABA-containing buffers, and returned to the blue/red light-emitting diode array for 1 h (Figures 6–8) or 2 h (Figure 3). The epidermal chamber was returned to the viewing chamber, and a second image was stored of the same field of view as in the first image. Widths of the stomatal apertures were measured using imagej software (National Institutes of Health, http://rsbweb.nih.gov/ij), compared on a pairwise basis as the percentage of the initial aperture of the same stomata, and all pairwise comparisons for a treatment were averaged. The experimenter was blind with respect to the ABA concentrations used. When mutants were used in experiments, the experimenter was also blind with respect to the genotype.
Viability staining. At the end of experiments, after having acquired stomatal aperture images, fluorescein diacetate was added to the buffer, and cells exhibiting green fluorescence within 5 min were designated as viable, as previously described (Young et al., 2006).
Time course analyses of stomatal closure. Epidermal strips were prepared and imaged as described above. After imaging, buffer was perfused (e.g. ABA added) at time 0. A second image was acquired within 30 sec, and then images were collected at 2 min intervals from time 0. The widths of individual mapped stomatal apertures were measured and analyzed in pairwise comparisons as described above, and the results at each time point were compared to the original image of the same stomate.
Each experiment was repeated in at least three independent trials. Calculation of standard error for stomatal apertures is based on the number of independent trials as reported in previous studies (Allen et al., 2000; Murata et al., 2001), rather than the larger total number of independent stomata, which would give a much lower estimate of the error of the measurements as previously described (Li et al., 2006).
All Ca2+ imaging experiments in this study were performed using a TE300 inverted microscope with a TE-FM epi-fluorescence attachment (Nikon Inc.). Excitation using a 75 W xenon lamp (Osram, http://www.osram.com) was attenuated by 97% (3% light transmission) by using both 4× and 8× neutral density filters to reduce exposure of the fluorescent reporters and cells to epifluorescence excitation. Wavelength specificity was achieved using a cameleon filter set [440/20 nm excitation, 485/40 nm emission 1, 535/30 nm emission 2, 455 nm DCLP (dichroic long pass) dichroic; filter set 71007a (Chroma Technology, http://www.chroma.com)]. The filter wheel, shutter and CCD camera were controlled using metafluor software (MDS Inc., http://www.mdsinc.com). For observation of imposed Ca2+ oscillations, intact epidermes from leaves of pGC1:Col plants were prepared for microscopy as previously described (Mori et al., 2006; Yang et al., 2008). On the microscope stage, intact epidermes were perfused with depolarizing buffer (25 mm K2-iminodiacetate, 100 μm BAPTA, 10 mm MES/Tris buffer pH 6.1) for 10 min to obtain a background value. Subsequently, hyperpolarizing buffer containing Ca2+ (1 mm K2-iminodiacetate, 1 mm CaCl2, 10 mm MES/Tris buffer pH 6.1) was applied for 2 min, followed by 5 min of depolarizing buffer (Figure 1a).
Arabidopsis guard cell protoplasts were isolated from rosette leaves of 4–6-week-old plants using a protoplast isolation solution containing 1.0% Cellulase R10, 0.5% Macerozyme R10 (Yakutt Horisha Co. Ltd, http://www.yakutt.co.jp/ypi/en/product.html), 0.5% bovine serum albumin, 0.1% kanamycin, 10 mm ascorbic acid, 0.1 mm KCl, 0.1 mm CaCl2 and 500 mm d-mannitol (buffered to pH 5.5 using KOH). Whole-cell patch-clamp experiments were performed as described previously (Allen et al., 2002). For analyses of S-type anion currents, the pipette solution contained 150 mm CsCl, 2 mm MgCl2, 6.7 mm EGTA, 5 mm Mg-ATP, 10 mm HEPES/Tris pH 7.1, and CaCl2 was added to give the desired free Ca2+ concentrations. Guard cell protoplasts were pre-exposed to ABA (20 μm) for 15 min prior to patch clamping, when indicated. The bath solution contained 30 mm CsCl, 2 mm MgCl2, 1 mm CaCl2 and 10 mm Mes/Tris pH 5.6. Note that gigaohm seal formation was performed in a 1 mm Ca2+ bath solution (Allen et al., 2002), not in a high-CaCl2 bath solution.
For analyses of inward-rectifying potassium channel currents, the pipette solution contained 30 mm KCl, 70 mm K-glutamate, 2 mm MgCl2, 5 mm Mg-ATP, 5 mm Tris-GTP and 10 mm HEPES/Tris pH 7.1. In addition, the free Ca2+ concentration was buffered with 6.7 mm EGTA and 2.61 mm CaCl2 (150 nm free Ca2+) or 6.03 mm CaCl2 (2 μm free Ca2+), and, for experiments in Figure 8(c,f), EGTA and CaCl2 were replaced by 4 mm BAPTA and 2.12 or 3.8 mm CaCl2 to give the indicated free Ca2+ concentrations. The bath solution contained 30 mm KCl, 2 mm MgCl2, 1 mm CaCl2 and 10 mm MES/Tris pH 5.5. When indicated, the pipette solution also contained 50 μm ABA. Guard cell protoplasts were pre-exposed to extracellular ABA (50 μm) for 20 or 120 min prior to patch clamping for K+in channel and S-type anion channel analyses, respectively.
We thank Gethyn Allen for discussions during early planning of this project, Jean Colcombet and Danessa Lerias for pilot experiments using external EGTA solutions, Jean Colcombet for comments on the manuscript, and Stephen Lee for additional stomatal movement analyses. We thank Zhen-Ming Pei for unpublished preliminary data on reduced [Ca2+]i regulation of K+in channels in Arabidopsis guard cells. We thank X. Wang for providing pldα1, and J. Giraudat and J. Leung for providing ost1 mutants. This research was supported by National Institutes of Health grant R01GM060396, by National Science Foundation grant MCB0417118, and in part by Department of Energy grant DOE-DE-FG02-03ER15449 to J.I.S and in part by Natural Science Foundation of China 2070/028 to S.X.