Regulation of reactive oxygen species-mediated abscisic acid signaling in guard cells and drought tolerance by glutathione

The phytohormone abscisic acid (ABA) induces stomatal closure in response to drought stress, leading to reduction of transpirational water loss. A thiol tripeptide glutathione (GSH) is an important regulator of cellular redox homeostasis in plants. Although it has been shown that cellular redox state of guard cells controls ABA-mediated stomatal closure, roles of GSH in guard cell ABA signaling were largely unknown. Recently we demonstrated that GSH functions as a negative regulator of ABA signaling in guard cells. In this study we performed more detailed analyses to reveal how GSH regulates guard cell ABA signaling using the GSH-deficient Arabidopsis mutant cad2-1. The cad2-1 mutant exhibited reduced water loss from rosette leaves. Whole-cell current recording using patch clamp technique revealed that the cad2-1 mutation did not affect ABA regulation of S-type anion channels. We found enhanced activation of Ca2+ permeable channels by hydrogen peroxide (H2O2) in cad2-1 guard cells. The cad2-1 mutant showed enhanced H2O2-induced stomatal closure and significant increase of ROS accumulation in whole leaves in response to ABA. Our findings provide a new understanding of guard cell ABA signaling and a new strategy to improve plant drought tolerance.

It has been suggested that guard cell ABA signaling involves redox regulation. Reactive oxygen species (ROS) including hydrogen peroxide (H 2 O 2 ) serve as a key mediator of ABA activation of I Ca channels (Pei et al., 2000;Murata et al., 2001;Kwak et al., 2003). Exogenous application of H 2 O 2 activates I Ca channels and evokes guard cell [Ca 2+ ] cyt increases (Pei et al., 2000). Plasma membrane NAD(P)H oxidases are responsible for ABA-induced ROS production in guard cells (Kwak et al., 2003). Arabidopsis glutathione peroxidase 3 (AtGPX3) was shown to function as both a ROS scavenger and a ROS signal transducer in ABA signaling (Miao et al., 2006). Emerging evidences suggest that ROS production by apoplastic enzymes such as cell-wall bound peroxidases is also involved in induction of stomatal closure Khokon et al., 2011;Hossain et al., 2013).
Glutathione (GSH) is the most abundant non-protein thiol compound in plants and a key regulator of cellular redox homeostasis. GSH is involved in various physiological processes including growth, development, and defense response to biotic and abiotic stresses (May et al., 1998;Noctor and Foyer, 1998). Previously we reported that ABA as well as methyl jasmonate (MeJA) decreases the GSH contents of guard cells (Akter et al., 2010;Okuma et al., 2011). Arabidopsis GSH-deficient mutants, chl-1 and cad2-1 exhibit enhanced ABA-induced and MeJAinduced stomatal closure and a membrane permeable derivative of GSH, GSH monoethyl ester (GSHmee) restored the stomatal phenotype of chl-1 and cad2-1 mutants Akter et al., 2010Akter et al., , 2012Akter et al., , 2013Okuma et al., 2011), demonstrating that GSH functions as a negative regulator of ABA and MeJA signaling in guard cells. However, the detailed mechanism of how GSH modulates the guard cell responses is still unclear.
In this study, we analyzed GSH regulation of ABA signaling in guard cells using the Arabidopsis GSH-deficient mutant cad2-1. The cad2-1 mutant is deficient in the first GSH biosynthesis enzyme, γ-glutamylcysteine synthetase. We found that the cad2-1 mutation causes enhanced ROS activation of I Ca channel and ABA-induced ROS accumulation in apoplast. A new signal model for regulation of ROS-mediated ABA signaling by GSH in guard cells is also proposed.

PLANT MATERIAL AND GROWTH
The Arabidopsis ecotype Columbia (Col) and cad2-1 mutant plants were grown on soil mixture of 70% (v/v) vermiculite (Asahi-kogyo, Okayama, Japan) and 30% (v/v) Sakata Supermix-A (Sakata Seed Corporation, Yokohama, Japan) in growth chambers at 21 • C under a 16-h-light/8-h-dark photoperiod with photon flux density of 80 μmol m −2 s −1 . Four-to six-week-old plants were used in all experiments.

WATER LOSS ASSAY
Three detached rosette leaves were placed on a plastic tray and the loss in fresh weight was monitored at the indicated times.

STOMATAL APERTURE MEASUREMENTS
Stomatal aperture measurements were performed as described previously (Munemasa et al., 2007;Okuma et al., 2011). Detached rosette leaves were floated on stomatal assay buffer containing 5 mM KCl, 50 μM CaCl 2 , and 10 mM MES-Tris (pH 5.6) for 2 h in the light to induce stomatal opening, followed by the addition of H 2 O 2 . After 2-h incubation in the light, the leaves were shredded and epidermal tissues were collected. At least 20 stomatal apertures were measured on each individual experiment.

DETECTION OF ROS ACCUMULATION IN WHOLE LEAVES
Accumulation of H 2 O 2 in whole leaves was monitored using 3,3diaminobenzidine (DAB) according to Maruta et al. (2010) with slight modifications. Detached rosette leaves were vacuum infiltrated with DAB assay buffer containing 1 mg mL −1 DAB, 5 mM KCl, 50 μM CaCl 2 , and 10 mM MES-Tris (pH 5.6). Infiltrated leaves are incubated in DAB assay buffer with or without ABA for 4 h in the light. The leaves were then decolorized by boiling in ethanol. Apoplastic ROS were visualized as a reddish-brown color and quantified using Image-J software (National Institutes of Health, USA).

STATISTICAL ANALYSIS
All statistical significance was analyzed by double-tailed Student's t-test. We regarded differences at the level of P < 0.05 as significant.

THE cad2-1 MUTANT SHOWED ENHANCED DROUGHT TOLERANCE
Previously we reported that GSH depletion by the cad2-1 mutation enhances ABA-induced stomatal closure in Arabidopsis . To assess effect of the cad2-1 mutation on drought tolerance, we monitored water loss from detached rosette leaves. As shown in Figure 1, compared to wild type, the cad2-1 mutant exhibited reduced water loss from detached rosette leaves (P < 0.018 for Col vs. cad2-1 at 120 min after leaf detachment). This result suggests that the cad2-1 mutation improves drought tolerance.

EFFECT OF THE cad2-1 MUTATION ON ABA REGULATION OF S-TYPE ANION CHANNELS.
The cad2-1 mutation does not affect ABA activation of I Ca channels, suggesting that GSH functions downstream of [Ca 2+ ] cyt FIGURE 1 | The cad2-1 mutant shows reduced transpirational water loss. Fresh weight loss of detached rosette leaves. The bars represent the mean ± SE values of five independent replicates.

Frontiers in Plant Science | Plant Physiology
November 2013 | Volume 4 | Article 472 | 2 elevation in guard cell ABA signaling . Activation of S-type anion channel is mediated by [Ca 2+ ] cyt elevation and considered as a crucial step for ABA-, MeJA-, and CO 2 -induced stomatal closure (Schroeder and Hagiwara, 1989;Pei et al., 1997;Munemasa et al., 2007;Hu et al., 2010). To confirm whether GSH depletion in the cad2-1 mutant affects ABA regulation of S-type anion channel activity, whole-cell patch-clamp analysis was performed. S-type anion currents were observed in wild-type GCPs pretreated with 10 μM ABA (P < 0.023 for Col Control vs. Col ABA at −135 mV; Figure 2). We also found that ABA evoked S-type anion currents in cad2-1 GCPs to the same extent as in wild-type GCPs (P < 0.045 for cad2-1 Control vs. cad2-1 ABA at −135 mV; Figure 2). Note that in our experimental condition, free Ca 2+ concentration in the pipette solution was buffered to 2 μM and no obvious S-type anion currents were observed in both wild-type guard cells and cad2-1 guard cells without ABA pretreatment. ABA activation of I Ca channels involves ROS as a second messenger (Pei et al., 2000;Murata et al., 2001;Kwak et al., 2003). Depletion of the major redox regulator, GSH, in the cad2-1 mutant might affect ROS-mediated ABA signaling in guard cells.
Previously we demonstrated that ABA activation of I Ca channels is not enhanced in cad2-1 guard cells . However, ROS activation of I Ca channels in cad2-1 guard cells was not yet analyzed. We examined effects of the cad2-1 mutation on ROS regulation of I Ca channels in guard cells. H 2 O 2 at 1 mM activates hyperpolarization-activated currents in both wild-type GCPs and cad2-1 GCPs (Figure 3).  Figure 4) but not wild type (P = 0.13 for Col Control vs. Col H 2 O 2 ; Figure 4). Moreover, we found that H 2 O 2 -induced stomatal closure in the cad2-1 mutant was significantly attenuated by a membrane permeable derivative of GSH, GSHmee (P < 0.047 for H 2 O 2 without GSHmee vs. H 2 O 2 with GSHmee; Figure S1). These results suggest that decreased GSH contents in cad2-1 guard cells confer enhanced stomatal response to H 2 O 2 . Accumulation of ROS in guard cell cytosol occurs during ABAinduced stomatal closure (Pei et al., 2000;Kwak et al., 2003). Previously we revealed that ABA-induced ROS accumulation in guard cell cytosol was not altered by GSH depletion (Jahan et al., 2008;Okuma et al., 2011;Akter et al., 2012). In addition to ROS produced by plasma membrane NAD(P)H oxidases, ROS produced by apoplastic enzymes such as cell-wall bound peroxidases participate in regulation of stomatal movement Khokon et al., 2011;Hossain et al., 2013). To examine whether apoplastic ROS accumulation contributes to guard cell ABA signaling, we performed DAB staining experiments, which allows us to monitor ROS produced by apoplastic enzymes as well as ROS produced by NAD(P)H oxidases (Bindschedler et al., 2006;Khokon et al., 2011;Hossain et al., 2013). Wild-type leaves did not exhibit apoplastic ROS accumulation even when treated with high concentration of ABA (50 μM) (P = 0.72 for Col Control vs. Col ABA; Figure 5). However, apoplastic ROS accumulation was significantly increased by ABA in cad2-1 leaves (P < 0.04 for cad2-1 Control vs. cad2-1 ABA; Figure 5). This result suggests that the accumulation of apoplastic ROS contributes to enhanced ABA response of cad2-1 guard cells.

DISCUSSION
Depletion of GSH enhances ABA-induced stomatal closure (Jahan et al., 2008;Okuma et al., 2011;Akter et al., 2012). However, the mechanism of how GSH regulates guard cell ABA signaling and involvement of GSH in controlling transpirational water loss have been unclear. In this manuscript, we tested involvement of GSH in controlling water loss from leaves and performed the detailed analysis of GSH regulation of guard cell ABA signaling using the GSH-deficient cad2-1 mutant. We confirmed that the cad2-1 mutation caused not only enhanced ABA-induced stomatal closure  but also reduction of water loss from leaves (Figure 1). Hence manipulation of GSH level might provide a new strategy to improve plant drought tolerance. ABA activation of S-type anion channels is not altered in the cad2-1 guard cells (Figure 2). It has been suggested that the elevated [Ca 2+ ] cyt is required for ABA activation of S-type anion channels and ABA "primes" Ca 2+ sensitivity of S-type anion channels (Siegel et al., 2009). In this study, we used the pipette solution where free [Ca 2+ ] was buffered to 2 μM (See Materials and Methods). No obvious S-type anion current was observed in cad2-1 guard cells as well as wild-type guard cells without ABA pretreatment (Figure 2), suggesting that the cad2-1 mutation does not affect the priming state of Ca 2+ sensitivity of S-type anion channels.
ABA induces guard cell [Ca 2+ ] cyt elevation via activation of plasma membrane I Ca channels (Hamilton et al., 2000;Pei et al., 2000). Previously we found identical ABA activation of I Ca channels in wild-type and cad2-1 guard cells . ROS mediates ABA activation of I Ca channels (Pei et al., 2000;Murata et al., 2001;Kwak et al., 2003). In this study, we found enhanced response of cad2-1 I Ca channels to exogenous H 2 O 2 (Figure 3). It was reported that ozone, an elicitor of ROS, induces biphasic [Ca 2+ ] cyt elevation in seedlings and depletion of GSH by the cad2-1 mutation and buthionine sulphoximine, an inhibitor of γ-glutamylcysteine synthetase, enhances first peak of the ozone-induced biphasic [Ca 2+ ] cyt response (Evans et al., 2005). These results suggest that GSH pools control sensitivity of plasma membrane Ca 2+ permeable channels to ROS and downstream Ca 2+ signals in plant cells.
Consistent with the enhancement of I Ca channel response to ROS, the cad2-1 mutant exhibits enhanced H 2 O 2 -induced stomatal closure (Figure 4). Note that depletion of GSH in guard cell cytosol is induced by ABA , but not by H 2 O 2 (Akter et al., 2013). These results imply that ABA sensitizes guard cells to ROS by decreasing GSH content via a pathway distinct from the ABA-mediated ROS production pathway (Figure 6).
Previously we reported identical cytosolic ROS accumulation induced by ABA in cytosol of wild-type and cad2-1 guard cells using 2 ,7 -dichlorodihydrofluorescein diacetate . It has been suggested that apoplastic ROS signals are also involved in regulation of stomatal movement Khokon et al., 2011;Hossain et al., 2013). In this study we monitored ROS accumulation in whole leaves using DAB. ABA induced ROS accumulation in whole leaves of the cad2-1 mutant but not in leaves of wild type (Figure 5). These results suggest that GSH depletion by the cad2-1 mutation affects ROS homeostasis in apoplastic space rather than that in guard cells during ABA-induced stomatal closure. Based on the findings obtained in this study, we present one simplified signal model shown in Figure 6. ABA decreases guard cell GSH content via ROS-independent pathway (Akter et al., 2013). The decreased GSH content causes significant ROS accumulation in apoplast ( Figure 5) and also sensitizes guard cell I Ca channels to apoplastic ROS (Figure 3) by unknown mechanism, resulting in enhanced stomatal response to ABA. In wild-type leaves, ABA decreases GSH content  but does not induce significant apoplastic ROS accumulation (Figure 5), suggesting that the apoplastic ROS signal is employed to modulate ABA responsiveness of guard cells by other stimuli rather than by ABA signaling itself. Molecular mechanisms of how GSH regulates ROS sensitivity to plasma membrane Ca 2+ permeable channels and apoplastic ROS homeostasis during ABA-induced stomatal closure would be investigated in the future.

ACKNOWLEDGMENTS
We thank Dr. Eiji Okuma (Okayama University) for critical reading of the manuscript.