Inhibition of Blue Light-Dependent H⁺ Pumping by Abscisic Acid through Hydrogen Peroxide-Induced Dephosphorylation of the Plasma Membrane H⁺-ATPase in Guard Cell Protoplasts

Abstract

Blue light (BL)-dependent H⁺ pumping by guard cells, which drives stomatal opening, is inhibited by abscisic acid (ABA). We investigated this response with respect to the activity of plasma membrane H⁺-ATPase using Vicia guard cell protoplasts. ATP hydrolysis by the plasma membrane H⁺-ATPase, phosphorylation of the H⁺-ATPase, and the binding of 14-3-3 protein to the H⁺-ATPase stimulated by BL were inhibited by ABA at 10 μm. All of these responses were similarly inhibited by hydrogen peroxide (H₂O₂) at 1 mm. The ABA-induced inhibitions of BL-dependent H⁺ pumping and phosphorylation of the H⁺-ATPase were partially restored by ascorbate, an intracellular H₂O₂ scavenger. A single-cell analysis of the cytosolic H₂O₂ using 2′,7′-dichlorofluorescin revealed that H₂O₂ was generated by ABA in guard cell protoplasts. We also indicated that H⁺ pumping induced by fusicoccin and the binding of 14-3-3 protein to the H⁺-ATPase were inhibited slightly (approximately 20%) by both ABA and H₂O₂. By contrast, H₂O₂ at 1 mm did not affect H⁺ pumping by the H⁺-ATPase in microsomal membranes. From these results, we concluded that inhibition of BL-dependent H⁺ pumping by ABA was due to a decrease in the phosphorylation levels of H⁺-ATPase and that H₂O₂ might be involved in this response. Moreover, there are at least two inhibition sites by ABA in the BL signaling pathway of guard cells.

The opening of stomata is mediated by an accumulation of K⁺ in guard cells, and K⁺ accumulation is driven by an inside-negative electrical potential across the plasma membrane (Assmann and Shimazaki, 1999; Schroeder et al., 2001). The electrical potential is created by the H⁺ pump in the plasma membrane in response to blue light (BL; Assmann et al., 1985; Shimazaki et al., 1986; Schroeder et al., 2001). Recent investigation has demonstrated that the H⁺ pump is the plasma membrane H⁺-ATPase, and BL activates the H⁺-ATPase through the phosphorylation of Thr residues in the C terminus (Kinoshita and Shimazaki, 1999). Phosphorylation then induces the binding of 14-3-3 protein to a penultimate residue of Thr in the H⁺-ATPase, which acts as a positive regulator for the H⁺-ATPase (Kinoshita and Shimazaki, 2002). Recently, phototropins (phot1 and phot2) have been identified as BL receptors that mediate the H⁺-ATPase activation in stomatal guard cells (Kinoshita et al., 2001). Phototropins are plant-specific BL receptors and are autophosphorylating Ser/Thr protein kinases with two LOV (light, oxygen, and voltage) domains (Briggs et al., 2001; Kagawa et al., 2001; Kinoshita et al., 2001; Sakai et al., 2001; Briggs and Christie, 2002).

The abscisic acid (ABA) is a key signal molecule in plants, mediating responses to various environmental stresses, and has been demonstrated to induce stomatal closure via the activation of signaling cascades, thereby preventing water loss (Assmann and Shimazaki, 1999; Schroeder et al., 2001; Hetherington, 2001). The ABA-induced stomatal closure is driven by the effluxes of Cl⁻, malate²⁻, and K⁺ from guard cells through Ca²⁺- and voltage-dependent anion channels and outward-rectifying K⁺ channels in the plasma membranes. Activations of these channels require membrane depolarization, and the depolarization can be achieved at least partly by the inhibition of the plasma membrane H⁺-ATPase. ABA inhibits BL-dependent H⁺ pumping by the H⁺-ATPase and maintains the membrane potential to depolarized state, allowing the continuous effluxes of these ions (Shimazaki et al., 1986; Goh et al., 1996; Schroeder et al., 2001). It is thus important to elucidate this ABA action in molecular levels because ABA is expected to act on guard cells under the light when water is easily lost from the leaves through stomata.

Since ABA induces increases both in cytosolic Ca²⁺ and pH in guard cells (Assmann and Shimazaki, 1999; Schroeder et al., 2001), several mechanisms can be suggested to account for the ABA-induced inhibition. Hydrogen peroxide (H₂O₂), a form of reactive oxygen species, emerged recently as a second messenger in plants (Neill et al., 2002). It also acts as an intermediate in ABA-induced stomatal closure, including the activation of plasma membrane Ca²⁺ channels (Pei et al., 2000; Murata et al., 2001), the inactivation of plasma membrane inward-rectifying K⁺ channels (Zhang et al., 2001b), and the induction of cytosolic alkalization (Blatt and Armstrong, 1993; Zhang et al., 2001a). Therefore, it is likely that ABA inhibits BL-dependent H⁺ pumping through H₂O₂.

In this study, we investigated the effects of ABA on BL-dependent H⁺ pumping in biochemical levels and found that ABA inhibited the pumping through a suppression of BL-dependent phosphorylation of the plasma membrane H⁺-ATPase in guard cell protoplasts (GCPs). We also reported that H₂O₂ mimicked the inhibitory effect of ABA on the BL-dependent H⁺ pumping.

RESULTS

ABA Inhibits ATP Hydrolysis by the Plasma Membrane H⁺-ATPase in Vicia GCPs

We confirmed a typical inhibition of BL-dependent H⁺ pumping by ABA in GCPs (Goh et al., 1996). The rate of H⁺ pumping was inhibited by 62% when 10 μm ABA was added to GCPs (Fig. 1). Since BL-dependent H⁺ pumping is mediated by the plasma membrane H⁺-ATPase (Kinoshita and Shimazaki, 1999), ATP hydrolysis was determined in GCPs that had been incubated with 10 μm ABA. Without ABA, ATP hydrolysis increased 30 s after the pulse of BL, and the hydrolysis reached a maximum at 2.5 min, decreasing gradually to the basal level within 20 min after the pulse (Fig. 2A). In the presence of ABA, the increase in ATP hydrolysis was inhibited by 60% to 70% for most sampling times, and the basal level of ATP hydrolysis under red light was also slightly decreased (Fig. 2B). BL-stimulated ATP hydrolysis was inhibited by ABA in a concentration-dependent manner (Fig. 3A). All instances of ATP hydrolysis were strongly inhibited by the addition of 100 μm vanadate, an inhibitor of the plasma membrane H⁺-ATPase (Figs. 2 A and B, and 3A). These results indicate that the inhibition of BL-dependent H⁺ pumping by ABA is due to the inhibition of plasma membrane H⁺-ATPase.

Figure 1.

Inhibition of ABA on BL-dependent H⁺ pumping in GCPs from Vicia. ABA was added to the protoplast suspension 30 min before the application of BL pulse (100 μmol m⁻² s⁻¹, 30 s). ABA was dissolved in DMSO, and the final concentration of DMSO in the reaction mixture was 0.05%. The amount of acid equivalents was determined by addition of 10 nmol of H⁺ at the end of each experiment. A typical result of three independent experiments is presented.

Figure 2.

Inhibition of BL-dependent ATP hydrolysis, phosphorylation of H⁺-ATPase, and the binding of 14-3-3 protein to H⁺-ATPase by ABA. A and B, ATP hydrolytic activities at the indicated times were measured by determining the Pi released from 2 mm ATP for 30 min. GCPs were preincubated under red light (600 μmol m⁻² s⁻¹) for 40 min in the absence (A) and presence (B) of ABA at 10 μm, then illuminated with a pulse of BL (100 μmol m⁻² s⁻¹, 30 s). Vertical bars represent ses of three separate experimental values. Vanadate was added at 100 μm. C and D, Phosphorylation levels of H⁺-ATPase in the absence (C) and presence (D) of ABA. Phosphorylation level was determined by autoradiography for the proteins immunoprecipitated by antiserum against the H⁺-ATPase. Far-western-blot analysis of the H⁺-ATPase in the absence (E) and presence (F) of ABA using recombinant glutathione S-transferase-14-3-3 protein as a probe. Western-blot analyses of H⁺-ATPase in the absence (G) and the presence (H) of ABA. ABA was added to the assay medium 30 min before the start of BL pulse. Measurement was done at 24°C. Arrowheads indicate the positions of H⁺-ATPase.

Figure 3.

Dose-dependent inhibition of BL-dependent ATP hydrolysis, phosphorylation of H⁺-ATPase, and the binding of 14-3-3 protein to H⁺-ATPase by ABA. A, ATP hydrolytic activities by H⁺-ATPase at various concentrations of ABA. B, Phosphorylation levels of the H⁺-ATPase with and without ABA. C, Far-western-blot analysis of H⁺-ATPase with and without ABA. D, Western-blot analysis of H⁺-ATPase with and without ABA. Other conditions are the same as shown in Figure 2.

ABA Reduces the Phosphorylation Levels of Plasma Membrane H⁺-ATPase and Inhibits the Binding of 14-3-3 Protein to the Phosphorylated H⁺-ATPase

Since the H⁺-ATPase activity is increased by protein phosphorylation at the C terminus of the enzyme in guard cells (Kinoshita and Shimazaki, 1999), the levels of H⁺-ATPase phosphorylation were determined after the application of ABA. Autoradiograms revealed that phosphorylation levels of the H⁺-ATPase with a molecular mass of 95 kD increased by BL in parallel with the ATP hydrolytic activities (Fig. 2, A and C). This BL-induced increase in phosphorylation was strongly suppressed by 60% to 70% in response to 10 μm ABA (Fig. 2, C and D), and this suppression was found to be concentration dependent (Fig. 3B). The results indicate that the inhibition of H⁺-ATPase by ABA was due to a decrease in the levels of phosphorylation.

The 14-3-3 protein has been shown to bind to the phosphorylated C termini of H⁺-ATPases, and this binding is required for the activation of H⁺-ATPase in guard cells (Kinoshita and Shimazaki, 1999, 2002; Emi et al., 2001). Far-western-blot analysis confirmed that 14-3-3 protein was bound to the phosphorylated H⁺-ATPase and that the amount of binding paralleled the phosphorylation levels of H⁺-ATPase (Fig. 2, C and E). ABA inhibited the binding of 14-3-3 protein to the H⁺-ATPase by decreasing the phosphorylation (Fig. 2, D and F). Inhibition was dependent on the ABA concentration (Fig. 3, B and C). Western-blot analysis revealed that the amount of H⁺-ATPase was not changed by ABA (Figs. 2, G and H, and 3D). The results indicate that the inhibition of H⁺-ATPase by ABA was due to a decrease in phosphorylation, thus preventing the binding of 14-3-3 protein to the H⁺-ATPase.

H₂O₂ Mimics ABA in Terms of the Inhibition of BL-Dependent H⁺ Pumping

H₂O₂ has recently been suggested to play a role as a second messenger in ABA signaling (Pei et al., 2000; Zhang et al., 2001c; Park et al., 2003). The finding that ABA inhibits BL-dependent H⁺ pumping prompted us to investigate the effect of H₂O₂ on the H⁺ pumping. As shown in Figure 4, H₂O₂ strongly inhibited BL-dependent H⁺ pumping, and this inhibition was concentration dependent. The rate of BL-dependent H⁺ pumping in GCPs was inhibited by 70% in the presence of H₂O₂ at 1 mm.

Figure 4.

Effect of H₂O₂ on BL-dependent H⁺ pumping in GCPs from Vicia. H₂O₂ was added to the protoplast suspension at indicated concentrations 30 min before the application of BL pulse. A typical result of three independent experiments is presented. Other conditions are the same as shown in Figure 1.

Since the amount of 14-3-3 protein bound to the plasma membrane H⁺-ATPase is proportional to the level of H⁺-ATPase phosphorylation, it is possible to estimate the phosphorylation level by the amount of bound 14-3-3 proteins. An exogenous H₂O₂ inhibited the binding of 14-3-3 protein to the H⁺-ATPase in a concentration-dependent manner (Fig. 5A), suggesting that the phosphorylation of plasma membrane H⁺-ATPase in response to BL was inhibited by H₂O₂. Western-blot analysis revealed that H₂O₂ had no effect on the amount of plasma membrane H⁺-ATPase (Fig. 5B). These results indicate that exogenous H₂O₂ mimics ABA action in inhibiting BL-dependent H⁺ pumping.

Figure 5.

Changes in levels of 14-3-3 protein binding to H⁺-ATPase in GCPs in response to BL in the presence of H₂O₂. H₂O₂ was added to the protoplast suspension 30 min before the application of BL pulse. GCPs were completely solubilized 2.5 min after BL pulse. A, Binding levels of 14-3-3 protein to H⁺-ATPase determined by far-western-blot analysis using recombinant glutathione S-transferase-14-3-3 protein as a probe. B, Western-blot analyses of H⁺-ATPase in GCPs. Arrowheads indicate positions of the H⁺-ATPase.

Ascorbate Partially Restores the Inhibition of BL-Dependent H⁺ Pumping by ABA

Since H₂O₂ mimics the inhibitory effect of ABA on BL-dependent H⁺ pumping, it is most likely that the H₂O₂ generated in GCPs by the ABA addition caused the inhibition of H⁺ pumping. If so, we can expect that the application of the H₂O₂ scavenger ascorbate to GCPs diminishes the effect of ABA. Ascorbate would act as the electron donor for the reduction of H₂O₂ to water with ascorbate peroxidase (Noctor and Foyer, 1998; Asada, 1999; Lee et al., 1999; Neill et al., 2002). As shown in Figure 6A, ascorbate suppressed the ABA-induced inhibition of BL-dependent H⁺ pumping by 32%. Ascorbate alone had no effect on the pumping. In accord with this finding, ascorbate partially increased the binding of 14-3-3 protein to the H⁺-ATPase in the presence of ABA (Fig. 6B). Ascorbate alone did not alter the binding. The amount of H⁺-ATPase was not changed by ascorbate (Fig. 6C). These results verified that the H₂O₂ generated in the guard cells in response to ABA inhibited the signaling cascades initiated by BL.

Figure 6.

Restoration of ABA-dependent inhibition of BL-dependent H⁺ pumping (A), and that of ABA-dependent decrease in the levels of 14-3-3 protein binding to H⁺-ATPase (B) by ascorbate (ASC). A, Effects of ascorbate on BL-dependent H⁺ pumping in the presence of ABA. Values are the means of four independent experiments. B, Effects of ascorbate on the levels of 14-3-3 protein binding to H⁺-ATPase in the presence of ABA determined by far-western-blot analysis. C, Western-blot analysis of the amount of H⁺-ATPase. Ascorbate at 10 mm and ABA at 10 μm were added to the protoplast suspension 40 and 30 min before BL pulse, respectively. Arrowheads indicate the H⁺-ATPase.

ABA Induces H₂O₂ Production from GCPs

Since the exogenous application of H₂O₂ has a similar effect as that of ABA in inhibiting BL-dependent H⁺ pumping, the endogenous generation of H₂O₂ by ABA was determined using GCPs, which lack cell walls. It is important to show the ABA-induced generation of H₂O₂ in the GCPs because all previous experiments had been carried out using intact guard cells in epidermal peels (Pei et al., 2000; Zhang et al., 2001c). We used 2′,7′-dichlorofluorescin (DCFH) to measure changes in the intracellular levels of H₂O₂. DCFH is thought to be oxidized selectively by H₂O₂ over other free radicals to the fluorescent form of dichlorofluorescein (Allan and Fluhr, 1997). As shown in Figure 7, the application of ABA enhanced the relative fluorescence intensity in GCPs after the application of ABA (Fig. 7, A and C). This effect was specific to ABA because the sole application of dimethyl sulfoxide (DMSO), a solvent of ABA, did not increase the fluorescence (Fig. 7, B and C). In the presence of ascorbate at 10 mm, H₂O₂ production by ABA was prevented strongly (data not shown). A bright-field analysis illustrated that ABA induced increases in dichlorofluorescein fluorescence intensity in the vicinity of guard cell chloroplasts, as reported previously (Zhang et al., 2001c). Since ABA and H₂O₂ exhibited very similar effects on BL-dependent H⁺ pumping and ABA led to the generation of H₂O₂ in GCPs, it is most likely that the inhibition of BL-dependent H⁺ pumping by ABA was mediated by H₂O₂.

Figure 7.

ABA-induced H₂O₂ production in GCPs. GCPs were suspended in a buffer containing 0.4 m mannitol, 1 mm CaCl₂, 50 mm KCl, 10 mm MES-KOH, pH 6.15, and 50 μm of DCFH-diacetate, and kept in the dark for 10 to 15 min. An aliquot of the suspension was transferred to the fresh medium. Then, ABA was added at 10 μm (A). As a control, DMSO without ABA was added (B). Confocal images of fluorescence intensity at 525 nm were taken at indicated times after the addition of ABA by excitation at 488 nm. A and B, Pseudocolor of fluorescence images (top sections) and transmission images (bottom sections) were indicated. Pseudocolor bars (A and B; top left) were indicated as pixel intensity values (0–255). Scale bars indicate 10 μm. C, Average intensities of fluorescence at indicated times. Three separate experiments were done on different occasions, and three to four GCPs were used for each experiment. All values of the fluorescence intensities were calculated and averaged (n = 11 for A; n = 10 for B). Asterisks show significant differences with a P < 0.005 (Student's t test) compared with the value at zero time.

Inhibition of FC-Dependent H⁺ Pumping by ABA and H₂O₂

Given the fact that the plasma membrane H⁺-ATPase is the terminal target of BL (Kinoshita and Shimazaki, 1999), the question arises of whether the inhibition of BL-dependent H⁺ pumping by ABA is due to the inhibition of H⁺-ATPase and/or BL signal transduction pathways. To address this question, a fungal phytotoxin fusicoccin (FC), an H⁺-ATPase activator (Kinoshita and Shimazaki, 2001), was further applied 20 min after the illumination of GCPs with a pulse of BL, in which the H⁺ pumping by BL ceased (Fig. 1). Since FC can activate the H⁺-ATPase via phosphorylation by bypassing BL signaling pathways (Shimazaki et al., 1992; Kinoshita and Shimazaki, 2001), effect of ABA on the H⁺-ATPase can be distinguished from those on the upstream components involved in BL signaling pathways.

As shown in Figure 8A, ABA at 10 μm revealed a decreased but nonetheless significant inhibitory effect on FC (10 μm)-dependent H⁺ pumping, with a 20% to 30% level of inhibition. Inhibition was concentration dependent, and the binding of 14-3-3 protein to the H⁺-ATPase was decreased by 30% at the same concentration of ABA (Fig. 8C). The amount of H⁺-ATPase was not changed by these treatments (Fig. 8E). Since an FC-dependent signaling pathway comprises part of the BL signaling pathway in guard cells (Kinoshita and Shimazaki, 2001), the inhibition shown here may be involved in the inhibition of BL signaling pathway by ABA. Since these inhibitions are expected to be mediated by H₂O₂, the effects of H₂O₂ on these two parameters were investigated. As shown in Figure 8, B, D, and F, the application of 1 mm H₂O₂ led to 20% decreases in both the rate of FC-dependent H⁺ pumping and the amount of 14-3-3 protein bound to H⁺-ATPase. From these results, we concluded that ABA inhibits FC-dependent H⁺ pumping and that this inhibition is mediated by H₂O₂. We tested the effects of ascorbate on the inhibition of FC-dependent H⁺ pumping by ABA and found a partial recovery from the inhibition (Fig. 9).

Figure 8.

Effects of ABA and H₂O₂ on FC-dependent H⁺ pumping (A and B) and on the binding of 14-3-3 protein to H⁺-ATPase (C and D) in GCPs. ABA (A) and H₂O₂ (B) were added 30 min before the application of BL pulse, respectively, and FC was added at 10 μm to GCP suspension 20 min after the application of BL pulse. GCPs were completely solubilized 5 min after FC application. Binding of 14-3-3 protein to H⁺-ATPase was determined by far-western-blot analysis, and effects of ABA (C) and H₂O₂ (D) on the binding were indicated. Western-blot analysis of the H⁺-ATPase in the presence of ABA (E) and H₂O₂ (F). Arrowheads indicate the H⁺-ATPase.

Figure 9.

Restoration of ABA-dependent inhibition of FC-dependent H⁺ pumping (A), and that of ABA-dependent decrease in the levels of 14-3-3 protein binding to H⁺-ATPase (B) by ascorbate. A, Effects of ascorbate on FC-dependent H⁺ pumping in the presence of ABA. GCPs were preincubated under background red light for 40 min and then illuminated with a pulse of BL. Ascorbate at 10 mm and ABA at 10 μm were added to the protoplast suspension 40 and 30 min before BL pulse, respectively. FC was added at 10 μm to GCP suspension 20 min after the application of BL pulse. Values are the means of three independent experiments. B, Effects of ascorbate on the levels of 14-3-3 protein binding to H⁺-ATPase in the presence of ABA determined by far-western-blot analysis. Other reaction conditions were the same as in A. GCPs were completely solubilized 5 min after FC application and solubilized GCP proteins were separated by SDS-PAGE. C, Western-blot analysis of the amount of H⁺-ATPase. Other reaction conditions were the same as in B. Arrowheads indicate the H⁺-ATPase.

H₂O₂ Does Not Inhibit H⁺ Pumping in Microsomal Fractions

We have demonstrated previously that ABA does not directly affect ATP-dependent H⁺ pumping mediated by the H⁺-ATPase in isolated microsomes from guard cells (Goh et al., 1996). In this study, we tested the effects of H₂O₂ on the H⁺ pumping using the same materials. As shown in Figure 10, H₂O₂ had no effect on H⁺ pumping in the microsomal fractions. The H⁺ pumping was inhibited by vanadate, a specific inhibitor of the plasma membrane H⁺-ATPase, confirming H⁺ was transported by the H⁺-ATPase. These results together with those described above suggest that H₂O₂ inhibits the signal component that transmits BL signal to the H⁺-ATPase and decreases the phosphorylation levels of H⁺-ATPase, but has no effect on H⁺ pumping in microsome.

Figure 10.

Effect of H₂O₂ on ATP-induced H⁺ pumping in microsome vesicles from Vicia GCPs. The basal reaction mixture (250 μL) contained membrane vesicles (10 μg of protein), 10 mm MOPS-KOH, pH 7.0, 0.25 m mannitol, 5 mm MgCl₂, 1 mm EGTA, 50 mm KNO₃, 5 μg mL⁻¹ oligomycin, and 1 μm quinacrine. ΔF/F, Change in fluorescence divided by the initial fluorescence.

DISCUSSION

ABA Inhibits ATP Hydrolysis by the H⁺-ATPase by Decreasing the Level of Phosphorylation

It is important for plants to adjust their stomata to appropriate pore sizes to avoid excessive transpiration and to retain photosynthetic CO₂ fixation activity under a variety of stresses, especially those that occur in the daytime. Recent investigations have demonstrated that BL stimulates stomatal opening by activating the H⁺ pump, which has been proved to be the H⁺-ATPase in the plasma membrane (Kinoshita and Shimazaki, 1999; Kinoshita et al., 2001). By contrast, ABA causes stomatal closure and prevents water loss from leaves by accelerating the effluxes of anions and cations from guard cells. Therefore, investigation of the effects of ABA on BL-dependent opening responses of stomata is necessary to gain a further understanding of the mechanism of ABA-induced stomatal closure in the daytime. In this context, our previous results have indicated that ABA inhibits BL-dependent H⁺ pumping (Shimazaki et al., 1986; Goh et al., 1996) and have suggested that the pump inhibition provides for membrane depolarization and maintains the continuous efflux of ions from guard cells, resulting in stomatal closure. However, the mechanism by which ABA inhibits BL-dependent H⁺ pumping has yet to be elucidated in biochemical levels. In this study, we provide evidence that the inhibition of H⁺ pumping by ABA is due to the inhibition of ATP hydrolytic activity of the plasma membrane H⁺-ATPase (Figs. 2, A and B, and 3A). We also indicate that inhibition is caused by a decrease in the level of phosphorylation, which is prerequisite for the activation of H⁺-ATPase (Figs. 2, C–F, and 3, B and C).

H₂O₂ May Mediate the Inhibition of Plasma Membrane H⁺-ATPase by ABA

Since the decreases in H⁺ pumping and ATP hydrolytic activity caused by ABA paralleled the decrease in the phosphorylation of H⁺-ATPase, the inhibition of H⁺ pumping can be accounted for by the inhibition of BL-dependent phosphorylation. However, the question of how ABA induces such inhibition remains. Recent biochemical and genetic studies have demonstrated that H₂O₂ is a signal molecule in plants, and the signaling role of H₂O₂ during ABA-mediated stomatal closure has become evident (Guan et al., 2000; Pei et al., 2000; Zhang et al., 2001c). It is thus likely that H₂O₂ affects BL signaling processes in stomatal guard cells. We applied H₂O₂ to GCPs at 1 mm in this study, as this concentration appeared to be realistic for guard cells in localized domain (Neill et al., 2002). We then determined the levels of both BL-dependent H⁺ pumping and the phosphorylation of H⁺-ATPase. We found that H₂O₂ mimics the action of ABA with regard to both parameters. From these results, we concluded that the inhibition of BL-dependent H⁺ pumping by ABA might be mediated by H₂O₂. If this is indeed the case, a decrease in inhibition would be expected by the application of H₂O₂ scavengers. In agreement with this interpretation, we indicated that the inhibition of BL-dependent H⁺ pumping by ABA was antagonized by ascorbate (Fig. 6A). Ascorbate also partially restored the amount of 14-3-3 protein bound to the plasma membrane H⁺-ATPase in guard cells (Fig. 6B).

Generation of H₂O₂ in GCPs

In guard cells of epidermal peels, H₂O₂ can be generated by ABA via several different pathways; guard cell chloroplasts and NADPH oxidase in the plasma membrane might be involved in these pathways (Zhang et al., 2001c; Kwak et al., 2003). Since H₂O₂ generation by ABA was not observed in isolated GCPs, we determined the H₂O₂ using GCPs under the same conditions as those for epidermal peels by Zhang et al. (2001c). ABA-induced H₂O₂ generation was shown in the GCPs, and this generation seemed to be localized in the vicinity of guard cell chloroplasts (Fig. 7). These results, when taken together, suggest that ABA induced the production of H₂O₂, and that the H₂O₂ might further inhibit BL-dependent H⁺ pumping by decreasing the phosphorylation levels of plasma membrane H⁺-ATPase in guard cells. However, since H₂O₂ can induce increases in both cytosolic pH (Zhang et al., 2001a) and the Ca²⁺ concentration via hyperpolarization-activated Ca²⁺ channels (Hamilton et al., 2000; Pei et al., 2000), it remains to be determined by which of these pathways the ABA effect is mediated.

Site of ABA Inhibition in BL Signaling Pathway in Guard Cells

Since the FC-dependent pathway is most likely to comprise a part of the BL signaling pathway in stomatal guard cells (Shimazaki et al., 1992; Kinoshita and Shimazaki, 2001), there may be at least two inhibition sites of ABA in BL signaling pathway. As shown above, BL-dependent H⁺ pumping was strongly inhibited (approximately 60%–70%) by ABA, whereas FC-dependent H⁺ pumping was slightly inhibited (approximately 20%) by ABA; the degree of inhibition of the BL signaling pathway by ABA may be sum of the two types of inhibition. One of these would be downstream of the site of FC action, and the others have reflected upstream of the site of FC action. At present, we were unable to identify the protein molecules in this study; the targets for ABA action may be signal proteins such as protein kinases and phosphatases. However, it is clear that H₂O₂ has no effect on the plasma membrane H⁺-ATPase itself, but instead affects the regulatory mechanisms of H⁺-ATPase (Fig. 10).

Site of ABA-Induced H₂O₂ Production

ABA was able to induce H₂O₂ generation in the vicinity of guard cell chloroplasts (Fig. 7). This finding is similar to the results of previous studies using epidermal peels (Zhang et al., 2001c; Park et al., 2003). However, NADPH oxidase is located in the plasma membrane of GCPs and has been demonstrated to be the source of H₂O₂ (Pei et al., 2000; Kwak et al., 2003). It is intriguing to determine the site of H₂O₂ generation in guard cells in response to ABA, and further investigation will be needed to elucidate this.

MATERIALS AND METHODS

Plant Materials and Isolation of GCPs

Plants of Vicia faba (cv Ryosai Issun) were cultured hydroponically in a greenhouse as described previously (Shimazaki et al., 1992). GCPs were isolated enzymatically from the lower epidermis of 4- to 8-week-old leaves according to a previous method (Kinoshita and Shimazaki, 1999). Isolated GCPs were stored in 0.4 m mannitol and 1 mm CaCl₂ on ice under dark conditions until use. Protein concentration was determined by the method of Bradford (1976) using bovine serum albumin as a standard.

Measurements of BL- and FC-Dependent H⁺ Pumping

BL-dependent H⁺ pumping by GCPs was determined with a glass pH electrode connected to a pH meter (Shimazaki et al., 1986, 1992). The reaction mixture (1.0 mL) consisted of 0.125 mm MES-NaOH, pH 6.0, 1 mm CaCl₂, 0.4 m mannitol, 10 mm KCl, and GCPs (50 μg of protein). BL (100 μmol m⁻² s⁻¹) was applied as a 30-s pulse 40 min after the onset of background red light (600 μmol m⁻² s⁻¹). Red light was obtained from a tungsten lamp (Philips EXR 300W; Eindhoven, The Netherlands) by passing the light through a red cut-off glass filter (Corning 2-61, >600 nm; Corning, NY). BL was obtained from a tungsten lamp (Sylvania EXR 150W; Danvers, MA) by passing the light through a blue band-path glass filter (Corning 5-60, peak 420 nm, half-band width 45 nm). Photon flux density was measured with a quantum meter (LI-COR model 185A; Lincoln, NE). FC-dependent H⁺ pumping by the GCPs was determined in the same reaction mixture under irradiation with red light (600 μmol m⁻² s⁻¹).

To examine the effects of ABA or H₂O₂ on BL-dependent H⁺ pumping, ABA or H₂O₂ was added to GCP suspensions 30 min before the application of the BL pulse. Ascorbate was added to the GCP suspensions at the onset of background red light. All of the measurements were performed at 24°C.

ATP Hydrolysis

ATP hydrolytic activity was measured by determining the Pi released from ATP, as described previously (Kinoshita and Shimazaki, 1999).

Determination of the Phosphorylation Levels of Plasma Membrane H⁺-ATPase

The phosphorylation levels of H⁺-ATPase were determined using ³²P-labeled GCPs, as described previously (Kinoshita and Shimazaki, 1999).

Electrophoresis and Immunodetection

For complete solubilization, an aliquot of withdrawn GCP suspension (25 μg of protein) was centrifuged in 10,000g for 10 s, and the pellet was mixed with a solution containing 10 mm MOPS-KOH, pH 7.5, 2.5 mm EDTA, 25 μg mL⁻¹ DNase (Sigma, St. Louis), 1 mm phenylmethylsulfonyl fluoride, 10 μm leupeptin, and 0.4% (v/v) Triton X-100. After 10 min of incubation at room temperature, the mixture was solubilized in the SDS cocktail at room temperature and subjected to SDS-PAGE (11.5 μg protein per lane). Immunological detection in the western blots was performed as described previously (Kinoshita and Shimazaki, 1999) with slight modifications. The reaction with rabbit antisera raised against the H⁺-ATPase was done at 2,000-fold dilution overnight at 4°C.

Far-Western-Blot Analysis

Protein-protein interactions were analyzed by far-western-blot analysis according to a previous method (Kinoshita and Shimazaki, 1999) with a slight modification. We omitted a procedure of denature and renature with guanidine-HCl.

Determination of H₂O₂ in GCPs

The GCPs were suspended in a loading buffer that contained DCFH-diacetate at 50 μm and kept in the dark for 10 to 15 min. An aliquot of the suspension was withdrawn and transferred to the fresh medium without DCFH-diacetate in a small vessel. Then, ABA was added at 10 μm to the GCPs and the medium was pipetted gently to facilitate the ABA diffusion to the GCPs. Fluorescence was measured at indicated times after the addition of ABA with a Bio-Rad MicroRadiance laser scanning confocal microscope (Bio-Rad Laboratories, Hercules, CA). The working conditions of the confocal microscope were Ex = 488 nm, Em = 525 nm, Power 3%, Zoom 4, mild scanning, and Frame 512 × 512.

Determination of H⁺ Pumping in Microsomal Membranes

A membrane fraction that contained vanadate-sensitive H⁺ transport activity was obtained according to previously described methods (Goh et al., 1996).