Reactive Oxygen Species Are Involved in Gibberellin/ Abscisic Acid Signaling in Barley Aleurone Cells 1[C][W][OA]

Reactive oxygen species (ROS) act as signal molecules for a variety of processes in plants. However, many questions about the roles of ROS in plants remain to be clariﬁed. Here, we report the role of ROS in gibberellin (GA) and abscisic acid (ABA) signaling in barley ( Hordeum vulgare ) aleurone cells. The production of hydrogen peroxide (H 2 O 2 ), a type of ROS, was induced by GA in aleurone cells but suppressed by ABA. Furthermore, exogenous H 2 O 2 appeared to promote the induction of a -amylases by GA. In contrast, antioxidants suppressed the induction of a -amylases. Therefore, H 2 O 2 seems to function in GA and ABA signaling, and in regulation of a -amylase production, in aleurone cells. To identify the target of H 2 O 2 in GA and ABA signaling, we analyzed the interrelationships between H 2 O 2 and DELLA proteins Slender1 (SLN1), GA-regulated Myb transcription factor (GAmyb), and ABA-responsive protein kinase (PKABA) and their roles in GA and ABA signaling in aleurone cells. In the presence of GA, exogenous H 2 O 2 had little effect on the degradation of SLN1, the primary transcriptional repressor mediating GA signaling, but it promoted the production of the mRNA encoding GAMyb, which acts downstream of SLN1 and involves induction of a -amylase mRNA. Additionally, H 2 O 2 suppressed the production of PKABA mRNA, which is induced by ABA:PKABA represses the production of GAMyb mRNA. From these observations, we concluded that H 2 O 2 released the repression of GAMyb mRNA by PKABA and consequently promoted the production of a -amylase mRNA, thus suggesting that the H 2 O 2 generated by GA in aleurone cells is a signal molecule that antagonizes ABA signaling.

The cereal aleurone is a secretory tissue whose metabolism is under hormonal control, and it is an important regulator of seed germination. The interaction between phytohormones, particularly GA and abscisic acid (ABA), is an essential factor controlling the metabolism in cereal aleurone cells. The responses of aleurone layers to GA involve the synthesis and secretion of hydrolytic enzymes, such as a-amylase. ABA blocks the GA response in aleurone cells through a mechanism that involves the expression of a distinct set of genes. Thus, cereal aleurone cells function as an excellent model system in which to explore the molecular mechanisms involved in hormonally regulated gene expression and particularly the antagonism between GA and ABA. So far, many studies of GA and ABA signaling in aleurone cells have been performed (Bethke et al., 1997;Lovegrove and Hooley, 2000;Ho et al., 2003;Sun and Gubler, 2004).
The current view is that GA binds to a soluble GID1 receptor that interacts with the DELLA receptor proteins in a GA-dependent manner; this interaction induces DELLA protein degradation via the E3 ubiquitin ligase SCF GID2/SLY1 (Ueguchi-Tanaka et al., 2005;Hirano et al., 2008;Schwechheimer, 2008). GAindependent changes in gene expression in aleurone cells require DELLA protein degradation. The barley (Hordeum vulgare) Slender1 (SLN1) protein contains a DELLA motif in the N-terminal region and is classified as a DELLA-type protein. The aleurone cells of sln1c mutants constitutively express a-amylase in the absence of GA induction Fu et al., 2002). Downstream of the DELLA proteins, GA regulates a-amylase synthesis in the aleurone via a GAinduced myb-like transcription factor (GAMyb) that binds to a specific region of the promoters of genes that encode a-amylase . GAMyb expression is stimulated by GA upstream of a-amylase gene action: Specifically, DELLA protein degradation stimulates the expression of GAMyb . In turn, GAMyb protein induces the expression of a-amylase genes . ABA both inhibits many GA-induced responses and promotes its own unique responses. The ABA-induced protein kinase ABA-responsive protein kinase (PKABA) is a component of the signal transduction pathway leading to ABA suppression of a-amylase gene expression. In aleurone cells, PKABA mRNA levels rise rapidly in response to ABA, and overexpression of PKABA1 can substitute for exogenous ABA in inhibiting the expression of GA-induced genes, such as GAMyb and a-amylase (Gó mez-Cadenas et al., 1999, 2001. Recently, it has been reported that reactive oxygen species (ROS) function as signal molecules in plants, acting as regulators of growth and development, programmed cell death, hormone signaling, and responses to biotic and abiotic stresses (Gapper and Dolan, 2006;Kwak et al., 2006;Van Breusegem et al., 2008). In aleurone cells, it is also known that ROS, especially hydrogen peroxide (H 2 O 2 ), are involved in the process of programmed cell death . In aleurone cells, GA stimulates cell death while ABA inhibits it. ABA increases the tolerance of cells to UV and H 2 O 2 and increases catalase activity; in contrast, GA causes a decrease in the activities of catalase, ascorbate peroxidase, and superoxide dismutase (SOD) and reduces the tolerance of cells to UV light and H 2 O 2 (Fath et al., 2001). This data suggest that the reduced ability to scavenge ROS in GA-treated cells is the cause of programmed cell death. It has also been reported that GA can change the redox status in aleurone cells (Maya-Ampudia and Bernal-Lugo, 2006). In seeds, ROS play a key role in various events, such as maturation, ripening, aging, and germination (Bailly et al., 1996(Bailly et al., , 2008Lehner et al., 2006;Mü ller et al., 2009aMü ller et al., , 2009b. We had reported in our previous study that NADPH oxidases, which produce the superoxide anion (a type of ROS), regulate germination of barley seeds (Ishibashi et al., 2010). H 2 O 2 also accelerates the germination of pea (Pisum sativum) seeds and stimulates the early growth of seedlings (Barba-Espin et al., 2010). In contrast, exogenously supplied antioxidant, which acts as an ROS scavenger, significantly suppressed the seed germination of several plant species (Ogawa and Iwabuchi, 2001;Ishibashi and Iwaya-Inoue, 2006). Production of H 2 O 2 during the early imbibition period has been demonstrated in seeds of soybean (Glycine max; Puntarulo et al., 1988), maize (Zea mays; Hite et al., 1999), wheat (Triticum aestivum; Caliskan and Cuming, 1998), and zinnia (Zinnia elegans; Ogawa and Iwabuchi, 2001); ROS produced after imbibition are assumed to play a role in seed germination. It was also reported that there was a strong relationship between sunflower (Helianthus annuus) seed dormancy alleviation and accumulation of ROS and peroxidation products in cells of embryonic axes (Oracz et al., 2007(Oracz et al., , 2009). In addition, ROS regulates the seed dormancy and germination of barley (Bahin et al., 2011) and Arabidopsis (Arabidopsis thaliana; Liu et al., 2010). Thus, these reports suggest that ROS might play a signaling role in seed germination and dormancy. Although several lines of evidence indicate that ROS affect seed germination, there is little information establishing a direct link between the change in levels of ROS and gene expression during the germination process.
The aims of this study were to examine whether ROS perform a crucial function in the induction of a-amylase production in barley aleurone cells and to confirm whether this effect is caused by an interaction with GA and ABA signaling. The results presented allow us to present a comprehensive view of the mechanisms of ROS-dependent a-amylase release in barley aleurone cells.

GA and ABA Regulate Production of H 2 O 2 in Barley Aleurone Cells
Production and distribution of H 2 O 2 were histochemically detected as the appearance of brown coloration after staining with 3,3-diaminobenzidine (DAB) as a substrate (Thordal-Christensen et al., 1997). To visualize H 2 O 2 accumulation in barley seeds, embryoless half-seeds were stained with DAB. When embryoless half-seeds were treated with GA, H 2 O 2 accumulation was observed in the aleurone layer but not in the endosperm (Fig. 1B). Additionally, the H 2 O 2 accumulation by GA was suppressed by ABA or ascorbic acid (AsA; Fig. 1C A bright fluorescent signal was detected in DHFDAloaded protoplasts treated with GA for 24 h compared with a culture medium treatment (control; Fig. 1, D and E). Additionally, the increase of fluorescence promoted by GA was suppressed by ABA treatment (Fig.  1F).
By using luminol-dependent chemiluminescence, we explored the possibility that GA and ABA regulate the production of H 2 O 2 in aleurone cells. In this test, the peak value of chemiluminescence reflects the relative abundance of H 2 O 2 . A peak was observed in aleurone protoplasts treated with culture medium (control), GA, or GA + ABA, but not with GA + AsA, a typical antioxidant (Fig. 2, A-D). These results indicated that H 2 O 2 was produced in aleurone cells at some level regardless of the hormone treatment. However, the relative strength of the peak chemiluminescence signal in GA-treated protoplasts was over twice that of the control (Fig. 2E). The peak chemiluminescence identified in protoplasts treated with GA + ABA was similar to the control value. The peak chemiluminescence identified in protoplasts treated with GA + AsA was very low. We also investigated H 2 O 2 content in aleurone layers treated with GA or ABA at different lengths of incubation (Fig. 3). The H 2 O 2 content in aleurone layers treated with GA was increased remarkably at 3 h after treatment, and the content was maintained until 24 h after treatment. Although the H 2 O 2 content in aleurone layers treated with GA + ABA or ABA peaked at 12 h after treatment, the content at all incubation times were less than GA treatment. These data strongly suggest that H 2 O 2 production in barley aleurone is upregulated by GA and downregulated by ABA.

H 2 O 2 Regulates a-Amylase Activity in Barley Aleurone Layers
Aleurone layers of barley were incubated in GA, H 2 O 2 , GA + H 2 O 2 , and GA + AsA, and a-amylase activity was determined after different lengths of incubation ( Fig. 4A). At 1 and 3 h after treatments, a-amylase activity in aleurone layers with GA + H 2 O 2 was significantly higher than that in aleurone layers treated with GA alone. Furthermore, the increase of a-amylase activity normally caused by GA was significantly suppressed in the presence of AsA. The a-amylase activities in barley half-seeds treated with GA in the presence of either low pH (2.8) or high osmolality (50 mM mannitol), conditions similar to those produced by the addition of AsA, were nearly identical to the a-amylase activity of the GA control (Supplemental Fig. S2). Additionally, AsA significantly reduced the level of H 2 O 2 in GA-treated barley aleurone cells (Fig. 2D). These results suggested that AsA suppresses the induction of a-amylase by scavenging H 2 O 2 in barley aleurone layers, not by changing the pH or osmolality of the solution. No a-amylase activity was detected in aleurone layers treated only with H 2 O 2 .
To examine whether the regulation of a-amylase activity caused by H 2 O 2 is accounted for by the level of a-amylase mRNA, we performed northern-blot analysis. Aleurone layers were incubated in GA, GA + H 2 O 2 , or GA + AsA for 24 h. The a-amylase1 and 2 mRNA levels in aleurone layers treated with GA reached its highest level at 12 to 18 h and 18 to 24 h after imbibition, respectively (Fig. 4B), as has been previously reported (Chandler and Jacobsen, 1991;Gubler et al., 1995). On the other hand, the a-amylase1 and 2 mRNA levels in aleurone layers treated with GA + H 2 O 2 were high at 0.5 to 3 h after imbibition. Furthermore, the highest level of a-amylase1 and 2 mRNA in aleurone layers treated with GA + AsA was at 24 h after imbibition. These results indicated that H 2 O 2 in barley aleurone cells is involved in signaling the production of a-amylase mRNA.

H 2 O 2 Is Involved in Expression of GAMyb mRNA after Degradation of SLN1 in Barley Aleurone Layers
GA is known to trigger degradation of SLN1, which is a DELLA-type transcription factor in the GRAS  family (Sun and Gubler, 2004) and a central component of GA signaling. In the presence of GA, SLN1 is degraded via the ubiquitin-proteasome system in a GA-dependent manner, leading to a-amylase production in the aleurone cells of barley seeds (Ikeda et al., 2001;Chandler et al., 2002;Sun and Gubler, 2004). Therefore, to address the role of H 2 O 2 in GAdependent degradation of SLN1 in barley aleurone cells, we performed an immunoblot analysis using anti-SlGAI (a specific antibody against DELLA proteins, including SLN1). GA treatment resulted in a rapid decrease in SLN1 protein level after 4 min ( Fig.  5A). GA + H 2 O 2 treatment caused a similar decrease in the SLN1 protein level. We also examined the relationship between H 2 O 2 and a-amylase activity in embryoless half-seeds of the barley ('Himalaya') sln1 mutant, sln1c, which has a G-to-A substitution in the codon corresponding to amino acid residue 602 (Trp, TGG to TGA), resulting in an early termination codon . In wild-type embryoless halfseeds, H 2 O 2 had an additive effect on the induction of the a-amylase activity induced by GA, whereas addition of AsA nearly eliminated a-amylase activity (Fig.  5B). In sln1c mutant plants, a-amylase activity was induced regardless of the addition of GA, and AsA suppressed the a-amylase activity as in the wild type (Fig. 5C). These results indicated that H 2 O 2 is not involved in degradation of SLN1 in aleurone cells.
Then, we examined whether H 2 O 2 affects the expression of GAMyb in aleurone cells. GAMyb, a transcription factor that acts downstream of SLN1, is a member of the MYB transcription factor family. GA-Myb promotes expression of a-amylase by binding to its promoter region (TAACAAA box). The expression of GAMyb in aleurone layers treated with GA gradually increased and peaked at 6 h after treatment, then decreased by 12 h after treatment (Fig. 6A). This result was consistent with a previous report . On the other hand, the expression of GAMyb in aleurone layers treated with GA + H 2 O 2 rapidly increased and peaked at 0.5 h, then decreased significantly by 3 h after treatment. Additionally, the expression of GAMyb in aleurone layers promoted by GA treatment was suppressed by AsA treatment. These results indicate that H 2 O 2 in barley aleurone cells regulates the amount of GAMyb mRNA. This may be due to increased production or reduced degradation (increased transcript stability). It was previously reported that the inhibitory effect of ABA on seed germination is reversed by ROS (Sarath et al., 2007). We obtained a similar result,  A, a-Amylase activity of aleurone layers treated with 1 mM GA, 1 mM H 2 O 2 , 1 mM GA + 1 mM H 2 O 2 , or 1 mM GA + 50 mM AsA; each treatment also contained 10 mM CaCl 2 , cefotaxime (150 mg mL 21 ), and nystatin (50 units mL 21 ). B, Northernblot analysis of total RNA from aleurone layers treated with 1 mM GA, 1 mM GA + 1 mM H 2 O 2 , or 1 mM GA + 50 mM AsA; each treatment also contained 10 mM CaCl 2 , cefotaxime (150 mg mL 21 ), and nystatin (50 units mL 21 ) using a specific DIG-labeled a-amylase genes probe. The reported values are the means and SD of five or three replications. Ishibashi et al. observing that H 2 O 2 restored a-amylase activity suppressed by ABA (Supplemental Fig. S3). It has been reported that an ABA-induced Ser/Thr protein kinase, PKABA, suppressed the induction of GAMyb expression in aleurone cells (Gó mez-Cadenas et al., 2001). Here, we examined the expression of PKABA in aleurone layers treated with GA, GA + H 2 O 2 , and GA + AsA (Fig. 6B). Although the expression of PKABA in aleurone layers at 0.5 to 3 h after treatment did not change in all treatments, the expression of PKABA in aleurone layers treated with GA + AsA at 6 to 24 h after treatment was significantly higher than that in GA-or GA + H 2 O 2 -treated aleurone layers. In addition, we investigated the relationship between H 2 O 2 and production of PKABA by ABA in barley aleurone cells. The expression level of PKABA in GA-treated embryoless half-seeds was the same as that seen for the control (Fig. 7A). When GA and ABA were applied together, the level of PKABA mRNA increased to almost 3 times that of the control and GA treatments. However, treatment with a combination of GA, ABA, and H 2 O 2 reduced the expression of PKABA to half that seen for the GA + ABA treatment. PKABA is a type of Suc nonfermenting1-related protein kinase 2 (SnRK2), and in addition to being regulated at the transcriptional level, SnRK2 was reported to be activated by autophosphorylation of its Ser or Thr residues, or both (Belin et al., 2006;Shin et al., 2007). Therefore, we also analyzed the effect of H 2 O 2 on autophosphorylation of PKABA in barley aleurone cells by a pull-down assay with anti-SlSnRK2C (a specific antibody against SnRK2 including PKABA; Supplemental Fig. S4). The autophosphorylation of PKABA in barley aleurone cells was detected in the GA + ABA treatment but not in the control, GA, or GA + ABA + H 2 O 2 treatments (Fig. 7B).

DISCUSSION
This study demonstrates that H 2 O 2 performs a crucial function in GA and ABA signaling in aleurone cells of barley. The first line of evidence supporting  this conclusion is the H 2 O 2 production detected in GAand ABA-treated aleurone protoplasts or aleurone layers, since the H 2 O 2 level in aleurone cells was regulated by GA and ABA (Figs. 1-3). ROS are components of programmed cell death in the aleurone of cereal seeds during germination and seedling establishment. GA-treated aleurone protoplasts are more susceptible to cell death from exogenously applied H 2 O 2 than are ABA-treated protoplasts . The activity of ROS-scavenging enzymes, such as catalase, ascorbate peroxidase, and SOD, is downregulated in GA-treated aleurone layers but unchanged in ABA-treated aleurone layers (Fath et al., 2001). Additional evidence comes from a study of Arabidopsis Cu/Zn-SOD isoforms, which are encoded by three genes (CSD1, CSD2, and CSD3; Kliebenstein et al., 1998). On the basis of microarray data, expression levels of CSD1 and CSD2 were lower in a ga1-3 quadruple DELLA mutant (a combination of four mutant DELLA genes [gai-t6 rga-t2 rgl1-1 rgl2-1] in a GA-deficient [ga1-3] background) than in a wild-type background (Achard et al., 2008). Those observations suggest that the DELLA protein restrains ROS accumulation by up-regulating genes for enzymes in the ROS scavenging system. In this study, the accumulation of H 2 O 2 by GA in aleurone cells may result from decreased ROS scavenging activity after degradation of the DELLA protein SLN1.
We then examined whether H 2 O 2 induced by GA affected the activity of a-amylase in aleurone cells. The a-amylase production in aleurone layers during GA treatment was markedly changed after H 2 O 2 production in aleurone layers (Figs. 3 and 4). Furthermore, exogenous H 2 O 2 enhanced early GA-induced aamylase activity in aleurone cells but had no effect in the absence of GA (Fig. 4A). In contrast, AsA significantly suppressed the a-amylase activity induced by GA treatment. These results imply that H 2 O 2 produced by GA is related to induction of a-amylase in aleurone cells by GA. So far, many studies of GA and ABA signaling in aleurone cells have been performed using the induction of a-amylase as an indicator of signaling and have discussed signaling in terms of the increase or decrease in a-amylase activity Lovegrove and Hooley, 2000). In this study, it is noteworthy that H 2 O 2 promoted the induction of a-amylase mRNA in the presence of GA but did not increase the level of induction, and AsA, which is antioxidant, suppressed the a-amylase activity in the presence of GA (Fig. 4).
To identify the target molecules of H 2 O 2 in GA and ABA signaling in aleurone cells, we focused on SLN1, GAMyb, and PKABA, all of which are master regulating molecules in the a-amylase gene induction process. In Arabidopsis, roots of ga1-3 quadruple DELLA mutant and wild-type plants treated with H 2 O 2 began to exhibit cell death within 30 min and 1 h of treatment, respectively; on the other hand, roots of a ga1-3 mutant began to exhibit cell death only after 2 h (Achard et al., 2008), suggesting that H 2 O 2 had no effect on DELLA protein stability. Here, H 2 O 2 did not affect the degradation of SLN1 protein in aleurone cells (Fig. 5A). Additionally, if H 2 O 2 regulates the degradation of SLN1 by GA, a-amylase activity in embryoless sln1c (SLN1-deficient) half-seeds would be expected to be unchanged by AsA treatment. However, a-amylase activity in sln1 mutant aleurone was markedly suppressed by AsA (Fig. 5C). These data indicated that the regulation of a-amylase by H 2 O 2 is unrelated to breakdown of the SLN1 protein and, therefore, that H 2 O 2 acts downstream of SLN1.
In a previous study, ROS induced the expression of several types of transcription factors, such as members of the bZIP, WRKY, and MYB families (Mittler et al., 2004). In this study, the induction of GAMyb mRNA was promoted but not increased by H 2 O 2 in the presence of GA (Fig. 6A); a similar pattern was seen for the induction of a-amylase mRNA in the presence of GA (Fig. 4B). Additionally, the relative expression level of GAMyb mRNA in GA + AsA-treated aleurone layers was lower than that in GA-treated aleurone layers (Fig. 6A). Furthermore, it was interesting that, in aleurone layer treated with GA, H 2 O 2 content increased at 3 h after treatment, expression of GAMyb mRNA increased at 6 h after treatment, and expression of a-amylase mRNA increased at 12 h after treatment (Figs. 3, 4B, and 6A). These results suggest that H 2 O 2 promoted the induction of a-amylase mRNA by regulating the expression of GAMyb mRNA in the presence of GA. The regulation of GAMyb mRNA induction by H 2 O 2 could be accounted for by suppression of PKABA because H 2 O 2 repressed the ABA-induced expression of PKABA mRNA (Fig. 7A). Additionally, the relative expression level of PKABA mRNA in aleurone layers treated with GA + AsA for 6 to 24 h was significantly higher than that in GA-treated aleurone layers (Fig.  6B). These results indicated that the suppression of PKABA expression by H 2 O 2 is crucial for the induction of GAMyb expression. However, at early times after treatment (0.5 to 3 h), PKABA expression in aleurone layers treated with GA + AsA was similar to GA or GA + H 2 O 2 treatment, although GAMyb expression induced by GA was promoted by GA + H 2 O 2 treatment and suppressed by GA + AsA treatment. It is known that activation of SnRK2 family, including PKABA, is attributed to autophosphorylation or phosphorylation by upstream kinases (Belin et al., 2006;Shin et al., 2007). In this study, autophosphorylation of PKABA was observed in barley aleurone cells treated with ABA ( Fig. 7B), suggesting that PKABA in barley aleurone cells was activated by ABA. Interestingly, the autophosphorylation of PKABA by ABA was suppressed in the presence of H 2 O 2 . It is well known that excessive production of ROS (e.g. caused by environmental stress) leads to DNA damage, lipid peroxidation, and protein oxidation (Miller et al., 2008), and we assumed here that there would be oxidative damage of PKABA by H 2 O 2 . However, PKABA in GA + ABA + H 2 O 2 -treated aleurone layers was phosphorylated in the presence of ATP (Supplemental Fig. S4), suggesting that the suppression of autophosphorylation of PKABA by H 2 O 2 was unrelated to oxidative damage of the protein. A tobacco (Nicotiana tabacum) homolog of Arabidopsis SnRK2 (ASK1/SnRk2.4) purified from BY-2 cells was inactivated by dephosphorylation with protein phosphatase 2A (PP2A; Mikołajczyk et al., 2000;Kelner et al., 2004), and alkaline phosphatase abolished salt activation of two rice SnRK2 proteins, SAPK1 and SAPK2 (Kobayashi et al., 2004). Additionally, PP2C also inactivated SnRK2, while in vitro biochemical studies revealed that H 2 O 2 directly inactivates the ABI1-and ABI2-type PP2C enzymes that function in ABA signaling (Meinhard and Grill, 2001;Meinhard et al., 2002). However, in human cells, H 2 O 2 -induced hypophosphorylation of retinoblastoma family proteins pRb, p107, and p130 was caused by the activity of PP2A (Cicchillitti et al., 2003). Although the interaction between PKABA and phosphatase is unclear at present, it is conceivable that the activation of PKABA is regulated directly or indirectly by H 2 O 2 .
From these results, we propose that, in aleurone cells, H 2 O 2 generated by GA promotes the induction of GAMyb mRNA by suppressing the gene expression and activity of PKABA; in this way, H 2 O 2 promotes the expression of a-amylase mRNA (Fig. 8). In this study, no a-amylase activity was detected in aleurone layers treated only with H 2 O 2 (Fig. 4A), although H 2 O 2 acts downstream of SLN1 and is related to the induction of GAMyb mRNA (Figs. 5 and 6A). The induction of GAMyb mRNA in aleurone cells was performed after the degradation of SLN1 in presence of GA . The SLN1 in aleurone layers treated only with H 2 O 2 is predicted to be not degraded because H 2 O 2 is not affected the degradation of SLN1 (Fig. 5). Therefore, aleurone layers treated only with H 2 O 2 could not induce the GAMyb and a-amylase mRNA, although H 2 O 2 suppress the gene expression and activity of PKABA. It has been reported that SLN1 regulates the expression of several genes in aleurone cells . However, little known is about factors that act between SLN1 and GAMyb on GA signaling. In addition, ROS in a cell is affected the intracellular redox state and led to diverse functional changes of a myriad proteins (Buchanan and Balmer, 2005). ROS in aleurone cells suppress the gene expression and activity of PKABA but also might regulate the induction of GAMyb mRNA by activating or inactivating the unknown factors, which are induced after degradation of SLN1, on GA signaling.
Recently, there have been several interesting reports related to this topic. First, members of the PYR/PYL/ RCAR protein family were found to bind to ABA and inhibit the activity of PP2C (Ma et al., 2009;Park et al., 2009). In the presence of ABA, the PYR/PYL receptor proteins can disrupt the interaction between SnRK2 and PP2C proteins, preventing the PP2C-mediated dephosphorylation of the SnRK2 proteins and resulting in the activation of the SnRK2 protein kinases (Sheard and Zheng, 2009;Nishimura et al., 2010). Role of ROS on GA/ABA Signaling in Aleurone Cells Second, the SnRK2 protein kinases are essential for the control of Arabidopsis seed development and dormancy (Nakashima et al., 2009). Finally, ROS are a key factor in cellular signaling in barley (Bahin et al., 2011), Arabidopsis (Liu et al., 2010), and sunflower (Oracz et al., 2007) seed germination and dormancy. Therefore, we speculate that the relationship between SnRK2 and ROS is a key factor in seed germination and dormancy. Although further analysis is necessary to clarify the detailed roles of ROS in GA and ABA signaling in barley aleurone cells, we demonstrated that at least a part of the ROS in GA signaling downregulates PKABA, both at the transcript level and by limiting its autophosphorylation because the transcript level does not imply transcriptional control (there may be a change of mRNA stability); thus, ABA signaling in barley aleurone cells is predicted to be negatively regulated by GA-induced ROS. In conclusion, the data presented here provide insight into the GA and ABA signal transduction pathways in aleurone cells and the roles of ROS in these pathways.

Plant Material
Barley (Hordeum vulgare 'Seijou17') was used as the test material. This cultivar was grown in a 30-m 2 plot in an experimental field of Kyushu University, Fukuoka, during the 2006 to 2007 growing season. Irrigation, fertilization, and pesticide treatment were performed following standard protocols for this species to ensure optimal plant growth. Ripe kernels were harvested on June 1, 2007 and stored at 4°C under low-humidity conditions.

Hormonal Treatment of Aleurone Layers and Embryoless Half-Seeds
Embryoless half-seeds were prepared to cut into two parts (embryoattached and embryoless) with a razor blade after sterilization. Aleurone layers were prepared from the embryoless half-seeds as described previously (Chrispeels and Varner, 1967). For each treatment, aleurone layers or embryoless half-seeds were treated with a solution containing 1 mM GA 3 , 1 (aleurone layers) or 10 (embryoless half-seeds) mM H 2 O 2 , or 50 mM AsA (alone or in combination, as specified) containing 10 mM CaCl 2 , cefotaxime (150 mg mL 21 ), and nystatin (50 units mL 21 ). Embryoless half-seeds were treated at 10 mM H 2 O 2 to postulate the degradation of H 2 O 2 at endosperm during treatment. After incubation at 22°C for the amount of time indicated, the samples were frozen until the measurements. a-Amylase Activity Assay a-Amylase activities in 20 aleurone layers and embryoless half-seeds treated with GA, H 2 O 2 , or AsA were measured using the Amylase HR Reagent kit (Megazyme International) according to the manufacturer's instructions. Extracts were diluted 1:6 with extraction buffer, and 200 mL of the diluted extract was incubated with 200 mL of HR reagent at 40°C for exactly 5 min. The reaction was stopped by adding 3 mL of the supplied stop reagent. Spectrophotometric measurements were carried out in 1-cm cuvettes at 410 nm. There were five replications of each treatment.

Visualization and Detection and Content of H 2 O 2 in Aleurone Protoplasts and Layers
Aleurone protoplasts were prepared from aleurone layers, as described previously (Bush et al., 1986;Bethke et al., 1999). Protoplasts were incubated in Gamborg's B-5 medium modified by the addition of 10 mM CaCl 2 . Where indicated, 1 mM GA 3 , 5 mM ABA, or 50 mM AsA were included in the incubation medium, and the protoplasts were incubated at 22°C for 24 h. Washed protoplasts were incubated in DHFDA for 15 min and then washed four times with baseline culture medium. Protoplasts were transferred to wells on microscope slides and illuminated using a mercury light source on a microscope fitted with optical filters (460-500-nm excitation, 505-nm dichroic mirror, 520-560-nm emission; Nikon). To detect H 2 O 2 in aleurone protoplasts, luminol was added to the incubation medium at a final concentration of 10 mM (Kawano and Muto, 2000). H 2 O 2 generation was monitored by measuring the luminol-dependent chemiluminescence using an automated luminometer (Luminescencer-PSN-R AB-2200-R; ATTO). The H 2 O 2 content in aleurone layers were measured by the modified colorimetrical method of Okuda et al. (1991). Twenty aleurone layers were homogenized at 0°C in 4 mL of 0.2 N HClO 4 and centrifuged at 10 ,000g for 15 min. The supernatant was adjusted to pH 7.5 with 4 N KOH, and the solution was centrifuged at 1,000g for 3 min. The reaction mixture contained 50 mL of the supernatant, 1.43 mL of 12 mM 3-(dimethylamino) benzoic acid in phosphate buffer and 1.3 mM 3-methyl-2benzothiazoline hydrazone and 20 mL of peroxides (0.25 units). After 10 min, the increase in absorbance at 590 nm was measured.

Northern-Blot Analysis
Total RNA was isolated from 50 aleurone layers by using the SDS/phenol/ LiCl method (Chirgwin et al., 1979). cDNA synthesis with total RNA, oligo (dT), and SS reverse transcriptase was performed according to standard procedures. Then, PCR was carried out with ExTaq DNA polymerase according to the manufacturer's manuals: The template was cDNA and the primers were gene-specific primers designed to target the open reading frame of barley a-amylase 1 and 2 (National Center for Biotechnology Information accession nos. M17126.1 and M17128.1; Supplemental Table S1). Digoxigenin (DIG) labeling of the PCR products was carried out with a nonisotopic DIG labeling kit (Roche Diagnostics). Total RNA (5 mg) was denatured in a mixture of 70% (v/v) formamide and 8% (v/v) formaldehyde and separated on a 1.5% agarose gel containing 1.8% (v/v) formaldehyde. After electrophoresis, RNAs were transferred to a Hybond-N + membrane (GE Healthcare). Prehybridization of the membrane was carried out at 65°C for 1 h in 0.3 M phosphate buffer containing 7% SDS followed by hybridization by incubating the membrane in the same buffer with DIG-labeled probes at 65°C for at least 15 h. Membranes were washed in 23 SSC containing 0.1% SDS (15 min) and then in 0.13 SSC containing 0.1% SDS (15 min) at 65°C. After incubation of the blots with anti-DIG antibody-horseradish peroxidase at 37°C for 1 h, DIG epitopes on the membranes were detected by FluorChem (a Innotech) using an ECL kit (GE Healthcare).

Quantitative Real-Time PCR
Total RNA was extracted from barley aleurone layers and embryoless halfseeds by the SDS/phenol/LiCl method (Chirgwin et al., 1979). The cDNA was synthesized from 1 mg total RNA with Rever TraACE reverse transcriptase (Toyobo) according to the manufacturer's protocol. The cDNA (1 mL) was amplified in a reaction medium containing 10 mL of SYBR Green real-time PCR master mix and 2 mL of plus solution (SYBR Green real-time PCR Master Mix Plus kit; Toyobo), 0.1 mL each of 50 mM forward and reverse primers, and 6.8 mL of water. The amplification was conducted in a real-time PCR machine (MJ Mini; Bio-Rad). A melting curve was obtained from the PCR product at the end of the amplification by heating from 50°C to 95°C. From the melting curve, the optimal temperature for data acquisition was determined. The primer sequences of HvGAMyb and HvPKABA were according to Moreno-Risueno et al. (2007) and Rønning et al. (2006), respectively. Relative amounts of each gene transcript were calculated by normalizing against the amount of mRNA for HvActin (Trevaskis et al., 2006) following the method of Pfaffl (2001). Primer sequences are shown in Supplemental Table S1.

Immunoblot Analysis and Phosphorylation Assay
Frozen embryoless half-seeds were homogenized in an ice-cooled mortar in an equal weight of lysis buffer containing 20 mM HEPES-NaOH, 50 mM Na 2b-glycerophosphate (pH 7.6), 5 mM EDTA, 5 mM EGTA, 30 mM NaF, 5 mM Na 3 VO 4 , 10% glycerol, 1% Triton X-100, 0.1% b-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 5 mM n-aminocaproic acid, 1 mM benzamidine, and 1 mM Na-bisulphite. The resultant homogenates were centrifuged at 15,000g for 20 min at 4°C. Protein concentrations in the supernatant were measured with the Bio-Rad protein assay kit with bovine serum albumin as the standard. For immunoblotting, polypeptides that had been separated by SDS-PAGE were electrotransferred to polyvinylidene difluoride membranes (Millipore) in blotting buffer containing 25 mM Tris, 0.05% SDS, and 20% methanol at 10 V cm 21 for 2 h. The membrane was incubated for 1 h in blocking buffer containing 13 Tris-buffered saline and 3% skim milk and then incubated for 2 h in blocking buffer supplemented with 0.05% Tween 20 and containing a 1/1,000 dilution of either anti-SlSnRK2C (Yuasa et al., 2007) or anti-SlGAI (T. Yuasa, unpublished data) as the primary antibody. The membrane was then incubated in blocking buffer supplemented with horseradish-peroxidase-labeled anti-rabbit antibody (1/5,000 dilution [v/v]; GE Healthcare) for 1 h. Immunoreactive signals were visualized with the ECL Plus kit (GE Healthcare) and a FluorChem imaging analyzer (a Innotech). A binding assay with GST-HvSLN1 or GST-HvPKABA recombinant protein was carried out to confirm the specificity of the antibody against the endogenous HvSLN1 or HvPKABA, respectively (Supplemental Fig. S5).
The phosphorylation of HvPKABA was examined by a pull-down assay with anti-SlSnRK2C as follows. The supernatant (containing 10 mg protein) was mixed with 20 mL anti-SlSnRK2C and 30 mL 50% (v/v) protein G-Sepharose CL-4B beads (GE Healthcare). The suspensions were rotated for 2 h at 4°C and then the beads were washed in 1 mL of washing buffer containing 25 mM Tris-HCl (pH 7.4), 1 M NaCl, 0.5 mM EDTA, 0.5 mM EGTA, 0.1% Triton-X, and 0.1% b-mercaptoethanol. After three more washes, the immunodecorated beads were mixed with reaction buffer containing 25 mM Tris-HCl (pH 7.5), 10 mM MgSO 4 , 2 mM EGTA, 10 mM Na 2 -b-glycerophosphate, 2 mM Na 3 VO 4 , and 0.1% b-mercaptoethanol, with or without 1 mM ATP. The reactions were performed at 30°C for 180 min and were then terminated by adding SDS-PAGE sample buffer. After the electrophoresis, acrylamide gels were stained with Pro-Q Diamond phosphoprotein gel stain (Molecular Probes) according to the instruction manual, and then phosphorylated polypeptides in the gel were visualized with a FluorChem imaging analyzer (a Innotech) with 354-nm excitation/595-nm emission.
Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers M17126, M17128, AF460219, X87690, AB058923, and AY145451.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Production of H 2 O 2 in barley aleurone cells.
Supplemental Figure S2. Effects of pH and osmotic stress on a-amylase activity in barley aleurone cells.
Supplemental Figure S3. H 2 O 2 restored a-amylase activity suppressed by ABA.
Supplemental Figure S4. Phosphorylation assay of PKABA in the presence of ATP.
Supplemental Table S1. Primers used for quantitative real-time PCR analysis.