Vacuolar Channels

As vacuoles can take up over 90% of the guard cell's volume, over 90% of the ions exported from the cell during stomatal closing must first be transported from vacuoles into the cytosol (MacRobbie, 1995, 1998). [Ca²⁺]_cyt elevation activates vacuolar K⁺ (VK) channels, which are proposed to provide a pathway for Ca²⁺-induced K⁺ release from the vacuole (Ward and Schroeder, 1994) ( Figure 3). At resting [Ca²⁺]_cyt, K⁺ efflux from guard cell vacuoles can be mediated by fast vacuolar (FV) channels, allowing for versatile vacuolar K⁺ efflux pathways (Allen and Sanders, 1996). The Arabidopsis genome contains five TPK/KCO genes which encode “two-pore” K⁺ channel proteins. A recent study with tpk1 knockout mutants showed that VK channel currents are absent in the mutant indicating that TPK1/KCO1 encodes a major component of the vacuolar VK channels (Gobert et al., 2007). Expression of TPK1 in yeast led to appearance of VK-like channels in yeast vacuoles (Bihler et al., 2005). 14-3-3 proteins physically interact with TPK1/KCO1 and activate TPK1/KCO1 in a dose-dependent manner (Latz et al., 2007). Furthermore, ABA-induced stomatal closure was impaired by the tpk1 mutation (Gobert et al., 2007), providing genetic evidence for the model that VK channels function in mediating vacuolar K⁺ release for stomatal closing (Ward and Schroeder, 1994; Gobert et al., 2007).

It has been proposed that calcium release is induced by inositol hexakisphosphate (Lemtiri-Chlieh et al., 2003), and that a protein phosphorylation event upstream of Ca²⁺ release is required for ABA signaling (Sokolovski et al., 2005). Slow vacuolar (SV) channels are voltage-dependent and slowly activated at elevated [Ca²⁺]_cyt (Hedrich and Neher, 1987). SV channels were first thought to mediate anion transport (Hedrich et al., 1986), but were later shown to be cation channels that are permeable to Ca²⁺ ions (Ward and Schroeder, 1994; Ward et al., 1995; Allen and Sanders, 1996) ( Figure 3). The vacuolar Ca²⁺-dependent channel TPC1 has been identified as an essential subunit of the vacuolar SV channels, because SV type currents are absent in the tpc1 knockout mutant (Peiter et al., 2005). ABA-induced stomatal closure was not affected in tpc1 mutants, whereas external Ca²⁺-inhibition of stomatal opening was impaired in tpc1 mutants (Peiter et al., 2005). Absence of Ca²⁺-activated SV currents was confirmed in an independent study (Ranf et al., 2008). SV channels are down-regulated by vacuolar Ca²⁺ (Pottosin et al., 1997; Pei et al., 1999), but can allow Ca²⁺ transport from the vacuole to the cytoplasm (Bewell et al., 1999), which is for example illustrated by “tail” currents recorded in pure CaCl₂ gradients (Ward et al., 1995). SV channels are ubiquitous and have been found in all plant vacuoles, but in spite of their broad distribution in plant vacuoles, their physiological functions are not yet understood. Further studies are required to address the physiological role for this SV channel protein in plants.

CLC family genes function as vacuolar anion (nitrate) uptake transporters (De Angeli et al., 2006), and Cl⁻ channels and malate channels can mediate low-affinity vacuolar anion uptake (Pei et al., 1996; Hafke et al., 2003; Kovermann et al., 2007), but the pathways for anion release from vacuoles still remain elusive.

K⁺_in Channels Mediate K⁺ Uptake During Stomatal Opening

Inward rectifying K⁺ channels were characterized and proposed to contribute to K⁺ influx during stomatal opening (Schroeder et al., 1984; Schroeder et al., 1987). The guard cell K⁺ channel subunit KAT1 and the root-expressed AKT1 homologue were the first cloned Arabidopsis K⁺_in channel genes (Anderson et al., 1992; Sentenac et al., 1992). KAT1 was characterized as the first eukaryotic inward-rectifying K⁺ channel gene (Schachtman et al., 1992). KAT1 is expressed in guard cells (Nakamura et al., 1995) and expression of dominant negative mutants, as well as knockout mutant analyses support the model that KAT1 and its homologues play an important role in stomatal opening (Kwak et al., 2001; Lebaudy et al., 2008). KAT2 is also expressed in guard cells (Pilot et al., 2001). Furthermore, expression analysis in Xenopus oocytes and yeast suggests that KAT1 and KAT2 form heteromultimeric channel proteins that are probably responsible for K⁺ influx into guard cells (Pilot et al., 2001). K⁺_in channel activity is regulated by various factors including [Ca²⁺]_cyt and protein de/phosphorylation. In addition, spermidine was shown to inhibit stomatal opening and induce stomatal closure (Liu et al., 2000) ( Figure 2). Spermidine is a polyamine that functions as a growth signal in both eukaryotes and prokaryotes. The levels of polyamines increase under stress conditions, suggesting their roles in stress responses (Slocum and Flores, 1991). Addition of spermidine inhibited K⁺_in channel currents in guard cells in a dose-dependent manner, and also inward K⁺ currents in mesophyll cells expressing KAT1, indicating a link between stress response and the regulation of K⁺_in channel activity (Liu et al., 2000).

Phopholipid kinases also modulate stomatal opening. Phosphatidylinositol (PI) 4,5-bisphosphate produced from either PI4P or PI5P was reported to inhibit both light-induced stomatal opening and ABA-induced stomatal closure (Jung et al., 2002) ( Figure 2). The PI3 kinase and PI4 kinase inhibitors wortmannin and LY294002 impaired ABA-induced [Ca²⁺]_cyt increases and stomatal closure (Jung et al., 2002). PI 4,5-bisphosphate was also shown to inhibit S-type anion channel current in guard cells (Lee et al., 2007). Furthermore, a mutation in the Arabidopsis PIP5 kinase 4 causes reduced stomatal opening in the light (Lee et al., 2007). Stomatal opening of the pip5k4 mutant was rescued by exogenous PI 4,5-bisphosphate, indicating that PIP5K4 is a positive regulator of light-induced stomatal opening (Lee et al., 2007) ( Figure 2).

The Arabidopsis genome encodes 11 ROP-type small G proteins that function as a molecular switch in various cellular responses in plants. It was recently shown that ROP2 is expressed in guard cells and translocates to the plasma membrane upon illumination (Jeon et al., 2008). Expression of a constitutively active form of ROP2 negatively modulated stomatal opening in both Vicia and Arabidopsis (Jeon et al., 2008) ( Figure 2). Furthermore, the guard cell-expressed RIC7 (ROP-Interactive CRIB motif-containing protein 7) physically interacts with ROP2 and translocates to the plasma membrane and cytosol from the nucleus upon illumination, which is dependent on ROP2 activation (Jeon et al., 2008). Expression of RIC7 in Vicia faba guard cells suppresses light-induced stomatal opening (Jeon et al., 2008). These results suggest that RIC7 negatively regulates light-induced stomatal opening by working downstream of ROP2 ( Figure 2).

Transcriptional modulation contributes to the regulation of guard cell signaling cascades, positively or negatively. Two MYB transcription factors have been identified as regulators of stomatal opening. MYB60 was shown to be specifically expressed in Arabidopsis guard cells, and the expression level of MYB60 was reduced by ABA and desiccation (Cominelli et al., 2005). A null mutation in MYB60 leads to a reduction in light-induced stomatal opening and to an enhanced drought tolerance, without affecting ABA sensitivity of guard cells (Cominelli et al., 2005). Furthermore, only 36 genes were shown to be down- or up-regulated in the atmyb60 mutant, suggesting that MYB60 specifically regulates a limited number of genes (Cominelli et al., 2005) ( Figure 2). In contrast, another MYB transcriptional factor, MYB61, was reported to be highly expressed in guard cells (Liang et al., 2005). Overex-pression of MYB61 results in a higher leaf surface temperature (indicative of smaller stomatal apertures), whereas atmyb61 null mutants have a lower leaf surface temperature (larger stomatal apertures), compared to wild type plants (Liang et al., 2005). Similar to the situation with atmyb60, neither overexpression or knockout of MYB61 appears to alter ABA sensitivity of the plants in stomatal movements (Liang et al., 2005). However, dark-induced stomatal closure is partially impaired in atmyb61 (Liang et al., 2005), suggesting that MYB61 is a negative regulator of light-induced stomatal opening and/or a positive regulator of dark-induced stomatal closure ( Figure 2).

In addition to the cytosolic factors, protein trafficking plays a role in regulating K⁺_in channel activity during stomatal movements. During swelling of guard cell protoplasts, an increase in the membrane surface area was reported to be associated with an increase in both K⁺_in channel and K ⁺_out channel currents without affecting the activation kinetics of K⁺ channels (Homann and Thiel, 2002). This result suggests that vesicles containing both K⁺_in and K ⁺_out channels are delivered to the plasma membrane during guard cell swelling, resulting in the increase in the membrane surface area and K⁺ currents. SYP121 is the Arabidopsis homolog of the tobacco syntaxin SYP121/Syr1 that was previously shown to function in ABA-activation of Cl⁻ currents and ABA inhibition of KAT1-mediated K⁺ currents in oocytes (Leyman et al., 1999). A cell biological study using the dominant-negative Sp2 fragment of SYP121 showed that the Arabidopsis SYP121 protein is responsible for delivery of KAT1 K⁺_in channel to the plasma membrane (Sutter et al., 2006). Furthermore, it was shown that ABA induces specific endocytosis of KAT1 K⁺_in channels in an endosomal pool in guard cells leading to a reduction in the number of KAT1 K⁺_in channel subunits residing in the plasma membrane (Sutter et al., 2007). The endocytosed KAT1 K+in channels were almost entirely recycled to the plasma membrane 8hr after removal of ABA (Sutter et al., 2007). These studies indicate that trafficking of K⁺ channels is involved in stomatal movements ( Figure 2).

In addition to K⁺_in channels, other transporters have been suggested to contribute to stomatal opening. AtCHX20, a proton antiporter that localizes to endomembranes, was reported to positively regulate light-induced stomatal opening (Padmanaban et al., 2007). The CHX family is comprised of 28 members, and their functions are largely unknown (Sze et al., 2004). AtCHX20 is highly and preferentially expressed in guard cells, and Atchx20 T-DNA insertional mutants exhibited a significant reduction in light-induced stomatal opening, implying that the endomembrane cation/H⁺ exchanger contributes to K⁺ uptake into vacuoles during stomatal opening (Padmanaban et al., 2007) ( Figure 2). During stomatal opening, K⁺ influx occurs which balances an influx of anions, mainly Cl⁻. Because nitrate is also abundant in soil and transported to plant cells by the nitrate transporter AtNRT1.1/CHL1, it was tested whether AtNRT1.1/CHL1 has a role in stomatal movements (Guo et al., 2003). chl1 mutant plants have reduced stomatal opening, reduced nitrate accumulation, and impaired nitrate-induced membrane depolarization in the light, suggesting that nitrate uptake is accompanying K⁺ influx and required for stomatal opening (Guo et al., 2003) ( Figure 2).

 Figure 2 illustrates the signaling cascade from activation of PHOT1, 2 to stomatal opening in terms of known genes and their relative position in the signal transduction pathway. For a detailed description of each gene please see the on-line version of this figure where the figure is clickable and linked to an updated version of (Schroeder et al., 2001).

ABA-Regulated Calcium Signaling

ABA triggers cytosolic calcium ([Ca²⁺]_cyt) increases ( Figure 3) (McAinsh et al., 1990; Allan et al., 1994; Grabov and Blatt, 1998; Allen et al., 2001; Webb et al., 2001; Klusener et al., 2002; Li et al., 2004; Israelsson et al., 2006; Nomura et al., 2008). Cytosolic Ca²⁺ elevations in guard cells down-regulate both K⁺_in channels (Schroeder and Hagiwara, 1989) and plasma membrane H⁺-ATPases (Kinoshita et al., 1995), providing a mechanistic basis for ABA and Ca²⁺ inhibition of K⁺ uptake during stomatal opening ( Figure 1).

Hyperpolarization-activated Ca²⁺-permeable “I_Ca” channels in the plasma membrane contribute to the increase in [Ca²⁺]_cyt (Hamilton et al., 2000; Pei et al., 2000; Kwak et al., 2003) ( Figure 3). ABA was shown to activate these plasma membrane Ca²⁺ channels in both Vicia faba and Arabidopsis (Hamilton et al., 2000; Pei et al., 2000; Murata et al., 2001). Protein phosphorylation plays a role in the activation of Ca²⁺ channels (Kohler and Blatt, 2002). Two Ca²⁺-dependent protein kinases, CPK3 and CPK6, are required for ABA activation of S-type anion and Ca²⁺-permeable channels (Mori et al., 2006). Two other CDPKs, CPK4 and CPK11 also help mediate ABA responses (Zhu et al., 2007). In addition, Ca²⁺ channels are regulated by hydrogen peroxide leading to increases in [Ca²⁺]_cyt (Pei et al., 2000; Kohler et al., 2003; Kwak et al., 2003; Lemtiri-Chlieh and Berkowitz, 2004) ( Figure 3). A hyperpolarization-activated Ca²⁺-permeable current was also activated by cyclic AMP (Lemtiri-Chlieh and Berkowitz, 2004). Further research is needed to determine whether the cAMP-activated Ca²⁺ channels are the same channels as ABA-regulated I_Ca channels.

[Ca²⁺]_cyt elevations activate two different types of anion channels in the plasma membrane of guard cells: Slow-activating sustained (S-type; Schroeder and Hagiwara, 1989) and rapid transient (R-type; Hedrich et al., 1990; Roelfsema et al., 2004) anion channels. Both mediate anion release from guard cells, causing depolarization ( Figure 1).

Early intracellular Ca²⁺ imaging studies showed that a significant number of guard cells do not show measurable [Ca²⁺]_cyt elevations in response to ABA (Schroeder and Hagiwara, 1990; Gilroy et al., 1991; McAinsh et al., 1992). These findings led to the model that both Ca²⁺-dependent and Ca²⁺-independent pathways mediate ABA responses (Allan et al., 1994). Based on cells showing no ABA-induced Ca²⁺ increases the Ca²⁺-independent pathway has also been proposed and discussed in further studies (Allan et al., 1994; Ward et al., 1995; Romano et al., 2000; Levchenko et al., 2005; Marten et al., 2007b). However, the relative contribution of the proposed Ca²⁺-independent pathway to ABA-induced stomatal closing has not yet been quantified, and questions of whether known ABA signaling mutations transduce or bypass this pathway remains unanswered. In addition, recent research on CO₂ and ABA signaling has led to the hypothesis that these physiological stomatal closing stimuli sensitize guard cells to cytosolic Ca²⁺ (Young et al., 2006; Yang et al., 2008). This proposed Ca²⁺ sensitivity priming might also mediate Ca²⁺-dependent signaling at lower cytosolic Ca²⁺ levels (Israelsson et al., 2006). Further work, beyond the “negative” results from multiple laboratories that led to the Ca²⁺-independent signaling model, is needed to characterize the Ca²⁺-independent pathway.

In addition to promotion of stomatal closure, ABA also inhibits stomatal opening by negatively regulating the plasma membrane H⁺-ATPases in a dose-dependent manner (Goh et al., 1996). Increases in [Ca²⁺]_cyt were shown to inhibit the ATP hydrolysis and H⁺ pumping by plasma membrane H⁺-ATPase, suggesting that [Ca²⁺]_cyt mediates this ABA inhibition of H⁺-AT- Pase activity (Kinoshita et al., 1995). Increases in [Ca²⁺]_cyt also inhibit K⁺_in channels and thus contribute to ABA inhibition of stomatal opening (Schroeder and Hagiwara, 1989; Wang et al., 1998).

Essential Function of Anion Channels in Stomatal Closing

Anion channels in the plasma membrane of guard cells were proposed to provide a central control mechanism for stomatal closing (Schroeder and Hagiwara, 1989). Genetic evidence for this model was recently obtained. SLAC1 encodes a plasma membrane protein in Arabidopsis guard cells and is required for the function of S-type anion channels and for CO₂-, ABA- and ozone-induced stomatal closing (Negi et al., 2008; Vahisalu et al., 2008) ( Figure 3). The SLAC1 protein has distant similarity to a bacterial malate transporter. Since S-type anion channels have a permeability to malate anions (Schmidt and Schroeder, 1994), and slac1 mutants lack S-type anion currents in guard cells (Vahisalu et al., 2008), SLAC1 encodes a long sought subunit of these guard cell anion channels. slac1 mutant alleles show residual stomatal closing responses to low-humidity, light-dark transitions and ABA, which correlates with the finding that R-type anion channels remain functional in slac1 guard cells (Vahisalu et al., 2008). R-type channels are thus modeled to function in parallel to S-type anion channels (Keller et al., 1989; Schroeder and Keller, 1992) ( Figure 1 and  Figure 3). slac1 mutant alleles show greatly reduced responses in ABA-, CO₂-, low-humidity, ozone- and Ca²⁺-induced stomatal closing (Negi et al., 2008; Vahisalu et al., 2008), supporting the model that S-type anion channels are major control mechanisms of stomatal closing (Schroeder and Hagiwara, 1989; Schroeder et al., 1993).

ABC transporters have been suggested to regulate S-type anion channels or to serve as the channel themselves (Leonhardt et al., 1997; Leonhardt et al., 1999). Two ABC transporters AtMRP4 and AtMRP5 were reported to have opposite regulatory roles in stomatal opening. Null mutations in AtMRP4 resulted in enhanced stomatal opening in response to light, increased transpirational water loss, and enhanced drought sensitivity, indicating that AtMRP4 is a negative regulator of stomatal opening (Klein et al., 2004). Another ABC transporter, AtMRP5, is expressed in guard cells and localizes to the plasma membrane (Suh et al., 2007) ( Figure 3). Atmrp5–1 knockout mutants show reduced stomatal opening, in response to the light, when compared wild type plant responses (Klein et al., 2003). Atmrp5–1 mutants also display other phenotypes consistent with the observed stomatal behavior, including a reduced transpiration rate, reduced water loss in detached leaves, enhanced drought tolerance, indicating that AtMRP5 positively regulates stomatal opening (Klein et al., 2003). Furthermore, the atmrp5 null mutant shows reduced sensitivity to ABA in stomatal closing (Gaedeke et al., 2001; Klein et al., 2003). atmrp5 gene disruption impairs Ca²⁺-activated S-type anion channel currents (Suh et al., 2007). However, in contrast to slac1 alleles, atmrp5 mutations impair both S-type anion channel and Ca²⁺ channel regulation (Suh et al., 2007) ( Figure 3), indicating that AtMRP5 may function as a transmembrane ion channel regulator, analogous to it's mammalian family members, the cystic fibrosis transmembrane conductance regulator (CFTR) and sulfonylurea receptor (SUR) (Ashcroft, 2006; Gadsby et al., 2006). Whether AtMRP5 directly interacts with ion channel proteins remains presently unknown.

Modulators of ABA Signaling Strength

The gene ERA1 encodes a β-subunit of farnesyltransferase, an enzyme that modifies specific target proteins by attaching a hydrophobic prenyl (farnesyl) group to their C terminus. ERA1 is a negative regulator of ABA signaling (Cutler et al., 1996) and a deletion of ERA1 causes hypersensitive ABA-induced stomatal closing and a reduced water loss of plants during drought stress (Pei et al., 1998). era1 guard cells show ABA hypersensitive activation of S-type anion channels and I_Ca Ca²⁺ channels, suggesting that ERA1 negatively regulates early ABA signaling and ion channels (Pei et al., 1998; Allen et al., 2002) ( Figure 3). Prenylation of target proteins leads to localization to the plasma membrane. Thus ERA1 may recruit negative regulators of ABA signaling to the plasma membrane of guard cells. Two small G proteins ROP10 and AtRac1 are putative targets of prenylation and show C-terminal sequences that may be prenylated. Interestingly, both small G proteins, ROP10 and AtRac1, are negative regulators of ABA signaling and stomatal closing (Lemichez et al., 2001; Zheng et al., 2002) ( Figure 3). Transgenic Arabidopsis plants expressing a constitutively active mutant of AtRac1 showed an impaired ABA-induced actin reorganization and stomatal closure (Lemichez et al., 2001). The null rop10 mutant shows enhanced ABA response in stomatal closure, and a dominant-negative mutant of ROP10 partially suppresses the abi2–1 mutant phenotype in seed germination and root elongation (Zheng et al., 2002). Cytosolic Ca²⁺ also mediates ABA-triggered actin reorganization, which in turn activates mechanosensitive Ca²⁺ channels (Hwang and Lee, 2001; Zhang et al., 2007a).

The extracellular Ca²⁺ sensor CAS was cloned from Arabidopsis using a functional expression assay, combined with Ca²⁺ imaging, in HEK293 mammalian cells (Han et al., 2003). CAS encodes a plant-specific protein that has a single transmembrane domain. The C-terminal region of CAS shares a sequence similarity with a rhodanese-like domain (Pfam00581) that might be involved in protein-protein interactions, and several acidic amino acids reside in the N-terminus that was shown to bind to Ca²⁺ (Han et al., 2003). CAS is mainly expressed in aerial parts of plants including guard cells and mediates external Ca²⁺-induced [Ca²⁺]_cyt increases by increasing IP3 concentration (Han et al., 2003; Tang et al., 2007; Nomura et al., 2008). In cas knockout mutants and CAS antisense plants, external Ca²⁺-induced stomatal closure is impaired whereas CAS overexpression triggers stomatal closure in the absence of external Ca²⁺ (Han et al., 2003; Nomura et al., 2008). The cellular targeting of CAS is presently a matter of debate, since there are data indicating plasma membrane and/or thy-lakoid membrane localizations (Han et al., 2003; Nomura et al., 2008; Weinl et al., 2008) ( Figure 3).

Phospholipids, phospholipid kinases, and phospholipid lipases function in guard cell ABA signaling. The phospholipase C (PLC) blocker U-73122 partially inhibits ABA-induced stomatal closure, suggesting that PLC positively regulates ABA signaling (Staxen et al., 1999). Transgenic tobacco with reduced levels of PLC showed a reduction in ABA inhibition of stomatal opening, further indicating that PLC is a positive regulator of the signaling cascade (Hunt et al., 2003) ( Figure 3). Phospholipase PLDα1 was reported to also function as a positive regulator by producing phophatidic acid (PA) that binds to and inhibits the negative regulator PP2C ABI1 (Zhang et al., 2004a) ( Figure 3). While ABI1 binding of PA would stimulate ABA promotion of stomatal closure, PA and PLDα1 interact with the GPA1 subunit (see below) to mediate ABA inhibition of stomatal opening, indicating a dual function of PA (Mishra et al., 2006). Pldα1 knock out mutants were reported to be disrupted in ABA-induced stomatal closing, and this disruption was alleviated in abi1 and plda1 double knockouts, suggesting that PLDα1 is not required for ABA responses in the absence of ABI1. The strong suppressor effect of abi1 knockout on the pldα1 phenotype appears interesting and unique, since many close homologues to the ABI1 protein phosphatase have been shown to have similar functions to ABI1 and are co-expressed in guard cells (Leonhardt et al., 2004; Saez et al., 2004; Kuhn et al., 2006; Saez et al., 2006; Yoshida et al., 2006b). Thus further investigation of the reported interaction of PLDα1 and ABI1 would be interesting.

Phosphatidylinositol 3- and 4-phosphate, sphingosine-1-phosphate (S1P), and phytosphingosine-1-phosphate (phyS1P) are additional phospholipids that positively regulate ABA signaling, indicating that PI3 kinase, PI4 kinase, and sphingosine kinase are positive effectors of the ABA signaling cascade (Ng et al., 2001; Jung et al., 2002; Coursol et al., 2003; Coursol et al., 2005). In the gpa1 mutant, S1P-induced stomatal closure and S-type anion channel activation are impaired, suggesting that S1P and phyS1P function upstream of GPA1 (Coursol et al., 2003). A pharmacological study combined with calcium imaging analysis suggests that PI3P and PI4P function upstream of [Ca²⁺]_cyt increases (Jung et al., 2002).

Genetic evidence suggests that the heterotrimeric G protein αsubunit (GPA1) mediates ABA signaling. Both ABA inhibition of inward K+ currents and stomatal opening are reduced by the gpa1–1 and gpa1–2 T-DNA insertional mutations (Wang et al., 2001) ( Figure 3). GPA1 physically interacts with GCR1, a G protein coupled receptor (Pandey and Assmann, 2004). In contrast to the gpa1 mutants, gcr1 mutant guard cells are hypersensitive to ABA and S1P, suggesting that GCR1 is a negative regulator of this ABA response (Pandey and Assmann, 2004). Analysis of ABA responses in seed germination, compared to guard cells, show differential effects of G protein mutants. Both gpa1 and gcr1 mutants show ABA hypersensitivity in seed germination and early seedling development, indicating that GPA1 negatively regulates this signaling pathway (Pandey et al., 2006). Thus G proteins affect ABA signaling and modulate ABA responses in a cell- and tissue-specific manner (Pandey et al., 2006; Pandey et al., 2007).

Reactive Oxygen Signaling Mechanisms in Guard Cells

Reactive oxygen species (ROS) and the gas nitric oxide (NO) function as second messengers in guard cell ABA signaling. ABA induces ROS production in guard cells (Pei et al., 2000; Zhang et al., 2001). ROS lead to activation of hyperpolarization-activated Ca²⁺ channels and stomatal closure (Pei et al., 2000; Murata et al., 2001). NADPH oxidases are multi spanning membrane proteins that produce extracellular superoxide (Keller et al., 1998; Sagi and Fluhr, 2001; Sagi and Fluhr, 2006). Guard cell microar-ray experiments showed that of the 10 NADPH oxidase genes in the Arabidopsis genome, AtRbohD and AtRbohF are highly expressed in guard cells (Kwak et al., 2003). Analysis of AtrbohD/F double knock out mutant guard cells, showed that these NADPH oxidases contribute to ABA-triggered ROS production, and function in ABA-activation of plasma membrane Ca²⁺ channels, and ABA-induced stomatal closure (Kwak et al., 2003) ( Figure 3). Compared to the mammalian counterpart, one of the unique structural features of plant NADPH oxidases is that there are two EF-hand Ca²⁺ binding domains in their N-terminus, suggesting the regulation of the protein by Ca²⁺ (Keller et al., 1998; Torres et al., 1998). Plant NADPH oxidases are activated by Ca²⁺ (Sagi and Fluhr, 2001; Ogasawara et al., 2008). Furthermore, NADPH oxidases appear to be positively regulated by protein kinases (Kobayashi et al., 2007; Nuhse et al., 2007; Ogasawara et al., 2008).

The ethylene receptor ETR1 was reported to mediate ethylene-and H₂O₂-induced stomatal closure, and AtrbohF functions in this pathway (Desikan et al., 2005; Desikan et al., 2006), indicating a signaling convergence between ABA and ethylene signaling pathways. However, ethylene was also shown to counteract and delay ABA-induced stomatal closure (Tanaka et al., 2005). Thus further research would be helpful to illuminate these apparently counter-acting physiological ethylene responses in stomatal regulation.

Hydrogen peroxide also functions in ABA inhibition of blue light-induced H⁺ pumping by triggering dephosphorylation of H⁺-ATPase in the plasma membrane (Zhang et al., 2004b). Hydrogen peroxide was reported to inhibit ATP hydrolysis by the plasma membrane H⁺-ATPase and blue light-induced phosphorylation of the H⁺-ATPase, whereas the cellular H₂O₂ scavenger ascorbate partially rescues ABA inhibition of H⁺ pumping, indicating that H₂O₂ is another molecular player mediating ABA inhibition of stomatal closure (Zhang et al., 2004b). Cytoplasmic alkalization and PI3P appear to function upstream of ROS, whereas ABA-induced NO production and stomatal closure rely on ABA-triggered ROS production (Park et al., 2003; Suhita et al., 2004; Bright et al., 2006). NO triggers Ca²⁺ release from intracellular stores contributing to an increase in the cytosolic Ca²⁺ concentration ( Figure 3), resulting in inhibition of K⁺_in channels and activation of anion channels (Garcia-Mata et al., 2003). The glutathione peroxidase AtGPX3 physically interacts with ABI1 and ABI2 and contributes to the regulation of redox states in guard cells (Miao et al., 2006) ( Figure 3). In addition, the redox state of the H₂O₂ scavenger ascorbate plays a role in stomatal movements. Overexpression of dehydroascorbate reductase results in increased ascorbate reduction, a lower H₂O₂ level in guard cells, and increased stomatal conductance and water transpiration (Chen and Gallie, 2004). Together, these studies show a key role for ROS in modulating the mechanisms that mediate stomatal closure, including Ca²⁺ channel activation (Pei et al., 2000; Murata et al., 2001; Kwak et al., 2003; Kwak et al., 2006), H⁺ pump inhibition (Zhang et al., 2004b), and, via NO production, K⁺_in channel de-activation and anion channel activation (Garcia-Mata et al., 2003).

Protein Kinases and Phosphatases in ABA Signaling

Protein de/phosphorylation plays a key regulatory role in guard cell ABA signaling. The dominant ABA insensitive abi1–1 and abi2–1 loci were isolated (Koornneef et al., 1984) and identified as closely related type 2C protein phosphatases (Leung et al., 1994; Meyer et al., 1994; Leung et al., 1997; Rodriguez et al., 1998). Hydrogen peroxide inhibits the activities of the negative regulatory type 2C protein phosphatases ABI1 and ABI2 (Meinhard and Grill, 2001; Meinhard et al., 2002). Localization of the mutant abi1-1 protein into the nucleus is required for the ABA-insensitive response (Moes et al., 2008). When the nuclear localization sequence in abi1-1 is disrupted, the transcription of genes regulated by ABA is rescued to wild-type levels, thus, suggesting that abi1-1 functions as a hypermorphic allele (Moes et al., 2008). Two additional PP2Cs, HAB1/AtP2C-HAand AtPP2CA, were identified to function as negative regulators of ABA signaling (Leonhardt et al., 2004; Kuhn et al., 2006; Yoshida et al., 2006b) ( Figure 3). Moreover, the stomatal response to ABA is enhanced when the two negatively regulating PP2C ABI1 and HAB1 are inactivated simultaneously (Saez et al., 2006). The dominant mutation abi1–1 and abi2–1 impair ABA activation of I_Ca Ca²⁺ channels (Murata et al., 2001) and S-type anion channels (Pei et al., 1997). ABI2 physically interacts with the SOS2 protein kinase, and this interaction is disrupted by the abi2–1 mutation (Ohta et al., 2003) ( Figure 3). Note that no pharmacological inhibitors for PP2Cs are known. Therefore, the dominant natures of the abi1–1 and abi2–1 alleles (Koornneef et al., 1984) do not allow to distinguish whether the phenotypes of these mutants occur through direct protein-protein interactions in early ABA signaling, or via less direct, possibly neomorphic or hypermorphic modulation of the early ABA signaling pathway (For detailed previous discussions, see (Pei et al., 1997; Murata et al., 2001; Yoshida et al., 2006b). A recent study showed that ABI1 physically interacts with the OST1/SnRK2E protein kinase, and ABA activation of OST1/SRK2E is inhibited in abi1–1 but not in abi2–1 mutants (Yoshida et al., 2006a). The PKS3 kinase interacts with the calcium binding protein ScaBP5/CBL1 and the PP2C ABI2 and ABI1, and RNAi-mediated silencing of ScaBP5 causes ABA hypersensitivity (Guo et al., 2002) ( Figure 3).

In addition to PP2Cs, type2A protein phosphatases function in ABA signaling. The RCN1 gene encoding the PP2A regulatory subunit A is a positive regulator of the ABA signaling cascade in guard cells and seed germination (Kwak et al., 2002). RCN1 was previously shown to increase PP2A enzyme activity (Deruère et al., 1999). T-DNA insertional mutation in RCN1 causes impairment in both ABA-induced stomatal closure and ABA activation of anion channels, a reduced sensitivity of ABA-triggered cytosolic Ca²⁺ increases, ABA-insensitive gene expression, and ABA-insensitive seed germination (Kwak et al., 2002) ( Figure 3). Furthermore, the PP2A inhibitor okadaic acid phenocopies the rcn1 phenotypes in wild type plants (Kwak et al., 2002).

The putative protein tyrosine phosphatase gene PHS1 was also suggested to negatively regulate ABA signaling (Quettier et al., 2006), suggesting a role for protein tyrosine phosphatases in stomatal movements (MacRobbie, 2002).

mRNA Metabolism and ABA Signaling

Transcriptional and post-transcriptional regulation provides cellular mechanisms through which ABA signaling is modulated. For reviews on transcription factors that function in ABA signaling, readers are referred to the following reviews (Giraudat, 1995; Finkelstein et al., 2002; Zhu, 2002; Kuhn and Schroeder, 2003; Yamaguchi-Shinozaki and Shinozaki, 2006) Mutations in either of the nuclear mRNA cap binding protein subunits ABH1 and CBP20 confer ABA hypersensitivity in Arabidopsis, indicating that specific mRNA processing events may be needed for expression of negative regulators of ABA signaling (Hugouvieux et al., 2001; Papp et al., 2004) ( Figure 3). Expression of the negative regulator AtPP2CA is greatly reduced in abh1 mutants (Hugouvieux et al., 2001) and atpp2ca mutants were identified in two independent forward genetic screens for causing ABA hypersensitivity (Kuhn et al., 2006; Yoshida et al., 2006b). Further analyses of AtPP2CA transcripts in abh1 showed that this effect can contribute to the abh1 phenotype, but detailed analyses also demonstrated that additional mechanisms are required for generating the ABA hypersensitivity of abh1 (Kuhn et al., 2006). The recent finding that the nuclear mRNA CAP binding proteins, including ABH1, function in generating small non-coding RNAs, may provide further insights into mechanisms by which abh1 causes ABA hypersensitivity (Gregory et al., 2008). Other mRNA binding proteins, for which mutations cause ABA hypersensitivity, include the dsRNA-binding protein HYL1, the poly(A)-specific RNase AHG2/AtPARN, and an Sm-like protein (Lu and Fedoroff, 2000; Xiong et al., 2001; Gregory et al., 2008). These results imply that RNA processing and turnover is utilized by plant cells to regulate ABA signaling or that rate-limiting genes that function in ABA signaling are affected in these mRNA processing mutants (for more details, see Kuhn and Schroeder, 2003).

Another guard cell-expressed RNA-binding protein GRP7 functions in mRNA export to the cytoplasm and positively regulates ABA-induced stomatal closure (Kim et al., 2008) ( Figure 3). The AP2 type AtERF7 interacts with the PKS3 kinase and a putative transcription repressor AtSin3 to negatively regulate ABA signaling (Song et al., 2005). Studies with transgenic plants showed that over-expression of the MYB44 transcription factor enhances ABA sensitivity in stomatal closure (Jung et al., 2008) ( Figure 3). Transgenic plants overexpressing MYB44 have reduced levels of negatively acting PP2Cs including ABI1, ABI2, AtPP2CA, HAB1, and HAB2 (Jung et al., 2008). Two calcium-dependent protein kinase genes CPK4 and CPK11 were also reported to positively regulate ABA signaling possibly through ABF1 and ABF4 transcription factors (Zhu et al., 2007) ( Figure 3).

The syntaxin protein OSM1/SYP61 and the RING finger E3 ligase protein SDIR1 were also found to positively control guard cell ABA signaling (Zhu et al., 2002; Zhang et al., 2007b).