SIZ1 deficiency causes reduced stomatal aperture and enhanced drought tolerance via controlling salicylic acid-induced accumulation of reactive oxygen species in Arabidopsis




Transpiration and gas exchange occur through stomata. Thus, the control of stomatal aperture is important for the efficiency and regulation of water use, and for the response to drought. Here, we demonstrate that SIZ1-mediated endogenous salicylic acid (SA) accumulation plays an important role in stomatal closure and drought tolerance. siz1 reduced stomatal apertures. The reduced stomatal apertures of siz1 were inhibited by the application of peroxidase inhibitors, salicylhydroxamic acid and azide, which inhibits SA-dependent reactive oxygen species (ROS) production, but not by an NADPH oxidase inhibitor, diphenyl iodonium chloride, which inhibits ABA-dependent ROS production. Furthermore, the introduction of nahG into siz1, which reduces SA accumulation, restored stomatal opening. Stomatal closure is generally induced by water deficit. The siz1 mutation caused drought tolerance, whereas nahG siz1 suppressed the tolerant phenotype. Drought stresses also induced expression of SA-responsive genes, such as PR1 and PR2. Furthermore, other SA-accumulating mutants, cpr5 and acd6, exhibited stomatal closure and drought tolerance, and nahG suppressed the phenotype of cpr5 and acd6, as did siz1 and nahG siz1. Together, these results suggest that SIZ1 negatively affects stomatal closure and drought tolerance through the accumulation of SA.


Stomata are natural pores in the epidermis, and are surrounded by a pair of guard cells that regulate the size of stomatal apertures (Bergmann and Sack, 2007). The main function of the stomata is to allow the entry of sufficient CO2 for optimal photosynthesis and the movement of water vapor out of the leaf though the process of transpiration. The stomata also play important roles in the control of leaf temperature by modulating the rates of transpirational water loss and the restriction of pathogen invasion through stomatal closure (Acharya and Assmann, 2009; Underwood et al., 2007; Neill et al., 2008). Multiple environmental factors, such as drought, CO2 concentration, light, humidity, biotic stresses and different plant hormones modulate stomatal apertures (Melotto et al., 2008; Kim et al., 2010; Wilkinson and Davies, 2010).

The opening and closing of stomata is conducted by the osmotic swelling or shrinking of the guard cells. An increase in the osmotic pressure in the guard cells, which is caused by an uptake of K+, results in the opening of stomata. This opening is essential for the supply of CO2 for photosynthesis, but it is also risky because plants can lose water through transpiration: approximately 90% of the water taken up by the plant is lost through transpiration (Schroeder et al., 2001a). Open stomata also allow bacterial invasion into the interior of the leaf (Lindow and Brandl, 2003; Melotto et al., 2006). Abscisic acid (ABA) is a phytohormone that triggers the closing of the stomata under conditions of insufficient water (Acharya and Assmann, 2009). ABA stimulates production of reactive oxygen species (ROS) in the guard cells, which is mediated by the guard cell-expressed NADPH oxidase catalytic subunits, AtrbohD and AtrbohF; the atrbohD/F double mutation impairs the ABA promotion of ROS production and ABA-induced stomatal closing (Kwak et al., 2003). The application of salicylic acid (SA) also induces ROS production (Dong et al., 2001; Mori et al., 2001), leading to stomatal closure (Manthe et al., 1992; Lee, 1998; Mori et al., 2001). It has been suggested that SA mediates ROS production not via NADPH oxidases but rather via a peroxidase-catalyzed reaction. As SA-induced ROS production may close stomata and inhibit pathogen invasion through the opening (Melotto et al., 2006), SA is required for defense against pathogens, which invade through stomata (Melotto et al., 2008). The stomatal closure induced by SA is accompanied by the inactivation of inward-rectifying potassium channels, which are regulated by peroxidase-mediated ROS production and extracellular Ca2+ fluctuation (Khokon et al., 2010a,b). Although both ABA and SA play positive roles in stomatal closure, treatment with ABA suppresses the induction of systemic acquired resistance (SAR), which is mediated by SA (Asselbergh et al., 2008; Yasuda et al., 2008; Bari and Jones, 2009). Because ABA suppresses SAR induction by the inhibition of the SA pathway, independently of the jasmonic acid or ethylene-mediated signaling pathways (Yasuda et al., 2008), ABA and SA antagonistically affect the development of SAR (Fujita et al., 2006). A further understanding of the complex communication between SA and ABA is required.

Some components of the SA-ABA network have been identified. AHG2 encodes a poly(A)-specific ribonuclease, and the mutant exhibits a hyper-response to ABA and higher levels of SA-inducible gene expression than the wild type (Nishimura et al., 2005). Genetic and molecular data have revealed that AHG2 independently modulates ABA, SA and mitochondrial functions (Nishimura et al., 2009). The activation-tagging line, myb96-1d, exhibits an enhanced drought tolerance (Seo et al., 2009) and enhanced disease resistance, with elevated SA levels (Seo and Park, 2010). Previously, we reported that the siz1 mutation, which is impaired in the SIZ-type small ubiquitin-related modifier (SUMO) E3 ligase in Arabidopsis (Miura et al., 2005), conferred ABA hypersensitivity (Miura et al., 2009; Miura and Hasegawa, 2009) and enhanced the accumulation of SA and SA-inducible gene transcripts (Lee et al., 2007). Sumoylation functions in plant development, stress responses and flowering (Miura et al., 2007a; Miura and Hasegawa, 2009), and SIZ1 SUMO E3 ligase plays an important role in the regulation of these functions (Miura et al., 2005, 2007b, 2009, 2010, 2011a,b; Yoo et al., 2006; Catala et al., 2007; Lee et al., 2007; Jin et al., 2008).

Here, we report that SIZ1 negatively affects stomatal apertures through the SA-induced ROS accumulation, independent of ABA hypersensitivity. The siz1 plants exhibited reduced stomatal aperture compared with wild-type plants, but the mutants still showed a response to exogenous ABA. The application of salicylhydroxamic acid (SHAM) and azide, inhibitors of peroxidase that reduce the generation of O2 induced by SA (Mori et al., 2001; Khokon et al., 2010a,b), recovered the stomatal aperture of siz1 to that of the wild type, but the application of diphenyl iodonium chloride (DPI), an inhibitor of NADPH oxidase, did not. Furthermore, the introduction of nahG into siz1, which reduces the SA level in siz1 (Lee et al., 2007), also restored the stomatal aperture, suggesting that the accumulation of SA promotes stomatal closure. Although the stomatal aperture was altered by the siz1 mutation, the stomatal index (stoma number divided by epidermal cell number) was not affected in the siz1 plants. Furthermore, the closure of the stomata in siz1, not the stomatal index, promoted the observed drought resistance; this drought resistance of siz1 was also suppressed by the introduction of nahG. To confirm the linkage between SA and stomatal closure-induced drought tolerance, we also investigated the stomatal apertures and drought resistance of cpr5 and acd6 mutants that accumulate SA (Bowling et al., 1997; Rate et al., 1999). Accordingly, cpr5 and acd6 exhibited reduced stomatal aperture and drought resistance, as did siz1, and cpr5 nahG and acd6 nahG showed a similar sensitivity to drought compared with the wild type. Together, these results suggest that SA accumulation enhances stomatal closing and drought tolerance.


The siz1 mutation causes stomatal closure

The opening and closing of stomata are controlled by environmental factors, such as light, humidity and CO2 concentration (Schroeder et al., 2001a; Fan et al., 2004), as well as by phytohormones, including ABA and SA (Schroeder et al., 2001a; Acharya and Assmann, 2009). Because the siz1 mutant accumulates SA (Yoo et al., 2006; Lee et al., 2007) and exhibits a hyper-response to ABA (Miura et al., 2009), we presumed that the stomatal movement of the siz1 plants may be altered in comparison with the wild type. The stomatal aperture of the wild-type plant was larger than that of siz1–3 under normal conditions (Figure 1a). In the dark, the apertures of both plants were similar, but the stomata of siz1–3 were not opened very wide compared with the aperture of the wild-type plant in the light period (Figure 1b).

Figure 1.

SIZ1 deficiency promotes SA-mediated stomatal closure, but less affects ABA-mediated stomatal closure. (a) The photographs are representative stomata of fifth rosette leaves of 4–week-old Col–0 and siz1–3 plants. Scale bars: 5 μm. (b) The stomatal aperture after light irradiation. Col–0 and siz1–3 plants were grown under long-day conditions (16 h of light/8 h of dark). After finishing the dark period, the stomatal aperture was measured. Values are means ± SEs (≥ 10); three independent experiments were performed and similar results were obtained. (c) The siz1 mutation inhibits fusicoccin-induced stomatal opening. Fusicoccin is a proton (H+)-ATPase activator, which induces H+-pumping and causes stomata to irreversibly open. Fusicoccin or DMSO (control) was added at 10 μm, and the leaves were incubated for 3 h with light irradiation. (d) ABA-mediated stomatal closure was observed in the siz1 mutant, but to a lesser extent than in wild-type plants. The stomatal aperture of Col–0 and siz1–3 without ABA treatment was 3.35 ± 0.039 and 2.58 ± 0.052 μm, respectively. Stomatal aperture of Col–0 and siz1–3 with the application of 10 μm ABA was similar (2.18 ± 0.072 and 2.14 ± 0.090 μm, respectively). (e) Peroxidase is involved in the inhibition of the stomatal opening of siz1. Four- to six-week-old leaves were treated with SHAM (salicylhydroxamic acid) and sodium azide, inhibitors of peroxidase, or DPI (diphenyl iodonium chloride), an inhibitor of NADPH oxidase. In all graphs, values are the means ± SEs (≥ 4). *P < 0.05 indicates a significant difference. (f) The accumulation of SA in siz1 causes stomatal closure, which was restored by nahG. The photographs are of representative stomata of 4–week-old Col–0, siz1–2, nahG, nahG siz1–2 and sid2–2 plants. Scale bar: 5 μm. (g) The stomatal apertures of Col–0, siz1–2, siz1–3, nahG siz1–2, nahG and sid2–2 were evaluated 3 h after light irradiation. Values are means ± SEs (≥ 10). *P < 0.05 indicates a significant difference from Col–0.

Fusicoccin, a fungal elicitor, activates the plasma membrane H+-ATPase and induces H+-pumping across the membrane, which provides the driving force for stomatal opening (Marré, 1979; Sze et al., 1999). Fusicoccin induced the stomatal opening of both Col–0 (wild type) and siz1–3 plants (Figure 1c), but there was still a difference noted between the Col–0 and siz1–3 plants. Because SA inactivates inward-rectifying potassium channel activity in guard cells (Khokon et al., 2010a,b), the accumulation of SA in siz1 may have inhibited the potassium influx, even though the driving force (i.e. the proton gradient) was provided.

Reduced Stomatal Apertures in siz1 are Caused by SA, but Not by ABA

ABA is a hormone that triggers the closing of the stomata when the available water in the soil (or growth medium) is insufficient to keep up with the transpiration rate (Kriedemann et al., 1972; Raschke and Zeevaart, 1976). The application of exogenous ABA induced the stomatal closure in both the wild-type and the siz1 plants, and the closed-stomatal aperture of both plants was similar (Figure 1d).

To elucidate whether the reduced stomatal aperture of siz1 is caused by SA or ABA, inhibitors were applied. SHAM and azide, an inhibitor of peroxidase, reduces the generation of inline image induced by SA (Mori et al., 2001; Khokon et al., 2010a,b); in contrast, DPI, an inhibitor of NADPH oxidase, removes H2O2 or reduces the generation of H2O2 induced by ABA (Levine et al., 1994; Alvarez et al., 1998; Lee et al., 1999; Zhang et al., 2001). The application of SHAM suppressed the reduced stomatal aperture of the siz1–3 plants, whereas DPI had no effect (Figure 1e). The application of another inhibitor, azide, which has a similar effect to SHAM (Khokon et al., 2010a,b), also suppressed the reduced stomatal aperture of the siz1–3 plants (Figure 1e). These results suggest that SA-induced peroxidase, but not ABA-induced NADPH oxidase, is involved in ROS production to close the stomata in siz1–3 plants.

To confirm the SA-induced stomatal closure in siz1, we measured the stomatal aperture in the nahG siz1–2 plants. The bacterial nahG gene encodes a salicylate hydroxylase that catalyzes the conversion of SA to catechol (Yamamoto et al., 1965), and the introduction of nahG into siz1 reduced SA accumulation (Yoo et al., 2006; Lee et al., 2007). The stomatal aperture in nahG siz1–2 was similar to that in nahG (Figure 1f). Because SA accumulation in nahG was less than that of Col–0 (Lee et al., 2007), the stomatal aperture in nahG was larger than in the wild type (Figure 1f,g). nahG plants accumulate catechol, instead of SA. To confirm that catechol has no effect on stomatal opening, sid2–2, which is impaired in isochorismate synthase for SA biosynthesis (Wildermuth et al., 2001), was used for the measurement of stomatal aperture (Figure 1f,g). The sid2–2 mutant exhibited a similar phenotype as nahG (Figure 1g), suggesting that stomatal closure is probably enhanced by SA, but not by catechol.

Expression of SIZ1 in guard cells for the regulation of stomatal aperture

Because SA induces ROS production, mediated by peroxidase (Mori et al., 2001; Khokon et al., 2010a,b), we hypothesized that the siz1 mutation would enhance ROS accumulation in the guard cells. Compared with the wild-type Col–0 plants, the siz1 mutants exhibited a higher accumulation of ROS (Figure 2a), whereas a reduction of SA accumulation in siz1 by the introduction of nahG (to produce nahG siz1–2) decreased the accumulation of ROS (Figure 2a). These results suggest that SIZ1 deficiency causes an decrease in stomatal aperture through the accumulation of SA, which enhances ROS production.

Figure 2.

Expression of SIZ1 in guard cells and ROS accumulation in siz1. (a) Accumulation of ROS was observed more in the guard cells of siz1, cpr5 and acd6 mutants. H2DCF-DA fluorescence levels were measured. The relative fluorescence level was calculated as the level of Col–0, which was normalized to 100%. (b, c) SIZ1pro::GUS expression in the guard cells of the abaxial side. Transgenic plants harboring SIZ1pro::GUS were used to determine SIZ1 expression by the histochemical staining of GUS activity before (b) and after (c) drought treatment. (d) The expression of SIZ1 in guard cells was also monitored by microarray data (Arabidopsis eFP Browser; Winter et al., 2007;

Analysis of transgenic plants expressing the reporter gene β–glucronidase (GUS), under the control of the SIZ1 promoter (from −2035 to −7 bp upstream of the first ATG; Miura et al., 2011a), showed that the guard cells were stained under normal conditions (Figure 2b), and that the expression level of SIZ1 was decreased 7 days after withholding water (Figure 2c). According to microarray data obtained through the Arabidopsis eFP browser (; Winter et al., 2007), mRNA transcripts of SIZ1 have been observed in guard cells (Figure 2d). These results suggest that the expression level of SIZ1 in the guard cells is decreased by drought treatment.

As described above, the siz1 mutation reduced the stomatal aperture (Figure 1). Therefore, we investigated whether the siz1 mutant also showed an alteration in the number of stomata (Figure 3a–c). Guard cells are produced by a series of asymmetric and symmetric divisions of the epidermal cells (Bergmann and Sack, 2007). The leaf abaxial stomatal density and epidermal cells were increased in siz1 (Figure 3d,f) because of a decrease in cell volume in siz1 (Miura et al., 2010). In addition, some of the guard cells in siz1 appeared to be arrested (Figure 3b,c,e). Because the epidermal cells in siz1 were smaller than those in the wild-type plants, leaf stomatal indices (stomata:epidermal cell ratio) of the siz1 and wild-type plants were calculated, and were found to be similar (Figure 3g). These results indicate that stomatal index is not altered in the siz1 mutant.

Figure 3.

The siz1 mutation does not affect the stomatal index. (a, b, c) The photographs are of representative abaxial epidermal layers from Col–0 (a), siz1–2 (b) and siz1–3 (c) leaves. Scale bars: 50 μm. The arrowhead indicates the arrested development of a guard cell. The epidermal cells from 4–week-old Col–0 and siz1–3 plants were used to measure the stomatal density (d), arrested stomata (e), epidermal cell density (f) and stomatal index (g) (mean ± SE,= 20–22). (d) The stomatal number was counted in the area from the microscope image, and the stomatal density was calculated. (e) Where the asymmetrical division of cells had not developed into stomata, the cells were counted as arrested stomata. The number of arrested stomata was excluded from the calculations of stomatal density and stomatal index. (f) The epidermal cell number was counted, and the density was calculated. (g) The stomatal index is the number of stomata divided by the total number of epidermal cells (including stomata).

siz1 exhibited drought resistance, which was suppressed by nahG

Even though siz1 did not display altered stomatal indices, compared with the wild type (Figure 3), the siz1 mutation reduced the stomatal aperture (Figure 1). Thus, we posited that the siz1 mutant loses less water than wild-type plants. The siz1–2 and siz1–3 mutations conferred an enhanced survival against the stress from water deficit imposed by withholding water (Figures 4a,b and 5a). More than 80% of the siz1 plants survived, compared with <20% of the wild type (Figures 4c and 5b). Like era1 (Pei et al., 1998), the stomatal aperture of siz1 was smaller than that of the wild type when plants were grown under conditions of sufficient water or in drought conditions (Figure 4d). In drought conditions the relative water content of the soil was similar in each pot (Figure 4e), indicating that evaporation is the major cause of soil moisture depletion. The stomatal aperture was similar in the siz1–2 plants transformed with SIZ1:GFP when compared with the wild-type plants (Figure 4d), because the expression level of SIZ1:GFP in siz1–2 was similar to that of SIZ1 in the wild type (Miura et al., 2009).

Figure 4.

The siz1 mutation exhibited drought tolerance. (a) Five-week-old wild-type (Col–0), siz1–2 and siz1–3 plants grown under long-day conditions at 23°C. (b) Three-week-old plants were withheld water for 2 weeks. (c) Watering was resumed and plants were incubated for 4 days after the 2–week drought treatment. (d) The stomatal aperture in siz1 was smaller than wild-type plants after the water-deficit treatment. The mean size of stomatal apertures of Col–0, siz1–2, siz1–3, era1 (a control; Pei et al., 1998) and siz1–2 transformed with SIZ1: GFP (Miura et al., 2009), were measured at the indicated times (= 50 or more ± SEs). The expression levels of SIZ1 in a transgenic line were similar to those of the wild type (Miura et al., 2009). (e) The relative soil water content after water-withheld was measured in each pot. The average was calculated with standard error (n = 12).

Figure 5.

nahG suppresses the drought-tolerance phenotype of siz1. (a) The photographs are of representative Col–0, siz1–2, siz1–3, nahG and nahG siz1–2 plants. The conditions were the same as described in Figure 4. (b) The survival ratio was measured. (c) Water loss per millimeter square was calculated in Col–0, siz1–2, siz1–3, nahG and nahG siz1–2 plants. (d, e) sid2–2 exhibited a similar phenotype to nahG.

Along with the restoration of the stomatal aperture in siz1 by the introduction of nahG (Figure 1g), the drought-resistant phenotype of siz1 was also suppressed in nahG siz1–2 plants (Figure 5a). The survival of nahG siz1–2 after drought treatment was approximately 20%, although more than 80% of siz1 plants survived (Figure 5b), suggesting that SIZ1 negatively affects drought resistance, as well as stomatal closure, through the control of the accumulation of SA. After leaves were detached, water loss from leaves was measured and the water loss was divided by leaf area. Water loss per leaf area was smaller in siz1 mutants than in wild-type or nahG plants (Figure 5c). The reduction of water loss was suppressed in nahG siz1–2 (Figure 5c). Thus, the siz1 mutation may enhance drought tolerance. Again, the sid2–2 mutant exhibited a similar drought phenotype as did nahG (Figure 5d,e). Because endogenous SA was increased up to fivefold during drought stress in Phillyrea angustifolia (Munne-Bosch and Penuelas, 2003), we observed SA-responsive gene expression. Drought induced expression of PR1 and PR2 (Figure 6a), which are SA-inducible genes, suggesting that SA accumulation may be required for drought tolerance, and that endogenous SA induction may lead to the induction of the protective mechanism (Hara et al., 2011). We also investigated the expression of the ABA synthesis gene, NCED3, and several drought-responsive genes, such as KIN1, RD29A and RD29B, before and after polyethylene glycol (PEG) treatment, which is commonly used to mimic drought treatment (Yu et al., 2008). The expression level of NCED3, KIN1, RD29A and RD29B was high in siz1–2, and it was suppressed in nahG siz1–2 (Figure 6b). These results suggest that the expression of drought-responsive genes is regulated by SIZ1.

Figure 6.

Drought enhances the expression of SA-responsive genes, and drought-responsive genes are upregulated in siz1. (a) The 2–week-old plants were withheld water for 10 or 14 days. The expression levels of PR1 and PR2 were measured by real-time RT-PCR analysis. (b) The expression levels of drought-responsive genes. The 7–day-old seedlings grown on MS medium at 23°C under a long-day photoperiod (16 h of light/8 h of dark) were transferred into liquid MS medium supplemented with 0 or 10% PEG6000, and incubated for 3 h. Quantitative RT-PCR was then performed. Data are means ± SDs (n = 3).

Because we assumed that genes in which expression is increased by SA, drought and H2O2 treatment, and that are also expressed in guard cells, may play a role in the regulation of SA and drought-mediated stomatal movement, our microarray data of siz1 (Lee et al., 2007) was compared with microarray data with SA, drought and H2O2 treatment, and SA-accumulating (cpr5) or SA-deficient (sid2) mutants, which were extracted from Genvestigator (; Zimmermann et al., 2004; Hruz et al., 2008). By using these data, cluster analysis was performed (Figure 7a). Genes in two clusters (surrounded by the yellow line in Figure 7a) were upregulated by SA, drought, and the SA-accumulating mutants siz1 and cpr5, but were not significantly different in sid2 (Figure 7b,c). Then, the expression of these genes in guard cells were investigated with the Arabidopsis eFP Browser (; Winter et al., 2007), which contains expression profiles of mesophyll and guard cell protoplasts, with or without ABA treatment (Yang et al., 2008). The genes, which have an expression level at 100 or more in guard cells, include genes encoding β–glucosidase, RNA-binding, ribonuclease, RING/U–box superfamily, dehydrin LTI30, methionine γ–lyase, alternative oxidase, auxin efflux carrier and an unknown (Figure 7d). Expression of these genes in wild-type, siz1, nahG and nahG siz1–2 plants grown under normal conditions, and in wild-type plants grown under drought treatment, was measured (Figure 8). The expression level of the genes encoding β–glucosidase, RNA-binding, ribonuclease, dehydrin LTI30, methionine γ–lyase, alternative oxidase and the unknown was increased by the siz1 mutation, and the expression was repressed in nahG siz1–2 (Figure 8). The genes described above, except for the gene encoding RNA binding, were also induced by drought treatment (Figure 8). RING/U–box gene was not significantly different among genotypes and treatment. Among these genes, dehydrin LTI30 is involved in freezing tolerance with LTI29 (Puhakainen et al., 2004). The overexpression of the chimeric genes LTI29 and LTI30 resulted in an accumulation of the corresponding dehydrins to levels higher than in cold-acclimated wild-type plants, and improved survival when exposed to freezing stress. The accumulation of dehydrin also contributes to an improved tolerance to salt and drought stress through osmotic adjustment (Brini et al., 2007). It is plausible that these genes are involved in the regulation of stomatal movement in response to SA, drought and H2O2.

Figure 7.

Functional categorization of SIZ1-regulated genes. (a) Cluster analysis of transcripts, which were up- or downregulated in siz1–2, with comparison to expression profiles of cpr5 and sid2 mutants, and SA, drought and H2O2 treatment. (b, c) Two of the clusters surrounded by yellow lines were magnified. The genes, the expression levels of which are high under all conditions, except for sid2, were included. (d) The genes that have expression levels at 100 or more in guard cells in the Arabidopsis e–FP browser were selected from the two clusters (b, c).

Figure 8.

Drought- and SIZ1-regulated gene expression. The relative expression level of eight genes picked up from the microarray data comparison (Figure 7) were measured. RD29A expression was measured as a marker for drought stress. Two-week-old wild-type (a), siz1–2 (c), siz1–3 (d), nahG (e) and nahG siz1–2 (f) plants, grown under normal conditions, and wild-type plants 5 days after water-withheld (b), were harvested, and relative expression levels were measured by real-time RT-PCR.

SA is one of the factors that enhance stomatal closing and drought tolerance

Because the reduced stomatal aperture and drought resistance caused by the siz1 mutation are more likely to be linked to the accumulation of SA in the siz1 mutant, we investigated whether other SA-accumulation mutants, cpr5 and acd6 (Bowling et al., 1997; Rate et al., 1999), exhibited stomatal closure and drought resistance. The stomatal apertures of 3–week-old plants were investigated before and after light irradiation. Similar to the siz1 mutants, both cpr5 and acd6 showed reduced stomatal aperture compared with the wild type, and the introduction of nahG into cpr5 and acd6 rescued this reduction of stomatal aperture (Figure 9a,b). Furthermore, both of these mutants (cpr5 and acd6) exhibited a drought-tolerant phenotype compared with wild-type plants, and both nahG cpr5 and nahG acd6 were suppressed in the drought-resistant phenotype (Figure 9c,d). These results indicate that SA accumulation caused by the siz1, cpr5 or acd6 mutation enhances reduced stomatal aperture and drought resistance.

Figure 9.

The cpr5 and acd6 mutants that accumulate SA also exhibited stomatal closure and drought tolerance, and the reduced SA levels caused by the introduction of nahG suppressed these phenotypes. (a) The photographs are of representative stomata of 4–week-old Col–0, cpr5, acd6, nahG, nahG cpr5 and nahG acd6 plants. Scale bar: 5 μm. (b) The stomatal apertures of Col–0, cpr5, nahG cpr5, acd6, nahG acd6 and nahG were evaluated 3 h after light irradiation. Values are the means ± SEs (≥ 20). *P < 0.05 indicates a significant difference from Col–0. (c) The photographs are of representative Col–0, cpr5, acd6, nahG, nahG cpr5 and nahG acd6. (d) The survival ratio was measured for conditions similar to those described in Figure 5.


In this study, we established that SIZ1 negatively affects stomatal movement by the repression of SA accumulation in a non-ABA-dependent manner. Probably because of reduced stomatal aperture in the siz1 mutant (Figure 1), the siz1 mutant also exhibited drought tolerance (Figure 5). Restoration of stomatal opening by nahG expression, i.e. nahG siz1–2, caused drought sensitivity (Figure 5). These results suggest that SIZ1 regulation of stomatal movement and drought response is caused by an accumulation of SA (Figure 10). Similar phenotypes were observed in the other SA-accumulating mutants, cpr5 and acd6 (Figure 9).

Figure 10.

A model for the control of stomatal closure. The accumulation of SA by siz1 enhanced the stomatal closure (Figure 1), which is suppressed by the reduction of SA, resulting from the introduction of nahG (Figure 1). Generally, SA induces ROS through peroxidase; thus, the application of SHAM and azide, inhibitors of peroxidase, to siz1 recovered the stomatal opening (Figure 1e). Lastly, SA-induced ROS production induces stomatal closure. ROS is highly accumulated in the guard cells of siz1 (Figure 2). This is a different mechanism from the ABA-mediated ROS production, because the application of DPI, an inhibitor of NADPH oxidase, did not suppress stomatal closure in siz1 (Figure 1e). Stomatal closure observed in siz1 mutants may cause drought tolerance, because transpiration is reduced. And it may also cause the inhibition of pathogen invasion, as the siz1 mutant exhibited an increased resistance to the bacterial pathogen Pseudomonas syringae pv. tomato DC3000 (Lee et al., 2007).

SA-mediated ROS accumulation through peroxidase promotes stomatal closure and drought tolerance

The size of the stomatal pore is tightly regulated (Roelfsema and Hedrich, 2005). When water is sufficient, plants open their stomata to transfer water and minerals from the roots to the shoots, and to exchange gases. The stomatal movement is affected by several environmental stimuli, such as humidity, CO2 concentration, light intensity and plant hormones, including ABA, SA and ethylene (Wilkinson and Davies, 2010). The role of ABA is well understood: it plays a central role in stomatal closure during drought stress (Wilkinson and Davies, 2010), and the ABA signal in guard cells results in the production of nitric oxide and H2O2, leading to stomatal closure (Schroeder et al., 2001a,b; Bright et al., 2006).

The stomata are also closed in response to live bacteria and purified pathogen/microbe-associated molecular patterns (PAMPs/MAMPs). This bacterium-induced stomatal closure is controlled by PAMP signaling and the homeostasis of the defense hormone SA (Melotto et al., 2006). Stomatal closing in response to bacterial infection has been shown to be compromised in nahG and the SA biosynthetic mutant eds16–2 (Melotto et al., 2006). Thus, it is inferred that the stomatal closure by SA is required for preventing bacterial entry (Melotto et al., 2008). Stomatal closure is important for the defense mechanism. Because the siz1 mutant exhibited resistance to the pathogen Pseudomonas syringae pv. tomato DC3000 (Lee et al., 2007), stomatal closure in the siz1 mutant is likely to contribute to blocking pathogen invasion (Figure 10). According to our results, stomatal closure was promoted by endogenous SA. The accumulation of SA enhanced the reduced stomatal aperture, but the reduction of stomatal aperture was compromised by the nahG-mediated SA reduction (Figures 1g and 9b). The exogenous application of SA induces ROS, H2O2 and Ca2+ accumulation, leading to stomatal closure (Dong et al., 2001; Mori et al., 2001; Liu et al., 2003; He et al., 2007).

The effect of SA on drought tolerance remains unclear (Hara et al., 2011). SA potentiates ROS generation in the photosynthetic tissues of Arabidopsis thaliana during salt and osmotic stresses (Borsani et al., 2001). In agreement with these results, the application of a high concentration of SA (0.5 mm) decreased the drought tolerance of maize plants (Németh et al., 2002). However, several reports have demonstrated that SA increased the drought tolerance of wheat, tomato and bean (Senaratna et al., 2000; Sakhabutdinova et al., 2003; Hara et al., 2011). Indeed, the endogenous SA content was increased during water deficit in Phillyrea angustifolia plants (Munne-Bosch and Penuelas, 2003). In Arabidopsis, plants carrying an activation-tagged allele of adr1 or myb96-1d exhibited both SA-dependent disease resistance and drought tolerance (Grant et al., 2003; Chini et al., 2004; Seo et al., 2009; Seo and Park, 2010). Overexpression of CBP60g encoding calmodulin-binding protein in Arabidopsis caused elevated SA accumulation and enhanced tolerance to drought stress (Wan et al., 2012). In addition, the expression of the pepper pathogen-induced gene, CAPIP2, in Arabidopsis confers disease resistance and drought tolerance (Lee et al., 2006). Our study demonstrates that SA induced stomatal closure, which resulted in an enhanced tolerance to drought (Figures 4, 5 and 9). The correlation between stomatal closure and drought tolerance has been well established. The loss of function of a zinc-finger protein DST (drought and salt tolerance) increases stomatal closure, leading to an enhanced drought tolerance in rice (Huang et al., 2008). However, in a previous report, the siz1 mutant showed a drought-sensitive phenotype (Catala et al., 2007). As described above, SA affects both drought sensitivity and tolerance. Because several biotic and abiotic stresses may cause an increase in the accumulation of SA (Janda et al., 2007), and the application of a high concentration of SA resulted in drought sensitivity (Németh et al., 2002), the siz1 mutant may suffer from other stresses than drought stress, and may accumulate high ROS, which may affect the growth of the siz1 plants in such conditions. We should note that the siz1 mutant also showed sensitivity to oxidative stresses.

Interaction between SA and ABA for stomatal closure and water deficit responses

Phytohormones ABA and SA have a broad effect on plant physiology, including developmental processes, abiotic responses and defense mechanisms (Lee et al., 2007). The complex regulatory and interaction network occurring between hormone-signaling pathways may be required for the activation of responses to different types of stimuli (Acharya and Assmann, 2009; Janda et al., 2007; Wilkinson and Davies, 2010). Generally, ABA antagonistically regulates SA-dependent defense response (Fujita et al., 2006), even though SA and ABA exhibit the same effects on stomatal closure. As we proposed (Figure 10), ROS accumulation may be an integrator between SA and ABA signaling on the regulation of stomatal closure. Hydrogen peroxide, one of the major ROS, plays an important role as a second messenger in ABA-induced stomatal closure (Khokon et al., 2010a,b), and is generated by the NADPH oxidase (Kwak et al., 2003). On the other hand, SHAM-sensitive guaiacol peroxidase, which is localized in both epidermal and guard cells, can also generate inline image , one of the ROS species, in guard cells induced by SA (Mori et al., 2001; Khokon et al., 2010a,b).

The detailed mechanisms for the interaction between SA and ABA on the regulation of water-deficit responses are unclear. One possibility is that stomatal closure induced by either SA or ABA causes a reduction of transpiration, resulting in the storage of water in leaves for survival under drought conditions. Another possibility is that SA stimulates ABA accumulation. SA treatment induces an accumulation of ABA, which may lead to a reduction in the damage caused by water deficit in seedlings of barley (Bandurska and Stroinski, 2005). Because several reports demonstrate that the application of SA increases stomatal closure and drought tolerance, also induced by ABA, SA and ABA may synergistically regulate these responses.

The regulation of SA signaling and sumoylation

The siz1 mutation causes constitutive SAR responses, including the expression of many SA-responsive genes (Lee et al., 2007). Some components in SA signaling are presumed to be targets for sumoylation. Proteome analyses have revealed that several kinds of proteins are SUMO substrates in Arabidopsis – they include transcription factors, homeodomain proteins, chromatin modifiers, RNA-related factors and cell cycle regulators (Budhiraja et al., 2009; Elrouby and Coupland, 2010; Miller et al., 2010). Several chromatin remodeling factors, histone acetylation/deacetylation enzymes and histones have also been identified as SUMO substrates (Budhiraja et al., 2009; Elrouby and Coupland, 2010; Miller et al., 2010). Because several SA-responsive genes are constitutively expressed in siz1, it is plausible that the formation of heterochromatin in the promoter of SA-responsive genes may be controlled by the SIZ1-mediated sumoylation of the chromatin-modifying factors and histones (van den Burg and Takken, 2009). Histones play an important role in the regulation of SA-responsive gene expression. Double mutations in HTA9 and HTA11, which encode histone H2A.Z, cause the constitutive expression of SAR marker genes, and increase the resistance to P. syringae pv. tomato DC3000 (March-Díaz et al., 2008). Overexpression of HDA19, encoding histone deacetylase, enhanced the resistance to P. syringae and expression of PR1 (Kim et al., 2008). It is plausible that the SIZ1-mediated sumoylation of histone or chromatin modifiers plays a role in the transcriptional repression of SA-responsive genes in plant cells.

In summary, the accumulation of endogenous SA caused by the siz1 mutations enhanced the production of ROS mediated by SHAM-sensitive peroxidases (Figure 1), and the accumulated ROS resulted in reduced stomatal aperture and drought tolerance (Figure 10).

Experimental Procedures

Plant materials, growth conditions and physiological analyses

The Arabidopsis T–DNA insertion mutants, siz1–2 and siz1–3 (Miura et al., 2005), sid2–2 (stock no. CS16438 in the Arabidopsis Biological Resource Center at Ohio State University,, and nahG plants (van Wees and Glazebrook, 2003) were in the A. thaliana Col–0 background. nahG siz1–2 plants were identified by diagnostic PCR (Lee et al., 2007). The era1 (Pei et al., 1998), cpr5 and nahG cpr5 (Bowling et al., 1997), and acd6 and nahG acd6 (Rate et al., 1999) mutants were kindly provided by Dr Peter McCourt (University of Toronto), Dr Xinnian Dong (Duke University) and Dr Jean T. Greenberg (University of Chicago), respectively. SIZ1Pro::SIZ1:GFP was transformed into siz1–2 (Jin et al., 2008). Arabidopsis plants were grown in soil at 23°C under a long-day photoperiod (16–h light/8–h dark).

For the drought tolerance test, when plants were 3 weeks old, watering was withheld for 2 weeks; plants were then watered. The relative humidity was ~60%. To quantitate the survival ratio, between six and 10 plants of each genotype were grown in the same pot. After water was withheld followed by re-watering, the survival ratio was calculated. The relative soil water content was calculated as follows. A container with soil and each plant was irrigated with water to saturation and weighed. This value minus the weight of the container was used as the initial weight. After water was withheld, the soil weight was measured periodically (fresh weight). Two weeks after water was withheld, the container was incubated at 60°C for 3 days to remove water completely, and then weighed (dry weight). The relative soil water content was calculated as (fresh weight−dry weight)/(initial weight−dry weight) × 100. Leaf water loss per unit area was calculated as (fresh weight just after being detached – fresh weight 1 or 2 h after being detached)/leaf area (mg mm−2). Leaf area was measured by ImageJ software.

Anatomical analysis

Leaves were incubated with a fixing solution containing 90% ethanol and 10% acetate overnight, and were then washed with 90, 70, 50 and 30% ethanol for 20 min at each step. Leaves were then incubated in 80% chloral hydrate and 10% glycerol (Miura et al., 2011c). Microscopic images were photographed with a DM RXA–6 microscope (Leica,, with differential interference contrast. Stomatal density (stomatal number per area), the number of arrested stomata, epidermal cell density (epidermal cell number per area) and stomatal index (stomatal number divided by epidermal cell and stomatal number) were measured as described by Yoo et al. (2010).

To observe stomatal aperture, the abaxial epidermis was obtained by transparent cellulose adhesive tape, and the epidermis was incubated in buffer as described by Yoo et al. (2010). The epidermis was placed on a slide. Images were obtained with a DM RXA–6 microscope (Leica).

Application of inhibitors

To examine the involvement of peroxidase and NADPH oxidase, SHAM (2 mm), sodium azide (1 μm), DPI (20 μm) and fusicoccin (10 μm) were employed. Each inhibitor was applied to the assay solution 30 min prior to light irradiation. Then stomatal apertures were measured after 2 h of incubation.

Quantitative RT-PCR

Total RNA was isolated using TRIZOL reagent (Invitrogen,, according to the manufacturer's protocol. Real-time PCR was performed as described previously (Miura et al., 2012a,b), with gene-specific primers (Table S1).

Histochemical analyses for GUS activity

The promoter region of SIZ1 was cloned into pCAMBIA1391Z, and transgenic plants were generated by the Agrobacterium-mediated, floral-dip method (Miura et al., 2011a). Three-week-old SIZ1Pro::GUS plants were subjected to drought conditions for 7 days. The leaves of the plants were incubated for 12 h at 37°C in a GUS reaction buffer (Miura et al., 2010). The abaxial side of the leaf was observed with a DM RXA–6 microscope (Leica).

Detection of ROS

The ROS production in guard cells was analyzed by using H2DCF-DA, as described previously (Murata et al., 2001; Munemasa et al., 2007). The epidermal peels were incubated for 3 h in a medium containing 5 mm KCl, 50 μm CaCl2 and 10 mm 2–(N–morpholino)ethanesulfonic acid (MES)-Tris (pH 6.15), and then 50 μm H2DCF-DA was added to this medium. The epidermal tissues were incubated for 30 min at 25°C, and then the excess dye was washed out. The dye-loaded tissues were incubated for 30 min at room temperature. The fluorescence of the guard cells was imaged and analyzed using aqua cosmos (Hamamatsu Photonics, Japan,

Cluster analysis

The microarray data of siz1–2, which is compared with wild-type seedlings (Lee et al., 2007), was used. The expression data of cpr5 and sid2 mutants, SA, drought and H2O2 treatments were obtained from the microarray database Genevestigator (; Zimmermann et al., 2004; Hruz et al., 2008). The conditions were as follows: cpr5, young rosette leaves (experiment AT-00175); sid2, leaf tissue samples of 4–week-old plants (experiment AT-00393); SA, 10 μm salicylic acid treatment for 3 h (experiment AT-00113); drought 1, adult leaves stressed by being completely deprived of irrigation for 10 days (experiment At-00290); drought 2, leaf samples with water withheld for 7 days (experiment AT-00292); light/drought, plants grown under approximately 250 μE m−2 s−1 light with no water for 3 days (experiment At-00319); and H2O2, treatment with 20 mm H2O2 for 1 h (experiment At-00185). Cluster analysis of the transcripts was performed as described by Miura et al. (2011a).

Statistical analysis

The significance of differences between sample means was calculated by a Student's t–test.

Accession numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative of GenBank/EMBL/DDBJ databases under the following accession numbers: ACD6 (At4g14400, AY344843); CPR5 (At5g64930, AY033229); and SIZ1 (At5g60410, AY700572).


We thank Ms Risa Osada and Ms. Ayaka Sato for technical assistance. We thank Drs Peter McCourt (University of Toronto), Xinnian Dong (Duke University), Jean T. Greenberg (University of Chicago) and Jane Glazebrook (Syngenta Research and Technology) for providing seeds. This research was supported, in part, by the following: a Special Coordination Funds for Promoting Science and Technology grant from the Ministry of Education, Culture, Sports, Science, and Technology, the Japanese Government (MEXT); a grant for Scientific Research on Innovative Areas from MEXT on ‘Environmental Sensing of Plants: Signal Perception, Processing Cellular Responses’ (23120503); Cooperative Research Grant of the Gene Research Center, the University of Tsukuba; Bilateral Program from the Japan Society for the Promotion of Science; and a US Department of Agriculture–National Research Initiative Competitive Grant (2008-35100-04529).