Reactive Oxygen Species-Dependent Nitric Oxide Production Contributes to Hydrogen-Promoted Stomatal Closure in Arabidopsis^,

Hydrogen-mediated stomatal closure in Arabidopsis and drought tolerance involves RbohF-dependent ROS production, subsequent NR-associated NO production, and GORK activation.

Abstract

The signaling role of hydrogen gas (H₂) has attracted increasing attention from animals to plants. However, the physiological significance and molecular mechanism of H₂ in drought tolerance are still largely unexplored. In this article, we report that abscisic acid (ABA) induced stomatal closure in Arabidopsis (Arabidopsis thaliana) by triggering intracellular signaling events involving H₂, reactive oxygen species (ROS), nitric oxide (NO), and the guard cell outward-rectifying K⁺ channel (GORK). ABA elicited a rapid and sustained H₂ release and production in Arabidopsis. Exogenous hydrogen-rich water (HRW) effectively led to an increase of intracellular H₂ production, a reduction in the stomatal aperture, and enhanced drought tolerance. Subsequent results revealed that HRW stimulated significant inductions of NO and ROS synthesis associated with stomatal closure in the wild type, which were individually abolished in the nitric reductase mutant nitrate reductase1/2 (nia1/2) or the NADPH oxidase-deficient mutant rbohF (for respiratory burst oxidase homolog). Furthermore, we demonstrate that the HRW-promoted NO generation is dependent on ROS production. The rbohF mutant had impaired NO synthesis and stomatal closure in response to HRW, while these changes were rescued by exogenous application of NO. In addition, both HRW and hydrogen peroxide failed to induce NO production or stomatal closure in the nia1/2 mutant, while HRW-promoted ROS accumulation was not impaired. In the GORK-null mutant, stomatal closure induced by ABA, HRW, NO, or hydrogen peroxide was partially suppressed. Together, these results define a main branch of H₂-regulated stomatal movement involved in the ABA signaling cascade in which RbohF-dependent ROS and nitric reductase-associated NO production, and subsequent GORK activation, were causally involved.

Stomata are responsible for leaves of terrestrial plants taking in carbon dioxide for photosynthesis and likewise regulate how much water plants evaporate through the stomatal pores (Chaerle et al., 2005). When experiencing water-deficient conditions, surviving plants balance photosynthesis with controlling water loss through the stomatal pores, which relies on turgor changes by pairs of highly differentiated epidermal cells surrounding the stomatal pore, called the guard cells (Haworth et al., 2011; Loutfy et al., 2012).

Besides the characterization of the significant roles of abscisic acid (ABA) in regulating stomatal movement, the key factors in guard cell signal transduction have been intensively investigated by performing forward and reverse genetics approaches. For example, both reactive oxygen species (ROS) and nitric oxide (NO) have been identified as vital intermediates in guard cell ABA signaling (Bright et al., 2006; Yan et al., 2007; Suzuki et al., 2011; Hao et al., 2012). The key ROS-producing enzymes in Arabidopsis (Arabidopsis thaliana) guard cells are the respiratory burst oxidase homologs (Rboh) D and F (Kwak et al., 2003; Bright et al., 2006; Mazars et al., 2010; Marino et al., 2012). Current available data suggest that there are at least two distinct pathways responsible for NO synthesis involved in ABA signaling in guard cells: the nitrite reductase (NR)- and l-Arg-dependent pathways (Desikan et al., 2002; Besson-Bard et al., 2008). Genetic evidence further demonstrated that removal of the major known sources of either ROS or NO significantly impairs ABA-induced stomatal closure. ABA fails to induce ROS production in the atrbohD/F double mutant (Kwak et al., 2003; Wang et al., 2012) and NO synthesis in the NR-deficient mutant nitrate reductase1/2 (nia1/2; Bright et al., 2006; Neill et al., 2008), both of which lead to impaired stomatal closure in Arabidopsis. Most importantly, ROS and NO, which function both synergistically and independently, have been established as ubiquitous signal transduction components to control a diverse range of physiological pathways in higher plants (Bright et al., 2006; Tossi et al., 2012).

The guard cell outward-rectifying K⁺ channel (GORK) encodes the exclusive voltage-gated outwardly rectifying K⁺ channel protein, which was located in the guard cell membrane (Ache et al., 2000; Dreyer and Blatt, 2009). Expression profiles revealed that this gene is up-regulated upon the onset of drought, salinity, and cold stress and ABA exposure (Becker et al., 2003; Tran et al., 2013). Reverse genetic evidence further showed that GORK plays an important role in the control of stomatal movements and allows the plant to reduce transpirational water loss significantly (Hosy et al., 2003) and participates in the regulation of salinity tolerance by preventing salt-induced K⁺ loss (Jayakannan et al., 2013). Due to the high complexity of guard cell signaling cascades, whether and how ABA-triggered GORK up-regulation is attributed to the generation of cellular secondary messengers, such as ROS and NO, is less clear.

Hydrogen gas (H₂) was recently revealed as a signaling modulator with multiple biological functions in clinical trails (Ohsawa et al., 2007; Itoh et al., 2009; Ito et al., 2012). It was previously found that a hydrogenase system could generate H₂ in bacteria and green algae (Meyer, 2007; Esquível et al., 2011). Although some earlier studies discovered the evolution of H₂ in several higher plant species (Renwick et al., 1964; Torres et al., 1984), it was also proposed that the eukaryotic hydrogenase-like protein does not metabolize H₂ (Cavazza et al., 2008; Mondy et al., 2014). Since the explosion limit of H₂ gas is about 4% to 72.4% (v/v, in the air), the direct application of H₂ gas in experiments is flammable and dangerous. Regardless of these problems to be resolved, the methodology, such as using exogenous hydrogen-rich water (HRW) or hydrogen-rich saline, which is safe, economical, and easily available, provides a valuable approach to investigate the physiological function of H₂ in animal research and clinical trials. For example, hydrogen dissolved in Dulbecco’s modified Eagle’s medium was found to react with cytotoxic ROS and thus protect against oxidative damage in PC12 cells and rats (Ohsawa et al., 2007). The neuroprotective effect of H₂-loaded eye drops on retinal ischemia-reperfusion injury was also reported (Oharazawa et al., 2010). In plants, corresponding results by using HRW combined with gas chromatography (GC) revealed that H₂ could act as a novel beneficial gaseous molecule in plant responses against salinity (Xie et al., 2012; Xu et al., 2013), cadmium stress (Cui et al., 2013), and paraquat toxicity (Jin et al., 2013). More recently, the observation that HRW could delay the postharvest ripening and senescence of kiwifruit (Actinidia deliciosa) was reported (Hu et al., 2014).

Considering the fact that the signaling cascades for salt, osmotic, and drought stresses share a common cascade in an ABA-dependent pathway, it would be noteworthy to identify whether and how H₂ regulates the bioactivity of ABA-induced downstream components and, thereafter, biological responses, including stomatal closure and drought tolerance. To resolve these scientific questions, rbohD, rbohF, nia1/2, nitric oxide associated1 (noa1; Van Ree et al., 2011), nia1/2/noa1, and gork mutants were utilized to investigate the relationship among H₂, ROS, NO, and GORK in the guard cell signal transduction network. By the combination of pharmacological and biochemical analyses with this genetics-based approach, we provide comprehensive evidence to show that H₂ might be a newly identified bioeffective modulator involved in ABA signaling responsible for drought tolerance, that HRW-promoted stomatal closure was mainly attributed to the modulation of ROS-dependent NO generation, and that GORK might be the downstream target protein of H₂ signaling.

RESULTS

ABA Rapidly Stimulates H₂ Release and Production

The phytohormone ABA plays an essential role in the adaptive responses to environmental stimuli, such as dehydration and high salinity (Cutler et al., 2010; Umezawa et al., 2010). To investigate whether ABA affects endogenous H₂ release, an H₂-specific microelectrode system (Unisense) was applied to monitor the released H₂, which is supposed to be dissolved in extracellular fluid of Arabidopsis leaf tissues. When ABA was applied for about 10 min, in comparison with the basal levels of H₂ in the control samples, a significant, rapid, and sustained increase of H₂ released by leaf tissues of Arabidopsis wild-type plants was detected (Fig. 1A).

Figure 1.

ABA-stimulated H₂ release and production, HRW-promoted H₂ production, stomatal closure, and drought tolerance in Arabidopsis. A, Real-time dynamics of H₂ release from leaves of Arabidopsis wild-type (Col-0) plants in response to ABA (100 μm). B, Time-course analysis of endogenous H₂ production in leaves of Arabidopsis wild-type (Col-0) plants in response to ABA (100 μm) or HRW (H₂ saturated in CO₂-free MES buffer; 100%). FW, Fresh weight. C, Time-course detection of stomatal aperture in response to HRW (H₂ saturated in MES buffer; 100%) or ABA (30 μm). D, Drought tolerance assay of wild-type (Col-0) plants irrigated with or without HRW (H₂ saturated in water; 100%). Plants were cultured in pots and irrigated with or without HRW for 7 d before stopping irrigation (drought). The photographs show plants at days 0 and 15 after drought. The survival rates were then measured 7 d after rewatering. Control (Con) means a treatment with CO₂-free MES buffer only (A–C) or irrigated with water (D). Data are presented as means ± se of three independent experiments (A and B) or 30 guard cells (C). Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

The results of ABA-triggered H₂ production were further verified by GC analysis (Fig. 1B; Supplemental Materials and Methods S1). When ABA was applied for 10 min, in comparison with the basal levels of H₂ in the control samples (about 50 nmol g⁻¹ fresh weight, 0.1 µL L⁻¹), an increase in the endogenous H₂ production of Arabidopsis leaves was detected (about 139 nmol g⁻¹ fresh weight, 0.278 µL L⁻¹). Importantly, it was found that Arabidopsis endogenous H₂ production was raised from 0.1 µL L⁻¹ to about 0.288 µL L⁻¹ (a physiological condition; in comparison with 0.278 µL L⁻¹ caused by ABA treatment) when 100% saturated HRW (1.56 µL L⁻¹ H₂ in MES buffer; Supplemental Table S1) was applied for 10 min. These results also indicated that exogenously applied HRW could cause the rapid induction of endogenous H₂ production in Arabidopsis.

HRW Promotes Stomatal Closure and Drought Tolerance

To verify whether H₂ had any effect on the regulation of stomatal closure in the plant drought stress response, stomatal movement assays were performed. ABA, the inducer of stomatal closure, was used as a positive control. Compared with the epidermal fragments of control samples, treatments with 25%, 50%, 75%, and 100% saturated HRW (about 0.30, 0.72, 0.88, and 1.56 µL L⁻¹ H₂ in MES buffer; Supplemental Table S1) induced stomatal closure in a dose-dependent fashion, with a maximal inhibition in 100% saturated HRW (1 h; Supplemental Fig. S1). It was further observed that 100% saturated HRW, mimicking the responses of ABA (30 μm), not only time dependently induced H₂ production (Fig. 1B) but also triggered Arabidopsis stomatal closure (Columbia-0 [Col-0]; Fig. 1C). Results of the washout experiment (Supplemental Materials and Methods S1) further revealed that these inhibitions on the stomatal apertures were completely reversible (Supplemental Fig. S2). Similarly, stomatal closure was promoted by 100% saturated HRW in the Landsberg erecta (Ler) ecotype (Supplemental Fig. S3). Therefore, 100% saturated HRW was mainly used in the following study.

It was further observed that the increased concentrations of H₂ in CO₂-free MES buffer (from 25% to 100% saturation) resulted in the distinct reduction of the dissolved oxygen (Supplemental Table S1; Supplemental Materials and Methods S1). For example, compared with the CO₂-free MES buffer only (control), the concentration of oxygen dropped from about 1.32 to 0.7 µL L⁻¹ when hydrogen was saturated (100% HRW). It was reported previously that changes of oxygen could affect stomatal movement (Akita and Moss, 1973). To investigate the possibility that HRW-promoted Arabidopsis stomatal closure might be partially due to the depletion of oxygen, we subsequently mixed CO₂-free MES buffer solutions separately saturated with H₂ or oxygen in different proportions, thus maintaining the oxygen concentration of HRW (25%–75% saturation) at around 1.3 µL L⁻¹, which is approximately that of the control treatment (CO₂-free MES buffer only). Similar to the HRW-alone treatments (Supplemental Fig. S1), it was found that 25%, 50%, and 75% saturated HRW with oxygen concentration at about 1.3 µL L⁻¹ also induced stomatal closure in a dose-dependent fashion (2 h), with a maximal induction triggered by 75% saturated HRW (Supplemental Table S1). Since a treatment with 75% saturated HRW alone for 2 h also effectively promoted Arabidopsis stomatal closure (about 0.95 μm) to a similar level, in comparison with that of the 100% saturated HRW alone (about 0.92 μm), 75% saturated HRW was chosen to further verify the participation of oxygen depletion in HRW-promoted Arabidopsis stomatal closure.

Meanwhile, HRW-induced stomatal closure was partially reversed when the concentrations of oxygen were increased progressively (Supplemental Table S2). For example, in the treatments with 75% saturated HRW, the stomatal aperture was differentially increased from about 0.95 to 1.29 μm when the corresponding dissolved oxygen concentrations were raised progressively from about 0.77 µL L⁻¹ (75% HRW) to approximately 1.33 µL L⁻¹ (75% HRW + oxygen [A], 75% HRW + oxygen [B], and 75% HRW + oxygen [C]). Similar responses were also found in 25% and 50% saturated HRW-treated samples (Supplemental Table S1). Comparatively, stomatal aperture dropped from about 3.57 to 0.95 μm (change of 2.62 μm) in 75% HRW-treated samples in comparison with about 13% caused by HRW-triggered oxygen decrease (stomatal aperture increased from 0.95 to 1.29 μm; change of 0.34 μm). Taken together, these results clearly indicated that the main factor of HRW-promoted stomatal closure was the dissolved H₂.

The phenotypes of 4-week-old wild-type (Col-0) plants with or without HRW irrigation followed by drought stress were also examined. Similar to the previous reports in Arabidopsis (Xie et al., 2012) and mung bean (Vigna radiata; Zeng et al., 2013), we observed that Arabidopsis plant growth was improved by HRW irrigation (Fig. 1D). Drought tolerance assays performed by withholding watering further revealed that the HRW-irrigated plants exhibited enhanced drought tolerance (only modest wilting symptoms) in comparison with the extreme wilting symptoms in HRW-free control plants. A further rewatering test sustained for 7 d brought about a better survival rate (94.9% ± 0.4%) compared with the control plants (17.4% ± 4.5%). A similar beneficial role was observed in the changes of the rate of water loss from the detached rosette leaves (Supplemental Fig. S4; Supplemental Materials and Methods S1).

NO Synthesis Is Induced by HRW in Guard Cells

An additive effect of HRW and ABA at low concentrations on the induction of stomatal closure (Supplemental Fig. S5) suggested that these actions might share downstream signaling components such as NO, a critical molecule involved in guard cell signaling (Bright et al., 2006). To confirm this possibility, first we examined whether HRW increases the production of NO in guard cells.

In this study, the dynamic changes of intracellular NO levels in guard cells were measured directly using a fairly specific NO fluorescent probe, 4-amino-5-methyl-amino-2′,7′-difluorofluorescein diacetate (DAF-FM DA). As expected, exogenously applied HRW or ABA for 2 h simultaneously enhanced DAF-FM DA fluorescence in guard cells (Supplemental Fig. S6), reaching the maximum values after 10 min of treatment and then gradually decreasing thereafter (Fig. 2A). Over a 20-min treatment period, the basic 4-amino-5-methyl-amino-2′,7′-difluorofluorescein (DAF-FM)-dependent fluorescence intensity in the control samples was relatively constant.

Figure 2.

HRW triggers NO production and stomatal closure in Arabidopsis. Epidermal fragments of wild-type (Col-0) plants were incubated with DAF-FM DA probe in MES buffer, and changes of DAF-FM fluorescence intensity were monitored at the indicated time points (A) or 10 min (B and top of C) after treatment with HRW (100% saturation), ABA (30 μm), SNP (50 μm), NONOate (50 μm), Old SNP (50 μm), Old NONOate (50 μm), l-NAME (200 μm), tungstate (100 μm), or cPTIO (300 μm; corresponding concentrations were consistent throughout all of the following experiments presented), alone or in combination. Representative confocal images are also provided (B). Meanwhile, the stomatal apertures were measured 2 h after various treatments (bottom of C). To confirm the above results, the NO content in plant leaves upon similar treatments (C) was determined by EPR (D). Control (Con) means a treatment with MES buffer only. Confocal data are displayed as mean pixel intensities. Data are presented as means ± se of 30 guard cells. Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

In our experimental conditions, the above DAF-FM-dependent fluorescence was specifically related to the endogenous NO level, because (1) treatments with two kinds of NO-releasing compounds (sodium nitroprusside [SNP] and diethylamine NONOate [NONOate]) could significantly induce DAF-FM-associated fluorescence in the wild type (Fig. 2, B and C), whereas their degradation products (Old SNP and Old NONOate), which were regarded as the negative controls of SNP and NONOate, respectively (Xie et al., 2013), had no such inducible behaviors; (2) the HRW-triggered induction of DAF-FM fluorescence intensity was markedly abolished by the application of the NO scavenger 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt (cPTIO; Fig. 2C); and (3) the obtained electron paramagnetic resonance (EPR) data (Fig. 2D) were in line with those of DAF-FM-associated fluorescence (Fig. 2, B and C), further demonstrating that, at least in our experimental conditions, the DAF-FM-dependent fluorescence was related to endogenous NO levels in Arabidopsis guard cells. Meanwhile, compared with 75% saturated HRW treatment alone, DAF-FM fluorescence was partially reversed when the concentration of oxygen was maintained at the control level (Supplemental Table S3). Together, the above results clearly confirmed that the rapid NO synthesis induced by HRW in guard cells mainly resulted from its dissolved H₂.

NR-Mediated NO Production Is Responsible for HRW-Induced Stomatal Closure

The role and source of NO in HRW-induced stomatal closure were further examined. In contrast with the effect of a higher concentration of HRW (100% saturation) plus SNP (50 μm; Fig. 2C), coapplication of HRW (50% saturation) with SNP (20 μm) did reveal an additive effect on DAF-FM-dependent fluorescence and stomatal closure (Supplemental Fig. S7). Unlike the significant induction of stomatal closure triggered by SNP or NONOate alone at both low and high concentrations, an application of Old SNP or Old NONOate failed to influence the stomatal response, all of which were consistent with the above-mentioned changes of DAF-FM-associated fluorescence and the EPR assay (Fig. 2, B–D; Supplemental Fig. S8).

Because both NR- and l-Arg-dependent pathways are mainly responsible for plant NO synthesis (Besson-Bard et al., 2008), we examined the effects of the NR inhibitor tungstate and the mammalian NO synthase-like enzyme inhibitor N_G-nitro-l-Arg methyl ester hydrochloride (l-NAME) on HRW responses. It was found that both NO accumulation and stomatal closure stimulated by HRW were greatly impaired by l-NAME or tungstate (Fig. 2C). Most importantly, their combined treatments (l-NAME + tungstate) exhibited the maximal reversing responses, mimicking the changes triggered by cPTIO. However, l-NAME or tungstate alone partly influenced DAF-FM-associated fluorescence, but the performance of the stomatal aperture was not altered.

To further discern the contribution of NR- and l-Arg-dependent pathways to the HRW-induced responses, we further adopted a genetic approach using Arabidopsis mutants exhibiting null NR activity (nia1/2; Zhao et al., 2009) and indirectly reduced endogenous NO level (noa1; Zhao et al., 2007; Xuan et al., 2010; Van Ree et al., 2011) as well as the triple mutant nia1/2/noa1 impaired in NR- and AtNOA1-dependent NO biosynthesis (Lozano-Juste and León, 2010; Xie et al., 2013). Meanwhile, ABA responses were recorded simultaneously as positive controls.

As shown in Figure 3A, the noa1 mutant was partially impaired in stomatal closure in response to HRW compared with wild-type plants, but to a much lesser degree than nia1/2 and nia1/2/noa1 mutants. These results suggested that stomatal closure in these mutants, especially nia1/2 and nia1/2/noa1, is largely insensitive to HRW supplementation. Consistently, in comparison with the wild type, guard cells (Fig. 3, B and C; Supplemental Fig. S8) and seedling leaves (Supplemental Fig. S9) of the nia1/2 mutant were unable to sustain the accumulation of NO when ABA or HRW was used individually or simultaneously. Correspondingly, stomatal closure in the nia1/2 mutant was also severely reduced when ABA was used alone or together with HRW (Fig. 3C). The above results suggested that endogenous NO mainly produced by NR is required for HRW-induced stomatal closure. By contrast, subsequent results revealed that SNP or NONOate, applied alone or together with HRW, not only enhanced DAF-FM-associated fluorescence intensity (with high concentration in particular; Fig. 3C; Supplemental Fig. S8) but also induced stomatal closure to an approximately similar degree in nia1/2. These results indicated that the linear relationship between DAF-FM-associated fluorescence and stomatal aperture might be restricted to a relatively narrow range and that NO production above the upper threshold might not influence stomatal closure. The opposite phenomenon occurred when NO was removed by cPTIO in HRW-treated samples. All these data further confirmed that NR might be responsible for the majority of NO production and stomatal closure in response to HRW.

Figure 3.

NIA1/2 mediates HRW-induced NO production and stomatal closure. Epidermal fragments of wild-type (Col-0), noa1, nia1/2, and nia1/2/noa1 mutant plants were incubated with HRW in MES buffer, and changes of the stomatal apertures were monitored at the indicated time points, taking the stomatal aperture of each genotype before HRW treatment as 100% (A). Epidermal fragments were also incubated with DAF-FM DA probe in MES buffer, and DAF-FM fluorescence intensity was monitored 10 min after treatments with HRW, ABA, SNP, NONOate, or cPTIO, alone or in combination (top of C). Some representative confocal images of the nia1/2 mutant are also provided (B). Meanwhile, the stomatal apertures were measured 2 h after various treatments (bottom of C). Control (Con) means a treatment with MES buffer only. Corresponding confocal data are provided, taking the fluorescence level for control samples as 100%. Data are presented as means ± se of 30 guard cells. Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

ROS Participates in HRW-Triggered Stomatal Closure

To gain insight into the mechanism of HRW-induced stomatal pore changes, we analyzed whether ROS might be involved in HRW functioning in stomatal movements, in a manner similar to the above-mentioned NO behavior.

In this study, we used the oxidatively sensitive fluorophore dichlorofluorescein (DCF) to directly measure fast changes in intracellular ROS level. As shown in Figure 4A, HRW rapidly induced the relative fluorescence intensity of DCF in guard cells. For example, an increased fluorescence intensity was observed as early as 5 min, then reached the maximal level at 10 min (Fig. 4, A and C) followed by a rapid decrease until 2 h of HRW treatment (Supplemental Fig. S10). A similar but more sustained and considerable change was observed in ABA-treated samples. Therefore, exogenous hydrogen peroxide (H₂O₂) was applied to mimic the function of HRW-triggered endogenous ROS production.

Figure 4.

HRW-induced ROS production and stomatal closure. Epidermal fragments of wild-type (Col-0) plants were incubated with H₂DCF-DA probe in MES buffer, and changes of DCF fluorescence intensity were monitored at the indicated time points (A) or 10 min after treatment with or without HRW in the presence or absence of AsA (1 mm), CAT (60 units mL⁻¹), DPI (30 μm), H₂O₂ (100 μm; these concentrations were consistent throughout all of the following experiments presented), or ABA (B). Representative confocal images are also shown (C). Meanwhile, the stomatal apertures were measured 2 h after various treatments (D). Control (Con) means a treatment with MES buffer only. Corresponding confocal data are provided, taking the fluorescence level for control samples as 100%. Data are presented as means ± se of 30 guard cells. Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

It was further revealed that 100 μm H₂O₂ was able to induce the relative fluorescence intensity and, thereafter, stomatal closure regardless of the addition of HRW (Fig. 4, B–D). When HRW (50% saturation) and exogenous H₂O₂ (50 μm) were used together, additive effects on the relative fluorescence intensity and stomatal movement were observed (Supplemental Fig. S11), which were similar to the corresponding results of cotreatment with a low saturation or concentration of HRW plus ABA or SNP (Supplemental Figs. S5 and S7). Furthermore, HRW-induced DCF-related fluorescence accumulation and stomatal closure were abolished partly by ascorbic acid (AsA), catalase (CAT), or diphenylene iodonium (DPI; Fig. 4, B and D), all of which either remove H₂O₂ or reduce ROS production by the inhibition of NADPH oxidase. Although these chemicals applied alone did decrease fluorescence intensity, no significant differences in stomatal responses were observed compared with untreated samples (Fig. 4D), suggesting the complexity of the signaling transduction pathway in guard cells. It was also observed that, in comparison with 75% saturated HRW treatment alone, ROS synthesis was largely reversed, but the corresponding stomata closed effectively (Supplemental Tables S2 and S3) when the concentration of oxygen was kept at the control level. This might be due to the fact that ROS, especially H₂O₂, are likely to be generated in localized hot spots within the subcellular compartment (Neill et al., 2002; Wilkins et al., 2011), thereby leading to localized effects of H₂O₂ on its targets and subsequent stomatal closure. Together, the above results clearly suggested that HRW might promote stomatal closure, at least partially, via an ROS-dependent pathway.

RbohF-Dependent ROS Production Might Be Required for HRW-Induced Stomatal Closure

As reported previously, two NADPH oxidase isoforms expressed in guard cells, RbohD and RbohF, are responsible for the ABA-induced ROS generation and stomatal closure (Kwak et al., 2003). To further validate the involvement of specific NADPH oxidases in HRW-stimulated increases of ROS level, we analyzed the effects of HRW on the transcript abundances of RbohD and RbohF in the wild type. Time-course analysis showed that the transcription profile of the RbohF gene was rapidly induced within 15 min after HRW treatment and then fell back to the control level during 60 to 120 min of treatment (Fig. 5A). By contrast, RbohD transcripts were not obviously altered. Comparatively, the expression of RbohD (in particular) and RbohF genes was significantly increased within 15 to 60 min of ABA treatment. Furthermore, with respect to 75% saturated HRW treatment alone, a slight but not significant decrease of RbohF expression was observed when the concentration of oxygen of 75% saturated HRW solution was returned to the control level (Supplemental Table S3).

Figure 5.

Rboh-dependent ROS production is required for HRW-induced stomatal closure. A, Time-course analyses of RbohD and RbohF transcripts in leaves of wild-type (Col-0) plants in response to HRW or ABA. The gene expression levels are presented as values relative to the corresponding control samples at 0 h. B, Time-course analysis for the stomatal apertures of wild-type (Col-0), rbohD, and rbohF mutant plants in response to HRW. The stomatal aperture of each genotype before HRW treatment was regarded as 100%. C, Epidermal fragments of wild-type (Col-0) and rbohF mutant plants were incubated with H₂DCF-DA probe in MES buffer, and changes of DCF fluorescence intensity were monitored 10 min after the indicated treatments. D, Corresponding stomatal apertures were measured 2 h after various treatments. Control (Con) means a treatment with MES buffer only. Data are presented as means ± se of three independent experiments (A) or 30 guard cells (B–D). Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

Time-course changes of the stomatal apertures confirmed that the HRW-induced stomatal closure was impaired in the rbohD mutant and, in particular, the rbohF mutant (Fig. 5B). However, with respect to the wild type, both HRW and ABA applied individually or simultaneously failed to trigger the accumulation of DCF-related green fluorescence in guard cells of the rbohF mutant (Fig. 5C). Stomatal responses to HRW or ABA individually or simultaneously were significantly reduced in rbohF, while exogenous H₂O₂ triggered a similar reduction in both the wild type and rbohF (Fig. 5D). Together, these results suggested that RbohF-dependent ROS production might be, at least partly, required for HRW-stimulated stomatal closure.

HRW-Induced NO Generation and Stomatal Closure Are ROS Dependent

Our next work was to establish a casual link between NR-associated NO production and RbohF-dependent ROS in HRW-stimulated stomatal closure. As shown in Figure 6A, HRW alone failed to induce DAF-FM-associated fluorescence in the rbohF mutant, which could be remarkably reversed by H₂O₂. These results were partially confirmed by using EPR examination (Fig. 6C). Meanwhile, the promotion of stomatal closure caused by H₂O₂ and SNP in rbohF was not significantly affected by cotreatment with HRW (Fig. 6B). Based on these data, we suggested that, in our study, some linearity may exist in guard cell ROS and NO signaling downstream of H₂. Additional results showed that the administration of HRW did not significantly alter the H₂O₂-induced accumulation of DCF-related fluorescence in the nia1/2 mutant (Fig. 7A). In this scenario, H₂O₂-induced DAF-FM-associated fluorescence intensity and stomatal closure were impaired in the nia1/2 mutant, which was not influenced by the cotreatment with HRW (Fig. 7, B and C). Taken together, the data presented above further place NR-associated NO production downstream of HRW-induced Rboh-dependent ROS synthesis, all of which causally led to stomatal closure.

Figure 6.

Changes of HRW-induced NO synthesis and stomatal closure in the rbohF mutant. Epidermal fragments of rbohF mutant plants were incubated with DAF-FM DA probe in MES buffer, and changes of DAF-FM fluorescence intensity were monitored 10 min after treatment with HRW, ABA, H₂O₂, or SNP, alone or in combination (A). Meanwhile, the stomatal apertures were measured 2 h after various treatments (B). To confirm the above results, the NO content in leaves of the rbohF mutant upon similar treatments (A) was determined by EPR (C). Control (Con) means a treatment with MES buffer only. Data are presented as means ± se of 30 guard cells, taking data of the control samples as 100%. Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

Figure 7.

HRW activation of NR-dependent NO production is dependent on H₂O₂. Epidermal fragments of wild-type (Col-0) or nia1/2 mutant plants were incubated in MES buffer in the presence of HRW or H₂O₂ alone or in combination. Control (Con) means a treatment with MES buffer only. Afterward, changes of DCF (A) or DAF-FM (B) fluorescence intensity were monitored for 10 min, taking the control value of each genotype as 100%. Meanwhile, the stomatal apertures were measured 2 h after various treatments (C). Data are presented as means ± se of 30 guard cells. Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

GORK Acts as the Downstream Target of HRW Signaling

Previous studies revealed the functional role of GORK in the regulation of stomatal behavior (Hosy et al., 2003). Keeping in mind that GORK is the unique depolarization-activated K⁺ outward-rectifying channel located in guard cells, the transcript abundance of GORK was then tested. As shown in Figure 8A, GORK expression was differentially induced by HRW, ABA (in particular), SNP, and H₂O₂ in the wild type. Further results confirmed that the up-regulation of GORK transcripts was weaker or blocked in nia1/2 and rbohF mutants in comparison with those in the wild type. It was also observed that, compared with 75% saturated HRW treatment alone, GORK activation was impaired to some extent when the concentration of oxygen was maintained at the control level (Supplemental Table S3).

Figure 8.

The GORK protein participates in HRW-promoted stomatal closure. GORK transcripts (A; leaves) and the stomatal apertures (B; epidermal fragments) of wild-type (Col-0), nia1/2, rbohF, or gork plants were determined upon HRW, ABA, SNP, or H₂O₂ treatment for 2 h The gene expression levels are presented as values relative to the corresponding control samples (A). Also, the stomatal apertures of wild-type (Col-0) or gork mutant plants under control conditions were regarded as 100% (B). Control (Con) means a treatment with MES buffer only. Data are presented as means ± se of three independent experiments (A) or 30 guard cells (B). Bars with different letters are significantly different at P < 0.05 according to Duncan’s multiple comparison.

To further delineate the potential role of GORK in HRW-induced stomatal closure, a gork mutant was used (Fig. 8B). As expected, the stomatal closure of gork in response to HRW, ABA, SNP, or H₂O₂ was partially suppressed (approximately 60% of the control levels), in comparison with a much greater degree in the wild type (approximately 26%–29%). Thus, our results preliminarily indicated that H₂-mediated ROS-dependent NO production operates downstream of ABA, promoting stomatal closure partially through GORK (Fig. 9).

Figure 9.

Schematic representation of the ABA signaling pathway involving H₂, Rboh-mediated ROS biosynthesis, NR-dependent NO production, and GORK-activated K⁺ efflux during stomatal closure and, thereafter, drought tolerance. The signaling cascade at top also might be mediated by an H₂-independent pathway (?), which might have an interaction with the H₂-dependent pathway, or by ROS/NO/GORK-independent pathways. Dashed lines denote indirect or still undescribed pathways.

DISCUSSION

H₂: A Novel Gas Intermediate Participating in the Promotion of Stomatal Closure and Drought Tolerance

Some biological gaseous molecules (including NO, carbon monoxide, and hydrogen sulfide) have recently emerged as gasotransmitters in guard cell signaling (García-Mata and Lamattina, 2013). In this study, we illustrated that H₂, the simplest gas in nature, is involved in the regulation of stomatal closure, which might lead to the enhancement of drought tolerance. This conclusion was supported by the following results.

First, using an H₂-specific microelectrode system developed for in situ noninvasive H₂ measurements, we observed that ABA triggered a rapid and sustained increase of H₂ release (Fig. 1A). The results of ABA-triggered H₂ production were also verified by GC analysis (Fig. 1B). These results were consistent with a recent report (Zeng et al., 2013) showing that H₂ production, detected by a portable H₂ gas detector, was increased and maintained for 36 h in rice (Oryza sativa) seedlings treated with ABA.

Second, the biological significance of the increased endogenous H₂ was ascertained by investigating the effects of HRW on the stomatal aperture. As expected, HRW triggered Arabidopsis stomatal closure in both dose- and time-dependent manners (Fig. 1C; Supplemental Fig. S1). Subsequent experiments confirmed that the main factor of HRW-promoted stomatal closure was the dissolved H₂, although the stomatal closure could be partially due to the HRW-triggered oxygen decrease in MES buffer (Supplemental Tables S1 and S2). Meanwhile, an additive effect of HRW and ABA on the induction of stomatal closure was observed at lower concentrations (Supplemental Fig. S5).

Third, concerning the role of stomatal control in evapotranspiration under stress conditions (Aubert et al., 2010), we performed a drought tolerance assay. Compared with untreated controls, Arabidopsis plants grown with HRW irrigation exhibited a better tolerance to water deficit, as evaluated by the enhanced drought tolerance phenotype, including survival rate and water-loss assays (Fig. 1D; Supplemental Fig. S4). Interestingly, we observed that Arabidopsis plant growth was improved by HRW irrigation (Fig. 1D), further implying that not only the stomatal responses but also some developmental and probably other metabolic pathways are also affected by HRW (Zeng et al., 2013; Hu et al., 2014; Lin et al., 2014).

It has been well documented that the hydrogenase is responsible for H₂ metabolism in bacteria and green algae, which is oxygen sensitive and functions only under anaerobic conditions (Forestier et al., 2003; Morimoto et al., 2005). Along with these studies, endogenous H₂ release was found in barley (Hordeum vulgare), rice, alfalfa (Medicago sativa), and Arabidopsis (Renwick et al., 1964; Xie et al., 2012; Jin et al., 2013; Zeng et al., 2013). The increasing significance of the physiological role of H₂ was reported in higher plants, in contrast with the limited knowledge about its biosynthesis. Especially, a very recent paper showed that H₂ production was induced by several phytohormones and abiotic stresses, which was consistent with the expression of putative hydrogenase genes in rice seedlings (Zeng et al., 2013). By contrast, it was also suggested that the Arabidopsis [FeFe]-hydrogenase-like GOLLUM proteins, homologs of the hypothetical rice hydrogenase HydA1 (accession no. BAF13185.1), did not generate H₂ but are essential for plant development in normoxic conditions and the modulation of energy metabolism (Cavazza et al., 2008; Mondy et al., 2014). Therefore, it is possible that the origin of H₂ might differ among plant species. Undoubtedly, the identification of the above enzymatic or nonenzymatic sources responsible for plant H₂ generation will promote the physiological study of H₂.

At present, similar to the research in animals and clinical trials (Ohsawa et al., 2007; Oharazawa et al., 2010), the use of exogenous HRW presented a powerful research approach to investigate the biological significance of H₂. This approach was also supported by the fact that exogenous application of 100% saturated HRW (1.56 µL L⁻¹ H₂ in MES buffer) not only triggered Arabidopsis stomatal closure (Col-0; Fig. 1C) but also led to a time-dependent increase of endogenous H₂ content (Fig. 1B), mimicking the physiological responses of ABA. The existing discrepancy of H₂ concentration between ABA treatment and 100% saturated HRW treatment might be attributed to the fact that ABA-induced H₂ production was restricted to hot spots within the cell; thus, the data were relatively underestimated in whole-cell extracts. Undoubtedly, the identification of cellular and subcellular localization of H₂ production will be important for future investigations. Similarly, the existence of hot spots for hydrogen sulfide and H₂O₂ has been reported previously in plants (Neill et al., 2002; García-Mata and Lamattina, 2010; Wilkins et al., 2011).

How Do NR-Associated NO and Rboh-Dependent ROS Collaborate in Relaying H₂ Signaling?

Understanding the H₂ signal transduction pathway is key to our knowledge of complicated physiological processes in plants. Besides the presentation of the empirical evidence, two questions must be asked. How does H₂ trigger stomatal closure? And what are the downstream components of H₂ involved in stomatal closing?

In this study, genetic and pharmacological evidence revealed that NR is responsible for the major production of NO accumulation, which participated in HRW-promoted stomatal closure. First, HRW triggered a rapid increase of NO production, as evaluated by the combination of DAF-FM fluorescence and EPR analysis, and the former probe sensed NO released by two NO-releasing compounds in both wild-type and nia1/2 mutant plants (Figs. 2 and 3C). Second, HRW-triggered NO production was markedly quenched by the addition of an inhibitor of NR (tungstate) or a scavenger of NO (cPTIO); meanwhile, corresponding phenotypes of stomatal closure were reversed (Figs. 2, C and D, and 3C). Third, unlike the wild type, HRW failed to stimulate NO accumulation in the NR-null mutant nia1/2 (Fig. 3, B and C; Supplemental Fig. S8). Comparatively, HRW-induced stomatal closure was markedly impaired in the NR-null mutant (Fig. 3, A and C; Supplemental Fig. S8). However, HRW-triggered NO production also might partially result from the oxygen deprivation of the treatment solution (Dordas et al., 2003) in an H₂-independent manner (Supplemental Table S3). Meanwhile, we cannot exclude the possibility that an NR-independent NO pathway might contribute to the HRW-induced stomatal closure. It was also reported that, in the NO-deficient mutant nia1/2/noa1, independent de novo NO biosynthesis mediated ABA-promoted stomatal closure and the inhibition of opening (Lozano-Juste and León, 2010).

ABA-induced NO generation is dependent on H₂O₂ synthesis during stomatal closure and lateral root development (Bright et al., 2006; Wang et al., 2010). Furthermore, multiple enzymatic origins of ROS in guard cells have been proposed, such as NADPH oxidases (Kwak et al., 2003), cell wall peroxidases (He et al., 2013), and copper amine oxidase (An et al., 2008). In this study, potential ROS scavengers (AsA and CAT) and an inhibitor of NADPH oxidase (DPI), all of which entirely suppressed HRW-induced ROS production, markedly abolished HRW-promoted stomatal closure (Fig. 4, B and D). Great advances have been made toward deciphering the specific functions of RbohD and RbohF in various signaling processes (Marino et al., 2012). For instance, ABA-induced stomatal closing was partially impaired in atrbohF and more strongly in atrbohD/F, while the response of atrbohD to ABA was the same as that of the wild type, indicating an overlap in the functions of RbohD and RbohF (Kwak et al., 2003). AtrbohF but not AtrbohD appears essential for ethylene-induced H₂O₂ production and stomatal closure (Desikan et al., 2006). In this study, therefore, we compared the relative contributions of RbohD and RbohF to ROS-mediated H₂ signaling. It had been noted that HRW transiently up-regulated RbohF transcripts (Fig. 5A), which was similar to the changes of DCF-dependent fluorescence (Fig. 4A; Supplemental Fig. S10), but had no significant influence on RbohD expression.

Besides activation at the transcriptional level, RbohD and RbohF proteins have been reported to be synergistically activated by direct phosphatidic acid binding to the Arg residues or by interaction with protein kinases (Zhang et al., 2009; Drerup et al., 2013). Therefore, we adopted a reverse genetics approach by using an RbohD-null or RbohF-null mutant. As expected, ROS generation and stomatal closure induced by HRW were greatly impaired in rbohF, which was markedly rescued by exogenous H₂O₂, a technical control (Fig. 5, B–D). Comparatively, a less significant impairment was observed in rbohD. We further concluded that the HRW-induced ROS increase might result mainly from the activation of RbohF. Meanwhile, RbohD and RbohF may interact to mediate the responses of HRW, and it is worthwhile to investigate any corresponding cooperating mechanism in relaying H₂ signal. The results further showed that the 75% saturated HRW-induced DCF fluorescence was largely impaired but that corresponding stomata closed effectively (Supplemental Tables S2 and S3) when the oxygen concentration was maintained around 1.3 µL L⁻¹, while corresponding RbohF expression was partially impaired. This might be due to the fact that ROS, especially H₂O₂, are likely to be generated in localized hot spots within the cell subcellular compartment (Neill et al., 2002; Wilkins et al., 2011), thereby leading to localized effects of H₂O₂ on its targets and subsequent stomatal closure. These observations not only implied the biological function of HRW-triggered oxygen decrease but also suggested the existence of H₂-independent HRW-promoted stomatal closure. Interestingly, it was previously reported that the unicellular green alga Chlamydomonas reinhardtii adapted to hypoxic conditions by the production of H₂ through [FeFe]-hydrogenase isoform1 (Pape et al., 2012). Regarding this issue, at this stage we could not fully rule out the contribution of oxygen decrease to endogenous H₂ production during the process of Arabidopsis stomatal closure triggered by HRW.

Both NO and ROS have been established to play crucial roles in the stomatal signaling network in response to ABA and ethylene (Bright et al., 2006; Desikan et al., 2006). To further reveal the possible interaction between NO and ROS involved in H₂-dependent signaling, we measured intracellular NO and ROS levels in the rbohF mutant and the NR-null mutant. In this study, a linear signaling pathway of ROS and NO controls the HRW-induced stomatal closure. For example, HRW failed to trigger NO accumulation in rbohF, whereas it was rescued by H₂O₂ (Fig. 6, A and C). Compared with the wild-type plants (Figs. 2C and 5D), the stomatal closure response to HRW or ABA was observed to be reduced in rbohF, but its response to exogenous H₂O₂ was not affected (Fig. 6B). These data strongly suggested that RbohF-generated ROS is essential for HRW-stimulated NO production and stomatal closure. As a positive control treatment, SNP- or NONOate-released NO could promote stomatal closure in both the wild type and the NR-null mutant (Fig. 3C; Supplemental Fig. S8). Unlike in the wild type, H₂O₂ or HRW when individually or simultaneously applied (Fig. 7, B and C) could not promote NO production in the NR-null mutant. Meanwhile, stomatal closure was largely impaired. These data imply that NR-mediated NO production is mainly responsible for stomatal closure in response to HRW as well as H₂O₂ (Bright et al., 2006). Consistently, the NR-null mutant generates ROS to approximately the same degree as the wild type in response to HRW (Fig. 7A). Altogether, the above-mentioned data provided genetic evidence that NR is localized downstream of HRW-induced ROS production, which was partially generated by RbohF, leading to NO production and, thereafter, stomatal closure. Certainly, it is worth determining whether NIA1 or NIA2 mediates HRW-activated NO generation in guard cells, and the protein-protein interaction of NR proteins (Kanamaru et al., 1999; Lambeck et al., 2010) might be important for the H₂-dependent signaling cascade.

In mammalian cells, the protective role of H₂ against oxidative damage has mainly been attributed to selectively scavenging activity for hydroxyl radicals, which are the most cytotoxic ROS. H₂ also reduced ONOO⁻ somewhat but did not affect H₂O₂, NO, and O₂⁻ (Ohsawa et al., 2007). In this report, however, we found that the HRW stimulated RbohF-dependent ROS synthesis, thereby activating NR-associated NO production, both of which causally resulted in stomatal closure (Figs. 2–7). These discrepancies may indicate the complexities of H₂ signaling in animals and plants. It was well known that both NO and ROS function as signaling molecules in higher plants, besides their reactive nature leading to phytotoxicity (García-Mata and Lamattina, 2013). Taking into account the fact that the levels of ROS and NO were tightly controlled by an intricate gene network, our results revealed a more complicated function and physiological significance of H₂ signaling in plants, which regulates ROS/NO-related metabolism, resulting in the modulation of stomatal closure.

An efflux of K⁺ across the plasma membrane to the extracellular space, mediated by GORK, directly causes a decrease in guard cell turgor and stomatal closing (Pandey et al., 2007). The knockout mutation of GORK resulted in the elimination of outwardly rectifying K⁺ currents and impaired ABA-induced stomatal closure (Hosy et al., 2003). In this study, GORK transcription in the wild type was up-regulated upon the onset of HRW, SNP, and H₂O₂ treatment, while it was impaired in nia1/2 and rbohF mutants (Fig. 8A), indicating that GORK appears to be induced by upstream components involving ROS and NO in H₂ signaling. It was interesting that the relative reductions of the stomatal aperture by HRW in the wild type and the gork mutant were 72% and 34%, respectively (Fig. 8B), indicating that the mutation of GORK could not fully diminish HRW-induced stomatal closure. Therefore, it is important to note that, although linearity in H₂, ROS, and NO signaling has been observed, it is likely that H₂-dependent guard cell signaling also remains divergent from ROS, NO, and GORK signaling (Fig. 9). Identification of these possibilities will be part of our investigations to gain a full understanding of H₂ signaling in plants.

A recent article reported that, similar to the response of auxin, HRW induced cucumber (Cucumis sativus) adventitious root development in a heme oxygenase1/carbon monoxide-dependent manner (Lin et al., 2014). Here, the results further support the idea that H₂ might also act as a hormone-like compound involved in different biological processes, which were similar to some hormone-like behaviors reported for carbon monoxide and NO in plants. Overall, the results of this study demonstrate the main branch of the H₂ signaling cascade in ABA-induced stomatal closure, which is causally attributed to the NR-associated NO and RbohF-dependent ROS production, followed by GORK activation (Fig. 9). The stomatal closure could also be partially due to an HRW-triggered oxygen decrease in the CO₂-free MES buffer (H₂-independent pathway) or the ROS/NO/GORK-independent pathways (Fig. 9). Furthermore, the possibility of oxygen deprivation resulting in the induction of endogenous H₂ production and its signaling could not easily be ruled out (Pape et al., 2012). This work not only identifies the previously unknown functions of H₂ in plants but also raises the possibility of the potential field utilization of HRW for sustainable agricultural development by improving crop stress tolerance.

MATERIALS AND METHODS

Chemicals

Unless stated otherwise, all chemicals were obtained from Sigma-Aldrich. The chemicals used for treatments were ABA, SNP (a well-known NO-releasing compound), NONOate (another well-known NO-releasing compound), l-NAME (a mammalian NO synthase-like enzyme inhibitor), tungstate (an inhibitor of NR), cPTIO (a scavenger of NO), and DPI (an inhibitor of NADPH oxidase; Bright et al., 2006; Ederli et al., 2006; Zhao et al., 2009; Chen et al., 2010; Desmard et al., 2012). H₂O₂, AsA (a scavenger of H₂O₂), and CAT (an H₂O₂-scavenging enzyme) were purchased from Shanghai Medical Instrument. In this study, the concentrations of these compounds were determined by pilot experiments from which significant responses were observed. Meanwhile, the Old SNP/Old NONOate solutions were used as negative controls, as described previously (Tossi et al., 2009; Xie et al., 2013).

Preparation of HRW

Purified H₂ (99.99% [v/v]) generated from an H₂ generator (SHC-300; Saikesaisi Hydrogen Energy) was bubbled into 100 mL of CO₂-free MES-KCl buffer solution (10 mm MES and 50 mm KCl, pH 6.15) or distilled water at a rate of 150 mL min⁻¹ for 30 min. Then, the corresponding HRW was rapidly diluted with CO₂-free MES-KCl buffer to the required saturations (25%, 50%, and 75% [v/v]; pH 6.13–6.15, stable during the 2-h experimental period). In our experimental conditions, the H₂ concentration in freshly prepared HRW (100% saturation) analyzed by GC was 781 μm (about 1.56 µL L⁻¹).

Plant Material and Growth Conditions

Arabidopsis (Arabidopsis thaliana) noa1 (CS6511, Col-0; Xie et al., 2013), nia1/2 (CS2356, Col-0; Xie et al., 2013), and gork (SALK_092448, Col-0; Jayakannan et al., 2013) mutants were purchased from the Arabidopsis Biological Resource Center (http://www.arabidopsis.org/abrc), and the nia1/2/noa1 mutant was generated previously (Xie et al., 2013). The rboh homozygous mutants were provided by W.H. Zhang (Department of Plant Science, Nanjing Agricultural University) and S.J. Neill (Center for Research in Plant Science, University of the West of England). The Ler ecotype seeds were provided by Dr. Chuanyou Li (Institute of Genetics and Developmental Biology, Chinese Academy of Sciences). Four-week-old plants were cultured in a growth chamber with a 16/8-h light/dark cycle (23°C/18°C) and 120 μmol m⁻² s⁻¹ irradiation.

Measurement of H₂ Content

The H₂ release was measured using a needle-type Hydrogen Sensor (Unisense) according to the manufacturer’s instructions. The H₂-specific electrode had a tip with a diameter of 50 μm and was polarized for 4 h prior to use. First, Arabidopsis leaves were placed on 1.5% agarose gel pieces to avoid damage to the electrode tip. When performing electrode analysis, this needle punctured leaf tissues (about 200 μm below the leaf surface; the tip was able to move up and down accurately in 50-μm steps as controlled by a micromanipulator equipped with a binocular microscope). After the basal line of H₂ signal was stable, treatment solution was added to immerse the electronic tip. Corresponding data were then recorded. A treatment with MES buffer only was regarded as a control. As a result, the released H₂, which is supposed to be dissolved in extracellular fluid of Arabidopsis leaf tissues, was measured. A standard solution of HRW was prepared by saturating H₂ gas in distilled water (781 μm at 25°C) at atmospheric pressure, while distilled water was used as a negative control. All manipulations were performed at 25°C ± 1°C.

According to our previous method (Jin et al., 2013) with minor modifications, endogenous H₂ production was measured by GC (Agilent 7890A device equipped with a thermal conductivity detector). Approximately 0.3 g of Arabidopsis leaves was homogenized with 7 mL of distilled water for 1 min and then placed in a vial, followed by the addition of 5 μL of octanol and 0.5 mL of sulfuric acid (5 m). Afterward, pure carbon dioxide was bubbled into the vial to fully displace the air. After it was capped and shaken vigorously for 1 min, the vial was heated at 70°C for 1 h to liberate H₂ from plant tissues and allowed to cool at room temperature before the headspace was analyzed.

Stomatal Bioassay

To examine stomatal movement, epidermal fragments were obtained from 4-week-old plants and incubated in CO₂-free MES-KCl buffer (10 mm MES and 50 mm KCl, pH 6.15) for 2.5 h under light conditions (100 μmol m⁻² s⁻¹) in a growth chamber. Afterward, epidermal fragments were immersed in the presence of various compounds, alone or in combination, with the indicated concentrations as indicated in the figure legends. At the indicated time points, the stomatal apertures were captured using a light microscope equipped with an imaging camera (model Stemi 2000-C; Carl Zeiss) and analyzed with ImageJ software (supplied by the National Center for Biotechnology Information [http://rsb.info.nih.gov/ij]) to obtain the stomatal aperture. Within each time point or treatment, 30 stomata were randomly selected and recorded.

Survival Assays

The survival rates of 4-week-old plants were measured according to the method described by Lozano-Juste and León (2010) and repeated at least three times. Representative photographs were taken.

Confocal Laser Scanning Microscopy

Endogenous NO and ROS production were monitored by the fairly specific NO fluorescent probe DAF-FM DA (He et al., 2013; Xie et al., 2013) and 2′,7′-dichlorofluorescein diacetate (H₂DCF-DA; Bright et al., 2006; Xie et al., 2011), respectively.

Epidermal fragments isolated from 4-week-old plants were preloaded with 10 μm DAF-FM DA or 50 μm H₂DCF-DA for 30 min before washing in MES buffer three times for 5 min each. Afterward, fragments were treated with various reagents, alone or in combination, as indicated in the figure legends. Within each sampling time, 30 guard cells in three independent epidermal fragments were observed using a TCS-SP2 confocal laser scanning microscope (excitation at 488 nm, emission at 500–530 nm; 25°C ± 1°C). To measure the relative fluorescence intensity of guard cells, acquired images were then analyzed via region of interest analysis, provided by the Leica software (Sieberer et al., 2009; Liesche and Schulz, 2012). Data were calculated as means ± se of pixel intensities.

NO Detection by EPR

Endogenous NO production was also quantified by using EPR as described previously (Huang et al., 2004; Sun et al., 2012; Xie et al., 2013). EPR was carried out on a Bruker A300 spectrometer (Bruker Instrument) under the following conditions: modulation frequency, 100 kHz; microwave frequency, 9.854 GHz; modulation amplitude 4 G; microwave power, 63.496 mW.

Real-Time Reverse Transcription-PCR Analysis

Real-time reverse transcription-PCR was performed using the Mastercycler ep realplex real-time PCR system (Eppendorf) with SYBR Premix Ex Taq (TaKaRa Bio). Using specific primers (Supplemental Table S4), relative expression levels of genes are presented as values relative to the corresponding control samples at the indicated times or under the indicated conditions, after normalization to actin2/7 (accession no. NM_121018) transcript levels.

Statistical Analysis

Data are means ± se from at least three independent experiments or 30 guard cells. For statistical analysis, Duncan’s multiple range test (P < 0.05) was chosen.

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers RbohD (At5g47910), RbohF (At1g64060), Gork (At5g37500), and Actin2/7 (NM_121018).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Dose dependence of HRW (H₂ saturated in MES buffer)-promoted stomatal closure.

• Supplemental Figure S2. Promotion of stomatal closure by ABA and HRW, and its reversibility.

• Supplemental Figure S3. Stomatal apertures of the wild type (Col-0 and Ler) in response to ABA and HRW.

• Supplemental Figure S4. Water-loss assays.

• Supplemental Figure S5. Additive response of HRW and ABA at low concentrations.

• Supplemental Figure S6. Time-course analysis of DAF-FM fluorescence in guard cells of the wild type in response to HRW and ABA.

• Supplemental Figure S7.Additive response of HRW and SNP at low concentrations.

• Supplemental Figure S8. Effects of 100% HRW, 100 μm SNP, and 100 μm NONOate, alone or in combination, on DAF-FM fluorescence and stomatal closure.

• Supplemental Figure S9. Determination of NO contents in the nia1/2 mutant by EPR.

• Supplemental Figure S10. Time-course analyses of relative DCF fluorescence in guard cells of the wild type in response to HRW and ABA.

• Supplemental Figure S11. Additive response of HRW and H₂O₂ at low concentrations.

• Supplemental Table S1. Changes of stomatal apertures of Arabidopsis.

• Supplemental Table S2. Changes of stomatal apertures of Arabidopsis.

• Supplemental Table S3. Relative changes of DAF-DM and DCF fluorescence, as well as RbohF and GORK expression of Arabidopsis.

• Supplemental Table S4. Sequences of PCR primers for real-time reverse transcription-PCR.

• Supplemental Materials and Methods S1. Washout experiment, water loss, and measurement of H₂ and O₂ contents.

Glossary

ABA

abscisic acid

ROS

reactive oxygen species

NO

nitric oxide

NR

nitrite reductase

H₂

hydrogen gas

HRW

hydrogen-rich water

GC

gas chromatography

Col-0

Columbia-0

Ler

Landsberg erecta

DAF-FM DA

4-amino-5-methyl-amino-2′,7′-difluorofluorescein diacetate

DAF-FM

4-amino-5-methyl-amino-2′,7′-difluorofluorescein

SNP

sodium nitroprusside

cPTIO

2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt

EPR

electron paramagnetic resonance

NONOate

diethylamine NONOate

Old SNP

degradation product of sodium nitroprusside

Old NONOate

degradation product of diethylamine NONOate

l-NAME

N_G-nitro-l-Arg methyl ester hydrochloride

H₂O₂

hydrogen peroxide

DCF

dichlorofluorescein

AsA

ascorbic acid

CAT

catalase

DPI

diphenylene iodonium

H₂DCF-DA

2′,7′-dichlorofluorescein diacetate