- © 2020 American Society of Plant Biologists. All rights reserved.
Abstract
Hydrogen sulfide (H2S) is a gaseous signaling molecule that regulates diverse cellular signaling pathways through persulfidation, which involves the post-translational modification of specific Cys residues to form persulfides. However, the mechanisms that underlie this important redox-based modification remain poorly understood in higher plants. We have, therefore, analyzed how protein persulfidation acts as a specific and reversible signaling mechanism during the abscisic acid (ABA) response in Arabidopsis (Arabidopsis thaliana). Here we show that ABA stimulates the persulfidation of l-CYSTEINE DESULFHYDRASE1, an important endogenous H2S enzyme, at Cys44 and Cys205 in a redox-dependent manner. Moreover, sustainable H2S accumulation drives persulfidation of the NADPH oxidase RESPIRATORY BURST OXIDASE HOMOLOG PROTEIN D (RBOHD) at Cys825 and Cys890, enhancing its ability to produce reactive oxygen species. Physiologically, s-persulfidation-induced RBOHD activity is relevant to ABA-induced stomatal closure. Together, these processes form a negative feedback loop that fine-tunes guard cell redox homeostasis and ABA signaling. These findings not only expand our current knowledge of H2S function in the context of guard cell ABA signaling, but also demonstrate the presence of a rapid signal integration mechanism involving specific and reversible redox-based post-translational modifications that occur in response to changing environmental conditions.
INTRODUCTION
Through evolution and diversification, land plants have adopted robust mechanisms to perceive and relay signals that convey information about water stress. The stomatal aperture in the epidermis of plant leaves is formed by a pair of specialized guard cells, which can sense multiple stimuli and, consequently, control gas exchange and transpirational water loss, which is critical for plant growth and sustenance in stressful and ever-changing environments (Blatt, 2000; Kim et al., 2010). Importantly, under drought stress, the hyperosmotic signal causes the accumulation of the phytohormone abscisic acid (ABA), which is regulated transcriptionally and post-translationally and integrates into a complex signaling network that regulates stomatal aperture (Zhu, 2016).
Reactive oxygen species (ROS) act as second messengers in guard cell responses to most of the stimuli that induce stomatal closure (Wang and Song, 2008). NADPH oxidase (respiratory burst oxidase homolog, RBOH) is a homolog of a mammalian 91-kD glycoprotein subunit of phagocyte oxidase (gp91phox), and a key enzyme in the production of ROS from the outer leaflet of the plasma membrane (Anderson et al., 2011; Marino et al., 2012). Among the 10 RBOH genes found in the Arabidopsis (Arabidopsis thaliana) genome, the RBOH D and F isoforms play pivotal roles in ROS production, which is rate-limiting to guard cell ABA signal transduction (Kwak et al., 2003). Similarly to mammalian RBOHs, these plant polytopic membrane proteins possess a core C-terminal region containing the functional NADPH-oxidizing dehydrogenase domain responsible for superoxide generation (Suzuki et al., 2011). This C terminus also functions as a toggle switch that affects the access of the NADPH substrate to the enzyme (Magnani et al., 2017). Additionally, plant RBOHs contain an N-terminal intracellular region that carries two EF-hand motifs and a number of phosphorylation sites, which can be activated via Ca2+ binding and protein phosphorylation events that occur individually or synergistically and involve protein kinases (Marino et al., 2012). Nevertheless, mutagenesis experiments have indicated that Ser343 and Ser347 residues are necessary but not sufficient for elicitor-induced full activation of AtRBOHD (Nühse et al., 2007). Apart from these canonical post-translational modifications of the N-terminal, the Cys890 amino acid residue in the RBOHD C-terminal is susceptible to modification by nitric oxide, to form a Cys s-nitrosylation, which suppresses ROS production during the defense response to infection by Pseudomonas syringae (Yun et al., 2011). Despite a long-established correlation between ABA-induced stomatal closure and a ROS burst, through the activation of RBOHD/F, the mechanisms underpinning this activation remain unclear.
Over recent years, tremendous progress has been made in unveiling the biological relevance of hydrogen sulfide (H2S), a small gaseous molecule that functions as an important developmental and stress-responsive signal in prokaryotes and eukaryotes (Lai et al., 2014; Guo et al., 2016, 2017). H2S regulates many physiological processes throughout the plant life cycle, including seed dormancy and germination, root growth, cell senescence, autophagy, stomatal aperture/closure, and immunity (Xie et al., 2013, 2014; Aroca et al., 2018; Corpas et al., 2019). H2S signaling has been implicated in plant stress responses to high salinity, drought, heavy metals, high temperature, osmotic stress, and oxidative stress (Gotor et al., 2019). A considerable number of reports highlight the importance of H2S and the pathways to its production in plants (Xie et al., 2013; Guo et al., 2016; Gotor et al., 2019; Shen et al., 2019). Although H2S production occurs predominantly via the photosynthetic sulfate-assimilation pathway in chloroplasts, most chloroplastic sulfide dissociates to its ionic form, HS−, as the pH is basic and H2S is unable to cross the chloroplast membrane. Therefore, the largest proportion of endogenous cytosolic H2S is generated from l-cysteine by cysteine-degrading enzymes (Gotor et al., 2019), of which l-cysteine desulfhydrase1 (DES1) is the first and most characterized (Álvarez et al., 2010). Recently, a number of studies have reported that H2S produced by DES1 is an important player in guard cell ABA signaling and plant drought tolerance (García-Mata and Lamattina, 2010; Jin et al., 2013; Du et al., 2019). In wheat (Triticum aestivum), ABA biosynthesis and signaling are activated by H2S upon drought stress (Ma et al., 2016). DES1 activity is required for ABA-dependent stomatal closure, and the enzyme acts early in ABA-mediated signal transduction (Scuffi et al., 2014; Du et al., 2019; Zhang et al., 2019). The induction of stomatal closure by sulfurating molecules is impaired in rbohD and rbohF mutants, indicating that NADPH oxidase acts downstream of H2S in ABA-induced stomatal closure (Scuffi et al., 2018). However, the biochemical and molecular mechanisms by which H2S regulates downstream targets involved in guard cell ABA signaling have been elusive.
Signaling by H2S is proposed to occur via persulfidation—the post-translational modification of protein Cys residues (R-SHs) by covalent addition of thiol groups to form persulfides (R-SSHs; Aroca et al., 2018). Similar to but more widespread than s-nitrosylation (Hancock, 2019), protein persulfidation is a redox-based modification that regulates diverse physiological and pathological processes. This action provides the framework on which to build an understanding of the physiological effects of H2S (Paul and Snyder, 2012; Filipovic and Jovanović, 2017). The covalent modification that occurs through persulfidation can be reversed by reducing agents such as DTT. Persulfidation modulates protein activities by a range of mechanisms, including alterations to subcellular localization, biochemical activity, protein–protein interactions, conformation, and stability (Aroca et al., 2017b; Filipovic et al., 2018). As a typical example of the biological relevance of persulfide modification, increased expression of H2S-producing enzymes and concomitant H2S production induce persulfidation of Cys38 in the p65 subunit of NF-κB, which enhances the binding of NF-κB subunits to the co-activator ribosomal protein S3. The activator complex then migrates to the nucleus, where it upregulates the expression of several anti-apoptotic genes (Sen et al., 2012).
In Arabidopsis, a number of persulfidated proteins involved in a variety of biological pathways have been functionally characterized (Aroca et al., 2015, 2017a, 2018). For instance, H2S-triggered persulfidation disturbs actin polymerization, resulting in stunted root hair growth (Li et al., 2018). Persulfidation regulates the activities of key enzymes involved in the maintenance of ROS homeostasis and redox balance, including ascorbate peroxidase1 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) isoform C1 (GAPC1). The nuclear localization of GAPC1 was found to be modulated by DES1-produced H2S (Aroca et al., 2015, 2017b). Therefore, it is reasonable to infer that the intracellular dynamic processes of persulfidation and persulfidation oxidation may be modulated by the redox state in plant cells. The spatio-temporal coordination of H2S and ROS production is critical to the initiation, amplification, propagation, and containment of H2S/persulfidation signaling.
In this study, we report the fine-tuned regulation of guard cell redox homeostasis and ABA signaling through persulfidation. In the presence of ABA, DES1 itself was activated by H2S through persulfidation at Cys44 and Cys205, which led to the transient overproduction of H2S in guard cells. This could facilitate the overaccumulation of ROS by persulfidation of NADPH oxidase RBOHD on Cys825 and Cys890 residues, thereby inducing stomatal closure. The overaccumulated endogenous ROS may prevent continuous activation of ABA signaling in guard cells, which was achieved by a negative feedback mechanism through persulfide-oxidation of DES1 and RBOHD.
RESULTS
ABA Triggers Stimulation of Activity and Persulfidation of DES1
DES1 is a component of the ABA signaling pathway in guard cells and responsible for intracellular H2S levels and proteome-wide persulfidation (Scuffi et al., 2014; Aroca et al., 2017b, 2018). Proteomic analysis of persulfidated proteins in Arabidopsis leaves showed that DES1 is susceptible to modification by persulfidation (Aroca et al., 2017a). We hypothesize that the activity of DES1 might be regulated by H2S through persulfidation. To test this possibility, we measured the activities of DES1 recombinant protein treated with either NaHS or DTT followed by dialysis. As expected, NaHS stimulated the enzymatic activity of DES1, indicating that the effects of H2S on DES activity may rely on the thiol-based redox status (Figure 1A). Also, DES1 activity was decreased significantly by DTT treatment. To determine whether DES1 is persulfidated in the presence of its product, H2S, we used a tag-switch assay in which persulfidated Cys was labeled with cyan-biotin and specifically detected by anti-biotin immune-blot analysis (Supplemental Figure 1; Aroca et al., 2017a; Filipovic et al., 2018). NaHS induced persulfidation of DES1 recombinant protein. This was then fully abolished by DTT, which is consistent with the occurrence of a reversible thiol modification (Figure 1A).
Regulation of DES1 Protein by Persulfidation
(A) Sulfurating molecule-induced persulfidation and activity of the DES1 recombinant protein. Purified DES1 recombinant protein was treated with sulfurating molecule NaHS (1 mM) for 1 h, dialyzed and then divided into two aliquots; one was used to measure the DES enzyme activity and the other was then subjected to a tag-switch assay for the analysis of the protein persulfidation (DES1-SSH). Treatment with DTT (1 mM) is shown as negative control. The DES activity data represent mean ± sd (n = 3) from three biological replicates. Different letters indicate significantly different at P < 0.05 according to one-way ANOVA (post-hoc Tukey’s HSD test).
(B) Analysis of persulfidation of DES1-GFP in guard cells. Total proteins were extracted from pMYB60:DES1-GFP des1 transgenic plants after exposure to different concentrations of NaHS or DTT (1 mM) for 1 h, and subjected to the tag-switch assay.
(C) ABA-induced persulfidation of DES1 in guard cells. Total proteins extracted from pMYB60:DES1-GFP des1 transgenic plants were subjected to the tag-switch assay after exposure to ABA treatment (10 μM) for the indicated times.
For in vitro assay, the persulfidated recombinant protein was detected by an anti-biotin antibody after tag-switch labeling. Con, control; CBB, Coomassie Brilliant Blue protein stain; IB, immunoblotting. For in vivo assay, 4-week–old plants grown in the soil were treated and harvested at the indicated times. The persulfidated DES1-GFP protein in guard cells was detected by an anti-GFP antibody after tag-switch labeling and streptavidin purification. The bottom representations show that the total DES1-GFP used for the tag-switch assay. The numbers above the bottom representations indicate the relative abundance of the corresponding persulfidated protein compared with that of the control sample (1.00). Signals from two independent experiments were quantified.
To determine whether DES1 could be persulfidated in guard cells, we generated a GFP-tagged DES1 transgenic line pMYB60:DES1-GFP des1, in which DES1 was expressed specifically in mature guard cells under the control of the guard cell-specific promoter of the MYB60 transcription factor (Supplemental Figure 2; Chater et al., 2015; Zhang et al., 2019). Transgenic Arabidopsis plants were treated with a wide range of sulfurating molecules, slow-release H2S donors, or DTT. Subsequently, endogenous persulfidated proteins were labeled with cyano-biotin, and purified with streptavidin beads, which were then immunoblotted with an anti-GFP antibody. DES1-GFP expression was detected with an anti-GFP antibody. Moreover, the labeled enzyme was found to be persulfidated in vivo, and further strengthened or fully abolished by a wide range of sulfurating molecules or slow-release H2S donor concentrations or DTT, respectively (Figure 1B; Supplemental Figure 3). However, there was a low but detectable persulfidation signal found in the GFP itself (Supplemental Figure 4). Next, we investigated whether DES1 was persulfidated during guard cell ABA signaling. Time course results revealed that DES1-SSH formation was promoted in the early stage of the ABA response, i.e., 10 min after initiation of the response, in pMYB60:DES1 des1 plants, indicating the biological relevance of DES1 persulfidation in guard cell ABA signaling (Figure 1C).
Redox-based Regulation of Persulfidation and Enzymatic Activity of DES1
Apart from the alterations of enzyme activity, another feature of the nucleophilicity of persulfides is that they react strongly with two-electron oxidants such as H2O2 (Filipovic et al., 2018). Thus, persulfidated proteins undergo oxidation reactions to form perthiosulfenic acids in the presence of excess oxidant (Gotor et al., 2019), thereby reducing the effective level of persulfidated proteins. To determine whether DES1 persulfidated residues can be oxidized in the presence of excess ROS, we analyzed the persulfidated levels of DES1 recombinant protein after treatment with H2O2 at a wide range of concentrations. The results demonstrated that H2O2 led to a dose-dependent decrease in persulfidation of DES1 recombinant protein in vitro (Figure 2A) and DES1-GFP protein in guard cells (Figure 2B), concurrently with a dose-dependent decrease in DES1 recombinant protein activity. Time-dependent oxidation of persulfidated DES1 recombinant protein at both low and high concentrations of H2O2 was also observed (Figure 2C; Supplemental Figure 5). The DES1-SSH protein level in guard cells increased by 12.9-fold in the presence of diphenyleneiodonium (DPI; Figure 2D), an inhibitor of plant flavoenzymes, including the ROS-producing enzyme NADPH oxidase, implying that persulfidated DES1 might be oxidized by ROS produced from NADPH oxidase. To further verify this possibility, the persulfidation levels of guard cell DES1 proteins were measured in the background of the RBOHD mutation, which has been proved to be an important source of ROS in guard cell signaling (Kwak et al., 2003). Under normal conditions, the level of persulfidated DES1 in guard cells was higher in the rbohD mutant background than in the parental line pMYB60:DES1 des1. Both lines were further enhanced by ABA treatment, indicating that the persulfidation of DES1 was negatively regulated by RBOHD-derived ROS in ABA signaling (Figure 2E). It is plausible that the DES1 on-off switch, controlled by persulfidation and persulfide-oxidation, is reversible. Our time-course experiments revealed that the enhanced persulfidation of DES1 recombinant protein by NaHS pretreatment was diminished by the application of H2O2, and vice versa (Figures 2F and 2G; Supplemental Figure 6).
Regulation of DES1 Protein by Persulfide-Oxidation
(A) and (C) H2O2-triggered oxidation and downregulation of activity of the persulfidated DES1-His recombinant protein in vitro. Persulfidated DES1-His recombinant protein was treated with H2O2 at a wide range of concentrations for 1 h (A) or increasing time period (10 μM of H2O2 for C), and then divided into two aliquots; one was used to measure the DES enzyme activity and the other was subjected to a tag-switch assay for the analysis of the level of persulfidation of DES1 (DES1-SSH). The DES1 activity data represent mean ± sd (n = 3) from three biological replicates. Different letters indicate significantly different at P < 0.05 according to one-way ANOVA (post-hoc Tukey’s HSD test).
(B) H2O2-triggered oxidation of persulfidated DES1 in guard cells. Total proteins extracted from pMYB60:DES1-GFP des1 transgenic plants were subjected to the tag-switch assay after exposure to H2O2 at different concentrations for 1 h. Two images of the same blot with different times of exposure are shown.
(D) Enhancement of ABA-induced DES1 persulfidation by DPI in guard cells. Total proteins extracted from pMYB60:DES1-GFP des1 transgenic plants were subjected to the tag-switch assay after exposure to ABA at 10 μM for 30 min, with or without pretreatment of DPI at 10 μM for 2 h.
(E) ABA- and H2O2-regulated persulfidation of DES1 in guard cells. Total proteins extracted from pMYB60:DES1-GFP des1 and pMYB60:DES1-GFP des1 rbohD transgenic plants were subjected to the tag-switch assay after exposure to ABA treatment at 10 μM for 30 min.
(F) and (G) Reversible persulfidation and persulfide-oxidation of DES1-His recombinant protein. Purified DES1 recombinant protein was NaHS-pretreated at 10 μM for 1 h, followed by dialysis and then treated with H2O2 at 10 μM for the indicated times (F), or H2O2-pretreated at 10 μM for 1 h, followed by dialysis and treated with NaHS at 10 μM for the indicated times (G), and then subjected to a tag-switch assay for the analysis of the persulfidation level (DES1-SSH).
For in vitro assay, the persulfidated recombinant protein was detected by an anti-biotin antibody after the tag-switch labeling. CBB, Coomassie Brilliant Blue protein stain; IB, immunoblotting. For in vivo assay, 4-week–old plants grown in the soil were treated and harvested at the indicated times. The persulfidated DES1-GFP protein in guard cells was detected by an anti-GFP antibody after tag-switch labeling and streptavidin purification. The bottom representations show that the total DES1-GFP used for the tag-switch assay as loading control. The numbers above the bottom panel indicate the relative abundance of the corresponding persulfidated protein compared with that of the control sample (1.00). Signals from two independent experiments were quantified.
Persulfidation of DES1 at Cys44 and Cys205 Is a Prerequisite for ABA-stimulated Activity and Stomatal Closure
Three Cys residues in DES1 are putative targets for persulfidation (Cys30, Cys44, and Cys205; Supplemental Figure 7). Mass spectrometric analysis further unequivocally confirmed that C44 and C205 underwent persulfidation in purified recombinant DES1 protein upon NaHS treatment (Figures 3A and 3B). To determine which Cys would affect H2S-mediated persulfidation and the stimulation of DES1 activity, we mutated each Cys to Ala separately or simultaneously. As shown in Figures 3C and 3D, two of the mutant proteins, the Cys44Ala and Cys205Ala mutations, significantly lost the inducing effect of NaHS on persulfidation of DES1 recombinant protein, with Cys205Ala presenting dominant function. Meanwhile, the Cys44Ala mutation decreased the level of DES1 recombinant protein persulfidation, which could be partially rescued by treatment with NaHS, but eliminated by the addition of DTT (Supplemental Figure 8). In contrast, after the substitution of Cys205 with Ala, virtually no persulfidation of DES1 recombinant protein was detected, even after NaHS treatment. Subsequently, we prepared protoplasts from plants transiently expressing the different DES1 protein versions fused to GFP and examined the level of persulfidation of the DES1 mutant proteins in vivo (Figure 3E). While DES1-GFP was found to be persulfidated in protoplasts, the amount of persulfidated protein (DES1-GFP-SSH) was reduced partially in protoplasts expressing the mutated version of Cys40Ala and was totally undetectable in Cys205Ala and Cys30/44/205Ala mutants, which was in agreement with the results from the in vitro observations of mutant DES1 recombinant protein. Moreover, the formation of DES1-GFP-SSH could be enhanced by ABA or NaHS treatment of wild-type plants (Figure 3F).
Persulfidation of DES1 on Cys44 and Cys205 Regulates ABA-Induced Stomatal Closure
(A) and (B) Fully annotated MS/MS spectra of IAM-derivatized DES1 peptides containing persulfided cysteines. Purified recombinant DES1 protein was treated with NaHS (1 mM for 1 h), alkylated with IAM, and analyzed by LC-MS/MS. The localization of IAM-derivatized persulfides was unequivocally annotated according to the modification-specific b- or y-ions .
(C) and (D) In vitro persulfidation analysis of DES1 and mutant recombinant proteins by the tag-switch method. The persulfidated protein was detected by an anti-biotin antibody.
(E) In vivo persulfidation analysis of the mutated versions of DES1 protein. Total proteins from Arabidopsis protoplasts transiently expressed DES1 and its mutant derivatives fused to GFP were subjected to the tag-switch assay for the analysis of the level of persulfidation of different versions of DES1(DES1-SSH). The bottom representation shows the total DES1-GFP used for the tag-switch assay as loading control.
(F) Effect of ABA and NaHS on the persulfidation level of DES1 in vivo. Total proteins from Arabidopsis protoplasts transiently expressed unmutated DES1 fused to GFP were treated with or without ABA (10 μM for 30 min) or NaHS (1 mM for 1 h) before being subjected to the tag-switch assay for the analysis of the level of persulfidation (DES1-SSH). The bottom representation shows the total DES1-GFP used for the tag-switch assay as loading control.
(G) DES enzyme activity of different versions of the DES1-His recombinant proteins in the presence or absence of NaHS or H2O2. Recombinant proteins were pretreated with either NaHS (10 μM) or H2O2 (10 μM) for 1 h. Con, untreated conditions. The data represent mean ± sd (n = 3) from three biological replicates. Bars denoted by the different letters were different significantly at P < 0.05 according to one-way ANOVA (post-hoc Tukey’s HSD test).
(H) Reversibility of the enzyme activity of DES1-His recombinant protein (wild type) and its mutant derivative Cys44/205Ala. Recombinant proteins were pretreated with either NaHS (10 μM) or DTT (1 mM) for 1 h, followed by dialysis and then treated with the H2O2 (10 μM) at the indicated times. The data represent mean ± sd (n = 3) from three biological replicates. Different letters indicate significantly different values at P < 0.05 according to one-way ANOVA (post-hoc Tukey’s HSD test).
(I) and (J) Effects of ABA and NaHS on the guard cell H2S production and stomatal closure of wild type, the des1, and the des1 with different mutated varieties. H2S production was monitored using the fluorescent dye 7-azido-4-methylcoumarin. Epidermal strips of each line were preincubated for 3 h in opening buffer (10 mM of MES at pH 6.15, and 10 mM of KCl) under light (120 µE m−2 s−1) and loaded with 7-azido-4-methylcoumarin (15 µM) for 40 min before washing in MES buffer three times for 15 min each. Subsequently, samples were treated for 30 min with ABA (10 µM), NaHS (100 µM), or H2O2 (10 µM), respectively. For (G), the value 100% corresponds to the fluorescence of wild type in control conditions. R.U., relative units. The corresponding stomatal aperture sizes were also measured. The data represent mean ± sd from three biological replicates (n ≥ 90 in each replicate). Bars denoted by the different letters indicate significantly different values at P < 0.05 according to two-way ANOVA (post-hoc Tukey’s HSD test).
(K) Stomatal closure and H2S production rate of different plant lines under ABA treatment for the indicated times. Treatment and measurement details are described in (I) and (J). The data represent mean ± sd from three biological replicates (n ≥ 90 in each replicate). Asterisks represent significant differences between pMYB60:DES1 des1 and pMYB60:DES1Cys44/205Ala des1 according to Student's t test (*P < 0.05, **P < 0.01, ***P < 0.001).
Four-week–old plants grown in the soil were treated and harvested at the indicated times. For the immunoblot, the numbers upon the bottom panels indicate the relative abundance of the corresponding persulfidated protein compared with that of the control sample (1.00). Signals from two independent experiments were quantified. Con, control; IB, immunoblotting; wild type, unmutated version of DES1 protein.
In this study, we focused on the functional characterization of Cys44 and Cys205, because a mutation at Cys30 did not significantly affect the persulfidation of DES1 recombinant protein. Treatment with NaHS led to marked induction of DES1 recombinant protein activity, but did not affect DES activity in the Cys44/205Ala double mutant and Cys30/44/205Ala triple mutant (Figure 3G). These mutants displayed a slight reduction in DES1 activity when untreated compared with unmutated DES1 recombinant protein. Similarly, treatment with NaHS or DTT did not alter the persulfidation of either the Cys44/205Ala- or Cys30/44/205Ala-mutated versions (Supplemental Figure 8). Taken together, these results indicated that both Cys44 and Cys205 are key to DES1 protein function and affect persulfidation-induced DES1 activity.
To explore the regulatory role of persulfide-oxidation at Cys residues on the activity of DES1, we further examined the effects of H2O2 on its activity. H2O2 treatment of wild-type DES1 protein pretreated with NaHS resulted in a significant decrease in activity, whereas DES1 activity in the Cys44/205Ala mutant was unchanged (Figure 3H). This suggests that the two cys residues coordinate to regulate DES1 activity by persulfide-oxidation.
To examine the biological consequences of DES1 persulfidation, the impact of this modification on ABA-induced stomatal closure was investigated. To do this, we generated the transgenic lines pMYB60:DES1 des1, in which DES1 and its mutated versions were specifically expressed in mature guard cells. When treated with ABA, guard cells of pMYB60:DES1 des1, pMYB60:DES1Cys30Ala des1, and pMYB60:DES1Cys44Ala des1 plants showed increased H2S production and stomatal closure, similar to that of the wild type (Figures 3G and 3H; Supplemental Figure 9). However, these ABA-induced responses were abolished in pMYB60:DES1Cys205Ala des1 and pMYB60:DES1Cys44/205Ala des1 plants (Figures 3I and 3J). Consistent with this, the ABA treatment time course showed that the guard cell-impaired ABA responses in pMYB60:DES1Cys44/205Ala des1 were similar to those of des1, which can be restored by application of NaHS (Figures 3K and 3L; Supplemental Figure 10).
RBOHD-derived ROS in Guard Cells Are Genetically Required for DES1-regulated Stomatal Closure
RBOHD is responsible for the autopropagation wave of ROS production in each cell along its systemic path, which is required for required for system-acquired acclimation (Miller et al., 2009; Mittler et al., 2011; Gilroy et al., 2014). To elucidate the tissue-specific function of RBOHD, we expressed RBOHD protein in the rbohD background under the control of ProML1, ProCAB3, and ProMYB60 promoters, which have been used widely to express proteins of interest in epidermal (including guard cells), mesophyll, and guard cells, respectively (Supplemental Figure 11; Chater et al., 2015; Kirchenbauer et al., 2016; Bernula et al., 2017). A stomatal bioassay was conducted using various tissue-specific complementation lines together with corresponding wild-type and rbohD mutant plants, under ABA, NaHS, or H2O2 treatments. Interestingly, the impaired production of ROS triggered by ABA in the guard cells of rbohD was rescued when RBOHD was expressed in either epidermal or guard cells (Figure 4A, top; Supplemental Figure 12). Accordingly, the rbohD, stomatal responses of tissue-specific complementation lines were induced by ABA and NaHS (Figure 4A, bottom).
H2S Requires RBOHD Activity for the Induction of Stomatal Closure
(A) ROS production in guard cells and stomatal responses of wild-type, rbohD, and rbohD rescued lines. RBOHD was rescued by expressing RBOHD under the control of tissue-specific promoters, in which RBOHD was expressed in guard cells (pMYB60: RBOHD), epidermal (pML1: RBOHD), or mesophyll cells (pCAB3: RBOHD) individually. ROS production was monitored using florescent dye H2DCFDA (upper representation). Epidermal strips of each line were preincubated for 3 h in opening buffer (10 mM of MES at pH 6.15, and 10 mM of KCl) under light (120 μE m−2 s−1) and loaded with H2DCFDA (15 μM) for 15 min before washing in MES buffer three times for 15 min each. Subsequently, samples were treated for 1 h with ABA (10 μM), NaHS (100 μM), or H2O2 (10 μM), respectively. The value 100% corresponds to the fluorescence of wild type (Col-0) in control conditions. Con, control; R.U., relative units. The corresponding stomatal aperture sizes were also measured (lower representation). The data represent mean ± sd from three biological replicates (n ≥ 90 in each replicate). Bars denoted by the different letters were significantly different at P < 0.05 according to two-way ANOVA (post-hoc Tukey’s HSD test).
(B) ABA- and NaHS-induced net H2O2 influxes in guard cells of rbohD and pCAB3:RBOHD rbohD plants. Leaves from 4-week–old plants were preincubated for 3 h in opening buffer (10 mM of MES at pH 6.15, and 10 mM of KCl) under light (120 µE m−2 s−1) and washed in MES buffer three times for 15 min each. The net H2O2 influxes were observed after ABA (10 µM) or NaHS (100 µM) treatment during the indicated times (n = 6).
(C) ROS production in guard cells and stomatal responses of wild-type, des1, des1/rbohD, and des1/rbohD rescued lines. Treatment and measurement details are shown in (A). Con, control.
(D) to (F) Effects of sulfurating molecules on NADPH oxidase activity of wild-type, des1, rbohD, and rbohD rescued lines. For NADPH oxidase activity assay, the crude membrane extracts from each line (4 weeks old) were extracted and treated with NaHS or Na2S at the indicated concentrations (0 mM to 1 mM; for F, NaHS at 100 μM) for 30 min. NADPH oxidase activity was determined after dialysis. The data represent mean ± sd (n = 3) from three biological replicates. Different letters indicate significantly different values at P < 0.05 according to two-way ANOVA (post-hoc Tukey’s HSD test).
Although there were a number of mesophyll cells attached to the strips after peeling, we observed that the mesophyll transgenic line behaved similarly to the guard cell complementation line. Importantly, a significant net ROS-associated influx was found in either ABA- or NaHS-treated guard cell of pCAB3:RBOHD rbohD plants, while a small net ROS-associated influx was detected in rbohD mutants (Figure 4B). These results suggested that RBOHD activity in both epidermal and mesophyll cells contributed to ABA-induced ROS production in guard cells and, hence, stomatal closure.
Previous pharmacological results showed that stomatal closure was impaired in wild-type epidermal strips treated with NaHS in the presence of DPI, an inhibitor of plant flavoenzymes, including the NADPH oxidase, indicating the interplay between H2S and H2O2 (Scuffi et al., 2018). The NADPH oxidase isoforms RBOHD and RBOHF are responsible for the majority of guard cell H2O2 production (Kwak et al., 2003). To further evaluate the contribution of RBOHD to H2S-induced stomatal closure, we crossed des1 with rbohD mutant plants and a homozygous des1 rbohD line was identified by genotyping-based PCR (Supplemental Figure 13). Our genetic data further demonstrated that, compared with the des1 mutant, NaHS-induced ROS production and stomatal closure were almost fully abolished in des1 rbohD (Figure 4C), indicating that RBOHD is required genetically for DES1/H2S-induced stomatal closure. The stomatal response and ROS production within des1 rbohD guard cells were less ABA-sensitive, but the corresponding responses to H2O2 were not altered.
Our results indicated that the regulation between intercellular ROS production and ABA signaling may occur during stomatal closure (Figures 4A to 4C). To further investigate whether intercellular ROS signaling was regulated by DES1, we expressed RBOHD protein in the des1 rbohD background under the control of tissue-specific promoters. ABA treatment failed to induce stomatal closure when RBOHD was expressed in either epidermal, mesophyll, or guard cells of the line carrying the des1 rbohD background (Supplemental Figure 14). Together with the results shown in Figure 4A, our observations suggested that functional DES1 activity is involved in ABA-triggered guard cell ROS production and intercellular ROS relay.
H2S-triggered Stimulation of NADPH Oxidase Activity of RBOHD
Next, to examine whether NADPH oxidase activity could be influenced directly by H2S in vitro, we performed enzyme activity assays of NADPH oxidase. Total crude membrane proteins from wild-type and des1 mutant plants were extracted and treated with different concentration of sulfurating molecules, followed by dialysis, before determination of NADPH oxidase activity. In both wild-type and des1 plants, we clearly observed direct and significant activation of NADPH oxidases by two sulfurating molecules, in a dose-dependent manner (Figure 4D). NADPH oxidase in the des1 mutant was present in much lower amounts than in the wild type (Columbia-0 [Col-0]) at basal level, suggesting the important role of endogenous H2S produced by DES1 in regulating NADPH oxidase activity in vivo.
To verify the contribution of RBOHD function to H2S-induced NADPH oxidase, changes to NADPH oxidase activity in the presence or absence of NaHS were determined in a rbohD mutant and rescued line pRBOHD:RBOHD rbohD. In this case, while low basal levels of NADPH oxidase activity in the rbohD mutant were still detectable compared with the wild type (Col-0), the NaHS-induced activity of NADPH oxidase observed in the wild type was almost fully abolished in this mutant (Figure 4E). This low level of NADPH oxidase activity in the rbohD mutant was partially rescued in pRBOHD:RBOHD rbohD, indicating the predominant contribution of RBOHD to NaHS-induced NADPH oxidase activity in vivo.
ABA Induces Persulfidation of RBOHD on Cys825 and Cys890
RBOHD protein contains 10 cysteines that might serve as potential sites for persulfidation-based modification by H2S in the N-terminal portion (N-terminal; Cys208), carboxy-terminal portion (C-terminal; Cys825 and Cys890), extracellular portion (Cys410, 412, 432, 651, and 695), and transmembrane portion (Cys387 and Cys480). To determine whether the cytosolic sides of RBOHD might be persulfidated, the C- and N-terminal portions of this protein were expressed and exposed to sulfurating molecules typically used to score for persulfidation in vitro. A tag-switch analysis revealed that persulfidation of the C-terminal portion of RBOHD recombinant protein was induced by two sulfurating molecules individually (Figure 5A), which was in agreement with the results from the analysis of the persulfidation level of RBOHD-C-GFP in protoplast (Figure 5B; Supplemental Figure 15), and treated with two H2S donors as well (Supplemental Figure 16). However, no persulfidation of the N-terminal portion was detected, regardless of the addition of sulfurating molecules (Figure 5C).
ABA-Induced Persulfidation of RBOHD at Cys825 and Cys890
(A) The C terminus of RBOHD is persulfidated in the presence or absence of two sulfurating molecules. RBOHD–C-terminal recombinant protein was treated with NaHS or Na2S at 1 mM for 1 h, followed by the tag-switch assay.
(B) Analysis of persulfidation of RBOHD–C-terminal in vivo. Total proteins from Arabidopsis wild-type protoplast transformed with 35S:RBOHD-C-GFP were extracted and exposed to NaHS for 1 h at the indicated concentrations, then subjected to the tag-switch assay.
(C) The N terminus of RBOHD is not persulfidated. RBOHD-N-terminal recombinant protein was treated with NaHS 1 mM for 1 h, followed by the tag-switch assay.
(D) ABA-induced persulfidation of RBOHD–C-terminal in vivo. Total proteins from Arabidopsis wild-type (Col-0) and des1 mutant protoplasts transformed with 35S:RBOHD-C-GFP were extracted and treated with or without ABA at 10 μM for 30 min, then subjected to the tag-switch assay.
(E) and (F) H2O2-triggered persulfide-oxidation of the RBOHD–C-terminal recombinant protein in vitro (E) or RBOHD-C-GFP protein in vivo (F). Samples were treated with H2O2 at indicated concentrations for 1 h, then subjected to the tag-switch assay.
(G) Enhancement of ABA-induced RBOHD-C-GFP persulfidation by DPI. Total proteins from Arabidopsis protoplast transformed with 35S:RBOHD-C-GFP were subjected to the tag-switch assay after exposure to ABA at 10 μM for 30 min, with or without pretreatment of DPI at 10 μM for 2 h.
(H) to (J) Persulfidation analysis of the C terminus of wild-type and mutant RBOHD derivatives in vitro (H) and in vivo (I) and (J).
For in vitro assay, the persulfidated recombinant proteins were detected by an anti-biotin antibody after the tag-switch labeling. Proteins were detected by immunoblotting using an anti-his antibody served as a loading control. Con, control.
For in vivo assay, total proteins from Arabidopsis protoplasts transiently expressed the RBOHD C terminus and its mutant derivatives fused to GFP were treated with or without ABA (10 μM for 30 min) or NaHS (1 mM for 1 h) before being subjected to the tag-switch assay and streptavidin purification. The persulfidated proteins were detected by an anti-GFP antibody. The bottom representations show the total GFP fused proteins used for the tag-switch assay as loading control. Wild type, the unmutated version of RBOHD-C protein; Con, untreated conditions. The numbers above the bottom representations indicate the relative abundance of the corresponding persulfidated protein compared with that of the control sample (1.00). Signals from two independent experiments were quantified.
We next determined whether the RBOHD-C-terminal was persulfidated during the plant response to ABA. After ABA treatment, the level of persulfidated RBOHD-C-SSH protein was increased, but this was largely impaired in des1 plants, further suggesting a regulatory role for DES1 in modulating the persulfidation the RBOHD C-terminal (Figure 5D). Subsequently, we analyzed RBOHD-C-SSH formation after treatment with H2O2 at a wide range of concentrations. The results suggested that H2O2 led to a time and dose- dependent decrease in the persulfidation of RBOHD-C-SSH protein both in vitro and in vivo (Figures 5E and 5F; Supplemental Figure 17). The level of RBOHD-C-SSH was consistently higher in the presence DPI, regardless of the presence of ABA (Figure 5G). Moreover, similarly to the persulfidation of DES1 recombinant protein, persulfidation and persulfide-oxidation processes were reversible on the thiol groups of RBOHD-C-terminal protein (Supplemental Figure 18).
The Cys residues within the RBOHD-C-terminal were, therefore, mutated either individually or in combination and the resulting proteins were expressed, treated with sulfurating molecules, and analyzed by tag-switch assay. Both the mutations of Cys825Ala or Cys890Ala significantly impaired persulfidation of RBOHD-C-terminal recombinant protein, particularly Cys825Ala (Figure 5H), but this could be reversed by treatment with two sulfurating molecules, separately (Supplemental Figure 19). Similarly, in vivo results demonstrated that persulfidation was additive on Cys825 and Cys890, and was almost completely eliminated in the Cys825/890Ala double mutant, which was not affected by treatments with either ABA or NaHS (Figures 5I and 5J). Unfortunately, we failed to identify any persulfided sites on RBHOD-C recombinant protein, probably due to the low abundance and/or poor purity of the purified protein. Collectively, these findings imply that RBOHD is specifically persulfidated at both Cys825 and Cys890, with Cys825 being the preferred target.
The Arabidopsis genome contain 10 genes that encode NADPH oxidase, of which RBOHDA, B, and C contain Cys residues that are homologous to both Cys825 and Cys890 of RBOHD, indicating that different RBOH proteins might also be persulfidated in this species. We therefore expressed recombinant RBOHC-C-terminal mutated versions and exposed them to either NaHS or DTT. As was the case with the RBOHD-C-terminal, the unmutated RBOHC-C-terminal was specifically persulfidated, but mutations of the conserved cysteines blocked this process of protein persulfidation (Supplemental Figure 20). The conservation of Cys residues and the persulfidation-dependent regulation of RBOH family proteins suggest that persulfidation-mediated activation is a general regulatory mechanism adopted by plants.
RBOHD Persulfidation Controls ABA-stimulated ROS Production and Stomatal Closure
We investigated whether the persulfidation-mediated activation of RBOHD plays a role in ABA-stimulated ROS production and stomatal responses. To explore the possible biological consequences of Cys825 and Cys890 persulfidation on RBOHD activity, we monitored NADPH oxidase activities in mutant rbohD plants expressing wild-type RBOHD, Cys825Ala, Cys890Ala, or Cys825/890Ala driven by the native RBOHD promoter (Supplemental Figure 21). NaHS-induced leaf NADPH oxidase activity was significantly impaired in the pRBOHD:RBOHDCys825Ala rbohD and pRBOHD:RBOHDCys890Ala rbohD lines, compared with pRBOHD:RBOHD rbohD (Figure 6A). Inducible NADPH oxidase activity and ROS accumulation in guard cells were fully abolished in the pRBOHD:RBOHDCys825/890Ala rbohD line, implying an additive effect of RBOHD persulfidation at Cys825 and Cys890 during ABA/H2S-triggered NADPH oxidase activity, which results in increased ROS generation in guard cells (Figure 6B, top).
DES1-Triggered Persulfidation of RBOHD at C825 and C890 Is Involved in ABA-Induced ROS Activation and Stomatal Closure
(A) Effects of NaHS or DTT on the activity of NADPH oxidase of mutant rbohD plants expressing either a wild-type RBOHD or Cys825Ala, Cys890Ala, and Cys825/890Ala mutant derivatives driven by its native promoter. Treatment and measurement details are represented in Figures 4D to 4F. Con, control.
(B) and (C) Effects of ABA or NaHS on ROS production and stomatal closure of wild type (Col-0), rbohD, des1 rbohD, and lines expressing either a wild-type RBOHD or Cys825Ala, Cys890Ala, and Cys825/890Ala mutant derivatives driven by its native promoter. Treatment and measurement details are shown in Figure 4A. Con, control; R.U., relative units.
Next, we analyzed the responsiveness of the transgenic rbohD lines (expressing either wild-type RBOHD or mutant derivatives) to ABA or NaHS during stomatal responses. When treated with NaHS or ABA, pRBOHD:RBOHD rbohD exhibited stomatal closure similar to that of the wild type (Figure 6B, bottom). However, this inducible stomatal closure was markedly reduced in the pRBOHD:RBOHDCys825Ala rbohD, pRBOHD:RBOHDCys890Ala rbohD, and pRBOHD:RBOHDCys825/890Ala rbohD lines.
To confirm that ABA-activated RBOHD function in stomatal closure and ROS production in guard cells are consequences of DES1-triggered persulfidation, we introduced RBOHD and its substitutional mutant derivatives under the control of the native RBOHD promoter into the des1 rbohD mutant, and tested the ROS production and guard cell responses to either ABA or NaHS in the resulting transgenic plants. All of the transgenic plants with the des1 rbohD background exhibited reduced ROS production and stomatal responses to ABA, which further reinforces the conclusion that DES1 is genetically required for ABA-induced RBOHD activation and stomatal closure (Figure 6C). When treated with NaHS, pRBOHD:RBOHD des1 rbohD guard cells exhibited an increase in ROS and stomatal closure, similar to that of the wild type. However, NaHS-induced ROS activation and stomatal closure were abolished in pRBOHD:RBOHDCys825Ala des1 rbohD, pRBOHD:RBOHDCys890Ala des1 rbohD, and pRBOHD:RBOHDCys825/890Ala des1 rbohD plants. Collectively, these results suggested that DES1/H2S-triggered RBOHD persulfidation at Cys825 and Cys890 plays a critical role in modulating ABA-induced guard cell ROS production and stomatal closure.
DISCUSSION
In dry conditions, plants rapidly synthesize the drought stress hormone ABA at high levels to trigger signal transduction events that lead to stomatal closure and reduced water loss. Although ABA is known to activate stress responses, the mechanisms by which ABA signals are relayed in guard cells remain unresolved. In this study, we unraveled the molecular framework for DES1/H2S-triggered protein persulfidation, which links ABA signaling pathways with stomatal closure.
Persulfidation of DES1 and RBOHD during ABA-induced Stomatal Closure
Maintaining the redox status is a conspicuous feature of signaling cascades and precisely controlled in plant cells (Hancock, 2019; Smirnoff and Arnaud, 2019). Interestingly, the reducing and oxidizing molecules H2S and H2O2, which are generated by DES1 and NADPH oxidase, are both critically involved in the ABA response (Kwak et al., 2003; Scuffi et al., 2018). However, the molecular mechanisms that mediate the synthesis and function of these signaling molecules are poorly understood. Perhaps the best example of the convergent action of DES1 and NADPH oxidase is the control of stomatal aperture, where H2S is suggested to serve as a gasotransmitter in stomatal closure (Honda et al., 2015; Papanatsiou et al., 2015). Our data are entirely consistent with these observations (Scuffi et al., 2018). While closure was not evident in the des1 rbohD mutant line after treatment with NaHS, the guard cell response to H2O2 was the same as that of wild type. The impairment in ABA/NaHS sensitivity of des1 rbohD, in terms of stomatal closure, was not altered by the introduction of DES1 driven by guard cell promoter MYB60, which illustrated a positive regulatory role for RBOHD downstream of DES1/H2S in guard cell ABA signaling. RBOHD was shown to mediate long-distance signaling via a dynamic autopropagating wave of ROS that can travel at a rate of up to 8.4 cm min−1 through the apoplast (Miller et al., 2009). In support of this finding, both ABA and NaHS triggered a ROS-associated influx into guard cells and stomatal closure when RBOHD was tissue-specifically expressed in the rbohD mutant background. Our results infer that plants retain the ability to produce ROS signals by RBOHD and facilitate cell-to-cell communications to trigger downstream events in guard cells in response to environmental changes.
H2S physiologically persulfidates a diverse range of proteins. As persulfidation occurs frequently, this post-translational modification has been proposed to affect a variety of biological pathways (Aroca et al., 2017a, 2017b, 2018). Here, we demonstrated that ABA-activated DES1-endogenous H2S physiologically persulfidated DES1 to regulate its activity, with Cys44 and Cys205 as the main sites of this modification. Our results revealed that the activity of persulfidated DES1 can return to the basal state quickly, depending on the dynamic redox status, e.g., endogenous H2O2 homeostasis, which can lead to reversible persulfide-oxidation of persulfidated DES1, thus fine-tuning the strength and duration of the ABA signal. Therefore, ROS accumulation may function as a negative feedback mechanism to prevent the continuous triggering of ABA signaling in guard cells, which is achieved by persulfide-oxidation of persulfidated DES1. Moreover, persulfides may serve a protective function of protein thiols in highly oxidized environments (Gotor et al., 2019). It is therefore a matter of importance as DES1-produced H2S contributes far more to the ABA signaling than previously imagined. The regulatory mechanism is physiologically related to the guard cell ABA response. In pMYB60:DES1Cys44/205Ala des1, ABA was less effective in inducing guard cell H2S production and stomatal closure than in wild-type plants within the treatment period. This observation supports a positive regulatory role for ABA in the persulfidation and activation of DES1, which may facilitate the downstream H2S sensing, which is essential in relaying the guard cell ABA signals.
The rapid production of ROS in guard cells acts as systemic secondary messenger to trigger stomatal closure (Song et al., 2014). Plant NADPH oxidases belong to the RBOH family, and catalyze the production of superoxide that can then be converted into H2O2 spontaneously or by superoxide dismutase or peroxidases (Smirnoff and Arnaud, 2019). In many cases, RBOHs regulation occurs on the N-terminal extension that carries EF-hand motifs and potential phospho-sites through Ca2+ binding and phosphorylation, which are thought to be important for activation of NADPH oxidase (Ogasawara et al., 2008; Chen et al., 2017). Notwithstanding the long-established correlation between ABA signaling and ROS burst activation in guard cells, the mechanisms that underpin the activation of a ROS burst, and are associated with the RBOHD C-terminal, are largely unknown. In this study, we discovered that H2S regulates guard cell ABA signaling through persulfidation of two cysteines in the C-terminal of RBOHD. Persulfidation triggers RBOHD to facilitate a burst of ROS, which, in turn, causes stomatal closure. This might represent a mechanism for raised ABA-triggered ROS signaling in guard cells. des1 plants showed reduced NADPH oxidase activity and low persulfidation at the C-terminal of RBOHD, supporting a positive regulatory role for H2S in the ABA-activated RBOHD.
Our findings are reminiscent of those in plant immunity, where S-nitrosylation at a conserved Cys (890, not 825) in RBOHD promotes cell death during plant immune function (Yun et al., 2011). While s-nitrosylation on Cys890 disrupts the side-chain position of Phe921, impeding FAD binding affinity, persulfidation enhances the reactivity of the Cys890 thiol and may render it more nucleophilic, and thus may facilitate its binding with FAD. Importantly, under our conditions, Cys825 was predominantly persulfidated and additional effects of Cys825 and Cys890 on guard cell ROS production and responses were seen. Thus, we could not fully reject the hypothesis that nitrosylation of Cys890 affects the guard cell response. Further investigations using the RBOHD native protein and 3D structure will help to unveil the mechanisms of persulfidation-regulated RBOHD activity. Notably, we observed that the RBOHDCys825/890Ala mutation abolished the H2S and ABA responses, suggesting that the persulfidation of RBOHD plays a regulatory role in guard cell ABA responses. It has been reported that persulfidation precedes nitrosylation on the same Cys residue in several instances (Sen et al., 2012; Vandiver et al., 2013). This opens the possibility of crosstalk between persulfidation and nitrosylation in the regulation of ROS production and ABA signaling in guard cells. Further clarification of the temporary sequences of persulfidation and nitrosylation would offer new insights into the sophisticated mechanisms of ABA-controlled guard cell signaling.
Redox-based Reversible Persulfidation Controls Guard Cell ABA Signaling
The on–off switch for RBOHD function controlled by persulfidation and persulfide-oxidation may fine-tune the strength or duration of RBOHD activation and ROS signals in response to ABA. Our results indicate that H2S and ROS produced by DES1 and RBOHD, respectively, can fulfill this role. Although the kinetic and structural details of DES1/RBOHD persulfidation and persulfide-oxidation remain limited, our findings suggest the following: ABA causes rapid and strong activation of DES1 in guard cells, which can persulfidate downstream functional proteins, including itself on Cys44 and Cys205, thus amplifying guard cell ABA signals. The inducible DES1/H2S also persulfidates NADPH oxidase RBOHD on Cys825 and Cys890 to facilitate the rapid induction of a ROS burst and stomatal closure (Figure 7 and the graphic abstract). When ROS accumulates, it triggers persulfide-oxidation of RBOHD and DES1 to form perthiosulfenic acid, leading to ABA desensitivity. The reducible S–S bond present in the formed perthiosulfenated Cys can be reduced by disulfide reductases such as thioredoxin, which has been demonstrated in the formed s-sulfocysteine bond in the active site of 3′-phosphoadenosine-5′-phosphosulfate reductase (Palde and Carroll, 2015) The stimulated RBOHD activity may be returned to the basal state quickly through this redox-based posttranslational modification, based on dynamic redox homeostasis. The finding that ROS-triggered persulfide-oxidation of persulfidated RBOHD and DES1 suggests that plants possess yin-and-yang mechanisms to orchestrate redox signaling coupled with hormonal response, which may feed back to prevent the continual activation of ABA signals in guard cells. As H2S and ROS are ubiquitous, this yin-and-yang interplay is most likely to be a rapid regulatory mechanism that primes the paradigmatic ABA-triggered regulation of plant RBOHs and ROS production in guard cells. Correspondingly, tipping the DES1/H2S-RBOHD/ROS balance toward increased endogenous ROS production can trigger protein persulfidation oxidation, further leading these activated proteins to return to their basal states. Plants endowed with this negative feed-back loop could sense the redox status and control ROS homeostasis under everchanging environments.
Working Model Shown in This Study
When drought stress signal was sensed by the plant, ABA was synthesized and triggered rapid and strong expression of DES1 in guard cells by an unknown mechanism. Inducible DES1 expression can persulfide many downstream signaling proteins, including itself, on Cys44 and Cys205—leading to the amplification of guard cell H2S signals. The inducible DES1/H2S also persulfidates NADPH oxidase RBOHD on Cys825 and Cys890 to facilitate the rapid induction of ROS burst, which in turn causes stomatal closure. When ROS accumulates to high levels, it further causes persulfide-oxidation of RBOHD and DES1 (−SSOnH), leading to ABA desensitivity. The oxidized persulfide (−SSOnH) may subsequently be reduced to a thiol group (−SH) by thioredoxin, which formed a feedback prevention of ABA signaling continuously activated in guard cells.
METHODS
Plant Materials and Growth Conditions
The Arabidopsis (Arabidopsis thaliana) des1 (SALK_103855; Col-0) mutant was obtained from the Arabidopsis Biological Resource Center (http://www.arabidopsis.org/abrc); rbohD was a gift from Prof. Wenhua Zhang, Department of Plant Science, Nanjing Agricultural University.
The des1 rbohD double mutant line was obtained by crossing rbohD with des1 (Supplemental Data Set 1). Homozygous mutants were identified by sequencing, combined with PCR-based genotyping and corresponding phenotypes (including reduced plant size and vigor).
Seeds were surface-sterilized and washed three times with sterile water for 20 min, then cultured in Petri dishes on solid half strength Murashige and Skoog (MS) medium (pH 5.8). Plants were grown in a growth chamber with a 16-h/8-h (23°C/18°C) day/night regime using bulb-type fluorescent light at a light intensity of 100 μmol photons m−2 s−1 irradiation.
Molecular Cloning
For Expression in Escherichia coli
PCR-amplified fragments of AtRBOHD-N (1 bp to 1,068 bp), AtRBOHD-C (2,269 bp to 2,763 bp), and full-length AtDES1 were introduced into PET28a vector (for His fusion) using homologous recombination technique (Vazyme), with the enzyme digestion site of NdeI and XhoI.
Site-directed mutagenesis was performed with Mut Express II Fast Mutagenesis Kit (Vazyme). All procedures followed the manufacturer’s manual. The specific primers used for AtRBOHD-N, RBOHD-C, and full-length AtDES1 are listed in Supplemental Data Set 1.
For Transient Expression in Arabidopsis Protoplasts
PCR-amplified fragments were cloned into PAN580 vector (from Aying Zhang’s lab, College of Life Sciences, Nanjing Agricultural University) at the sides of XbaI and BamHI using a homologous recombination technique (Vazyme).
Arabidopsis protoplasts were isolated according to the method described previously by Yoo et al. (2007). The transient expression vector PAN580 was delivered into Arabidopsis protoplasts using a PEG-calcium–mediated approach and protoplasts were cultured for 12 h. Protoplasts expressing the targeted GFP driven by the constitutive 35S enhancer under fluorescence microscopy were used for transient expression analysis.
For Expression In Planta
Promoter cloning of CAB3, ML1, MYB60, and RBOHD was done using Arabidopsis wild-type DNA as template. The promoter regions of CAB3 (2,000 bp), ML1 (1,500 bp), MYB60 (1,500 bp), and RBOHD (1,500 bp) were amplified by PCR (Supplemental Data Sets 1 and 2). CAB3, pML1, and MYB60 fragments were cloned into pCAMBIA 1305 vector (from Aying Zhang’s lab, College of Life Sciences, Nanjing Agricultural University). The RBOHD products were cloned into 130221-flag vector (from Yiqun Bao’s lab, College of Life Sciences, Nanjing Agricultural University).
The constructed plasmids were transferred into an Agrobacterium tumefaciens strain, and then the inflorescence of Arabidopsis was affected by dipping. Transgenic plants were selected on half strength MS medium containing Hygromycin B (50 μg/mL). Identification of transgenic plants was conducted by genotyping combined with PCR, fluorescence observation, and immunoblot analysis.
Expression and Purification of Recombinant Protein
Expression and purification of the recombinant protein was performed in BL21 cells (Vazyme). A quantity (0.1 mM) of IPTG was added and the bacteria grown to OD600 = 0.4 to 0.6, for 12 h at 16°C. After enriching the bacterial solution, the precipitation was suspended in PBS buffer, the protein broken by ultrasonic crusher, and was centrifuged at 12,000g for 30 min, with the extract then taken for purification. The HIS-tagged proteins were purified using a NI-NTA prepacked gravity column (Sangon Biotech); the procedure for protein purification was performed in accordance with this column specification.
Direct Mass Spectrometry Analysis of Persulfidated Cys Sites on DES1
Direct mass spectrometry analysis of persulfidated Cys sites on DES1 was performed with modifications as described by Bogdándi et al. (2019) and Fu et al. (2019). In brief, purified recombinant proteins (∼0.1 mg/mL, 120 μL) in 50 mM of HEPES at pH 7.6 were incubated with 1 mM of NaHS at room temperature for 1 h. Protein samples were then alkylated with 10 mM of iodoacetamide (IAM) in the dark at room temperature for 30 min. Next, alkylated proteins were resolved by SDS-PAGE gel and stained with Coomassie Brilliant Blue. Gel bands corresponding to DES1 (∼40 kd) were sliced, destained with 50 mM of NH4HCO3 in 40% methanol, and subjected to in-gel digestion as described by Shevchenko et al. (2006). The resulting peptides were extracted, desalted, and evaporated to dryness until liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. LC-MS/MS was performed with the same settings as previously described by Akter et al. (2018) and Huang et al. (2019) on a Q Exactive Plus (Thermo Fisher Scientific) instrument operated with an Easy-nLC1000 System (Thermo Fisher Scientific). The resulting mass spectrometry data were searched against the protein sequence of DES1 (UniProt accession number: F4K5T2) using the software pFind 3 (Chi et al., 2018). The precursor-ion mass and fragmentation tolerance were 10 microliters per liter and 20 microliters per liter, respectively. A specific-tryptic search was used with a maximum of three missed cleavages allowed. Mass shifts of +15.9949 D (Met oxidation, M), +57.0214 D (C2H3N1O1, carbamidomethylation, C), and +88.9935 D (C2H3N1O1S1, IAM-derivatized persulfide, C) were searched for dynamic modifications.
Enzyme Activity Assay of DES1 Recombinant Protein
The DES activity for DES1 were determined using the methods described in Álvarez et al. (2010). The DES activity was measured by the monitoring the release of sulfide from l-cysteine in a total volume of 3 mL containing 0.8 mM of l-cysteine, 100 mM of Tris-HCl at pH 9.0, and 300 μL of purified DES1 protein. The reaction was initiated by adding l-cysteine. After incubation at 37°C for 15 min, the reaction was terminated by the addition of 300 μL of 30 mM of FeCl3 dissolved in 1.2 M of HCl and 300 μL of 20-mM n,n-dimethyl-p-phenylenediamine dihydrochloride dissolved in 7.2 M of HCl. The formation of methylene blue was determined at 670 nm in a spectrophotometer. The DES enzymatic activity was calculated by comparison to a standard curve of Na2S. The amount of total protein was determined by the method of Bradford (1976).
SDS-PAGE and Immunoblotting
Protein extracts were separated by 12% SDS-PAGE. After electrophoresis, the gel was transferred to a polyvinylidene difluoride membrane (Roche) at 100 V for 45 min at 4°C. The membrane was blocked with 5% skimmed blocking solution (5% [w/v] BSA in TBS with 0.5% [v/v] Tween-20) and incubated for 2 h with shaking at room temperature or overnight at 4°C. Immunoblot analysis was performed with antibodies (Supplemental Data Set 2) diluted in blocking solution at the following dilutions: anti-GFP-HRP-DirecT, 1:4,000 (MBL); and anti-biotin-HRP, 1:10,000 (Abcam). For quantification of protein abundance, the software ImageJ (https://imagej.nih.gov/ij/) was used and signals from two independent experiments were quantified. Full-sized membrane scans are presented in Supplemental Figure 22.
Immunochemical Detection of s-Persulfidated Proteins
s-persulfidated proteins were detected using a modified Tag-switch method as described by Aroca et al. (2017a).
For in vitro assays, the purified recombinant proteins were treated with NaHS, DTT, or H2O2 at the indicated concentrations, respectively. The proteins were blocked with methylsulfonylbenzothiazole, and the s-persulfidated cysteines were labeled by CN-biotin. The s-persulfidated proteins were detected by immunoblot using an anti-biotin antibody (anti-biotin-HRP, 1:10,000; Abcam).
For in vivo assay, Arabidopsis leaves (∼1 g) were ground in a mortar under liquid nitrogen. Extract and Arabidopsis protoplasts were homogenized in buffer (25 mM of TRIS, 100 mM of NaCl, and 0.2% [v/v] Triton X-100, at a pH of 8.0) containing a complete protease inhibitor cocktail (Sigma-Aldrich). The extract was centrifuged at 4°C for 30 min. Blocking buffer consisting of 50 mM of methylsulfonylbenzothiazole that was dissolved in tetrahydrofuran was added to the extract and the solution was incubated at 37°C for 1 h to block free sulfhydryl groups. Proteins was precipitated with acetone and resuspended pellet in buffer (50 mM of TRIS, 2.5% [w/v] SDS, and 20 mM of CN-biotin, at a pH of 8.0) and incubated 3 h to 4 h 37°C. Protein was precipitated with acetone, washed with acetone 70% (w/v) twice, and resuspended pellet in buffer (50 mM of TRIS and 0.5% [w/v] SDS, at a pH of 8.0). The biotinylated proteins were purified by immunoprecipitation for 1 h at 25°C with 80 μL of streptavidin magnetic particles (Roche). Particles were washed three times and bound proteins were eluted with a suitable buffer, and separated by 12% (w/v) SDS-PAGE, transferred to a polyvinylidene fluoride membrane. Proteins was detected with an antibody against GFP (anti-GFP-HRP-DirecT, 1:4,000; MBL).
Measurement of H2S and ROS in Guard Cells and Stomatal Aperture
Epidermal strips of each ecotype were preincubated for 3 h in opening buffer (10 mM MES at pH 6.15, and 50 mM of KCl) under light (100 µE m−2 s−1) and loaded with a fluorogenic probe useful for the detection of H2S (7-Azido-4-methylcoumarin; Sigma-Aldrich) for 20 min, or loaded with 2′,7′-Dichlorofluorescein diacetate (an oxidation-sensitive fluorescent probe to measure ROS; Sigma-Aldrich) for 15 min before washing in MES buffer three times for 15 min each, subsequently treated for 1 h with 10 µM of ABA, 100 µM of NaHS (the sulfurating molecule), or 10 µM of H2O2 and then analyzed using a TCS-SP2 Laser Scanning Confocal Microscope (Leica; excitation at 405 nm, emission at 454 nm to 500 nm for the detection of H2S; excitation at 488 nm, emission at 500 nm to 530 nm for the detection of ROS). Stomatal aperture measurements were performed as previously described by Xie et al. (2016); epidermal strips were soaked in to the opening buffer, which was 10 µM of ABA, 100 µM of HT (the scavenger of H2S), 100 µM of NaHS (the sulfurating molecule), or 10 µM of ABA+100 µM HT (the scavenger of H2S) for 1 h. All manipulations were performed at 25 ± 1°C. Data were analyzed using the software ImageJ. Each value represents the mean of at least 90 stomata taken from different leaves, bars (when present) represent the se of each treatment.
Noninvasive Microtest Technique Measurement
The basic principles of noninvasive microtest technique (NMT) measurement are described by Newman (2001). The net guard cells ROS-associated fluxes were measured using the Scanning Ion-Selective Electrode Technique System (BIO-001A; Younger USA). A ROS-sensitive microsensor was purchased from the Xuyue Beijing NMT Service Center. The experimental voltage is +600 mV. The primary position (M1) of an ROS-sensitive microsensor was placed 10 μm from the guard cells’ surface, and the further-away position (M2) was 45 μm. Net guard cell ROS-associated fluxes were calculated by Fick’s law of diffusion: J (pmol/cm2/s) = −Do × (dc/dx), where Do indicates the diffusion constant, dc the concentration gradient (calibration with H2O2 standard curve), and dx the distance (30 μm). Experiments were conducted at room temperature (24 ± 1°C) under ambient light. Data (means ± sd, n = 6). Leaves from a 1-month–old plant were placed in a Petri dish and immersed in MES (at pH 6.15) assay solution, after preconditioning, steady-state ion fluxes were recorded over a period of 5 min. Then, an ABA/NaHS-containing stock solution was applied and mixed to reach the required final concentration of 10 μM and 100 μM, respectively.
Determination of NADPH Oxidase Activity
Crude membrane extracts were isolated. Two grams of leaf tissues were ground in liquid nitrogen and dissolved in four volumes of the extraction buffer (25 mM of HEPES at pH 7.5, 5 mM of EDTA, 2 mM of EGTA, 1 mM of DTT, 150 of mM KCl, 250 mM of mannitol, and 0.5 mM of PMSF). The crude extract was filtered with four layers of gauze and centrifuged at 10,000g for 20 min and the resulting supernatant was ultracentrifuged at 50,000g for 30 min. The resulting pellet was resuspended in suspension buffer (2.5 mM of HEPES at pH 7.5, 5 mM of EDTA, 1 mM of DTT, 150 mM of KCl, 250 mM of mannitol, and 1 mM of PMSF), and used as the membrane fraction to measure NADPH oxidase activity.
The NADPH-dependent O2− generating activity in isolated crude membrane was determined by following the reduction of sodium 3′-(1-[phenylamino-carbonyl]-3,4-tetrazolium)-bis(4-methoxy-6-nitro) benzenesulfonic acid hydrate (XTT) by O2−. The assay mixture of 1 mL contained 50 mM of Tris–HCl buffer at pH 7.5, 0.5 mM of XTT, 100 µM of NADPH, and 15 µg to 20 µg of membrane proteins. The reaction was initiated with the addition of NADPH, and XTT reduction was determined at 470 nm. Corrections were done for background production in the presence of 50 units SOD. Rates of O2– generation were calculated using an extinction coefficient of 2.16 × 104 M−1 cm−1.
Statistical Analysis
Statistical analysis and plotting of graphs were performed using the software SPSS 16.0 (https://www.ibm.com/products/spss-statistics). To determine statistical significance, we used one-way or two-way ANOVA (corrected with post-hoc Tukey’s HSD test). Differences were considered significant at P < 0.05. Statistical data are provided in Supplemental Data Set 3.
Accession Numbers
DNA and derived protein sequence data from this article are in the Arabidopsis Genome Initiative database under the following accession numbers: DES1 (At5g28030), RBOHD (At5g47910), RBOHF (At1g64060), MYB60 (At1g08810), ML1 (At5g61960), CAB3 (At1g29910).
Supplemental Data
Supplemental Figure 1. Analysis of persulfidation of DES1-GFP in vivo (supports Figure 1).
Supplemental Figure 2. Relative expression of DES1 in wild type (Col-0), des1, and lines expressing either a wild-type DES1 or Cys30Ala, Cys44Ala, Cys205Ala, and Cys30/44/205Ala mutant derivatives driven by MYB60 promoter (supports Figures 1 and 3).
Supplemental Figure 3. Detection of persulfidation of the DES1 protein in vivo (supports Figure 1).
Supplemental Figure 4. Analysis of persulfidation of DES1-GFP in vivo (supports Figure 1).
Supplemental Figure 5. Time-dependent persulfidation oxidation of DES1-His recombinant protein (supports Figure 2).
Supplemental Figure 6. Persulfidation of DES1-His recombinant protein by H2O2 pretreatment (supports Figure 2).
Supplemental Figure 7. Amino acid sequence of DES1 protein. DES1 protein has 323 amino acids in total (supports Figure 3).
Supplemental Figure 8. The persulfidation formation of DES1-His recombinant protein and its derivatives Cys44Ala, Cys205Ala, Cys44/205Ala, and Cys30/44/205Ala (supports Figure 3).
Supplemental Figure 9. Representative images of ABA on the guard cell H2S production of des1 rescued lines shown in Figure 3G (supports Figure 3).
Supplemental Figure 10. The effects of NaHS on the guard cell H2S production (A) and stomatal closure (B) of wild type, des1, and des1 rescued lines (supports Figure 3).
Supplemental Figure 11. Relative expression of RBOHD in wild type (Col-0), rbohD and lines expressing RBOHD under the control of tissue-specific promoters, in which RBOHD was expressed in mesophyll cells (pCAB3:RBOHD), epidermal (pML1:RBOHD), or guard cells (pMYB60:RBOHD) individually (supports Figure 4).
Supplemental Figure 12. Representative images of guard cell ROS production for rbohD and its rescued lines shown in Figure 4A (supports Figure 4).
Supplemental Figure 13. Genotyping for the des1 rbohD mutant (supports Figure 4).
Supplemental Figure 14. Stomatal responses of wild-type, des1 rbohD, and des1 rbohD rescued lines (supports Figure 4).
Supplemental Figure 15. Subcellular localization of RBOHD-C-GFP (supports Figure 5).
Supplemental Figure 16. Detection for the persulfidation of the RBOHD-C-terminal in vivo (supports Figure 5).
Supplemental Figure 17. A time-dependent decrease of persulfidation of RBOHD-C recombinant protein in vitro (supports Figure 5).
Supplemental Figure 18. Reversible persulfidation of RBOHD-C recombinant protein in vitro (supports Figure 5).
Supplemental Figure 19. Persulfidation of RBOHD-C recombinant protein and relative derivatives (supports Figure 5).
Supplemental Figure 20. Persulfidation of the RBOHC-C recombinant proteins and related derivatives (supports Figure 5).
Supplemental Figure 21. Relative expression of RBOHD in wild type (Col-0), rbohD, and lines expressing either a wild-type RBOHD or Cys825Ala, Cys890Ala, and Cys825/890Ala mutant derivatives driven by its native promoter individually (supports Figure 6).
Supplemental Figure 22. Immunoblot membrane scans for this study (supports Figures 1–3, and 5).
Supplemental Data Set 1. Oligonucleotide primers used in this study.
Supplemental Data Set 2. Materials used in this study.
Supplemental Data Set 3. Detailed statistical analyses of the results in this study.
DIVE Curated Terms
The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:
Acknowledgments
The authors thank Honghong Wu from Huazhong Agricultural University for his helpful advice on the NMT experiment. Y.X. thanks his wife Dan Chen, daughter Mengmeng, and family for their unwavering support and encouragement, without which this article could not have been published. The work was supported by the National Natural Science Foundation of China (grants 31670255 and 21922702), the Fundamental Research Funds for the Central Universities (grant KYZ201859), the National Natural Science Foundation of Jiangsu Province (grant BK20161447), and theEuropean Regional Development Fund through the Agencia Estatal de Investigación of Spain (grant BIO2016-76633-P).
AUTHOR CONTRIBUTIONS
J.S., J.Z., H.Z., B.C., J.Y., and Y.X., designed the study; J.S., J.Z., M.Z., H.Z., and L.F., performed the experiments; B.C., Q.P., C.G., L.C.R., L.F., J.Y., W.S., and Y.X. analyzed the data; C.G., L.C.R., C.H.F., and Y.X. wrote the article.
Footnotes
↵1 These authors contributed equally to this work.
- Received October 22, 2019.
- Revised January 17, 2020.
- Accepted February 5, 2020.
- Published February 5, 2020.
References
