Hydrogen sulfide - cysteine cycle system enhances cadmium tolerance through alleviating cadmium-induced oxidative stress and ion toxicity in Arabidopsis roots

Cadmium (Cd2+) is a common toxic heavy metal ion. We investigated the roles of hydrogen sulfide (H2S) and cysteine (Cys) in plant responses to Cd2+ stress. The expression of H2S synthetic genes LCD and DES1 were induced by Cd2+ within 3 h, and endogenous H2S was then rapidly released. H2S promoted the expression of Cys synthesis-related genes SAT1 and OASA1, which led to endogenous Cys accumulation. The H2S and Cys cycle system was stimulated by Cd2+ stress, and it maintained high levels in plant cells. H2S inhibited the ROS burst by inducing alternative respiration capacity (AP) and antioxidase activity. H2S weakened Cd2+ toxicity by inducing the metallothionein (MTs) genes expression. Cys promoted GSH accumulation and inhibited the ROS burst, and GSH induced the expression of phytochelatin (PCs) genes, counteracting Cd2+ toxicity. In summary, the H2S and Cys cycle system played a key role in plant responses to Cd2+ stress. The Cd2+ tolerance was weakened when the cycle system was blocked in lcddes1-1 and oasa1 mutants. This paper is the first to describe the role of the H2S and Cys cycle system in Cd2+ stress and to explore the relevant and specificity mechanisms of H2S and Cys in mediating Cd2+ stress.

Cadmium (Cd 2+ ) is a common toxic heavy metal ion in the environment. It greatly affects the growth and development of plants and is harmful to human health through the food chain 1,2 . Because of its carcinogenic properties and its detrimental effects on the growth of organisms, Cd 2+ contamination of agricultural soil has become a critical concern. Preventing reduced growth and accumulation of Cd 2+ in harvested organs of plants growing on Cd 2+ -contaminated soils has become an urgent task as it can contribute to food safety. Thus, it is important to explore plant stress defense mechanisms and to find ways to reduce the Cd 2+ accumulation in grains.
As a heavy metal not participating in redox reactions, Cd 2+ can easily dissolve in water and quickly be taken up by plant roots 3,4 . The physiological consequences of Cd 2+ toxicity in plants are chlorosis, stunted growth, and cell death, among others [5][6][7] . At the cellular level, Cd 2+ can alter protein structure and inhibit enzyme activity by binding to sulfhydryl and carbonyl groups and replacing essential co-factors of enzymes [7][8][9] . The overproduction of reactive oxygen species (ROS) is the primary response of plants to Cd 2+ with negative impact on cell function 10 . Further damage can be caused by ROS-independent, secondary mechanisms. Lipid peroxidation is the most deleterious effect caused by Cd 2+ -induced ROS 4 . Malondialdehyde (MDA), one of the decomposition products of lipid peroxidation, can modify active substrates in plant cells, including nucleic acids, proteins and saccharides 11 . To become resistant to Cd 2+ toxicity, plants have developed several strategies, such as inducing the

Results
Effect of Cd 2+ on root elongation, MDA and EL in Arabidopsis roots. Arabidopsis seedlings (7-d-old) were transferred aseptically to 1/2 MS medium-containing CdCl 2 , and the lengths of the primary roots were measured 5 d later. Cd 2+ stress led to toxicity symptoms and inhibited the elongation of Arabidopsis roots in a dose-dependent manner. As shown in Fig. 1a and b, root growth was slightly inhibited under 25 μ M Cd 2+ , but root elongation was significantly inhibited under 50 to 150 μ M Cd 2+ , exhibiting 53.5% to 34.9% inhibition, respectively. Malondialdehyde (MDA) and electrolyte leakage (EL) are considered as good indicators of stress-induced cell damage. Cd 2+ stress caused lipid peroxidation and MDA accumulation. When plants were treated with 50 to 150 μ M Cd 2+ , the MDA content of the roots increased by 117% to 200%, respectively (Fig. 1c). In the presence of 50 to 150 μ M Cd 2+ , EL increased by 131% to 217%, respectively (Fig. 1d). The 50 to 150 μ M Cd 2+ treatment had significant effects on Arabidopsis roots. 100 and 150 μ M Cd 2+ concentrations were too violent for plant growth and 150 μ M Cd 2+ concentrations lead to the seedling etiolation in Fig. 1a. Thus, 50 μ M Cd 2+ was chosen for further study of Cd 2+ stress.
Effect of Cd 2+ , NaHS and Cys on the H 2 S and Cys cycle system. To explore the working mechanisms of the H 2 S and Cys cycle system's response to Cd 2+ stress in Arabidopsis roots, a time-course analysis of endogenous H 2 S and Cys contents was performed. Endogenous H 2 S and Cys contents undulated along with the time of Cd 2+ stress. H 2 S content was rapidly induced after treatment with Cd 2+ for 3 h, reached the highest level at 9 h, and then decreased at 12 h, but had another increase at 36 h (Fig. 2a). Treatment with Cys could enhance the H 2 S level and maintain H 2 S content at a high level in Cd 2+ stress (Fig. 2a). The Cys level slightly decreased in the initial stage under Cd 2+ treatment, but it increased after treatment with Cd 2+ for 12 h and maintained a high level from 24 to 36 h (Fig. 2b). Treatment with NaHS promoted Cys accumulation and a high level of Cys was maintained during Cd 2+ stress (Fig. 2b). When plants were treated with 50 or 100 μ M Cd 2+ for a long time, both H 2 S and Cys levels were enhanced in Arabidopsis roots ( Fig. 2c and d). As mentioned above, H 2 S and Cys contents were elevated by Cd 2+ stress, and H 2 S appeared to be an important mediator in the Cd 2+ -induced increase of Cys, and the H 2 S and Cys cycle system was enhanced.
The effects of Cd 2+ , NaHS and Cys on synthetic genes of H 2 S and Cys. To study the direct effects of Cd 2+ , NaHS and Cys on genes regulating the synthesis of H 2 S and Cys, the Arabidopsis seedlings were exposed to various treatments for 3 h. The expressions of H 2 S synthetic genes LCD and DES1 were markedly induced by Cd 2+ and Cys ( Fig. 3a and b), but the expression of D-CDES was not significantly affected by Cd 2+ and Cys treatments (Fig. 3c). SATs and OASs are the important synthetic genes of Cys, but they had different responses to Cd 2+ , NaHS and Cys. The expression levels of SAT1 and OASA1 were slightly increased in Cd 2+ stress, but they were Scientific RepoRts | 6:39702 | DOI: 10.1038/srep39702 markedly induced by treatment with NaHS ( Fig. 3d and g). Additionally, the expression of SAT5 was inhibited by Cys (Fig. 3f). SAT3, OASB and OASC did not respond to Cd 2+ , NaHS or Cys (Fig. 3e,h and i).
The effects of Cd 2+ , NaHS and Cys on root elongation, MDA and EL in lcddes1-1 and oasa1 mutants. Five-day-old Arabidopsis seedlings were transferred aseptically to Cd 2+ -containing 1/2 MS medium, and the lengths of the primary roots were measured one week later. The root elongation of lcddes1-1 mutant was shorter compared to WT root elongation under control conditions, but the root elongation of oasa1 was the same as WT ( Fig. 4a and b). The lcddes1-1 and oasa1 mutants were more sensitive to Cd 2+ stress. Application of NaHS or Cys recovered the Cd 2+ -induced growth inhibition in WT. NaHS markedly recovered the effect of Cd 2+ in lcddes1-1, but it only partly recovered the effect of Cd 2+ in oasa1 ( Fig. 4a and b). On the contrary, treatment with Cys slightly recovered the effect of Cd 2+ in lcddes1-1, but it significantly recovered the effect of Cd 2+ in oasa1 ( Fig. 4a and b). Cys-mediated partial recovery the root length may be due to an independent physiological action of Cys in lcddes1-1 because H 2 S production was blocked in the double mutant. NaHS or Cys treatment markedly decreased the EL level and the content of MDA under Cd 2+ stress in WT ( Fig. 4c and d). NaHS strongly reduced the MDA content and the EL in lcddes1-1 and oasa1 ( Fig. 4c and d). Cys prevented the effects of Cd 2+ on the MDA content and EL in oasa1 but partly weakened the effects of Cd 2+ on the MDA content and EL in lcddes1-1 ( Fig. 4c and d).
The effects of Cd 2+ , NaHS and Cys on the alternative respiratory pathway. The alternative respiratory pathway plays an important role in plant stress resistance by limiting the ROS burst 38 . In this study, we sought to elucidate the roles of NaHS and Cys in the alternative respiratory pathway under Cd 2+ stress. In general, the alternative pathway operates at a low level under normal conditions, but it can be significantly induced when plants are stimulated by various environmental stresses 23 . We first checked the expression of AOX genes after Cd 2+ , NaHS and Cys treatments for 3 h. The expression of AOX1A, AOX1C and AOX2 were increased in Cd 2+ stress (Fig. 5a,b and c). Interestingly, NaHS and Cys treatments also markedly enhanced the expression levels of AOX1A, AOX1C and AOX2 in both control and Cd 2+ stress conditions (Fig. 5a,b and c). Furthermore, the total respiration capacity (TP), cytochrome respiration capacity (CP) and alternative respiration capacity (AP) were analyzed in WT and mutant plants. TP was slightly enhanced by 25 or 50 μ M Cd 2+ , but it was inhibited by 100 or 150 μ M Cd 2+ in WT (Fig. 5d). Under Cd 2+ stress, CP was inhibited in a dose-dependent manner; however, AP was increased under Cd 2+ stress, and AP achieved its maximum induction with the 50 μ M Cd 2+ treatment (Fig. 5e). Similar to the pattern of expression of the AOX genes in response to NaHS or Cys treatments under Cd 2+ stress, AP was increased by NaHS or Cys under Cd 2+ stress (Fig. 5f). However, it was different in the mutant plants. Under Cd 2+ stress, the effects of Cys on AP were not observed in lcd and des1-1, and they were especially decreased in lcddes1-1 (Fig. 5g,h and i). In oasa1, the effects of NaHS and Cys were the same as in WT under Cd 2+ stress (Fig. 5g).
The effects of Cd 2+ , NaHS and Cys on antioxidant enzyme activity and GSH level, and the relationship among AP, antioxidant enzyme activity, and GSH level in Cd 2+ stress. Antioxidant enzymes depress the level of ROS. A previous study showed that H 2 S could enhance antioxidase activity in rice 39 . In addition, many studies suggested that AOX was important in maintaining the homeostasis of the redox state 22,38 . Therefore, the effects of Cd 2+ , NaHS, Cys and AP on antioxidant enzyme activity were analyzed. As shown in Fig. 6a and c, after 12 h of 50 μ M Cd 2+ treatment, the activities of SOD and CAT in plants were significantly higher than in the control plants in WT. NaHS or Cys treatments could enhance the antioxidase activity under unstressed conditions ( Fig. 6b and d), and this enhancement was further strengthened under Cd 2+ stress in WT ( Fig. 6a and c). However, treatment with n-propyl gallate (nPG) had no significant effect on the antioxidase activity of the plants either under Cd 2+ stress or under unstressed conditions. Furthermore, nPG did not affect the elevated antioxidase activity of the NaHS-and Cys-treated plants under Cd 2+ stress ( Fig. 6a and c). The effects of Cd 2+ and Cys were altered in lcddes1-1; treatment with Cd 2+ or Cys did not enhance the antioxidase activity in lcddes1-1 ( Fig. 6b and d). Additionally, the effects of Cd 2+ on the SOD activity were also weakened, and CAT activity was negligible in oasa1 (Fig. 6d). Contrarily, treatment with NaHS still enhanced the antioxidase activity in mutant plants ( Fig. 6b and d).
GSH is the product of sulfur metabolism, and it has positive biological functions in plant responses to heavy metal stress and oxidative stress 25 . As shown in Fig. 6e, the GSH content was increased in Cd 2+ stress. NaHS and Cys also enhanced the GSH level in WT ( Fig. 6e and f). Specially, Cys had a significant promoting effect on GSH content. The oasa1 mutant did not respond to Cd 2+ and NaHS, and even had a reduced GSH level, but Cys still increased the GSH content in oasa1 (Fig. 6f). Additionally, the effect of Cd 2+ , NaHS and Cys on the GSH content in lcddes1-1 was the same as WT plants (Fig. 6f).

Effect of NaHS and Cys on ROS, and the relationship between AP and ROS in Cd 2+ stress.
To estimate the potential role of the H 2 S and Cys cycle in ROS homeostasis, we visualized the production of H 2 O 2 in the roots under Cd 2+ stress. Over-accumulation of H 2 O 2 was visualized by fluorescence labeling in the roots subjected to Cd 2+ stress (Fig. 7a). Conversely, NaHS or Cys treatment considerably diminished the accumulation of H 2 O 2 in Cd 2+ stress ( Fig. 7a and b). Additionally, inhibiting the alternative respiratory pathway with nPG caused an over-accumulation of H 2 O 2 under Cd 2+ stress. The effects of NaHS and Cys were partly averted and slightly weakened by nPG in Cd 2+ stress, respectively ( Fig. 7a and b). As shown in the time-course of H 2 O 2 , the ROS burst occurred during the early phase of Cd 2+ stress. Then, high levels of ROS were maintained from 4 to 8 h and declined after 12 h. H 2 S supplementation could maintain H 2 O 2 at a low level during Cd 2+ stress. Treatment with Cys did not alter the burst of H 2 O 2 in the early phase, but it prevented the over-accumulation of H 2 O 2 after 6 h (Fig. 7c).
Effect of NaHS and Cys on Cd 2+ accumulation. The role of H 2 S and Cys in Cd 2+ homeostasis was investigated by measuring the percentage of Cd 2+ in the root. The results in Fig. 8a show that Cd 2+ accumulation increased in roots under Cd 2+ stress in WT and in the mutants, but the mutants accumulated more Cd 2+ than the WT. NaHS or Cys supplementation had inhibitory effects on Cd 2+ uptake and accumulation in WT and oasa1. Nevertheless, lcddes1-1 did not respond to the effect of NaHS under Cd 2+ stress, but the effect of Cys on Cd 2+ uptake and accumulation was only partially reduced in lcddes1-1.  NaHS promoted the expression of MT1A, MT1B and MT2B (Fig. 8b). To further study the effect of the H 2 S and Cys cycle system on the heavy metal chelator genes, the time-course of PCS1 and MT1A gene expression was investigated. Cd 2+ was found to up-regulate the expression of PCS1 and MT1A genes at 0.5 h, which then remained at a high expression level. Cys enhanced the expression of the PCS1 gene at 0.5 h, which reached a maximum by 1 h, but NaHS enhanced the expression of the PCS1 gene at 6 to 12 h. The expression of MT1A was different from PCS1. After 3 h of Cys supplementation, the expression of MT1A started to increase, reaching a maximum at 6 h, and NaHS enhanced the expression of MT1A gene at 0.5 h (Fig. 8c and d).

Discussion
The root is the primary organ that plants deploy to accumulate most of the heavy metals to which they are exposed 40,41 . Sulfur metabolism is required for the growth and development of plants, and the production of sulfur metabolites also plays a critical role in plant responses to heavy metal-induced stress 25 . H 2 S and Cys are important sulfur metabolism products that participate in suppressing heavy metal stress in plants 40 . In previous reports, H 2 S and Cys were always studied separately in plant responses to abiotic stress 39,42 . Recently, H 2 S and H 2 S-induced Cys accumulation were reported to be critical in imparting Cr 6+ tolerance in Arabidopsis 43 . Therefore, the H 2 S and Cys cycle is an important system for regulating H 2 S and Cys functions in heavy metal stress. In this study, we used the lcddes1-1 and oasa1 Arabidopsis mutants to block the H 2 S and Cys cycle system. Then, we intensively researched the relevant and specificity roles of H 2 S and Cys in Cd 2+ stress. Our results indicated that Cd 2+ can rapidly accumulate in Arabidopsis roots and inhibit the primary root growth in a Cd 2+ concentration-dependent manner (Figs 1a and 7a), suggesting that Cd 2+ is easily absorbed and highly toxic. We observed that endogenous H 2 S and Cys levels undulate from 3 to 48 h under Cd 2+ stress (Fig. 2). However, the endogenous patterns of change were different for H 2 S and Cys levels. Endogenous H 2 S was first induced by Cd 2+ stress, and then Cys levels increased. On this account, we suppose that H 2 S is produced rapidly under Cd 2+ stress and that it acts as second messenger to activate the synthesis of Cys, implying that Cd 2+ stress could be the direct cause of endogenous H 2 S release but that Cys accumulation is a secondary effect of Cd 2+ stress. Data for the expression of H 2 S and Cys synthetic genes supports this hypothesis. The expression of H 2 S synthetic genes was directly induced by Cd 2+ , and then, exogenous H 2 S supplementation induced the upregulation of Cys synthetic-related genes (Fig. 3). Additionally, exogenous H 2 S or Cys supplementation during Cd 2+ stress could rapidly induce mutual endogenous levels of the other contents in Cd 2+ stress. These results suggested that the H 2 S -Cys cycle system was triggered by Cd 2+ and that H 2 S and Cys could promote the production of each other, forming a cycle of activation. Finally, treatment with Cd 2+ for 5 d, H 2 S and Cys contents increased significantly in Arabidopsis roots.
The expression levels of the Cys synthesis-related genes OASA1 and SAT1 were up-regulated significantly by H 2 S treatment, and the H 2 S synthesis genes LCD and DES1 were up-regulated significantly by Cys treatment (Fig. 3). OASA1 directly regulated Cys synthesis, and LCD and DES1 directly regulated H 2 S synthesis; thus, lcd, des1-1, lcddes1-1 and oasa1 were used to study the H 2 S and Cys cycle system in Cd 2+ stress. The Cd 2+ -induced root shortening and increases in MAD and EL were markedly enhanced in mutant plants, suggesting that the Cd 2+ resistance was weakened when the H 2 S and Cys cycle was blocked. Exogenous H 2 S or Cys supplementation only partly restored the root growth, MAD and EL levels, suggesting that H 2 S or Cys alone could not replace the function of the H 2 S and Cys cycle in plant cells. Additionally, the H 2 S and Cys system is also important for stress caused by other heavy metals, such as Cr 6+ ; it was reported that NaHS treatment increases the expression levels of the Cys synthesis-related genes 43 . However, different heavy metal stress condition lead to the difference genes expression include MTs genes 43 , thus Cd 2+ and Cr 6+ condition also could lead to the difference MTs genes expression. The details regarding the mechanism of H 2 S in heavy metal resistance requires further study.
Excessive Cd 2+ can induce the production of ROS, which is highly toxic to biomembranes, nucleic acids and proteins 11 . The alternative respiratory pathway plays an important role in stress conditions by repressing the production of ROS 22,23,42 . Our study also found that the CP and AP activities were altered by Cd 2+ stress (Fig. 5e). Plant signaling molecules, such as nitric oxide, can regulate the alternative respiratory pathway in stress conditions 44 . Whether an H 2 S signal or Cys could affect AP activity was not previously known; our analysis found that exogenous H 2 S or Cys supplementation could further induce the activity of AP in Cd 2+ stress. However, in H 2 S synthesis mutants, the effect of Cys was negligible, and in Cys synthesis mutants, the effect of H 2 S was not altered. These data imply that H 2 S is a direct trigger of AP activity and that Cys might play an indirect role in Cd 2+ stress. The expression of AOX genes was also induced by H 2 S within 3 h, but not by Cys.
Antioxidases are also one of the central elements in maintaining ROS levels in plant cells 45 . We investigated the connection between the alternative respiratory pathway and antioxidases, but we found that the activities of SOD and CAT were not altered when the alternative respiratory pathway was inhibited by nPG ( Fig. 6a and c), suggesting that the alternative respiratory pathway and antioxidases have independent functions in response to Cd 2+ stress. The activities of SOD and CAT were induced by Cd 2+ and increased Cd 2+ resistance (Fig. 6). H 2 S or Cys biosynthesis was necessary for the increase in SOD and CAT activities in response to Cd 2+ stress because Cd 2+ -induced activities of SOD and CAT were weakened in H 2 S and Cys synthesis mutants. We further studied the relationship of H 2 S and Cys in this physiological process. H 2 S supplementation could remedy the deficiency of Cys biosynthesis and increase the activities of SOD and CAT in oasa1 mutants, but Cys supplementation could not. These data suggest that the activities of SOD and CAT are directly regulated by H 2 S and that Cys indirectly affects the activities of SOD and CAT by promoting the generation of H 2 S.
GSH performs numerous physiological functions in the plant response to heavy metal stress 46 . Cys is a precursor of GSH, which stores and transports GSH via the γ -glutamyl cycle 47 . In this study, supplementation with exogenous H 2 S or Cys strengthened Cd 2+ -mediated GSH elevation in WT plants (Fig. 6f). It is interesting that the effects of Cd 2+ and H 2 S were reversed in oasa1, but the effects of Cd 2+ and H 2 S were not altered in lcddes1-1 (Fig. 6f). These results suggest that Cys is a direct regulatory factor of GSH, and H 2 S affects GSH levels indirectly. Additionally, the GSH content was not altered by nPG (Fig. 6e), suggesting that the alternative respiratory pathway and GSH are not related in their responses to Cd 2+ stress.
Cd 2+ enrichment was also observed in this study (Fig. 8a). Inhibiting Cd 2+ uptake and enhancing Cd 2+ efflux are the main defense strategies that plant cells use to prevent Cd 2+ toxicity. Exogenous H 2 S or Cys supplementation effectively inhibited the accumulation of Cd 2+ (Fig. 8a). When endogenous H 2 S or Cys synthesis was blocked, Cd 2+ over-accumulation occurred (Fig. 8a), suggesting that the H 2 S and Cys cycle system is important for inhibiting Cd 2+ uptake or enhancing Cd 2+ efflux. Additionally, the effect of Cys was partly inhibited in the lcddes1-1 mutant, implying that the role of H 2 S in the H 2 S and Cys cycle might be to directly regulate Cd 2+ uptake or efflux.
The generation of chelators is also an effective pathway in plant cells for avoiding Cd 2+ toxicity. PCS1, PCS2, MT1A, MT1B and MT2B are mainly expressed in roots and regulate PCs and MTs synthesis; the expression of these chelators is generally induced by numerous heavy metal ions 42,43,48 . Interestingly, the expression of PCS1 and PCS2 was found to be induced by Cys in a very short time, and the expression of MT1A, MT1B and MT2B was induced by H 2 S (Fig. 8b). However, only long-term supplementation of Cys or H 2 S induced the expression of PCS1 and MT1A (Fig. 8c,d). These data suggest that the generation of chelators can be regulated differently in plant cells. Cys and H 2 S played different roles in the physiological process, but when combined Cys and H 2 S mutually promoted the expression of chelator synthesis genes to a level higher than when they were used as separate supplements.
Based on the data described above, a signal pathway model was developed and is depicted in Fig. 9. It shows the specific roles of H 2 S and Cys in regulating plant responses to Cd 2+ stress and their interaction. H 2 S is activated much earlier than Cys in plant responses to Cd 2+ stress, acting as a secondary messenger to increase Cys accumulation by regulating the transcription levels of SAT1 and OASA1. In addition, the production of H 2 S might deplete the endogenous Cys pool, which might subsequently increase the expression of SAT1 or OASA1. Furthermore, once the H 2 S and Cys cycle is initiated, it works to maintain elevated H 2 S and Cys levels. H 2 S inhibits the ROS burst by promoting CP and antioxidase activities, and it weakens Cd 2+ ion toxicity by inducing the gene expression of MTs. Cys acts as a precursor of GSH to promote GSH accumulation, which then contributes to inhibiting the ROS burst. GSH also induces genes expression of PCs, leading to raised PC activity, which counteracts Cd 2+ ion toxicity. In sum, the H 2 S and Cys cycle system is a key regulator of the response to Cd 2+ stress in plants that acts to induce and maintain levels of bioactive molecules (H 2 S, Cys, GSH, PCs, and MTs) that improve plant resistance to Cd 2+ stress.

Materials and Methods
Plant material and chemical treatments. This study was carried out on Arabidopsis thaliana, including wild ecotypes Columbia (Col-0) and the lcd (SALK_082099), des1-1 (SALK_103855), lcddes1-1 and oasa1 (SALK_074242c) mutants. Seeds were surface sterilized with 70% ethanol for 30 s and 15% sodium hypochlorite for 15 min and were washed at least five times with sterilized water before sowing on solid 1/2 Murashige and Skoog (MS) medium (pH 5.7), which contained 1% (w/v) sucrose, and 0.8% (w/v) agar. After that, the seeds were vernalized for 48 h at 4 °C. Then, the seedlings were grown in a growth room, which had the temperature set at 22 °C and a 14/10 h light/dark photoperiod under a photon flux density of 120 mmol m −2 s −1 . The Arabidopsis plants used throughout this work were grown routinely in a growth chamber under 50-60% humidity.

Root elongation assays.
Seven-day-old Arabidopsis seedlings grown on the vertical 1/2 MS agar plates were transferred to the 1/2 MS agar medium containing various chemicals for the different treatments. Root elongation was measured after 5 d of various treatments. All experiments were repeated at least three times, with photographs collected at 7 d from one representative experiment being shown. The root length was measured with ImageJ.
Electrolyte leakage assay. Measurement of ion leakage was determined according to Sairam and Srivastava (2002) with some modifications 43 . The 7-d-old Arabidopsis seedlings were treated for 5 d on the 1/2 MS agar medium containing different chemicals. Following the treatments, the roots were collected and washed in deionized water three times to remove surface-adhered electrolytes. Then, they were immersed in 10 ml deionized water for 3 h at 25 °C in test tubes. After the incubation, the conductivity in the bathing solution was determined (C 1 ), and the conductivity of deionized water was also determined (C 0 ). The samples were heated in boiling water for 1 h before the total conductivity was measured in the bathing solution (C 2 ). Relative ion leakage was expressed as a percentage of the total conductivity after heating in boiling water [relative ion leakage = ( MDA and GSH content assays. The chemical treatments were the same as the measurements of ion leakage.
Following the treatments 49 , Arabidopsis roots were collected. Lipid peroxidation of the roots was measured by estimating the MDA content according to the method of Heath and Packer. The GSH content was measured based on a previously described method 49 .
Measurement of H 2 S content. H 2 S quantification was performed as described by Nashef et al. 50 . The chemical treatments were the same as the methods of ion leakage. Following the treatments, the seedling roots were collected with liquid nitrogen and ground into fine powder with mortar and pestle; 0.3 g of frozen tissue was homogenized in 1 ml 100 mM potassium phosphate buffer (pH = 7), which contained 10 mM ethylenediaminetetra-acetic acid (EDTA). The homogenates were centrifuged at 15,000 × g for 20 min at 4 °C, and 100 μ l of supernatant was used for the quantification of H 2 S in an assay mixture containing 1,880 μ l extraction buffer and 20 μ l of 20 mM 5,5′ -dithiobis (2-nitrobenzoic acid), for a total volume of 2 ml. The assay mixture was incubated at room temperature for 2 min, and the absorbance was read at 412 nm. H 2 S was quantified based on a standard curve of known concentrations of NaHS.

Measurement of the Cys content.
The chemical treatments were the same as the measurements of ion leakage. Following the treatments, Arabidopsis roots were collected. Cys can react specifically with acid ninhydrin, and the product was extracted by methylbenzene, which has a maximum absorbance at 560 nm. The reaction is highly sensitive for Cys determination. Thus, the Cys content could be determined as described previously 51 .
RNA isolation and qRT-PCR. Seven-day-old Arabidopsis seedlings were transferred to the 1/2 MS agar medium containing different chemicals and treated for 0-12 h. Following the treatments, roots of Arabidopsis were harvested to extract total RNA for real-time polymerase chain reaction (RT-PCR). Total RNA was extracted using an RNAprep pure plant kit (Tiangen, Beijing) and was treated with RNase-free DNase reagent (RNase-free DNase kit, Tiangen). The total RNA was reverse-transcribed into first-strand cDNA using PrimeScript ™ Reverse Transcriptase (Takara, Japan) and Oligo (dT) 15 primer (Takara) following the manufacturer's instructions. The samples were amplified using SYBR Green I (SYBR ® Premix Ex Taq ™ Kit, Takara). The housekeeping gene EF1A was used as an internal control. The thermal cycle used was as follows: 95 °C for 10 s, and 40 cycles of 95 °C for 5 s and 59 °C for 25 s. This was followed by 80 cycles of 10 s during the time elapsed at 55-95 °C. The PCR amplifications for each gene were performed in triplicate. The results were analyzed by Rotor-Gene Real-Time Analysis Software 6.1 (Build 81).
Extraction and assay of antioxidant enzymes. Seven-day-old Arabidopsis seedlings were transferred to the 1/2 MS agar medium containing different chemicals and treated for 6 h. Following the treatments, Arabidopsis roots were collected and enzymes extracted according to the method of Mostofa et al. 51 . Activities of antioxidase and glyoxalase were determined by the standard methods reported in Mostofa and Fujita 52 for SOD (EC 1.15.1.1) and CAT (EC 1.11.1.6). The protein standard was bovine serum albumin (BSA), which was employed to determine the protein content.
Determination of H 2 O 2 contents. H 2 O 2 was visualized using the specific H 2 O 2 fluorescent probe dichlorofluorescein diacetate (H 2 DCF-DA) according to the method described by Maffei et al. 53 . Seven-day-old Arabidopsis seedlings were transferred to the 1/2 MS agar medium containing different chemicals and treated for 0-24 h. Following the treatments, Arabidopsis seedlings were incubated in the reaction buffer containing 10 mM 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid (HEPES)-NaOH (pH 7.5) and 10 μ M H2DCF-DA for 15 min at 25 °C. Thereafter, the roots were washed three times with the HEPES-NaOH buffer (pH 7.4) prior to visualization using a laser confocal scanning microscope (Leica SM IRBE Multisync FE 1250). Excitation was at 480 nm and emission was at 520 nm. Images were processed and analyzed using the Leica Tcs SP2 software.