The Early Oxidative Stress Induced by Mercury and Cadmium Is Modulated by Ethylene in Medicago sativa Seedlings

Cadmium (Cd) and mercury (Hg) are ubiquitous soil pollutants that promote the accumulation of reactive oxygen species, causing oxidative stress. Tolerance depends on signalling processes that activate different defence barriers, such as accumulation of small heat sock proteins (sHSPs), activation of antioxidant enzymes, and the synthesis of phytochelatins (PCs) from the fundamental antioxidant peptide glutathione (GSH), which is probably modulated by ethylene. We studied the early responses of alfalfa seedlings after short exposure (3, 6, and 24 h) to moderate to severe concentration of Cd and Hg (ranging from 3 to 30 μM), to characterize in detail several oxidative stress parameters and biothiol (i.e., GSH and PCs) accumulation, in combination with the ethylene signalling blocker 1-methylcyclopropene (1-MCP). Most changes occurred in roots of alfalfa, with strong induction of cellular oxidative stress, H2O2 generation, and a quick accumulation of sHSPs 17.6 and 17.7. Mercury caused the specific inhibition of glutathione reductase activity, while both metals led to the accumulation of PCs. These responses were attenuated in seedlings incubated with 1-MCP. Interestingly, 1-MCP also decreased the amount of PCs and homophytochelatins generated under metal stress, implying that the overall early response to metals was controlled at least partially by ethylene.


Introduction
The release of cadmium (Cd) and mercury (Hg) to the environment, caused mainly by mining, metallurgy, and other industrial activities, represents a major problem to the environment and human health [1,2]. Cadmium and Hg are extremely persistent pollutants and highly toxic to most living organisms, even at low concentrations [3]. Plants exposed to Cd and Hg suffer from growth inhibition, damages to the cellular metabolism, and, ultimately, cell death [4]. One of the major alterations caused by the accumulation of these metals is the induction of oxidative stress, through the generation of reactive oxygen species (ROS), such as superoxide (O 2 •− ) and hydrogen peroxide (H 2 O 2 ) [5,6]. This oxidative stress can be early induced by Cd and Hg, which is detected as rapid accumulation of H 2 O 2 in roots after only 1 to 3 h of treatment [7,8]. Several cellular sources of ROS are envisaged, among them are the activation of plasma membrane NADPH oxidases that release apoplastic H 2 O 2 [9,10], and mitochondria, where intracellular ROS are produced [11]. In turn, it is thought that apoplastic H 2 O 2 generated under metal stress feeds extracellular basic POXs, which enzymatic activity augments under Cd and Hg stress [12,13]. These enzymes catalyse lignification and cross-linking of cell wall components, enhancing cell wall stiffness arresting cell elongation [14]. Additionally, the accumulation of ROS leads to drastic alterations in cellular components, causing lipid peroxidation and protein oxidation [15,16]. Oxidative damage of proteins under harmful environmental conditions affect fundamental metabolic triggered by Cd and Hg that are mediated by ethylene, which could help to optimize plant tolerance to these contaminants.

Plant Material and Hydroponic System
Seeds of Medicago sativa 'Aragon' were purchased from "Semillas Mur S.L." (VAT B50773605), sterilized, and grown as described by Ortega-Villasante et al. [9] in Murashige-Skoog (MS) nutrient solution. After 24 h acclimation, seedlings were treated with CdCl 2 or HgCl 2 (3, 10, and 30 µM) for 3, 6, and 24 h. In the ethylene signalling inhibition experiments, alfalfa seedlings were grown in identical conditions, but were pre-incubated with 10 µM 1-methylecyclopropene (1-MCP) 24 h prior to metal exposure. 1-MCP was also added to the corresponding growing media during the selected Hg and Cd treatments. We used a microscale hydroponic system where 25 seedlings were grown per biological replicate of each sample, with at least four full independent experiments [7]. Shoots and roots were frozen in liquid nitrogen and stored at −80 • C until analysis (see Supplementary Information for further details).

Redox Enzymatic Activities in Gel
Enzymatic extracts were prepared with 0.5 g of frozen material by homogenisation in an ice-cooled mortar and pestle with 1 mL of extracting mixture (10 mL 30 mM MOPS at pH 7.5, 5 mM EDTA-Na 2 , 10 mM DTT, 10 mM ascorbic acid, 0.6% PVP, 10 µM PMSF and protease inhibitor cocktail). After 15 min centrifugation at 12,000× g at 4 • C, the supernatant was stored at −80 • C in single-use aliquots of 100 µL. Protein concentration was determined using the Bio-Rad Protein Assay, and protein loading was re-adjusted using denaturing polyacrylamide gel electrophoresis (SDS-PAGE; [27]) and Coomassie blue staining, prior to specific in gel enzymatic activity staining after non-denaturing polyacrylamide gel electrophoresis (ND-PAGE).

Root Extracellular H 2 O 2 Generation
Extracellular H 2 O 2 release was measured in root segments (1 cm) according to the method of Ortega-Villasante et al. [9]. After washing and equilibration in MS medium buffered with 2 mM MES at pH 6.0 in the dark for 1 h, root segments were individually placed in a 96-well microtitre plate containing the same medium supplemented with 50 µM Amplex Red (Molecular Probes, Eugene, OR, USA). Fluorescence was recorded at λ exc = 542 nm and λ em = 590 every 5 min for 6 h using a Synergy HT Biotek plate reader (Winooski, VT, USA).

Dye Loading and Confocal Laser Scanning Microscopy
Cellular oxidative stress degree was visualised by staining with 10 µM 2 ,7dichlorofluorescein diacetate (H 2 DCFDA), and cell death was detected with 25 µM propidium iodide (PI), which were observed with Leica TCS SP2 confocal microscope (Wetzlar, Germany) as described by Ortega-Villasante et al. [7]. Microscopy images are representative observations of at least three independent experiments.

Image and Statistical Analysis
Densitometry analysis of the bands were performed using the ChemiDoc™ XRS+ System and ImageLab Software (BioRad), according to manufacturer's specifications. Only relevant differences are presented, and representative gels of three independent assays are shown. A one-way ANOVA statistical analysis with post hoc Duncan test was performed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). Results were expressed as mean ± standard error, and differences were considered significant at p < 0.05.

Early Stress Responses to Cd and Hg
Alfalfa seedlings exposed to Cd and Hg suffered remarkable inhibition of growth concomitantly with metal concentration and time of treatment (Supplementary Figure S1). The strongest growth inhibition was observed in seedlings treated with Hg, after 3 h with 30 µM Hg, reaching the maximum inhibition (40%) after 24 h ( Figure 1A). On the other hand, the growth inhibition was below 15% in seedlings treated with 30 µM Cd for 24 h ( Figure 1B). Metal concentration increased in parallel to metal dose and time of exposure. However, Hg accumulation reached high values already after 3 h of treatment (50 to 75% of the concentration found after 24 h; Figure 1C), while Cd concentration augmented remarkably in seedlings exposed for 24 h, paralleling the Cd dose supplied in the medium ( Figure 1D).

Early Stress Responses to Cd and Hg
Alfalfa seedlings exposed to Cd and Hg suffered remarkable inhibition of growth concomitantly with metal concentration and time of treatment (Supplementary Figure S1). The strongest growth inhibition was observed in seedlings treated with Hg, after 3 h with 30 µ M Hg, reaching the maximum inhibition (40%) after 24 h ( Figure 1A). On the other hand, the growth inhibition was below 15% in seedlings treated with 30 µ M Cd for 24 h ( Figure 1B). Metal concentration increased in parallel to metal dose and time of exposure. However, Hg accumulation reached high values already after 3 h of treatment (50 to 75% of the concentration found after 24 h; Figure 1C), while Cd concentration augmented remarkably in seedlings exposed for 24 h, paralleling the Cd dose supplied in the medium ( Figure 1D). The amount of HSP70, sHSP17.7 Class II, and sHSP17.6 Class I was studied as biomarkers of cellular damage [21,23] Figure 2c,f,i) was found under metal stress. The amount of both sHSPs increased sharply even after only 3 h of treatment, indicating that the accumulation of these types of stress-related chaperones is promoted early by metal stress. However, this increase was more remarkable for Hgtreated seedlings, independently of the isoform and dose supplied, showing saturation profile. Nonetheless, Cd induction was delayed and milder, especially for sHSP 17.6 ( Figure 2h). The amount of HSP70, sHSP17.7 Class II, and sHSP17.6 Class I was studied as biomarkers of cellular damage [21,23] induced by toxic elements by Western-blotting ( Figure 2). The amount of HSP70 was not affected at any exposure time (Figure 2a,d,g), while a remarkable induction of sHSP 17.7 (Figure 2b,e,h) and sHSP17.6 ( Figure 2c,f,i) was found under metal stress. The amount of both sHSPs increased sharply even after only 3 h of treatment, indicating that the accumulation of these types of stress-related chaperones is promoted early by metal stress. However, this increase was more remarkable for Hg-treated seedlings, independently of the isoform and dose supplied, showing saturation profile. Nonetheless, Cd induction was delayed and milder, especially for sHSP 17.6 ( Figure 2h).
The greatest changes in redox enzymatic activities were found in the roots of alfalfa seedlings (Figure 3), rather than cotyledon seedlings (Supplementary Figure S3). APX activity was severely inhibited in seedlings treated with 30 µM Hg after just 3 h, albeit APX protein levels did not change appreciably (Figure 3a,b). Additionally, an earlier and stronger inhibition of GR activity was detected after only 3 h of treatment with 3 µM (Figure 3c). GR inhibition was almost complete after 6 to 24 h of exposure to 10 and 30 µM Hg, with an extremely faint band observed in 3 µM Hg-treated seedlings (Figure 3g,k). Interestingly, the inhibition of GR activity by 30 µM Hg was accompanied with a higher accumulation of GR protein (Figure 3h The greatest changes in redox enzymatic activities were found in the roots of alfalfa seedlings (Figure 3), rather than cotyledon seedlings (Supplementary Figure S3). APX activity was severely inhibited in seedlings treated with 30 µ M Hg after just 3 h, albeit APX protein levels did not change appreciably (Figure 3a  . Induction of small Heat Shock Proteins (sHSPs) in roots of Medicago sativa seedlings treated with Hg or Cd (0, 3, 10, and 30 µM) for 3, 6 and 24 h in the microscale hydroponic system, detected by immunostaining: α-HSP70 (a,d,g), α-sHSP17.7 Class II (b,e,h), and α-sHSP17.6 Class I (c,f,i), with apparent molecular (K; KDa) weight of bands of interest. Band intensity was normalised against Coomassie blue general protein staining after denaturing PAGE, corresponding to L (protein loading) bands. The numbers show the fold change relative to control samples, and green coloured figures represent fold-changes above 30%.
Cd exposure did not appreciably change alkaline POX activity (Figure 3m,o,q), but NADPH-oxidases were activated after 3 to 6 h with 10 and 30 µM Cd (Figure 3n,p). Interestingly, NADPH-oxidase activity also increased in seedlings given the lowest dose of Hg (3 µM) after 6 or 24 h treatment (Figure 3p,r). On the contrary, POX and NADPHoxidase activities were strongly inhibited in seedlings grown with 10 and 30 µM Hg, even after only 3 h of treatment (Figure 3m  Redox enzymatic activities determined in gel after ND-PAGE, of alfalfa seedling treated with 0, 3, 10, and 30 µ M Hg or Cd for 3, 6, and 24 h. Identification of each band labelle lower case letters: APX activity (a,e,i), α-APX (28 KDa apparent molecular weight) immuno tion (b,f,j), GR activity (c,g,k), α-GR (60 KDa apparent molecular weight) immunodetection alkaline POX after isoelectric focusing at the immobilised pH (IP) 7.0 to 10.0 range (m,o,q plasmalemma NADPH-oxidase (n,p,r). Band intensity was normalised against Coomassie blu eral protein staining after denaturing PAGE, corresponding to L (protein loading) bands. The bers represent the fold change relative to control samples. Green (up) and red (down) rep fold-changes greater than 30%.
Cd exposure did not appreciably change alkaline POX activity (Figure 3m,o,q NADPH-oxidases were activated after 3 to 6 h with 10 and 30 µ M Cd (Figure 3n,p). estingly, NADPH-oxidase activity also increased in seedlings given the lowest dose The exposure of alfalfa seedlings to Cd and Hg led to root extracellular H2O2 g ation and induction of oxidative stress and cell death in root epidermal cells, in a and metal dose dependent manner ( Figure 4). All metal treatments increased H2O2 g ation after 3 h. Interestingly, the high production of H2O2 in roots of seedlings expo 10 and 30 µ M Hg decreased in prolonged (6 and 24 h) treatments, even below the c seedlings ( Figure 4A). On the other hand, the highest fluorescence intensity occur . Redox enzymatic activities determined in gel after ND-PAGE, of alfalfa seedlings roots treated with 0, 3, 10, and 30 µM Hg or Cd for 3, 6, and 24 h. Identification of each band labelled with lower case letters: APX activity (a,e,i), α-APX (28 KDa apparent molecular weight) immunodetection (b,f,j), GR activity (c,g,k), α-GR (60 KDa apparent molecular weight) immunodetection (d,h,l), alkaline POX after isoelectric focusing at the immobilised pH (IP) 7.0 to 10.0 range (m,o,q), and plasmalemma NADPH-oxidase (n,p,r). Band intensity was normalised against Coomassie blue general protein staining after denaturing PAGE, corresponding to L (protein loading) bands. The numbers represent the fold change relative to control samples. Green (up) and red (down) represent fold-changes greater than 30%.
The exposure of alfalfa seedlings to Cd and Hg led to root extracellular H 2 O 2 generation and induction of oxidative stress and cell death in root epidermal cells, in a timed and metal dose dependent manner ( Figure 4). All metal treatments increased H 2 O 2 generation after 3 h. Interestingly, the high production of H 2 O 2 in roots of seedlings exposed to 10 and 30 µM Hg decreased in prolonged (6 and 24 h) treatments, even below the control seedlings ( Figure 4A). On the other hand, the highest fluorescence intensity occurred in seedlings exposed to 30 and 10 µM Cd for 24 h, with intermediate extracellular H 2 O 2 generation in those treated with 3 µM Hg for 24 h ( Figure 4A). The acute phytotoxicity of Hg was confirmed in vivo by oxidative stress (H 2 DCFDA, pseudo-green colour) and cell death (IP, pseudo-red colour) fluorescent markers, in seedlings treated with 10 µM Cd or Hg for 6 h ( Figure 4B). Scattered oxidative stress was detected in several root epidermal cells exposed to 10 µM Cd (white arrows), with limited numbers of dead cells (blue arrows). In the presence of 10 µM Hg, the number of necrotic cells augmented substantially, and oxidative stress (H 2 DCFDA staining) was only noticed in cellular layers below the epidermis (white  The treatments of alfalfa seedlings with Hg and Cd also resulted in differential responses of the biothiols profile in roots, which depended on metal dose and time of exposure ( Figure 5, Supplementary Table S1, Supplementary Figure S3). However, minor changes were found in cotyledons, probably due to the limited exposure of the seedlings' aerial part to Cd and Hg (Supplementary Table S2). Three biothiols were identified in alfalfa seedlings grown in control nutrient solution: cysteine (Cys), glutathione (GSH), and homoglutathione (hGSH). Homologous to GSH, these biothiols are present in many leguminous species [7]. The concentrations of Cys and hGSH were similar after 3 and 6 h, although after 24 h, hGSH became the major biothiol (over 45%, Figure 5 and Supplementary Table S1). Biothiol concentration decreased remarkably at the highest dose of Hg (30 µM), becoming virtually undetectable after 24 h (Figure 5), probably due to the strong binding of this metal with sulfhydryl group that impedes their detection with Ellman's Reagent [35]. PCs were detected only after 24 h of metal treatments, specifically under moderate Hg doses (3 µM), with neat peaks of hPC 2 ((γ-Glu-Cys) 2 -Ala) and hPC 3 ((γ-Glu-Cys) 3 -Ala) (Supplementary Figure S3). Under extreme Hg phytotoxicity attained with long (24 h) exposure, the amount of PCs decreased remarkably in 10 µM Hg treated alfalfa, and disappeared almost completely with 30 µM Hg ( Figure 5, Supplementary Figure S3). On the other hand, Cd promoted clearer changes in the biothiol profile at the highest dose (30 µM), in terms of PCs concentration ( Figure 5, Supplementary Table S1) and variety, since PC 2 ((γ-Glu-Cys) 2 -Gly)), hPC 2 , PC 3 ((γ-Glu-Cys) 2 -Gly)), hPC 3 , PC 4 ((γ-Glu-Cys) 4 -Gly)), and hPC 4 ((γ-Glu-Cys) 4 -Ala)) were found (Supplementary Figure S3).

Impact of Ethylene Signalling Inhibition on Metal Induced Stress
The influence of ethylene in the early responses of alfalfa to Cd and Hg was studied in seedlings pre-incubated with the ethylene signalling inhibitor 1-MCP (10 µ M) for 24 h in the microscale hydroponic system. According to the results shown previously, Hg was more toxic than Cd, so we assessed ethylene regulation in metal doses that caused similar (moderate) phytotoxic effects: i.e., 3 µ M HgCl2 and 30 µ M CdCl2, after 6 and 24 h of treatment. Plants pre-incubated with 1-MCP suffered lower growth inhibition than their counterparts only treated with Hg and Cd ( Figure 6).

Impact of Ethylene Signalling Inhibition on Metal Induced Stress
The influence of ethylene in the early responses of alfalfa to Cd and Hg was studied in seedlings pre-incubated with the ethylene signalling inhibitor 1-MCP (10 µM) for 24 h in the microscale hydroponic system. According to the results shown previously, Hg was more toxic than Cd, so we assessed ethylene regulation in metal doses that caused similar (moderate) phytotoxic effects: i.e., 3 µM HgCl 2 and 30 µM CdCl 2 , after 6 and 24 h of treatment. Plants pre-incubated with 1-MCP suffered lower growth inhibition than their counterparts only treated with Hg and Cd ( Figure 6). tathione; hGSH, homoglutathione; and PCs, phytochelatins. The numbers in the brown box represent the average total concentration of biothiols (nmol g −1 FW) in each treatment. Absolute values and statistics are shown in Supplementary Table S1.

Impact of Ethylene Signalling Inhibition on Metal Induced Stress
The influence of ethylene in the early responses of alfalfa to Cd and Hg was studied in seedlings pre-incubated with the ethylene signalling inhibitor 1-MCP (10 µ M) for 24 h in the microscale hydroponic system. According to the results shown previously, Hg was more toxic than Cd, so we assessed ethylene regulation in metal doses that caused similar (moderate) phytotoxic effects: i.e., 3 µ M HgCl2 and 30 µ M CdCl2, after 6 and 24 h of treatment. Plants pre-incubated with 1-MCP suffered lower growth inhibition than their counterparts only treated with Hg and Cd ( Figure 6).   Similar behaviour was found by the induction of several redox enzymatic activities upon exposure to Hg and Cd, which were remarkably attenuated when ethylene signalling was blocked (Figure 7). In particular, the specific inhibition of GR activity by Hg was reduced in 1-MCP pre-incubated seedlings, while the activation evoked by 30 µM Cd decreased when ethylene perception was blocked (Figure 7a). Additionally, 1-MCP prevented the activation of plasma membrane NADPH-oxidase in seedlings treated with 3 µM Hg and 30 µM Cd for 6 and 24 h (Figure 7e). However, APX activity did not change in response to metal stress or 1-MCP preincubation (Figure 7c), following the same pattern found in previous observations. Western-blot analysis using α-GR, α-APX and α-HSP70 did not show differences between treatments (Figure 7b,d,f), as observed previously. However, 3 µM Hg and 30 µM Cd caused strong induction of sHSP17.7 and sHSP17.6 after 6 and 24 h following the pattern already shown, which was nonetheless minimised in seedlings preincubated with 1-MCP (Figure 7g,h).  (Figure 7e). However, APX activity did not change in response to metal stress or 1-MCP preincubation (Figure 7c), following the same pattern found in previous observations. Western-blot analysis using α-GR, α-APX and α-HSP70 did not show differences between treatments (Figure 7b,d,f), as observed previously. However, 3 µ M Hg and 30 µ M Cd caused strong induction of sHSP17.7 and sHSP17.6 after 6 and 24 h following the pattern already shown, which was nonetheless minimised in seedlings preincubated with 1-MCP (Figure 7g,h).  Extracellular H 2 O 2 release increased concomitantly with time in seedling roots treated with 30 µM Cd and 3 µM Hg for 3, 6, and 24 h, with stronger response triggered by 30 µM Cd compared to 3 µM Hg ( Figure 8A). Remarkably, AmplexRed fluorescence was attenuated in seedlings preincubated with 1-MCP for both metals, indicating that the production of H 2 O 2 was limited when ethylene signalling was blocked. Similarly, intracellular oxidative stress augmented in Cd-and Hg-treated seedlings, with a scattered pattern of epidermal cells labelled by H 2 DFCDA. Interestingly, the oxidative stress and cell damage induced by both metals was again mitigated when seedlings were pre-incubated with 1-MCP and marked by the respective negligible signal of H 2 DFCDA and propidium iodide fluorescence in root epidermal cells ( Figure 8B).
Antioxidants 2023, 12, x FOR PEER REVIEW 12 of 20 intracellular oxidative stress augmented in Cd-and Hg-treated seedlings, with a scattered pattern of epidermal cells labelled by H2DFCDA. Interestingly, the oxidative stress and cell damage induced by both metals was again mitigated when seedlings were pre-incubated with 1-MCP and marked by the respective negligible signal of H2DFCDA and propidium iodide fluorescence in root epidermal cells ( Figure 8B). Regarding the total concentration of biothiols, they augmented in Cd-treated roots when exposure was prolonged from 6 to 24 h, whereas it remained close to control values under Hg after a transient depletion at 6 h ( Figure 9A; Supplementary Table S3). As found previously, Hg and Cd led to a remarkable accumulation of PCs after 6 and 24 h of treatment, with hPC2 and hPC3 being the most abundant ones. The profile of biothiols detected was not affected by the inhibition of ethylene signalling, and only changes in their concentration were observed ( Figure 9B). Pre-incubation with 1-MCP reduced the concentration of hGSH slightly under control conditions, while it augmented compared to single metal treatments when combined Cd or Hg plus 1-MCP were applied ( Figure 9A, Supplementary Table S3). On the other hand, 1-MCP significantly diminished the amount of PCs that were produced under Cd and Hg stress: from 57% to 39%, and from 42% to 28% of total biothiol concentration, respectively ( Figure 9A, Supplementary Table S3). Regarding the total concentration of biothiols, they augmented in Cd-treated roots when exposure was prolonged from 6 to 24 h, whereas it remained close to control values under Hg after a transient depletion at 6 h ( Figure 9A; Supplementary Table S3). As found previously, Hg and Cd led to a remarkable accumulation of PCs after 6 and 24 h of treatment, with hPC 2 and hPC 3 being the most abundant ones. The profile of biothiols detected was not affected by the inhibition of ethylene signalling, and only changes in their concentration were observed ( Figure 9B). Pre-incubation with 1-MCP reduced the concentration of hGSH slightly under control conditions, while it augmented compared to single metal treatments when combined Cd or Hg plus 1-MCP were applied ( Figure 9A, Supplementary Table S3). On the other hand, 1-MCP significantly diminished the amount of PCs that were produced under Cd and Hg stress: from 57% to 39%, and from 42% to 28% of total biothiol concentration, respectively ( Figure 9A, Supplementary Table S3).  In red, peaks detected in seedlings exposed to Hg or Cd: 4, PC 2 (γ-GluCys) 2 -Gly); 5, hPC 2 (γ-GluCys) 2 -Ala); 6, PC 3 (γ-GluCys) 3 -Gly); 7, hPC 3 (γ-GluCys) 3 -Ala); and 8, PC 4 (γ-GluCys) 4 -Gly). I.S.: internal standard of N-acetyl cysteine (25 nmol per injection).

Discussion
Short-term exposure of alfalfa to Cd and Hg led to rapid root growth inhibition, accompanied by onset oxidative stress, as described previously [7,9,52]. Mercury was more toxic than Cd and triggered an oxidative burst at lower doses and time of treatment than Cd, in agreement with our previous results [9]. This inhibition could be related to alterations in the function and stability of biological membranes, in which different process alterations could modify the flow of H 2 O required for cell elongation [53]. For example, toxic metals impair ion homeostasis, leading to ion leakage and membrane depolarisation [54]. Under short-term treatments with Cd and Cu, plasma membrane H + -ATPase activity increased, possibly to restore the membrane potential required for proper transport across this membrane [55]. Such an alteration may lead to NADPH-oxidase activation and apoplastic H 2 O 2 release in a process mediated by secondary messengers such as Ca 2+ [56,57]. In this sense, we observed remarkable H 2 O 2 production in roots of alfalfa seedlings upon short-term exposure to both Cd and Hg (Figure 3), as previously found in alfalfa and soybean seedlings [8,9]. Apoplastic ROS release is probably directed by plasma membrane NADPH-oxidases ( Figure 3) which were activated under mild metal stress (Figure 2), following the pattern described by Montero-Palmero et al. [40]. However, under acute stress (i.e., 30 µM Hg) NADPH-oxidase (and peroxidase) activities were visibly inhibited, possibly reflecting extensive damages in membrane linked proteins as observed in roots of GSH-depleted genotypes of Arabidopsis sensitive to Cd-and Hg [37]. In addition, this is clearly visible in the Coomassie-stained protein band pattern under acute Hg stress ( Figure 2L). Different toxicity of Cd and Hg was also reflected in the enzymatic antioxidant and biothiol profile levels. Mercury caused stronger cellular damages than Cd in alfalfa seedlings, resulting in a remarkable inhibition of GR and APX activities in roots (Figure 2). The differential inhibition of GR activity, even at low doses of Hg, was firstly described in 15 day old alfalfa seedlings, possibly through an Hg-specific mechanism [28]. Interestingly, the strong inhibition of GR was accompanied by an accumulation of GR (Figure 2d,h,l), possibly as a compensatory mechanism. On the other hand, GR activity increase by Cd was also found in alfalfa [12] and wheat roots [33]. This differential response to Hg and Cd was also observed in the profile of GSH and PCs, with more diverse and enhanced accumulation of several PCs under Cd stress, in agreement with previous results in alfalfa and Arabidopsis [7,37].
Concomitant with oxidative stress and membrane integrity, we found a strong induction of sHSP17.7 and sHSP17.6 accumulation even at the lowest doses of Cd and Hg, while HSP70 amount did not change ( Figure 1). Similar induction of those classes of sHSPs was found shortly after exposure of Cucumis sativus roots to Cd and Cu [58]. Conversely, heat shock pretreatments caused redox imbalance and activation of ROS scavenging enzymes, which, in turn, lessened Cd toxicity in rice seedlings, possibly through a H 2 O 2 priming effect [20]. Cytosolic sHSP 17.2 was also induced by Cd in pea leaves, which was partially dependent on H 2 O 2 [22]. H 2 O 2 seems to activate several redox sensitive Heat Shock Factors (HSFs), transcription factors that promote the expression of sHSPs and ROS scavenging enzymes genes in several abiotic stresses, where NADPH-oxidases plays a relevant promotion role [59]. Interestingly, Shim et al. [60] found that HSF Class A4 is important for Cd tolerance, as it directs the expression of detoxifying genes such as metallothioneins. In recent years, it has been becoming clear the existence of a multi-level interplay between HSFs and ROS accumulation, with complex connections between protein chaperonins and ROS scavengers, two faces of the defence barriers triggered to limit cellular damage under oxidative stress [61]. Interestingly, these defence responses are apparently modulated by phytohormones such as salicylic acid or ethylene [62].
To test the hypothesis that ethylene may be involved in the early responses of alfalfa plants to Cd and Hg, we incubated a set of seedlings with the ethylene signalling blocker 1-MCP, known to delay senescence and oxidative stress symptoms under abiotic stress [63]. Although 1-MCP blocked the expression of ethylene responding genes and attenuated H 2 O 2 release by alfalfa roots under metal stress [40], a comprehensive analysis of the antioxidant and tolerance responses mediated by ethylene was required in alfalfa seedlings. Thus, the early cellular oxidative stress, H 2 O 2 apoplastic release, and NADPH-oxidase activity induced by Cd and Hg were remarkably attenuated by 1-MCP ( Figures 5 and 6), similar to previous results [40]. Likewise, the addition of aminoethoxyvinylglycine (AVG), an ethylene synthesis blocker, lessened the cell death rate in tomato suspension cells exposed to Cd [48]. Similarly, tomato plants with limited ethylene perception suffered lower oxidative stress symptoms caused by Cd [46]. Induction of oxidative stress arsenite (As(III)) toxicity was also overcome in ethylene signalling Arabidopsis plants, while the addition of the ethylene precursor ACC enhanced lipid peroxidation and As(III) damages [64]. Concerning Hg, GR activity may be considered to be a specific bioindicator of Hg toxicity in roots, as it declines severely in the presence of Hg (Figure 2). This coincides with previous results of Hg-sensitive Arabidopsis genotypes (cad2-1 or pad2-1 mutants defective in GSH synthesis), which showed stronger GR inhibition [27]. Therefore, the alleviation of Hg-dependent GR inhibition by 1-MCP may be due to attenuated stress symptoms occurring when ethylene signalling is blocked. Contribution of ethylene in the early oxidative stress responses to moderate doses of Cd, such as the overexpression of NADPH-oxidases, GSH synthesis and GR genes, were also described by Schellingen et al. [65] using Arabidopsis acs2-1/acs6-1 double mutant defective in ethylene synthesis, and in Arabidopsis etr1-1 and ein2-1 ethylene signalling mutants [42]. Oxidative stress symptoms caused by Cd were also attenuated in the aquatic macrophyte Nelumbo nucifera G., and treated with the ethylene signalling blocker silver thiosulfate [66]. NADPH-oxidases contribute at least partially to the generation of ROS under Cd stress [67], which may explain the increased NADPH-oxidase activity detected in alfalfa seedlings exposed to Cd and Hg (Figure 2; [40]). Ethylene seems to facilitate the upregulation of NADPH-oxidase genes in the presence of Cd, which, in turn, modulates the expression of ALTERNATIVE OXIDASE1a, as observed in Arabidopsis acs2-1/acs6-1 mutant and ethylene insensitive mutant ein2-1. This enzyme may contribute to generating ROS at the mitochondria, perhaps amplifying the oxidative stress signalling [68]. In agreement with these results, defective ethylene signalling also prevented the generation of H 2 O 2 and O 2 •− , which are products of NADPH-oxidases caused by the exposure of Zn nanoparticles in leaves of ein2-1 and etr1-3 Arabidopsis plants [69]. Finally, our results show that the attenuation of the oxidative stress obtained by 1-MCP under Cd and Hg stress also concurs with the limited expression of oxidative stress hallmark genes in ein2-1 mutants in the presence of Cd [65].
Recent evidence supports the idea that H 2 O 2 is an important component of metal stress signalling and important for the activation of detoxification mechanisms [70]. ROS prompt a series of redox-derived changes in protein function, leading to the quick alteration of gene transcription and activation of several signalling pathways that involve phosphorylation, Ca 2+ release, and/or Cys residues redox shifts [71]. Among other processes, the overexpression of sHSPs depends on HSFs activation through a H 2 O 2 -mediated signalling cascade [72]. Conversely, the defence mechanisms triggered by HSF comprise the enhancement of ROS-scavenging systems, such as APX, which would contribute to maintaining the cellular redox balance under stress [73]. Additionally, ethylene prevents heat shock cellular damages, possibly through the activation of the heat shock pathway [74]. Ethylene seems to mediate in the basal thermotolerance of Arabidopsis, perhaps intertwined with ABA and other stress phytohormones, along with several antioxidants such as ASA and GSH, to prevent the oxidative stress caused by heat [75]. In this sense, the blocking of ethylene signalling with 1-MCP attenuated premature leaf senescence, ROS production and oxidative stress damages observed in soybean seedlings grown under high temperature stress [63]. This coincides with our data, as we found that 1-MCP decreased the early induction of ROS accumulation, oxidative stress, and enhanced accumulation of a sHSPs under Cd and Hg stress (Figures 5 and 6). Therefore, our working model considers that ethylene enhances ROS production, and the subsequent redox unbalance may induce sHSPs expression under Cd and Hg stress.
Biothiols are keystones of plants' tolerance to toxic elements [30], and crosstalk between GSH and ethylene is highly probable, since this phytohormone is synthesised from methionine and depends highly on S metabolism [49], and ethylene enhances sulphur assimilation [76]. For example, Arabidopsis mutant plants strongly defective in GSH synthesis (i.e., rml1-1) overexpressed ERF (Ethylene Response Factor) transcription factors ERF11, ERF2, and ESE3 [77]. On the other hand, ethylene, and other stress related phytohormones such as jasmonate and salicylic acid, seem to promote the biosynthesis of GSH or modify the GSH redox status [78]. Interestingly, we observed that incubation of alfalfa seedlings with the ethylene signalling blocker 1-MCP decreased the levels of GSH and PCs treated with Cd and Hg (Figure 7, Supplementary Table S3). In this respect, the activation of GSH synthesis promoted by Cd was blocked in mustard plants incubated with the ethylene synthesis inhibitor AVG [45]. Similarly, AVG repressed the early accumulation of GSH induced by Cd in Lycium chinense, which also affected the expression of GSH synthesis genes, such as γ-glutamylcysteine synthetase (γ-ECS) and glutathione synthetase [50]. Schellingen et al. [65] found lower expression of GSH metabolism genes in the ethylene defective Arabidopsis mutants acs2-1/acs6-1 subjected to 24 and 72 h treatments with Cd. Additionally, the concentration of GSH in leaves also decreased in acs2-1/acs6-1 mutant plants compared to the WT after 24 h, though the difference was less pronounced when Cd treatments were extended to 72 h. Similar results were observed in the ethylene signalling of etr1-1, ein2-1, and ein3-1 Arabidopsis mutants, where GSH levels were lower than the WT under Cd stress, which corresponded with the limited expression of glutathione synthase gene in the etr1-1 genotype [42]. To sum up, ethylene seems to transiently mediate the canonical biothiol synthesis pathway under toxic metal stress. Although the signalling mechanisms are still unknown, we hypothesise that ethylene exerts its control on biothiols synthesis through an indirect pathway via a redox switch. Several GSH metabolic enzymes (i.e., γ-ECS and GR) required for GSH biosynthesis and homeostasis are post-translationally activated via Cys thiol/disulfide redox shifts under oxidative stress and redox cellular imbalance [79], which are apparently promoted by ethylene in the presence of toxic elements such as Cd and Hg.
It is becoming apparent that ethylene mediates the early activation by Hg and Cd of NADPH-oxidase and H 2 O 2 production, as well as the differential changes in GR activity, i.e., inhibition by Hg and modest activation by Cd, and the synthesis of GSH and PCs; these are symptoms that were transiently delayed in alfalfa seedlings incubated with 1-MCP and exposed for 6 h to Hg and Cd. However, the exact mechanism by which ethylene promotes the induction of NADPH-oxidase activity and the alteration of the redox balance remains unknown, which could involve transcriptional regulation as well as post-translational activation, for example via calcium and calcium dependent kinases [80]. Future experiments should focus on analysing this aspect of the signalling process that occurs in the early moments of toxic metal exposure by tuning the response to ethylene supplied at different doses and times of incubation, and in the presence of 1-MCP to exploring the behaviour of plants at advanced phenological status (i.e., by comparing juvenile versus adult developmental phases). In addition, the attenuation of the metal-induced oxidative burst was less apparent after 24 h treatment, as Keunen et al. [68] observed that Arabidopsis plants lacking ethylene response (ein2-1 and ein2-5 mutants) suffered inexorably stress at higher Cd doses or longer treatments. These results imply that other mechanisms of toxicity perception, perhaps depending on other stress related phytohormones, such as jasmonate or abscisic acid [41], may overtake the ethylene dependent pathway as plants become poisoned by toxic elements, and this should also be a matter of future research. With the idea of improving metal tolerance, an interesting alternative to modify the endogenous levels of ethylene could be the inoculation of plants with bacteria producing ACC deaminase found to promote plant growth under Cd stress [81]. Therefore, such strategy could be combined with the selection of plants with limited sensitivity to and/or synthesis of ethylene.

Conclusions
Inhibition of ethylene signalling using 1-MCP resulted in attenuated induction of the oxidative stress triggered by Hg and Cd in alfalfa seedlings. Subsequently, ethylene seemed to modulate in the roots of juvenile alfalfa seedlings the generation of ROS, the accumulation of sHSPs (HSP17.7 and HSP17.6) chaperones, the induction of the pro-oxidant NADPH-oxidase activity, alteration of redox enzymatic activities needed to maintain the GSH redox balance, such as GR, and the ability to accumulate PCs under Cd and Hg stress. Characterisation of the role ethylene, and other stress-related phytohormones, will help to improving the tolerance of plants to toxic metals, with the aim of optimising phytoremediation strategies to clean up metal contaminated soils.