Introduction

Soil pollution is a serious problem that has turned into a worldwide environmental issue. This contamination is mainly due to mining, metallurgical, and chemistry activities that have occurred for decades and have deposited many toxic elements in the soil (Kabata-Pendias 2010; Mykolenko et al. 2018). These toxic elements are not biodegradable, and they are dangerous to plants and animals. The uptake of these toxic elements by plants not only reduces crop yields but also is a potential risk for human health through inputs into the food chain (Ali et al. 2013).

Arsenic (As) is a non-essential element that is highly toxic to humans and other living organisms causing serious effects, such as arsenicosis and carcinogenesis (Francesconi et al. 2002; Abioye 2011). It usually appears as a waste product of mining and refining of metals such as Au, Fe, Hg, or Sn and, as a consequence, high concentrations of As have been measured in abandoned mining sites (Otones et al. 2011; Ordóñez et al. 2013; Gallego et al. 2016; Fernández-Fuego et al. 2017; Pistelli et al. 2017). As a phosphate analog, As can severely inhibit plant growth and exert its toxicity by its capacity to interfere with the phosphate metabolism–altering processes such as ATP synthesis and oxidative phosphorylation. Furthermore, once absorbed by the plant, it is difficult to be eliminated (Tripathi et al. 2007; Finnegan and Chen 2012; Shukla et al. 2016).

Phytoremediation can be defined as the use of plants to remove or reduce contaminants from the environment thereby limiting the damage that the contaminants produce (Liang et al. 2016). Additionally, it is an environmentally friendly technique that maintains virtually unaltered biological activity of the soils (McGrath et al. 2001; Liang et al. 2016). The term phytoremediation includes a variety of techniques and strategies, including phytoextraction and phytostabilization. Phytoextraction involves the absorption of pollutants by the roots and their transport and accumulation in the aboveground plant parts. It is an effective technique when plants produce a large biomass and are able to accumulate high concentrations of heavy metals in their leaves (McGrath et al. 2001; Van der Ent et al. 2013). In contrast, phytostabilization focuses on the sequestration of contaminants within the root system, reducing their mobility and bioavailability in the soil (Ali et al. 2013).

It has been reported that the pH of the soil or the culture medium is a key factor that affects the availability of metal(loid)s to the plant, and the highest As bioavailability for some plant species has been observed at basic pH values (Kabata-Pendias 2010; Moreno-Jimenez et al. 2011). Therefore, the optimization of the pH values could be an important strategy in order to increase As uptake and accumulation, which may lead to better phytoremediation. The application of chelating agents is an innovative option to improve metal(loid) absorption and accumulation by plants. One of the most effective and commonly used chelating compounds is EDTA (Wuana et al. 2010). However, nowadays, the use of this compound is controversial, since it is able to reach and contaminate the groundwater (Muhammad et al. 2009). Consequently, other harmless chelating compounds that are naturally present in plants, such as citric acid (CA), have been assayed without showing toxic effects to the environment (Farid et al. 2017). In addition, it has also been reported that the presence of CA increases metal extraction in soil at acid pH (Gaber et al. 2011) and that it assists metal(loid) elimination from soil and/or sludge (Liu and Lin 2013).

When metal(loid)s reach the cell cytoplasm, plants use different mechanisms to detoxify and tolerate them (Benavides et al. 2005, Evangelou et al. 2007; Mishra et al. 2008). One of the most important detoxification mechanisms is the chelation of free metal ions with low-molecular-weight ligands, such as phytochelatins (PCs) and other thiolic compounds, and the subsequent compartmentalization of these less toxic complexes in the vacuole (Verbruggen et al. 2009; Ali et al. 2013; Fernández et al. 2014b). PCs are non-protein thiols (NPTs) formed by repetitions of the peptide, γ-glutamyl-cysteine with glycine as terminal residue. They have the basic structure of (γ-Glu-Cys)n-Gly (n ranging from 2 to 11) and, in most plants, the most abundant PCs are those in which n is between 2 and 5 (Clemens 2006). PCs are synthesized from glutathione (GSH) by the enzyme PC synthase, and they can be present in the plant cells in their final form, but without the terminal glycine residue (desGly-PCs) or with an extra cysteine (Cys-PCs) (Cobbett and Goldsbrough 2002; Zagorchev et al. 2013). The synthesis of these thiol compounds is highly induced by the presence of metal(loid)s, especially Cd and As (Fernández et al. 2014b; Mesa et al. 2017).

The understanding of the biochemical and physiological processes controlling metal(loid) accumulation is crucial in order to design more practical and effective phytoremediation projects. In this way, the use of native plant species, already adapted to the local soil and climatic conditions, can be a key success factor. Therefore, several plant species found growing naturally in metal(loid)-polluted soils in Asturias (Spain) were cultured in vitro in the presence of As (Table 1) in order to select the highest arsenic accumulator plant. Among these plant species, Eupatorium cannabinum L. was selected based on its remarkable As tolerance and accumulation capacity (Table 1). This plant belongs to the order Asterales, family Asteraceae, which shows high biomass (reaching up to 1.5 m in height and 1.2 m in width) and grows well in wet and flooded places of Europe, Asia, USA, and North Africa, where it can be easily found growing in metal(loid)-polluted sites. This characteristic, along with its high biomass, makes it suitable for phytoremediation. Considering all the abovementioned, the aim of this paper is to study the effect of the pH of the culture medium and the addition of As and/or CA on the growth and As accumulation of E. cannabinum in order to develop future phytoremediation strategies. Additionally, considering that As detoxification in plants is a complex process involving several mechanisms whose relative importance is still unclear and that depend on many factors (such as plant species, the metal(loid) involved, or its concentration), we also studied the role of PCs and other NPTs in the tolerance and accumulation of As of E. cannabinum.

Table 1 Arsenic concentration (mg kg−1 dry wt.) in different plant species cultured for 30 days at pH 5.7 in 1/4 MS medium plus 11 mg As L−1. Different letters within the same column denote significant differences between plants at p < 0.05. TF, translocation factor
Full size table

Material and methods

Plant material and culture conditions

Seeds of E. cannabinum were collected from a heavy metal–polluted area of Asturias, Spain, sterilized and introduced into culture tubes (Fernández et al. 2011). Each germinated seed turned into a seedling and apical shoot segments (15–25 mm) of these seedlings were transferred individually into a vessel with Murashige and Skoog medium (MS) (Murashige and Skoog 1962) with 0.2 g L−1 of sequestrene 138-Fe (Ciba-Geigy AG) as an iron source, 30 g L−1 sucrose, and 7 g L−1 agar. The pH of the medium was adjusted to 5.7. Explants of each seedling constituted a clone. Apical shoots of each clone were cultured with different As concentrations. The highest accumulating clone (EC-J) was selected by comparing plant’s in vitro growth and As accumulation in the presence of As. The explants were then grown in a growth chamber at 25 °C, and a 16-h photoperiod (150 μmol m−2 s−1) for 20 days. After this period, plantlets were rinsed in double-deionized water and their fresh weights and shoot and root lengths were measured. Then, plantlets were transferred to Magenta vessels (Magenta Corp. USA) of 600-ml capacity containing 10 cellulose plugs and 80 ml of 1/4 strength MS medium. In order to test the effect of the pH, three different pH values were selected (3.7, 5.7, and 7.7). To study the effect of CA addition on the growth and As accumulation, two different CA concentrations were tested (10 and 20 mg L−1). Arsenic was added to the medium as Na2HAsO4·7H2O at a concentration of 11 mg L−1. CA and As concentrations were selected according to previous assays in the laboratory. Therefore, the treatments for each pH were as follows: 1/4 MS only (control treatment); 1/4 MS + 10 CA; 1/4 MS + 20 CA; 1/4 MS + As; 1/4 MS + As + 10 CA; and 1/4 MS + As + 20 CA.

After 20 days of culture in the growth chamber under the conditions described above, the plants were carefully extracted from the vessels and rinsed with tap water followed by three 5-min rinses with distilled water. The plants were weighed, their aboveground parts and roots were separated, and their lengths were measured. To obtain dry weights, the aboveground parts and roots of a group of plants were oven-dried at 35 °C for 48 h and then powdered using liquid nitrogen. Water content (WC) was calculated as follows:

$$ \mathrm{WC}\ \left(\%\right)=\left[\mathrm{fresh}\ \mathrm{weight}-\mathrm{dry}\ \mathrm{weight}/\mathrm{fresh}\ \mathrm{weight}\right]\times 100 $$

For NPT analysis, samples of fresh aboveground parts and roots were collected from a different group of plants; they were also powdered using liquid nitrogen and stored at − 80 °C until use.

Analysis of arsenic content

Analysis of the As content was carried out using oven-dried plant material by inductively coupled plasma-mass spectrometry (ICP-MS). Powdered leaf and root samples (100 mg) were digested with 3 ml of concentrated, high-purity HNO3 (68%) in a microwave oven at 240 W for 6 min in order to completely liquidize the samples (Montaser 1998). After cooling, two consecutive dilutions were made with doubly deionized water (Mili-Q185 Plus System, 18 MΩ cm−1 resistivity), first a 1:20 dilution, and next a 1:10 dilution in which a solution of 10 μg rhodium kg−1 in concentrated HNO3 was added as internal standard. Blanks, in which the same procedure was followed without adding plant material, were also assayed. Finally, As content was analyzed using an ICP-MS (quadrupole type, HP-7500cc) as previously described in Fernández et al. (2008).

Analysis of non-protein thiols

The extraction and analysis of NPTs was carried out from leaf and root samples in cold room (4 °C) following the method previously described by Fernández et al. (2012). The high-performance liquid chromatography (HPLC) separation was performed using a Waters 600 (Waters Corporation) with a post-column derivatization with Ellman’s reagent (Ellman 1959). The sample (100 μl) was injected into a Kromasil 100 C18 5 μm (250 × 4.6 mm) (Scharlau) column and eluted with solvents A (acetonitrile:water, 2:98, v/v) and B (acetonitrile:water, 98:2, v/v), to which trifluoroacetic acid (TFA) 0.05% was added. Samples were separated using a linear gradient (0–25% in 25 min and 25–50% in 5 min) of solvent B at 1.5 mL min−1 flow for 30 min. The derivatized thiols were detected at 412 nm using a Waters 996 photodiode array detector and the peaks obtained were compared with external standards of GSH and a mix of PC2, PC3, PC4, and PC5. The quantitative changes in the thiolic compounds observed were calculated using the integration of their peak areas at 412 nm of absorbance converted into nanomole and quantified as GSH equivalents. The total content of NPTs, excluding GSH, was also calculated.

Statistical analysis

The study was conducted as three independent experiments, each with six different plants per treatment. The results of As accumulation and NPTs are shown as means ± standard errors of three independent replicates. A two-way (CA × As) analysis of variance (ANOVA) was used to evaluate quantitative differences in growth, As accumulation, and NPT content between treatments at the same pH value. A one-way ANOVA was used to evaluate quantitative differences in the abovementioned parameters between each treatment at different pH values. A log transformation was applied to approximate normality when necessary. When the F-ratio was significant (p ≤ 0.05), Tukey’s honestly significant difference test (HSD, p ≤ 0.05) was employed to compare between individual means. Data were analyzed using R software (version 3.3.1, http://www.r-project.org/) with the package agricolae (version 1.2_4. http://tarwi.lamolina.edu.pe/~fmendiburu).

Results and discussion

Effect of pH, arsenic, and citric acid on E. cannabinum growth

After 20 days of culture, no changes in shoot or root lengths were observed in control plants (1/4 MS) cultured at the different pH values, except for a reduction in root length at pH 7.7 in comparison to pH 3.7 (Fig. 1). In general, the presence of CA in the culture medium did not have any significant effect on the shoot or root length at any of the two concentrations assayed when compared with controls, except for an increase in root length observed at both CA concentrations at pH 7.7 (Fig. 1b). However, the addition of As to the culture medium reduced both shoot and root lengths in comparison with the controls (Fig. 1). It is interesting to note that this reduction was significantly higher in roots, which could be explained by the fact that the root is the first organ in contact with the culture medium and, therefore, it is not strange that it exhibits more sharply the negative effects caused by this metalloid. Similar results have also been described in other plant species exposed to Cd (Fernández et al. 2008) or Ag (Farid et al. 2018). The addition of CA to the culture medium with As significantly reduced the negative effect of this metalloid on shoot length but only at the two acidic pH values (Fig. 1a). However, no differences were observed in roots, apart from those in plants treated with 10 mg CA L−1 at pH 5.7 (Fig. 1b). Najeeb et al. (2011) working with Juncus effusus L. grown in a solution with CA and Cd reported an improvement in root growth, but a slight reduction in stem length when compared to the treatment without CA.

Fig. 1
figure 1

Shoot (a) and root (b) lengths of Eupatorium cannabinum cultured for 20 days at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control), or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, or 20 mg L−1). Bars represent means ± standard errors. Different letters denote significant differences between treatments at the same pH at p < 0.05. Different numbers denote significant differences among the different pH values at the same treatment at p < 0.05

Full size image

Neither the fresh nor dry weight of control plants were significantly affected by the pH of the medium (Fig. 2). However, the biomass accumulated by plants cultured with CA were variable when compared to control plants and depended on the pH of the culture medium and the concentration of CA used (Fig. 2). Reductions in fresh and dry weights (of about 65% and 50%, respectively) in comparison with the control occurred in 1/4 MS + As (Fig. 2) due to a decrease in the length of the stem and especially in the length of the root. However, it is worth mentioning that plants did not show any apparent toxicity symptoms. These reductions in fresh and dry weight were corroborated when the water content was measured at the 3 different pH values assayed (Table 2). Similar biomass and water content reductions in response to the presence of As were also reported by other authors in Phaseolus vulgaris L. (Stoeva et al. 2005), Oryza sativa L. (Shri et al. 2009), Vigna radiata (L.) Wilczek (Mumthas et al. 2010), and Brassica napus L. (Farooq et al. 2015). These authors concluded that the toxic effect of As decreased the photosynthetic rate and metabolic activity of plants, which is reflected in growth inhibition. For E. cannabinum, the addition of CA to the medium containing As led to an increased fresh weight in plants cultured at pH 3.7 and 5.7 but not at 7.7 (Fig. 2a) compared to As alone. With regard to dry weight, this parameter was enhanced by both CA concentrations at pH 3.7 but only when 20 mg CA L−1 were added at pH 7.7 (Fig. 2b). A similar improvement in plant growth after the addition of CA to the culture medium has also been reported in plants of B. napus or Helianthus annuus L. exposed to different heavy metals including Cd (Ehsan et al. 2014), Pb (Shakoor et al. 2014), Cu (Zaheer et al. 2015), or Cr (Afshan et al. 2015; Farid et al. 2017). In some of these studies, it is suggested that CA has an important role enhancing the tolerance of plants to metal(loid)s, since it reduces the presence of free metal ions, thereby improving plant growth through increased photosynthesis and the activities of antioxidant enzymes.

Fig. 2
figure 2

Fresh (a) and dry (b) weight of plants of Eupatorium cannabinum cultured for 20 days at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control), or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, or 20 mg L−1). Bars represent means ± standard errors. Different letters denote significant differences between treatments at the same pH at p < 0.05. Different numbers denote significant differences among the different pH values at the same treatment at p < 0.05

Full size image
Table 2 Water content in plants of E. cannabinum after 20 days of culture at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control) or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, and 20 mg L−1). Different letters within the same column denote significant differences between treatments at the same pH value. No significant differences were observed among the different pH values for the same treatment at p < 0.05
Full size table

Arsenic accumulation

After 20 days of culture, no differences in As accumulation were observed in leaves of plants grown in 1/4 MS + As at different pH values assayed, while in roots the concentration of As was lower at pH 3.7 than at 7.7 (Fig. 3). Moreno-Jimenez et al. (2011) also reported that the amount of As accumulated by several Mediterranean shrubs was higher when soil pH was higher than 5. This increased As accumulation can be explained by higher As mobility at pH 7–9 as pointed out by Kabata-Pendias (2010). Additionally, As accumulation was much higher in roots than in leaves, exceeding 3000 mg kg−1 in the roots (Fig. 3). It is also important to highlight that the amount of As accumulated by E. cannabinum L., both in roots and in leaves, exceeded the level for toxicity (1–20 mg kg−1) established by White and Brown (2010) for non-tolerant crop plants without showing apparent toxicity symptoms (such as wilting, necrosis, or malformation of the leaves), which suggest a great tolerance to As. In the same way, Mubarak et al. (2016) also observed that plants of Boehmeria nivea L. grown hydroponically were able to tolerate up to 15 mg As L−1 without any apparent signs of toxicity.

Fig. 3
figure 3

Arsenic accumulation (mg kg−1) in leaves (a) and roots (b) of Eupatorium cannabinum after 20 days of culture at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control) or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, and 20 mg L−1). Bars represent means ± standard errors. Different letters denote significant differences between treatments at the same pH at p < 0.05. Different numbers denote significant differences among the different pH values at the same treatment at p < 0.05

Full size image

In order to be considered an As hyperaccumulator, a plant needs to be able to accumulate at least 1000 mg As kg−1 dry weight in leaves (Van der Ent et al. 2013). Considering this, and according to our results, E. cannabinum could not be considered an As hyperaccumulator. However, it was able to accumulate As in its roots to levels three times higher than the hyperaccumulation threshold (Fig. 3b). This high As accumulation in roots may be probably responsible for the reductions in root length and both fresh and dry weights previously discussed. In the same way, it has been reported that other plant species, such as Agrostis castellana Boiss & Reut. or Rumex acetosella L., are able to accumulate large amounts of As in their roots (more than 200 mg kg−1) when growing in polluted mining sites (Otones et al. 2011). These plants with a higher accumulation of this metalloid in their roots, as is the case of E. cannabinum, are considered good candidates for use in phytostabilization. Additionally, a higher accumulation of As in roots than in leaves was also reported by Pérez-Sirvent et al. (2012) while working with the herbaceous plant, Dittrichia viscosa. In contrast, Pistelli et al. (2017) observed a higher As concentration in leaves (more than double than that measured in roots) while working with the same plant species (D. viscosa) in an ex-mining site. In our opinion, the high amount of As accumulated in roots of E. cannabinum L. together with the low root-to-shoot translocation rate observed may constitute an important As detoxification mechanism in this plant species. This mechanism would prevent As toxicity in leaves thereby avoiding the alteration of key physiological processes such as photosynthesis, as pointed out by Afshan et al. (2015) and Zaheer et al. (2015). In addition, we added As to the culture medium as As(V), a phosphate chemical analog that plants roots take up via phosphate transporters. Once inside the root cells, As(V) is rapidly reduced to As(III) (Xu et al. 2007; Verbruggen et al. 2009), which has a high affinity to form complexes with the sulfhydryl groups of thiols and could then be sequestered into vacuoles. Considering this, it is possible that plants, as a defense mechanism, keep this more toxic As form immobilized in cellular compartments of the roots and do not transport it to the aboveground parts, avoiding interference with their metabolism (Xu et al. 2007; Verbruggen et al. 2009).

It is described in the literature that CA increases plant metal(loid) accumulation due to its capacity to react with the metal(loid)s present in the culture medium or soil, forming soluble complexes that allow a better absorption by the plants (Kwak et al. 2013; Afshan et al. 2015; Zaheer et al. 2015; Gogoi et al. 2017). However, the effect of CA on metal(loid) accumulation is controversial since its effect seems to be dependent on the plant species and/or the metal(loid)s involved. Thus, while Anwer et al. (2012) observed that Cd accumulation and translocation in Zea mays L. was inhibited in presence of CA, other authors (Chen et al. 2003; Duarte et al. 2007; Almaroai et al. 2012, 2013; Ehsan et al. 2014) observed that plants of Raphanus sativus L., Halimione portulacoides (L.) Aellen, Z. mays L., or B. napus were able to accumulate higher amounts of Cd, Pb, or As in the presence of CA. In our case, the addition of 10 mg L−1 CA to the medium containing As decreased the accumulation of this metalloid in both leaves and roots (except in leaves at pH 5.7) (Fig. 3). In contrast, the addition of CA at a higher concentration (20 mg L−1) generally led to an increased As accumulation when compared to the 1/4 MS + As treatment. This effect was mainly observed at acidic pH values (pH 3.7 and 5.7 in leaves but only at pH 3.7 in roots), while As accumulation in roots was reduced at pH 7.7 (Fig. 3). This contrasting effect of CA at the two different concentrations assayed (10 and 20 mg L−1) is surprising and difficult to explain. However, taking into consideration our results and those of the other authors, we suggest that the abovementioned controversy in the literature around the effect of CA on metal(loid) accumulation could be at least partially explained by differences between the CA doses applied in each case.

Analysis of non-protein thiols

After 20 days of culture, changes in the concentration of thiols and de novo synthesis of NPTs were observed in leaves and roots of E. cannabinum among the different treatments assayed (Tables 3 and 4).

Table 3 Non-protein thiol concentrations in leaves of E. cannabinum after 20 days of culture at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control) or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, and 20 mg L−1). Different letters within the same column denote significant differences between treatments at the same pH value. Different numbers within the same column denote significant differences among the different pH values for the same treatment at p < 0.05. (GSH: glutathione)
Full size table
Table 4 Non-protein thiol concentrations in roots of E. cannabinum after 20 days of culture at different pH values (3.7, 5.7, and 7.7) in 1/4 MS medium (control) or 1/4 MS plus 11 mg As L−1 and with different concentrations of citric acid (CA) (0, 10, and 20 mg L−1). Different letters within the same column denote significant differences between treatments at the same pH value. Different numbers within the same column denote significant differences among the different pH values for the same treatment at p < 0.05. (Cys: cysteine; GSH: glutathione; nd: not detected)
Full size table

In leaves of control plants (1/4 MS), five different thiolic compounds were detected, including GSH, desGly-PC2, Cys-PC2, PC3, and PC4 (Table 3), GSH and PC4 being the most abundant forms at all the pH values assayed. Several authors (Vurro et al. 2011; Akhter et al. 2012) as well as our own previous studies (Fernández et al. 2014a, b; Fernández-Fuego et al. 2017) have also reported the presence of PCs in plants not exposed to metal(loid)s, suggesting that these compounds play an important role in the control of metal-ion homeostasis which is the mechanism that regulates the availability of metal ions in plant cells. In this way, the presence of NPTs in leaves of control plants could be explained by the fact that the medium used in this in vitro experiment, even though it was diluted, could be rich enough in mineral nutrients to activate the synthesis of NPTs. Additionally, Kühnlenz et al. (2015) have also hypothesized that PCs are not only synthesized under stress conditions, since the enzyme involved in PC synthesis, PC synthase, could also have an immune function.

In leaves, the addition of CA to the control medium induced neither changes in constitutive concentrations of NPTs nor de novo synthesis of new NPTs when compared with the control plants. However, as expected, the addition of As induced an increase in the total amount of NPTs at pH 5.7 and 7.7, mainly due to an increased synthesis of Cys-PC2, PC3, and PC4 (Table 3). When CA was added to a culture medium containing As, we observed a reduction in the total concentration of NPTs in leaves at pH 5.7 while at pH 7.7 it was only observed at the lowest CA concentration assayed (10 mg L−1) (Table 3). The decrease in the concentration of NPTs when CA is present, even though As accumulation was high (CA 20 mg L−1) (Fig. 3a), may suggest that the chelating effect of the CA allows the plants to synthesize lower amounts of PCs in order to tolerate the As toxicity (Oven et al. 2002; Zaheer et al. 2015).

In roots of control plants, only PC4 and small amounts of GSH and Cys-PC3 were measured regardless of the pH treatment (Table 4). Similar results were observed in plants treated with both CA concentrations (Table 4). The addition of As to the culture medium dramatically increased the concentration of NPTs in roots through an increased synthesis of the constitutive NPTs (GSH, Cys-PC3 and PC4) and de novo synthesis of Cys, PC2, desGly- PC2, PC3, desGly-PC3, and desGly-PC4. It is interesting to note that Cys and GSH, apart from being precursors of PC synthesis, also play a key role in metal(loid) detoxification (Oven et al. 2002; Gupta et al. 2010; Zagorchev et al. 2013) and, in the case of GSH, mediate protection against oxidative stress (Shri et al. 2009). Therefore, the presence of these compounds could be used as an indicator of a stress response to metalloid(s) (Hernández et al. 2015). The high synthesis of NPTs observed reflects a rapid response of the plants to As and, in the case of roots, this response was more pronounced, since roots provide the primary route for the penetration of metal(loid) ions in the plant (Piechalak et al. 2002; Fernández et al. 2012). The synthesis of new NPTs in response to metal(loid)s has also been observed in other plant species (Harada et al. 2002; Anjum et al. 2015; Shukla et al. 2016; Fernández-Fuego et al. 2017). It is known that larger PCs form more stable complexes with metal(loid)s, which makes them more effective at alleviating metal(loid) toxicity, thereby enhancing plant tolerance and allowing it to accumulate larger amounts of these elements (Gusmão et al. 2010). Our results support this hypothesis, considering that PC3, PC4, and their derivatives (desGly-PC3, Cys-PC3, and desgly-PC4) were the predominant polythiol forms involved in As chelation and detoxification in roots, regardless of the CA treatment. However, it is also noteworthy that the synthesis of larger PCs also involves a higher energetic cost for the plant (Zagorchev et al. 2013).

Interestingly, the total concentration of NPTs was significantly higher in roots than in leaves (Tables 3 and 4). Considering that, although controversially, some authors suggest that PC synthesis takes only place in leaves, this would imply that NPTs would later be actively transported to the roots in order to cope with the large amounts of As accumulated in this organ, as suggested by Chen et al. (2006). It is also important to note that roots contained the highest amount of NPTs but also the highest As accumulation (Fig. 3b). This high concentration of NPTs in roots of E. cannabinum could be one of the possible reasons for the low root-to-shoot As translocation observed, since the elevated presence of these chelating compounds would allow the plant to accumulate As in the root cells, as suggested by Akhter et al. (2012). Similar results have been observed in plants of Vetiveria zizanioides (L.) Roberty growing in Pb-polluted soils (Andra et al. 2009). In the same way, plants of Betula celtiberica Rothm & Vasc. cultured in vitro also increase the content of NPTs in their roots in the presence of As (Mesa et al. 2017). In contrast, other authors have observed higher PC concentrations in shoots despite the fact that the highest metal(loid) accumulation was measured in roots, as it was the case of B. pubescens Ehrh cultured in a polluted soil (Fernández-Fuego et al. 2017). This contradiction could be due to the fact that the roots of plants grown in pot or field conditions show a higher number of lignified and suberized cells that are less physiologically active than those grown in vitro or hydroponics. In any case, we cannot forget that thiols are not the only detoxification mechanism of plants, but that many others can act in parallel, e.g., the sequestration of metals to cell walls, already described for As (Fernández-Fuego et al. 2017) or Cd (Fernández et al. 2014a). These authors proposed the sequestration of metals to the cell wall as the main detoxification strategy followed by the plant, while the synthesis of NPTs reduces the toxicity caused by heavy metals once they get inside the cells. Due to the high toxicity of these metal(loid)s, the plant rapidly complexes them to the NPTs and stores them in the root vacuoles to avoid major damages in the plant.

When CA (10 mg L−1) was added to a culture medium containing As, we observed a decreased synthesis of NPTs in roots at all the three pH values assayed (Table 4). This reduction in the content of NPTs was coincidental with a reduction in As accumulation (Fig. 3b), which suggests that As detoxification in E. cannabinum is strongly dependent on NPTs. A similar correlation between PC synthesis and As accumulation was also observed in B. juncea (L.) Czem plants exposed to As stress (Gupta et al. 2009). The addition of CA (20 mg L−1) also diminished the synthesis of NPTs when compared to 1/4 MS + As, except to that at pH 7.7. It is also remarkable that this decrease in NPTs was concomitant with a higher or similar As accumulation, with the only exception of the abovementioned pH 7.7 where a slight reduction was observed (Fig. 3b). Taking these data into account, we hypothesize a different efficiency in the chelating activity of the CA, it being higher at acidic pH, as has also been observed by Suanon et al. (2016). So, in our case, in the presence of CA and at acidic pH, a lower synthesis of NPTs would be necessary in order to cope with the high amount of As accumulated in the plant roots.

Conclusions

Our results confirmed that E. cannabinum cultivated in vitro had good growth at the different pH values of the culture medium. In addition, plants accumulated high concentrations of As, especially in their roots. Although this metalloid reduced plant growth, they showed a great tolerance without presenting any toxicity symptoms such as yellowing or wilting. In leaves, the pH of culture medium did not affect As accumulation while in roots, a basic pH increased accumulation. On the other hand, the higher CA concentration (20 mg L−1) added to a medium with As reduced the toxic effect of this metalloid, promoting an increase in both plant biomass and As accumulation at acidic pH values. All these data suggest that E. cannabinum is an As-resistant plant with a potential use in phytostabilization programs in which the application of CA would be an interesting strategy to follow, especially at an acid pH.

The fact that the highest NPT concentration was measured in roots, the organ that had the highest As accumulation suggests that these thiolic compounds play a major role in As detoxification in E. cannabinum, although we cannot disregard the fact that other detoxification mechanisms could also be involved. Finally, considering facts that at acidic pH and, especially in roots, the addition of CA (20 mg L−1) to the culture medium led to an increased As accumulation without increasing the synthesis of NPTs, this could indicate that the presence of CA allowed the plant to overcome As toxicity with a lower NPT synthesis. This could constitute an advantage for the plant since it leads to a reduction in the elevated energetic cost associated with the synthesis of NPTs.