Rapid Hormetic Responses of Photosystem II Photochemistry to Cadmium Exposure

Five-day exposure of clary sage (Salvia sclarea) to 100 μM cadmium (Cd) in hydroponics was sufficient to increase Cd concentrations significantly in roots and aboveground parts and affect negatively whole plant levels of calcium (Ca) and magnesium (Mg), since Cd competes for Ca channels, while reduced Mg concentrations are associated with increased Cd tolerance. Total zinc (Zn), copper (Cu), and iron (Fe) uptake increased but their translocation to the aboveground parts decreased possible due to translocation barriers. Despite the substantial levels of Cd in leaves, without any observed defects on chloroplast ultrastructure, an enhanced photosystem II (PSII) efficiency was observed, with a higher fraction of absorbed light energy to be directed to photochemistry (ΦPSΙΙ). The concomitant increase in the photoprotective mechanism of non-photochemical quenching of photosynthesis (NPQ) resulted in an important decrease in the dissipated non-regulated energy (ΦNO), modifying the homeostasis of reactive oxygen species (ROS), through a decreased singlet oxygen (1O2) formation. Thus, when clary sage was exposed to Cd for a short period, tolerance mechanisms were triggered, with PSII photochemistry to be regulated by NPQ in such a way that PSII efficiency to be enhanced. However, exposure to a combination of Cd and high light or for longer duration (8 days) to Cd alone, resulted in an inhibition of PSII functionality pointing out towards Cd toxicity. Thus, the rapid activation of PSII functionality at short time exposures and the inhibition at longer duration suggests a hormetic response and describes these effects in terms of “adaptive response” and “toxicity”, respectively.


Introduction
Cadmium (Cd) is occurring in soils at low concentrations but can arise to high concentrations as a result of numerous human activities, while being not biodegradable in the soil is considered as one of the most toxic elements, and also a non-essential element for plants [1][2][3][4][5][6]. Cadmium is taken up by roots and its translocation from the roots to the shoots and leaves with subsequent accumulation in the chloroplasts, will eventually disturb photosynthesis [7]. However, many plant species have established numerous special effective mechanisms for Cd detoxification and tolerance [4][5][6].
Cadmium tolerance is related with declines of the internal Cd accumulation [8] but some plant species are able to sustain growth or even improve their growth and functioning under Cd exposure, despite a high Cd accumulation in roots and shoots [9][10][11]. Foliar Cd content above 0.01 % dry  respectively, but while Mg accumulation decreased in both roots (43%) and shoots-leaves (53%) (Figure 3b), Mn accumulation increased in roots (1.6-fold) and its translocation to the shoots decreased (65%) (Figure 2d). Magnesium, the next most negatively affected element after Ca, decreased in roots from 4184 ± 126 µg g -1 before Cd treatment to 2391 ± 72 µg g -1 , while in aboveground tissues from 5851 ± 175 to 2778 ± 83 µg g -1 . Manganese increased in roots from 62.01 ± 1.8 µg g -1 before Cd treatment to 98.1 ± 2.9 µg g -1 , while in aboveground tissues decreased from 64.2 ± 1.9 to 22.88 ± 0.69 µg g -1 . . Calcium (a) and Mg (b) content, in µg g -1 dry weight, of control (con) and 5 days Cd treated Salvia sclarea aboveground (shoots-leaves) tissues, roots and whole plants. Error bars are standard deviations (n=5). Means between the two treatments that are statistically different (p < 0.05) are marked by an asterisk (*).

The Efficiency of Photosystem II after Cadmium Exposure
In order to understand how PSII functionality is affected by exposure of plants to Cd, we measured the maximum efficiency of PSII photochemistry (Fv/Fm) ( Figure 4a) and the efficiency of the water-splitting complex on the donor side of PSII (Fv/Fo) (Figure 4b). Both parameters increased (p < 0.05) in S. sclarea plants exposed to Cd stress for 2 and 5 days compared to their respective controls, indicating an enhanced PSII functionality under Cd stress.

Changes in the Quantum Yields and the Fraction of Open Photosystem II Reaction Centers after Cadmium Exposure under Low Light
The quantum efficiency of PSII photochemistry (ΦPSΙΙ) measured at low light (LL, 220 μmol photons m -2 s -1 ) ( Figure 5a) increased (p < 0.05) in S. sclarea plants exposed to Cd stress for 2 and 5 days compared to their respective controls, indicating a higher fraction of absorbed light energy to be directed to photochemistry under Cd stress. The concomitant increase in the quantum yield of regulated heat dissipation in PSII (ΦNPQ) (Figure 5b) under 2-and 5-days Cd stress resulted in a significant decrease in the quantum yield of non-regulated energy dissipated in PSII (ΦNO) (Figure  5c), indicating a better use of the absorbed light energy in S. sclarea Cd stressed plants. The fraction of open PSII reaction centers (qP), at LL (Figure 5d), increased (p < 0.05) in S. sclarea plants exposed to Cd stress for 2 days, while at 5 days exposure remain the same, compared to controls.

Changes in Non-photochemical Fluorescence Quenching and Electron Transport Rate after Cadmium Exposure under Low Light
Non-photochemical quenching (NPQ) increased (p < 0.05) in plants exposed to Cd stress for 2 and 5 days compared to their respective controls (Figure 6a). Electron transport rate measured at 220 μmol photons m -2 s -1 (Figure 6b) increased (p < 0.05) in S. sclarea plants exposed to Cd stress for 2 and 5 days compared to their respective controls, following the pattern of ΦPSΙΙ (Figure 5a).

Changes in Excess Excitation Energy under Low and High Light after Cadmium Exposure
The excess excitation energy (EXC) at PSII in S. sclarea, at 220 μmol photons m -2 s -1 (LL) after 2 days Cd exposure, decreased (p < 0.05) compared to control while, after 5 days Cd exposure was at the same level to control ( Figure 7a). However, under high light (HL, 900 μmol photons m -2 s -1 ) after 5 days Cd exposure, EXC increased (p < 0.05) compared to control (Figure 7b), indicating that the synergistic effect of Cd stress and HL resulted in a lower efficiency of light energy use by PSII.

Changes in the Quantum Yields under High Light after Cadmium Exposure
The allocation of absorbed light energy to PSII photochemistry (ΦPSΙΙ) measured at HL in S. sclarea plants exposed to Cd for 5 days decreased (p < 0.05) compared to controls, while the regulated heat dissipation (ΦNPQ) increased (p < 0.05) in such a degree that the non-regulated energy dissipation (ΦNO) in S. sclarea plants exposed to Cd for 5 days did not differ compared to controls ( Figure 8).

Changes in Non-photochemical Fluorescence Quenching, Electron Transport Rate and the Fraction of Open Photosystem II Reaction Centers under High Light after Cadmium Exposure
Non-photochemical quenching (NPQ), measured at HL, increased (p < 0.05) in S. sclarea plants exposed for 5 days to Cd compared to control plants (Figure 9a), while PSII electron transport rate decreased (p < 0.05) compared to controls (Figure 9b) following the pattern of ΦPSΙΙ ( Figure 8). The fraction of open PSII reaction centers (qP) decreased (p < 0.05) in S. sclarea plants exposed for 5 days to Cd compared to control plants ( Figure 9c).

Chlorophyll a Fluorescence Images under Low and High Light
Chlorophyll a fluorescence images of the fluorescence parameters ΦPSΙΙ and ΦNO, measured at LL and HL, of control and 5-days Cd treated S. sclarea plants, revealed a spatial heterogeneity over the whole leaf area (  At longer duration exposure (8 days) to Cd under LL, an inhibition of PSII functionality is observed ( Figure S2). More specifically, at 8 days exposure of S. sclarea plants to Cd, the lower Fv/Fm values were found near the midvein while the lower ΦPSΙΙ values at the leaf area behind the leaf edge ( Figure S2).

Chloroplast Ultrastructure after Cadmium Exposure
Leaves from untreated plants exhibited rather electronically dense mesophyll chloroplasts (Figure 11a) that showed a typical internal membrane structure with well-organized grana and stroma thylakoids (Figure 11a). Both 2-day ( Figure 11b) and 5-day ( Figure 11c) Cd treated plastids did not show any noticeable structural disruption having a similar appearance to the control, with the 5-day Cd treated plastids to appear more electronically dense (Figure 11c). In both treated and untreated plastids starch grains were noticeable.

Discussion
Exposure of S. sclarea plants to 100 μM Cd for 5 days in hydroponics resulted in a high Cd uptake with a 2400-fold increase at the whole plant level but with Cd ions to be retained almost exclusively in the roots ( Figure 1b) and only 53.3 µg g -1 to be translocated to the aboveground tissues (Figure 1a). Low Cd accumulation in leaves may represent a tolerance mechanism that protects the photosynthetic equipment against additional oxidative stress [50][51][52]. In the tolerant plant species, the excess heavy metals in roots play a significant role by sequestrating and detoxifying the extreme amount of heavy metal in order to protect the delicate aboveground photosynthetic tissues [53,54]. Since leaf Cd contents greater than 5-10 μg g −1 have been characterized toxic to most plants [55], it seems that S. sclarea could have kept Cd concentration in the photosynthetic tissues in non-toxic forms. This, could be done by depositing it in the vacuoles of leaf epidermal cells [56] and/or by complexation with cellular ligands [7,[57][58][59]. Hyperaccumulators can accumulate Cd to levels above 100 μg g −1 of shoot dry weight, without showing any toxicity symptoms [7,12,60]. Our results agree with those of He et al. [61] that roots of Cd tolerant plants (non-hyperaccumulators) retain considerably higher Cd concentrations than the aboveground parts, and only minor Cd is translocated to the aerial parts. In comparison to the above-ground tissues, S. sclarea roots showed a higher bioaccumulation ability of Cd with translocation to shoots-leaves to be restricted.
Cadmium uptake is affected by Ca levels because Cd competes for Ca channels [7,[62][63][64] and the low Ca content of the hydroponic solution may enhance Cd uptake [63,65] resulting in enhanced Cd and decreased Ca in many plant species [7,64], as we also observed in S. sclarea experiments, with Ca being the most affected element (58% total uptake decrease, Figure 3a). Likewise, in Oryza sativa seedlings exposed to Cd, uptake of Ca was decreased, and Ca content in both roots and aboveground parts was significantly reduced [66]. Magnesium was the next most negatively affected element after Ca in our experiments, but low Mg status has been associated with increased Cd tolerance [67][68][69][70][71][72], indicating that plants regulate nutrient concentrations to mitigate Cd toxicity [72].
Antagonistic effects of Cd with Fe [52,64,73] and Zn [74][75][76] have been frequently reported. However, in S. sclarea exposed to 100 μM Cd for 5 days, total Zn, Cu, and Fe uptake increased but their translocation to the aboveground parts decreased possible due to translocation barriers. It seems that Cd uptake in S. sclarea is not taking place through Fe or Zn pathway, while conditions that lead to increased Cd uptake in plants may also favor increased Fe uptake [7]. Cd treatment has been frequently mentioned that it increases Fe retention in roots but obstructs its translocation to shoots, thus reducing Fe concentrations in aboveground parts [61,77,78]. In rice Cd has been shown to be taken up predominantly via the Mn pathway [79,80] but this was not the case in S. sclarea, since Mn was the less negatively affected element.
Cadmium contamination of soil has become a serious environmental alarm as it is estimated that around 30,000 t of Cd are released annually into the environment with a consequence to the food chain and a threat to human health [81,82]. The use of plants for heavy metals elimination from pollutant soils and water is a technique known as phytoremediation [83]. In phytoremediation, plants that absorb heavy metals from soils and translocate them to the harvestable shoots are used for phytoextraction, while those that stabilize metal contaminants through accumulation in the root zones are used for phytostabilization [82,84]. Plant species with high bioconcentration factor but relatively low translocation factor (< 1) may be considered as potential phytostabilizers [85]. Salvia sclarea exhibited high accumulation capacity for Cd and by limiting its translocation from roots to shoots may be considered a potential phytostabilizer that can be used in heavy metal contaminated environments. Other plant species that have been proposed as Cd phytostabilizers are Iris lactea [82] and Sesuvium portulacastrum [86,87].
Despite the significant levels of Cd in leaves, a higher fraction of absorbed light energy was directed to photochemistry (ΦPSΙΙ) under 2-and 5-days Cd stress, with a concomitant increase in ΦNPQ that resulted in a significant decrease in ΦNO. The non-regulated energy loss in PSII (ΦNO) encompasses internal conversions and intersystem crossing, that results to singlet oxygen ( 1 O2) creation via the triplet state of chlorophyll ( 3 chl * ) [31,[88][89][90][91]. To optimize photosynthesis and growth under stressful conditions plants have evolved a variety of mechanisms against photodamage and photoinhibition [92,93]. Non-photochemical quenching is the key photoprotective process that dissipates excess light energy as heat and protects photosynthesis [88,[94][95][96][97][98]. Thus, the increased non-photochemical quenching of photosynthesis (NPQ) decreased ROS through a 1 O2 formation. Consequently, in S. sclarea plants exposed to 100 μM Cd, ROS homeostasis could be regulated by NPQ in such a way so that plants can cope with Cd stress [21,24,28,29].
The potential PSII efficiency of S. sclarea plants exposed to 100 μM Cd estimated by the maximum efficiency of PSII photochemistry (Fv/Fm) (Figure 4a) and the efficiency of the watersplitting complex on the donor side of PSII (Fv/Fo) (Figure 4b) indicated an enhanced PSII functionality under Cd stress. In accordance, S. sclarea plants exposed to 100 μM Cd show an increased capacity to keep quinone (QA) oxidized, thus, to have a higher fraction of open PSII reaction centers (qP) compared to controls. In other words, S. sclarea plants exposed to 100 μM Cd show a low PSII excitation pressure associated with toxicity tolerance mechanisms [99,100]. High excitation pressure defines excess energy and consequently a disproportion between energy resource and requirement [101]. This discrepancy leads to an increase in the energy transmitted from chlorophyll to oxygen, resulting in 1 O2 generation [102]. Control S. sclarea plants that show increased excess excitation energy (EXC) at PSII (Figure 7a), show also increased 1 O2 creation via 3 chl * (Figure  5c), compared to plants exposed to Cd for 2 days. A basal level of ROS is needed for optimal growth [22,27], with a low level of ROS to be beneficial in accomplishing the maximum fitness of the biological plasticity, and a high level of ROS out of the boundaries to be harmful to plants [22,36]. Photosystem II responses of S. sclarea plants to short time Cd exposure can be described as a hormetic response ( Figure 12) representing an ''over-compensation'' response to a disruption in homeostasis [44]. Hormesis [103] is defined as the stimulatory effect of short exposure times of toxic constituents, e.g., Cd on a biological factor (photosystem II photochemistry), of a particular organism (S. sclarea). The hormetic effect is defined by an inverse U-shaped biphasic curve [8,36] in which short exposure time has a stimulatory effect; however, at longer exposure time, a toxic effect is evident.
Although excess Cd accumulation is detrimental to plants, different strategies of Cd tolerance and accumulation are adopted by plants [104]. Tang et al. [105] described a stimulation of plant growth, increase of photosynthesis and an up-regulation of the related genes in Sedum alfredii exposed to 5 μM Cd. Similar results with stimulation of growth when Noccea caerulescens was exposed to 100 μM Cd were reported by Lombi et al. [106]. A stimulatory effect of Cd on the photosynthetic apparatus of Arabidopsis halleri was also described recently [35]. Małkowski et al. [38] reported a stimulation of the photosynthetic rate by Cd only at low concentrations, whereas at higher Cd concentrations there was a significant decrease compared to controls. However, other studies have shown a dose dependent negative impact of Cd that increases with the generation of ROS and oxidative damage and the inhibition of photosynthetic rate to follow [107-116new 1, new 2]. On the other hand, protection to stress through the production of ROS [117,118] has been shown that it can be regulated by NPQ in such a way so that plants can cope with stress [24,28,29].
A negative impact of Cd on photosynthesis has been assigned to decreases in chlorophylls; ascribed to Cd-induced damage in chloroplasts' ultrastructure [119,120]. Cadmium has been reported to alter chloroplast ultrastructure, reduce photosynthesis [64,116,121], and inactivate enzymes involved in CO2 fixation [62]. Ultrastructural changes that are observed in Cd exposed leaves of sensitive plants (thylakoid dismantling, increase of lipid droplets, etc) [122][123][124][125][126], are similar to those occurring at leaf ageing [127]. In Cd tolerant species the only ultrastructural alteration observed in leaves was a reduction of starch grains in chloroplasts [52,128,129] that may be due to disorders in the photoassimilate transport or to nutrient deficiency [52,130]. In S. sclarea Cd-treated plants an increase in deposited electron-dense material was observed, as reported also by Mizushima et al. [52], but no other noticeable alteration, further consolidating the chlorophyll fluorescence imaging results. Thus, a hormetic response of PSII photochemistry to short term Cd exposure was observed, indicating an ''over-compensation'' response to Cd disruption in homeostasis, justifying the statement of Carvalho et al. [8] that Cd can be regarded from a toxic element, a beneficial one. Hormesis research data and data on priming (preconditioning), an expression of hormesis [17,36,[131][132][133], indicate that stimulatory response detection of the low-dose or short-time exposure are highly dependent on the study strategy, including dose range and number with duration exposure and endpoint selected [14,15,17,36,54].
Exposure of S. sclarea plants to a combination of Cd and high light (900 μmol photons m -2 s -1 ) resulted in an inhibition of PSII functionality (Figures 8,10), while the increased NPQ (Figure 9a [96,97,[134][135][136][137], as was observed under low light exposure of S. sclarea plants to Cd (Figure 5d). Thus, the combination of Cd and high light is pointing out to Cd toxicity. The same conclusion is reached at longer duration exposure (8 days) to Cd at LL, with an inhibition of PSII functionality to be observed ( Figure S2) [138].

Plant Material and Growth Conditions
Seeds of Salvia sclarea L. used for the experiments were collected from the Rose Valley (Bulgaria). After germination on soil in a growth room for about a month, the seedlings were transferred to pots containing continuously aerated modified Hoagland nutrient solution (described in detail before) [6]. The nutrient solution was adjusted to pH 6.0 and changed every 3 days. The growth room conditions were 24±1/20±1 °C day/night temperature, 14/10 h day/night photoperiod with photon flux density 200±20 μmol photons m -2 s -1 .

Cadmium Treatment
Two-month-old S. sclarea plants in the hydroponic culture experiments were subjected to 0 or 100 μM Cd (as 3CdSO4 8H2O) for a period up to five days. The pots containing only Hoagland nutrient solution served as control, while all solutions were renewed every two days.

Determination of Elemental Concentration by Inductively Coupled Plasma Mass Spectrometry (ICP-MS)
After 5 days treatment with 0 (control) or 100 μM Cd, Salvia plants were harvested, separated in roots and aboveground (shoots-leaves) tissues, washed three times in deionized water and then dried at 65 °C to constant biomass, milled and finally sieved. Dried sieved samples of 0.3g were transferred in 10 mL quartz vessels with 65% (v/v) nitric acid (Suprapur, Merck, Germany) and 30% (v/v) hydrogen peroxide (Suprapur, Merck, Germany) in 3:1 ratio. Digestion was carried out in the microwave assisted digestion system Ethos One (Milestone Sorisole, BG, Italy). The process run out in 3 stages: ramp time -20 min to reach 200 °C and 1500W; hold time -30 min at 200 °C and 1500W; cooling -30 min. Next step was the quantitative transfer of digested samples into polypropylene tubes and dilution with demineralized water (Direct-Q 3 UV, Merck, Germany). All prepared samples were diluted immediately prior to inductively coupled plasma mass spectrometer (ICP-MS) analysis. Samples were analyzed in an ICP-MS model ELAN DRC II (PerkinElmer Sciex, Toronto, Canada) [139]. ICP-MS operational conditions, instrumental settings calibration solutions, data validation and validation parameters are given in Appendix A. Elemental analysis was performed for Cd, Cu, Ca, Mg, Mn, Fe and Zn.

Chlorophyll Fluorescence Imaging Analysis
Chlorophyll fluorescence measurements were conducted on dark adapted (20 min) leaves of S. sclarea plants, treated for 2-and 5-days with 0 (control) or 100 μM Cd, using an Imaging PAM M-Series system (Heinz Walz Instruments, Effeltrich, Germany) as described in detail previously [140]. Two light intensities were used for measurements of photosynthetic efficiency of S. sclarea leaves, a LL, similar to the growth light (220 μmol photons m −2 s −1 ), and a HL (900 μmol photons m −2 s −1 ). In each leaf representative areas of interest (AOIs) were selected so as to have measurements of the whole leaf area. The definitions of the five main chlorophyll fluorescence parameters (Fo, Fm, Fo', Fm' and Fs) measured by the Imaging PAM M-Series system are presented in Table S1, while a typical modulated fluorescence trace showing how the main five parameters are formed is presented in Figure S1. The chlorophyll fluorescence parameters calculated from the five main parameters with their definitions are described in Table 1. Representative results are also shown as color-coded images of ΦPSΙΙ and ΦNO, after 5 min illumination with 220 μmol photons m -2 s -1 (LL) and 900 μmol photons m -2 s -1 (HL) intensity.

Leaf Ultrastructure Observations by Transmission Electron Microscopy
To study leaf ultrastructure alterations after 2-and 5-days of Cd treatment, leaves from both Cdtreated and untreated plants were excised and segmented with a razor blade into small pieces of 0.5 x 1 mm. Leaf segments were fixed with 2% paraformaldehyde plus 4% glutaraldehyde, in 0.05 M sodium cacodylate buffer, pH 7.0 solution [140]. After a 5h fixation at room temperature the samples were washed with a 0.05 M sodium cacodylate buffer and post-fixed for another 3 h in a similarly buffered 2% osmium tetroxide solution (Agar Scientific, Essex, UK). Afterwards samples were dehydrated in an acetone series, treated with propylene oxide, and embedded in Durcupan ACM resin (Fluka Chemie AG, Buchs, Switzerland). Ultrathin sections (80-90 nm) were cut in a ULTROTOME III TYPE 8801A ultramicrotome (LKB, Stockhom, Sweden) equipped with a glass knife., collected on nickel grids. Ultrathin sections were stained with 2% uranyl acetate and 1% lead citrate and examined in a JEOL JEM 1011 transmission electron microscope equipped with a Gatan ES500W digital camera. Digital electron micrographs were obtained with the DigitalMigrograph 3.11.2 software according to the manufacturer's instructions.

Statistical analyses
Mean values were calculated from three independent treatments (biological replicates). Statistically significant differences among the means were determined using one-way analysis of variance or two-way ANOVA. Means (± SD) were considered statistically different at a level of p < 0.05.

Conclusions
Although surplus Cd accumulation is detrimental to all plants different strategies of Cd tolerance and accumulation are adopted by different plants [104]. When clary sage was exposed to Cd for a short time, tolerance mechanisms were triggered, with PSII photochemistry to be regulated by the photoprotective mechanism of NPQ in such a way that PSII functionality to be enhanced (Figures 4-6) without any defects to chloroplasts, as observed by transmission electron microscopy ( Figure 11). However, exposure to a combination of Cd and high light (Figures 8,10) or longer duration exposure to Cd alone (8 days, Figure S2) [138] resulted in an inhibition of PSII functionality pointing out to Cd toxicity. Thus, an activation of PSII function at short time exposures and an inhibition at longer duration suggests a hormetic response (Figure 12), and describes these effects in terms of "adaptive response" and "toxicity", respectively.

Supplementary Materials:
The following is available online at www.mdpi.com/xxx/s1, Table S1: The definitions of the five main chlorophyll fluorescence parameters (Fo, Fm, Fo', Fm' and Fs) measured by the Imaging PAM M-Series system. Figure S1: A typical modulated fluorescence trace showing how Fo, Fm, Fo', Fm' and Fs, are formed to measure photochemical and non-photochemical parameters. Figure

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

ICP-MS operational conditions
Operational conditions were optimized daily, using a solution of Mg, In, and U at a concentration of 1 mg L -1 and Ba at a concentration of 10 mg L -1 (Smart Tune Solution e Elan DRC II/plus, Atomic Spectroscopy Standard, Perkin Elmer Pure). Whilst tuning the ICP-MS, compromise conditions for maximum signal intensity of the analyte ( 24 Mg + , 115 In + , 238 U + ) and minimum ratio of oxide ( 140 Ce16O + / 140 Ce < 3%) and doubly charged ions ( 128 Ba 2+ / 128 Ba + < 3%) were found.

Calibration solutions
Calibration solutions were prepared by appropriate dilution of 10 mg L -1 multi-elemental stock solution in 5% HNO3 (Multi-Element Calibration Standard 3, PerkinElmer, MA, USA). The calibration curves were constructed in the concentration ranges: 0.1-100 µg L -1 for Cd, Cu, Mn, Zn and 50-1500 µg L -1 for Ca, Fe and Mg.

ICP-MS instrumental settings
The ICP-MS instrumental settings were as follows: (0.89-0.91) L min -1 sample gas flow, 16 L min -1 plasma gas flow, 1.2 L min -1 auxiliary gas flow, 1250W RF generator power, dual detector mode, autolens mode. Dynamic reaction cell (DRC) mode with ammonia as reaction gas was used in order to remove the polyatomic interferences. The non-spectral interferences were reduced by diluting the sample and using 10 µg L -1 of Ge and Rh as internal standard.

Data validation
For data validation of the applied analytical procedure, two CRMs were used: Trace elements in spinach leaves NIST SRM 1570a (NIST, USA) and TM-27.4 (National Water Research Institute, Canada).

Validation parameters
Validation parameters such as linearity, precision, LOD and trueness were evaluated. The calibration curves for the determined elements were linear in the range of calibration standards. Coefficient of correlation (R) values were estimated daily and were greater than 0.999 for all analytes. Residual plots showed a random distribution of residuals around the vertical axis. Precision values were calculated as coefficient of variation (CV) (%) ranged from 1.7% to 3.7% for all elements. Trueness was evaluated by applying the certified reference materials and expressed as recovery (%).
Recovery values ranged from 95% to 106% respectively. LOD values were estimated as 3.3 S/b, where S means the standard deviation of the result obtained for the blank sample and b is the sensitivity (n=5). The LODs were as follows: Cd 0.02 μg g -1 , Cu 0.05 μg g -1 , Ca 30 µg g -1 , Mg 0.8 μg g -1 , Mn 0.03 μg g -1 , Fe 40 μg g -1 and Zn 0.01µg g -1 . Measurement traceability was established by applying the certified reference materials.