Soil dynamics of Cr(VI) and responses of Portulaca oleracea L. grown in a Cr(VI)-spiked soil under different nitrogen fertilization regimes

The reduction potential of the highly toxic Cr(VI) to the inert Cr(III) in an alkaline soil was studied during a 50-day experiment with Portulaca oleracea L. grown in pots. We aimed at assessing whether our test species can be a phytoremediation candidate for Cr(VI)-contaminated soils. We measured the Cr(VI) reduction rate in soil, determined the Cr(VI) and Cr(III) concentrations in aerial and root P. oleracea tissues, and calculated the transfer coefficient (TC = metal in plant over metal in soil) and the translocation factor (TF = metal in aerial biomass over metal in roots) in order to assess Cr(VI) uptake and distribution in plant tissues, while we also studied the effect of added nitrogen in the studied parameters. We added five different Cr(VI) levels (from the unamended T-0 to the treatment of T-4 = 150 mg Cr(VI) kg−1 soil) and also had two N levels (equivalent to 0 and 200 kg ha−1). The results indicated that Cr in plant tissues was mainly found in its reduced form (Cr(III)) and only a minor fraction of Cr was detected in its oxidized form (Cr(VI)), with only 1.04% of plant Cr being hexavalent at T-4 with no added N and 1.30% at T-4 with added N. The main remediation mechanism was found to be that of the naturally occurring Cr(VI) reduction that effectively produced Cr(III), followed by the uptake of Cr(VI) from our test plants (at T-4 with no N, 58% of soil added Cr(VI) was reduced and 0.1% absorbed, while at T-4 with added N, 63% was reduced and only 0.4% absorbed by plant). We also found that Cr(VI) in P. oleracea tissues was mainly found in roots and relatively low Cr(VI) concentrations were found in the above-ground tissues. We concluded that P. oleracea is a tolerant plant species, especially if assisted with a sufficient level of N fertilization, although it failed to approach the threshold of being categorized as an accumulator species. However, as this is a rather preliminary experiment, before reaching more conclusive suggestions about P. oleracea as a potential phytoremediation species, further investigation is necessary in order to verify the gained results with naturally contaminated soils with Cr under field conditions.


Introduction
Metal ions can be introduced to surface soils by natural or anthropogenic processes, and their environmental impact and availability are greatly affected by soil mineralogical and geochemical properties (Dong et al. 2021). Chromium has several oxidation states, ranging from −2 to +6. In the environment, Cr occurs primarily in two valence states, +3 (chromite (Cr(III))) and +6 (chromate (Cr(VI))), and in natural soil conditions, Cr is found predominantly in its trivalent state (Sharma et al. 2021). Geogenic derivation of hexavalent Cr is unusual, and its presence is rather entirely anthropogenic. Thus, Cr(VI) can be found mainly in the locality of industrial zones, but if discharged through streams, it may be found even kilometers down the stream, polluting adjacent surface soils and groundwater (Lilli et al. 2015;Megremi et al. 2019).
Several soil factors, such as pH, Cr speciation, and attributes of the soil colloidal phase, have a major influence on Cr availability (Ertani et al. 2017;Shi et al. 2020). Cr(III) is of low mobility in soil. In pH<4, insoluble inorganic compounds are formed (i.e., complexes of Cr 3+ and Fe oxides), and as pH increases, trivalent Cr is mainly found in its hydrolyzed form (which species are of the general form of Cr(OH) n 3−n 1 − 3); these species tend to form organic and inorganic complexes with fluoride, ammonium, cyanide, thiocyanate, oxalate, and sulfate, with inorganic ligands being of much lower solubility compared to the organic (Ertani et al. 2017;Jobby et al. 2018;Shahid et al. 2017). On the other hand, Cr(VI) can be retained more strongly than Cl − and SO 4 2− ions and its retention strength can be compared to that of phosphates on hydrous Fe and Al oxides surfaces (Jobby et al. 2018;Shi et al. 2020).
Hexavalent chromium, even in well-aerated soils, is expected to be readily reduced to the inert, less toxic, and lower mobility trivalent form (Cr(III)). Cr(VI) reduction, thus, acts as a natural, self-remediation process, which takes place even in the presence of particularly weak electron donors, such as H 2 O (Antoniadis et al. 2017b;Antoniadis et al. 2018;Chen et al. 2015). Cr(VI) reduction is commonly encouraged by various reduced soil carbon compounds as follows: 1) Cr(VI) reduction from soil carbon compounds in the form of carbohydrates:  (Antoniadis et al. 2017b;Jobby et al. 2018) Some indices are frequently used to assess trace element toxicity in plants: (1)soil-to-plant element mobility (transfer coefficient, TC), equal to the ratio of metal concentration in plant tissues (C p ) over the total concentration in soil (C s ) (TC = C p / C s ), which shows the potential of an element to be transferred from soil to the plant tissues. Plant species, Cr bioavailability, and Cr soil concentration are the major factors governing plant tissue Cr content and affecting TC. This index is taken into consideration when assessing a plant species for its phytoremediation potential, with values close, or greater than, unity being desirable (Antoniadis et al. 2017a;Antoniadis et al. 2021;Nagarajan and Sankar Ganesh 2014). An often used variant of this index is BAI (bioavailability index), with the same numerator but extractable soil Cr(VI) as the denominator (instead of total soil Cr(VI)). (2) Translocation factor (TF), equal to the ratio of metal concentration in aerial plant tissues (C aerial ) over that in roots (C root ) (TF = C aerial /C root ). Translocation factor is indicative of the plant species' capacity to control toxic element translocation to the aerial biomass where metabolic activity is more intense. Cr accumulates preferentially in roots, and minimal Cr concentrations are found in above-ground plant tissues. Cr distribution in plant tissues (roots, stems, and leaves) follows a stable plant species-specific pattern that appears to be independent of the soil Cr concentration and bioavailability. Desirable TF values for a plant with phytoremediation potential are over unity (Antoniadis et al. 2017a;Antoniadis et al. 2021;Ertani et al. 2017;Wu et al. 2021). Our test plant, Portulaca oleracea L., is a plant of high added value when commercially cultivated and well-known for its tolerance towards harsh abiotic stresses, e.g., salinity and drought (Alam et al. 2014;Karkanis and Petropoulos 2017;Petropoulos et al. 2016;Ozturk et al. 2020). Such tolerant species may be candidates that would act as phytoremediation species in contaminated soils. However, tests assessing its potential as a possible phytoremediation species for Cr(VI)-laden soils are rare. Also, it is known that well-fertilized plants are more robust in addressing environmental stresses. The effects of fertilizers on chemical composition and plant growth have been evaluated (Alam et al. 2014;Disciglio et al. 2017;Montoya-García et al. 2018;El-Saadany et al. 2015), but these works focus mainly on P, an anionic element known to have a similar assimilation behavior to Cr(VI) in plants. However, N may also have a significant role: The scientific hypothesis tested in this work is that the improvement of plant functions due to adequate N nutrition may help the species not only tolerate higher Cr(VI) stress but also improve the plant's ability to uptake higher quantities of Cr(VI). To the best of our knowledge, the effect of sufficient levels of added N in the behavior of P. oleracea in addressing Cr(VI) stress is never before explored and needs to be investigated. Therefore, the aim of this work was to assess Cr(VI) dynamics in soil, and the responses of P. oleracea concerning its ability for Cr(VI) uptake, and the Cr speciation in plant tissues, as well as the distribution of Cr(VI) in plant tissues, under different N fertilization regimes. Based on the used phytoremediation indices, we also explored a practical issue, i.e., the number of P. oleracea harvests required to annihilate the spiked soil Cr(VI) levels, an approach that also indicates the novelty of this study.

Materials and methods
Soil properties, Cr(VI) spiking, and nitrogen application A pot experiment was established using an alkaline soil obtained from an area away from any known source of pollution, Velestino (39.394930 N,22.757112 E), near Volos, central Greece. The soil was sieved through a 2-mm mesh sieve, and three samples randomly acquired were analyzed for selected physiochemical parameters (Table 1) according to established protocols (Koutroubas et al. 2020). For our pot experiment, we prepared a Cr(VI)-spiking solution, prepared at a concentration of 10,000 mg L −1 , by dissolving 19.22 g of CrO 3 in 1000 mL of double distilled water. The spiking solution was added to soil, resulting in 5 different Cr(VI) soil concentrations, namely T-0: 0.0 mg Cr(VI) kg −1 soil (control), i.e., unamended soil; T-1: 20 mg Cr(VI) kg −1 soil, by adding 2 mL of the spiking solution per kg soil; T-2: 50 mg Cr(VI) kg −1 soil (with 5 mL of spiking solution kg −1 soil); T-3: 100 mg Cr(VI) kg −1 soil (10 mL of spiking solution kg −1 soil); and T-4: 150 mg Cr(VI) kg −1 soil (15 mL of spiking solution kg −1 soil) (Table 2).
Also, we added nitrogen to half of the replicates of each of the established treatments, while the other half did not receive any additional N. Nitrogen was added at rates equivalent to 200 kg per hectare (considering an effective rhizosphere depth of 15 cm and dry bulk density of 1.33 g cm −3 ), by applying 20 mL per kg of soil of a solution containing 14.3 g NH 4 NO 3 L −1 (equal to 5 g N L −1 ). The non-added-N treatments are thereafter named N-0, while those that received N, N-1. Overall, the experimental design resulted in 10 treatments (5 Cr(VI) rates × 2 nitrogen rates). The spiked soils were placed into 2-L pots, with each pot containing 1 kg of soil; each treatment consisted of 10 replicates. Pots were irrigated to 65% of their water holding capacity and left to equilibrate for 20 days. During the equilibration period, pots were thoroughly mixed every second day and water was added to compensate for moisture loss. At the end of the equilibration period (considered as day 0 of the experiment), four soil samples per Cr(VI) treatment were obtained from the pots, air dried, and passed through a 2-mm sieve in order to determine the initial hexavalent chromium (Cr(VI)) concentration. After this initial sampling, plants on day 0 were transplanted into the pots, as explained in the subsequent section.

Plant establishment, measurements, and soil and plant analyses
On day 0, P. oleracea plants, already sown in peat-filled seedling trays 25 days before day 0, were transplanted in the pots; when transplanted, they had reached a height of 12 cm. Pots were then placed in an unheated greenhouse. During the growth period, plants were watered according to their needs, and the positions of pots were exchanged regularly, to compensate for possible light and temperature differences. In the samples obtained on day 0, Cr(VI) was extracted using 0.01 M KH 2 PO 4 , color was developed with the diphenyl carbazide method, and absorption values were determined using a Biochrom Libra S11 spectrophotometer at 540 nm (Antoniadis et al. 2018). Soil total chromium (Cr(III) + Cr(VI)) was extracted with aqua regia (HCl/HNO 3 , 3/1) after digestion for 2 h at 180°C in a Velp DK 20 digestion unit (Golia et al. 2020), and total chromium concentrations were determined using a flame atomic absorption photometer (Perkin Elmer 3300). Trivalent chromium (Cr(III)) concentrations were calculated by subtracting the hexavalent chromium concentration from total chromium (Cr(III) = Cr(III+VI) -Cr(VI)) (Molla et al. 2012).
The growth period lasted for 50 days, from 14 October 2019 (day 0-commencement of the experiment) to 4 December 2019 (harvest day). On the harvest day, we measured the weight of stems and the leaf area per plant. The plants were cut 2 cm above the soil surface, and the weight of fresh leaves was recorded. Then, the aerial plant tissues were washed with deionized water and placed in a draughtforced oven at 70°C. Roots were meticulously washed so that no soil particles remained attached, rinsed with deionized water, and likewise placed at 70°C. Plant tissues remained in the oven until no further weight loss was recorded, i.e., for 96 h, after which both aerial and root tissues were weighed and pulverized. Then, 1.00 g of plant tissue was dry-ashed at 500°C for 4 h and extracted with 10 mL of 20% HCl. For plant tissues, Cr(VI) and Cr (III+VI) content was estimated.
Also, on the harvest day, four soil samples per pot were obtained; the samples were obtained throughout the whole depth of the pots and mixed into one composite sample per pot, so that samples may be as much representative for each particular pot as possible. Samples were then extracted for Cr(VI) and Cr(III+VI) concentrations. Available Cr(III) concentrations were determined with extraction with DTPA-TEA-CaCl 2 pH 7.3 (Lindsay and Norwel 1978), and residual soil N (assumed to be in the form of NO 3 -N) was determined (Norman et al. 1985). As for the determination of Cr(VI) and Cr(III+VI) in soil, this was performed as described for the soil samples of day 0. The uptake of Cr(VI) and Cr(III) by plant, measured as the quantity of Cr in plant per pot, i.e., mg Cr in plant pot −1 , was then calculated by multiplying concentration

Indices
Apart from soil and plant tissue analysis, transfer coefficient, translocation factor, and bioavailability index were assessed (Antoniadis et al. 2017a;Antoniadis et al. 2018;Buscaroli 2017;Levizou et al. 2019). Full details concerning their calculation are reported in Appendix (Table 5).

Quality assurance and statistical analysis
For data quality control assurance, certified soil (CRM051 and CRM042 -Labmix24 GmbH, Germany) and in-house plant and soil reference materials were used. Recovery rates of the reference materials concerning total (aqua regia) Cr were within the range of 95 to 105% of the certified value. In every extraction batch, blank samples were included, in order to rule out any case of cross-contamination. For Cr calibration curves, standard solutions from Merck were used. Every sample was measured in triplicate, and samples with coefficient of variation of greater than 15% were discarded and measured again. Statistical analysis of the data was performed using IBM SPSS Statistics 26 and Excel 2019. To identify statistical significance among differences of all treatments, two-way ANOVA and post-hoc Duncan's multiple range tests were performed.

Soil characteristics and Cr(VI) reduction rate
In this experiment, we used an alkaline (pH 7.8) soil with medium CaCO 3 content (10.4%) and intermediate clay content of 16%. Indigenous Cr(III) concentration of geogenic origin was 221.3 mg kg −1 and that of Cr(VI) was below the detection limit (Table 1). The evolution of Cr(VI) concentration in soil throughout the growing period is presented in Table 2. In aerated soils, Cr(VI) reduction to Cr(III) is favored as part of a dynamic system affected by a series of parameters (Antoniadis et al. 2018;Bartlett 1991;Ertani et al. 2017), the most important of which are pH and redox potential. Because of the naturally occurring Cr(VI) reduction process, at the end of our 50-day growing period, Cr(VI) soil concentrations were reduced to 9.72 at T-1, to 15.95 at T-2, to 45.47 at T-3, and to 50.71 at T-4 (units in mg Cr(VI) kg −1 of soil; Table 2). In order to illustrate the efficiency of Cr(VI) reduction, we calculated the percentage of Cr(VI) remaining in soil at the end of the experiment in relation to the Cr(VI) concentration on day 0. We found that in the lower added Cr(VI) rate (T-1, no N), almost half of the added Cr(VI) remained in soil, while the other half evolved to the reduced Cr(III). With increasing added Cr(VI), the percentage of Cr(VI) remaining in soil in the hexavalent species decreased, to the extent that almost 2/3 of added Cr(VI) was reduced to Cr(III). In the N-0 treatments, the percentage of decrease in Cr(VI) from the initial to the last experimentation day increased gradually with enhanced added Cr(VI); in the added-N treatments, this was rather not observed. As for the added-N treatments, no significant differences (p = 0.118) were noticed concerning soil Cr(VI) dynamics, indicating that added N in the form of NH 4 + (NH 4 NO 3 ) was not in any way in sufficient concentration to become effective electron donor for Cr(VI) reduction, and thus, it did not affect Cr(VI) reduction rates.

Plant uptake of Cr(VI)
In order to assess the potential of P. oleracea to be used as a phytoremediation species, we measured the overall Cr(VI) uptake per individual plant. If our assumptions were based solely on the absolute concentration of Cr(VI) in plant tissues, Values are reported as mean ± standard error of all measured replicates. Different letters within columns denote significant (p<0.05) differences among means according to Duncan's multiple range test leaving out of the discussion the produced plant biomass under Cr(VI) stress, the uptake potential of the plant species would not be securely assessed. Thus, we assessed the product of the multiplication of Cr(VI) concentration and plant biomass, i.e., Cr(VI) uptake (units in μg Cr(VI) in plant per pot or per kg of soil). As rather expected, the increasing soil Cr(VI) concentrations resulted in higher Cr(VI) uptake both for aerial and root biomass. More specifically, in the treatments of added Cr(VI) beyond T-2, the increase in Cr(VI) uptake was remarkable and nitrogen application at the highest Cr(VI) level (T-4) resulted in a further significant increase (p = 0.004), both for aerial (225.9 vs. 125.1 μg Cr(VI) pot −1 ) and root biomass (p < 0.001) (610 vs. 166.8 μg Cr(VI) pot −1 ) ( Table 3).

Remediation time
The time required for the studied plant to annihilate the added soil Cr(VI) was also calculated. For simplicity and in order to make more conservative estimations, the evolved Cr(III) was not taken into consideration and we thus proceeded in estimating the number of harvests required for purslane to fully

Indices of soil-to-plant mobility, bioavailability index, and translocation factor
We calculated the soil-to-aerial plant tissue Cr(VI) transfer coefficient. We found that the upward transfer from soil to the aerial tissues of purslane was significantly affected by Cr(VI) soil concentrations (p = 0.004), with the highest index values being observed at T-1 and the lowest at T-3 (Fig. 1a); the effect of nitrogen was not significant (p = 0.854). The values of soil-to-root transfer index were 10-to 100-fold greater than those of the aerial tissues (TC aerial ); with increasing Cr(VI) soil concentrations, TC root gradually increased (p < 0.001), irrespective of added N (p = 0.302). On the combined effect of Cr(VI) and N on TC root , it is evident that at T-4, roots absorbed more Cr(VI) and upon nitrogen application TC root values were almost doubled (p < 0.001) (Fig. 1b). These results indicate that Cr(VI) uptake from roots is concentrationdependent and nitrogen addition has a positive effect on Cr(VI) root absorbance, as it evidently increases plant vigor and thus its ability to absorb Cr(VI). Bioavailability index (BAI) indicated that neither Cr(VI) soil concentration (p = 0.116) nor nitrogen addition (p = 0.153) exerted significant effect in it. Our data suggest a decreasing trend with increasing Cr(VI) concentrations from T-0 to T-3, indicating that plants effectively limited the uptake of Cr(VI) up to the level of added Cr(VI) corresponding to T-3, while in higher Cr(VI) levels, plant mechanisms that limit Cr(VI) uptake proved rather insufficient (Fig. 1c). Translocation of Cr(VI) from roots to shoots (reported for clarity in values multiplied with 1000, similar to transfer coefficient) was notably lower at T-3 and T-4 compared to that at T-1 and T-2. In the highest Cr(VI) soil concentrations, TF was remarkably low, indicating that P. oleracea under high Cr(VI) stress manages to effectively limit Cr(VI) translocation from root to the aerial tissues. The effect of nitrogen on TF values was not significant (p = 0.128) (Fig. 1d).

Chromium speciation in plant
In aerial plant tissues, low values of the Cr(VI)/Cr(VI+III) ratio were observed, indicating that Cr was almost exclusively found in its trivalent form. In the high added Cr(VI) concentrations (T-3 and T-4), a decreasing trend was noticed for the Cr(VI)/Cr(VI+III) ratio, indicating that defense mechanisms resulting in Cr(VI) reduction were triggered in response to excessive Cr(VI) stress. On the other hand, the Cr(VI) to total Cr ratio in roots reached noticeably higher values. Contrary to what was found for the aerial parts, increasing added soil Cr(VI) concentrations resulted in higher Cr(VI)/ Cr(VI+III) ratio values (p < 0.001) and nitrogen amendment also resulted in significantly higher values when combined with the highest (T-4) Cr(VI) soil concentration (p < 0.001); otherwise, nitrogen had a non-significant effect (p = 0.117) on the plant Cr(VI)/Cr(VI+III) ratio (Fig. 2).

Soil residual nitrogen
As added Cr(VI) increased, soil concentrations resulted in lower N uptake (p < 0.001) (Table 4), resulting in the accumulation of N in soil at the end of the experiment. Low N uptake from plants under Cr(VI) stress may be due to the reduced plant growth rate that subsequently led to lower N demands.

Discussion
Cr(VI) reduction to the less toxic Cr(III) during our 50-day growing period of P. oleracea was as high as 79.2% of the initial added Cr(VI) soil concentration. This is in agreement with various works confirming that after spiking, Cr(VI) is reduced under normal soil conditions relatively fast. Soil Cr(VI) reduction rates have been reported to be largely dependent on soil conditions such as moisture, temperature, pH, and redox potential (Korai et al. 2021;Kozuh et al. 2000). The most prominent among soil properties is undoubtedly the organic matter content, which was 2.6% in the soil used in our work (equivalent to 1.5% of organic C, Table 1); this was also agreed by Raptis et al. (2018) who suggested that increasing organic matter content through soil amendments could be a useful and cost-effective practice to alleviate negative effects of Cr on crop production and quality and also protect soil quality.
Cr(VI) uptake in the aerial tissues increased with rising soil added Cr(VI) concentrations (Table 3) and this was also noted for Cr(VI) plant concentrations (also in Table 3-values in parentheses), and N addition seemed to have a positive effect in plant uptake of Cr(VI). The fact that root tissues of the well-fertilizedP. oleracea absorbed significantly higher Cr(VI) quantities from the soil compared to those of the non-fertilized may indicate an enhanced ability of the test plant to accumulate more Cr(VI) when nitrogen is applied (Marciol et al. 2007;Martinez-Trujillo and Carreόn 2015;Yao et al. 2020). The alleviating effects of Cr(VI) stress with added nutrients have also been reported in a work concerning Arabidopsis thaliana, although the effect of N was not found to be significant (Ortiz Castro et al. 2007). Cr(VI) is known to impede plant physiological functions, such as photosynthesis, as well as various enzyme and gene functions; these functions are boosted when N content is sufficient which indicates that negative effects from exposure to Cr are partly due to toxicity while it also interferes with the uptake of similar ions (Ca, K, Mg, P, B, and Cu). We thus assume that added N alleviates Cr(VI) toxicity symptoms through the restoration of nutrient uptake, as also indicated by Thalassinos et al. (2021). Uptake was high even at T-4, despite the fact that Cr(VI) stress significantly decreased growth rate (as noted in Thalassinos et al. (2021), a work reporting data from the same experiment), and this was achieved largely due to added N. These results are in agreement with the findings of other similar works (Marciol et al. 2007;Martinez-Trujillo and Carreόn 2015;Nagarajan and Sankar Ganesh 2014).
Compartmentalization of potentially toxic elements in plant tissues is of great importance, contributing to the plant capacity to address stress induced by toxic elements (Ertani et al. 2017;Shanker et al. 2005). In the present study, we observed that the translocation of Cr(VI) from roots to shoots (as noted in the TF) was notably lower at T-3 and T-4 compared to those at T-1 and T-2. In the highest Cr(VI) soil concentrations, low TF indicate that our test plant under high Cr(VI) stress manages to effectively limit Cr(VI) translocation from roots to the aerial tissues, while the effect of N on TF was found to be non-significant (Fig. 1).
This indicates that purslane defense mechanisms prohibited Cr(VI) translocation to aerial tissues, even though Cr(VI) was readily absorbed by root tissues. As for the TC for the aerial tissues, it was orders of magnitude lower than the corresponding values of roots. These findings indicate that plants effectively managed to limit the influx of Cr(VI) to the plant tissues and to retain the absorbed Cr(VI) to roots, although under severe Cr(VI) stress. Cr accumulation in roots and the minimal transport to aerial tissues has been reported as a defense mechanism for Cr toxicity in a number of other similar studies concerning other plant species (Antoniadis et al. 2017a;Marciol et al. 2007;Moral et al. 1995;Shanker et al. 2005;Singh et al. 2013). These findings render the use of P. oleracea as an attractive option for phytoremediation, although both TC and TF were much lower than unity, contrary to the requirement of some researchers demanding TC and TF values of higher than unity so that a plant species may be categorized as accumulators (Antoniadis et al. 2017a;Antoniadis et al. 2021;Buscaroli 2017;Ertani et al. 2017;Moral et al. 1995). For this purpose, the option of multiple harvests of the species should be also considered since it is a cultivation practice that could increase the total amounts of Cr removed from soil within the same growing period and better facilitate the remediation of polluted soils.
As for the Cr speciation in plant tissues, our results indicate that in aerial tissues increasing soil Cr(VI) concentrations probably activated plant enzymatic and non-enzymatic mechanisms that effectively reduced Cr(VI) to the less toxic Cr(III). The available concentration of Cr(III) was zero in the control soil (Table 2), while the total Cr concentration was 221.3 mg Cr kg −1 ( Table 1). The control plants had non-detectable Cr levels. The aforementioned low Cr(III) phytoavailability combined with the known minimal absorption of Cr(III) from  plants can lead us to the assumption that Cr(III) found in plant tissues was almost exclusively absorbed as Cr(VI) and subsequently reduced back to Cr(III). In root tissues, rising soil Cr(VI) concentrations resulted in higher Cr(VI) concentrations, indicating that the reduction capacity of roots was surpassed, resulting in an increased Cr(VI) to Cr(VI+III) ratio compared to those of the aerial parts. These findings are in accordance with a series of other similar works reporting that enzymatic and non-enzymatic mechanisms in plant cells readily reduce Cr(VI) (Antoniadis et al. 2018;Ertani et al. 2017;Levizou et al. 2019;Paiva et al. 2009). Concerning residual N, our results are in agreement with previous findings which suggest that Cr(VI) toxic effects result in reduced root growth that subsequently limit nutrient uptake. Furthermore, it is known that Cr(VI) interferes with nitrate transporters and through antagonistic effects may result in decreased N uptake from plants (Sundaramoorthy et al. 2010;Ortiz Castro et al. 2007). Indeed, we found that the N remaining in soil (residual) increased with Cr(VI) uptake, denoting decreased N uptake by plant with increased Cr(VI) uptake: the correlation between Cr(VI) aerial uptake and residual N was indeed positive, with r = 0.546 (significant at p < 0.01).

Conclusions
The potential of P. oleracea to absorb Cr(VI) increased with added N, likely as a result of the enhanced plant vigor due to sufficient fertilization. This was reflected in enhanced uptake (measured as quantity in μg of Cr(VI) in plant per pot, rather than tissue concentration); this meant that a complete site clean-up could be achieved within 256 harvests in the high Cr(VI) added rate at T-4 (vs. 405 harvests at non-N-added T-4). However, the test plant failed to be categorized as a hyperaccumulator, since TF was far from the threshold of 1.0, and so was the case with TC. However, the option of multiple harvests should be also considered when taking into account the total amount of Cr removed from contaminated soils. Moreover, it is not fully known if the rate of reduction of Cr(VI) towards Cr(III) will continue past our 50-day growing period due to stress symptoms. Thus, more research is necessary in order to assess the Cr(VI) long-term soil dynamics and to evaluate the required time needed for Cr(VI) remediation of a contaminated site. Taking into consideration (a) plant Cr(VI) uptake and (b) Cr(VI) reduction naturally occurring in soil, we conclude that the latter is the predominant mechanism of soil Cr(VI) elimination. We also conclude that P. oleracea may be used as a tolerant plant species in the process of remediating a Cr(VI)-contaminated soil, especially if assisted with a sufficient level of N fertilization, although the plant falls short of the threshold of being categorized as an accumulator species. We further recognize the fact that the results reported here derive from a small-scale pot experiment, and hence certain limitations do exist; thus, further studies are required to verify these findings based on experimentation with naturally Cr(VI)-contaminated soils and extrapolate the current findings to field conditions.