Root cell-type specific expressions of bacterial mercury transporter MerC and plant SNARE SYP121 fusion protein differentially affect cadmium accumulation patterns of Arabidopsis

ABSTRACT There is increasing demand for solutions against cadmium pollution to secure food safety, and phytoremediation is one of the potential tools. We previously found that a bacterial mercury transporter MerC possesses cadmium uptake activity and its overexpression as a fusion protein with a plasma-membrane resident SNARE protein, SYP121 enhances the cadmium uptake ability of Arabidopsis plants. In this study, we examined whether two different root cell-type specific expression systems of MerC-SYP121 fusion protein could efficiently enhance cadmium accumulation of Arabidopsis plants, compared to the p35S-driven ubiquitous expression system. Representative transgenic lines expressing MerC-SYP121 in root surface cells (pEpi lines) or root endodermal cells (pSCR lines), established in our previous studies, were subjected to different cadmium treatments along with the p35S line. A vertical agar plate assay showed that root surface-specific line pEpi, as well as the p35S line, showed about 15% higher cadmium accumulation in shoots after one-day 10 µM cadmium treatment, compared to the wild-type Col-0. On the other hand, the endodermis-specific line pSCR accumulated 30% less cadmium in its shoots. A similar cadmium accumulation pattern in shoots was observed under the environmentally relevant much lower cadmium treatment of 0.1 µM for 4 d, using the hydroponic culture system. To further examine the potential of the MerC-SYP121 expression system for cadmium phytoremediation, the transgenic plants were hydroponically exposed to 0.1 µM for 4 weeks. The cadmium accumulation after the 4 weeks of treatment was again 16% higher in the pEpi shoots compared to that of Col-0, whereas the p35S line only showed 6% higher Cd concentration. Shoots of the pSCR line accumulated slightly less cadmium compared to Col-0. Ionomic profiles in these plants were analyzed, however, pSCR-specific patterns were not evident. Nevertheless, in our previous studies, pEpi and pSCR lines both efficiently accumulated more mercury in shoots than in the wild-type. The presented results suggest that the effects of cell-type-specific MerC-SYP121 expression differ by the target metals and its expression in root surface cells rather than that in endodermis is suitable for enhancing root cadmium uptake and subsequent shoot accumulation.


Introduction
As infamously known as the causal pollutant of itai-itai disease which occurred in the 1950s, cadmium (Cd) is highly toxic and classified as a class I carcinogen.Due to various anthropogenic activities, a vast area of arable soils is moderately contaminated with Cd and the consequent Cd accumulation in plant-derived foods has increasingly been recognized as a potential 'slowpoisoning' risk to human health (Clemens and Ort 2019).A considerable amount of Cd has been found in cereal grains and other agricultural products worldwide (Meharg et al. 2013;Tsukahara et al. 2003;Wang et al. 2019).The detected values are controversial referring to the provisionally tolerable intake limit values suggested by the Joint FAO/WHO Expert Committee on Food Additives (JECFA), the European Food Safety Authority (EFSA), and the Agency for Toxic Substances and Disease Registry (ATSDR).Therefore, Cd accumulation in plant-derived foods should be mitigated to secure food safety, and remediation of soils using plants (phytoextraction) is one of the promising approaches.
merC gene was identified among mer operons that confer mercury (Hg) resistance to bacteria and MerC was characterized as an inward mercurial transporter (Inoue, Kusano, and Silver 1996;Kusano et al. 1990;Sahlman, Wong, and Powlowski 1997).Because efflux transporter-mediated metal(loid) extrusion from the cells is a common mechanism responsible for bacterial metal(loid) resistance, as represented by ArsB for arsenic resistance (Kuroda et al. 1997;Silver and Phung 1996), the influx activity of MerC is unique and can be a useful feature for the bioaccumulation/remediation of mercurials from the environment.In addition to MerC, three more Mer transporters MerE, MerF, and MerT were also characterized as mercurial uptake transporters, but we found the highest inward transport activity for MerC (Sone et al. 2013).Taking advantage of the mercury uptake activity of MerC, we generated a series of MerC expressing Arabidopsis, and the plant lines exhibited enhanced mercury (Hg) accumulation (Kiyono et al. 2013;Uraguchi et al. 2019aUraguchi et al. , 2019b) ) or enhanced Hg tolerance (Uraguchi et al. 2022) attributed to the ectopic MerC expression.In this engineering, a plasma-membrane resident plant SNARE SYP121 was fused with MerC to regulate membrane localization of the MerC in plant cells for boosting Hg accumulation.MerC-SYP121 was localized to the plasma-membrane, whereas the majority of MerC without SYP121 was stuck at ER-Golgi in Arabidopsis cells (Kiyono et al. 2012(Kiyono et al. , 2013)).The plasma-membrane localized MerC-SYP121 enhances Hg uptake in roots and consequently shoots Hg accumulation when driven by the ubiquitous p35S promoter (Kiyono et al. 2013), the root-surface specific promoter pEpi (Uraguchi et al. 2019a), and the root endodermis specific SCARECROW promoter pSCR (Uraguchi et al. 2019b).It should be noted that the different promoters' regulation equally enhances shoot Hg accumulation under various Hg conditions, suggesting that the epidermis and endodermis are equally important for root Hg uptake.More importantly, limited cell-type-specific expression of MerC-SYP121 in roots is enough for enhancing Arabidopsis Hg accumulation.The importance of epidermis and endodermis-specific expression of transporters/ channels in roots has been well demonstrated for nutritional element uptake (Alassimone, Naseer, and Geldner 2010;Sasaki et al. 2012;Takano et al. 2010;Ueno et al. 2015).Our results of the MerC-SYP121 transgenic lines suggest that these root cell types are also crucial for toxic metal uptake like Hg.
MerC has been characterized as a mercury transporter, however, previous in vitro studies found the Cd uptake activity as well (Ohshiro et al. 2020;Sasaki et al. 2005).This is likely attributed to the chemical similarities of Hg(II) and Cd(II) ions that both belong to group 12 of the periodic table.The Cd transport activity of MerC implies that MerC-expressing plants can be applied to the remediation of soils contaminated with Cd, and also the remediation of soils co-contaminated with Hg and Cd.Indeed, the p35S-driven ubiquitous expression of MerC-SYP121 increased Arabidopsis Cd accumulation in our previous study (Kiyono et al. 2012).In the present study, we examined whether different promoters' regulation of MerC-SYP121 would enhance plant Cd uptake and accumulation in shoots as in the case of Hg as reported previously (Uraguchi et al. 2019a(Uraguchi et al. , 2019b)).Since the previous p35S line's study (Kiyono et al. 2012) employed relatively high Cd treatments (10-100 µM), the present study examined the MerC-SYP121 expressing plant phenotypes under more environmentally relevant Cd conditions (0.1 to 10 µM).
Arabidopsis seeds were surface sterilized with 70% EtOH and sown on the control plates (without Cd).After 2 d stratification at 4°C, plants were grown vertically in a longday growth chamber (16 h light/8 h dark, 22°C) for 10 d.Uniformly grown seedlings were then transferred to the Hoagland plates containing 10 µM CdCl 2 and grown for an additional 1 d as a Cd treatment.At harvest, roots and shoots from each plate (normally 15 seedlings per line) were separately pooled as a single sample.At least three replicates were prepared for each experiment, and the experiments were independently repeated three times.

Hydroponic culture
The hydroponic culture was conducted according to the previous method with some modifications.For the hydroponic culture, the following essential microelements (4.63 µM H 3 BO 3 , 32 nM CuSO 4 , 915 nM MnCl 2 , 77 nM ZnSO 4 , 11 nM MoO 3 ) were supplemented to the one-tenth-strength modified Hoagland's No.2 solution described elsewhere but sucrose was not added to the solution.
200 µL standard PCR tubes were used as a holder for seedling establishment.A small hole (approximately 1 mm wide) was punched in a cap of a PCR tube for seed sowing and the bottom of the tube was cut off to prepare an approximately 3 mm-wide hole for root elongation.The tube was filled with the modified Hoagland medium supplemented with the microelements and 0.6% agar.Sucrose was not added to the media to avoid microbiological contamination.After solidification, an Arabidopsis seed was sown into the top hole of the agar-filled PCR tube.Prior to sowing, the surface sterilized seeds were stratified in 0.1% agar at 4°C for 1 d.After seed sowing, the PCR tubes were set in a plastic box for 200 µL pipette tips which were filled with the modified Hoagland solution.The tip boxes were covered with transparent plastic lids and set in a short-day growth chamber (8 h light/16 h dark, 22°C).Plants were grown for 2 weeks, and the medium was changed weekly.Uniformly grown seedlings together with the holed-PCR tubes were then transferred to 50 ml plastic tubes filled with the modified Hoagland solution.Prior to the seedling transfer, the lid of the 50 ml tube was drilled to prepare a 5-6 mm wide hole that fit PCR tubes.Plants transferred to the 50 ml tubes were then subjected to respective Cd treatments.

Short-term hydroponic Cd assay
The 2-week-old seedlings transferred to 50 ml tubes prepared as described elsewhere were grown for additional 3 weeks with a controlled hydroponic solution.The medium was changed weekly.The plants were then subjected to 0.1 µM CdCl 2 treatment for 4 d.Three to eight biological replicates were prepared for each experiment, and the experiment was independently repeated.Plants were harvested and subjected to Cd determination as described elsewhere.

Long-term hydroponic Cd assay
The 2-week-old seedlings transferred to 50 ml tubes prepared as described elsewhere were grown for additional 4 weeks with controlled or Cd-containing hydroponic solutions.First, to optimize Cd treatment conditions for assessing plant Cd accumulation ability, different Cd concentrations (0, 0.05, 0.1, and 0.2 µM) and different medium change frequencies (once a week or three times a week) were tested using Col-0.Aliquots of hydroponic solutions were sampled every week before the medium change and subjected to Cd determination by an inductively coupled plasma-optical emission spectroscopy (ICP-OES, iCAP7400Duo, Thermo-Fisher Scientific) to quantify the levels of remaining Cd in the hydroponic solutions.After 4-week treatments, plants were photographed and harvested for Cd determination as described elsewhere.
For the evaluation of the Cd accumulation ability of the MerC-SYP121 expressing transgenic lines, 2-week-old plants were subjected to a control condition and 0.1 µM Cd treatment.The hydroponic solutions were exchanged three times a week.After 4-week treatments, plants were harvested, and ionomic profiles including Cd were analyzed as described elsewhere.At least four biological replicates were prepared for each experiment, and the experiment was repeated independently.

Elemental analysis
The plant elemental analysis was conducted as described previously (Uraguchi et al. 2020).Separating shoot and root, shoot samples were washed with MilliQ water twice.Root samples were subjected to sequential washing procedures: roots were desorbed for 10 min each in ice-cold MilliQ water, 20 mM CaCl 2 (twice), 10 mM EDTA (pH 5.7), and MilliQ water again.Washed roots and shoots were dried at 50°C before acid digestion at least for a few days.Dried plant samples were wet-digested with 3 ml of HNO 3 .Elemental concentrations including Cd were quantified by ICP-OES.

Quantitative RT-PCR
To examine the transgene expression, 10-d-old plants grown on the control agar plates were transferred to the plates containing 10 µM CdCl 2 .After 1 d, the roots and shoots were harvested and subjected to RNA extraction.NucleoSpin RNA Plant (MACHEREY-NAGEL) was used for total RNA extraction from roots or shoots.DNase treatment was applied during the RNA extraction.PrimeScript RT Master Mix (Takara Bio) was used for cDNA synthesis, and quantitative RT-PCR was performed with PowerUp SYBR Green Master Mix (Thermo Fisher Scientific).In addition to the well-used housekeeping gene AtEF1a (At5g60390) (Kühnlenz et al. 2014;Remans et al. 2008;Uraguchi et al. 2022), a SAND family gene (At2g28390), and an F-box protein gene (At5g15710) served as internal control genes.These two genes were suggested as more stable internal control genes, less affected by Cd and other heavy metal stress (Remans et al. 2008).merC primers were used to detect the transgene transcript.A relative quantification method using standard curves was applied.The primer sequences used for expression analyses are listed in Supplementary Table S1.

Confocal laser microscopy
The expression of mTRQ2-MerC-SYP121 in the pEpi line was examined by a laser scanning confocal microscope (FV-3000, Olympus).Prior to the observation, roots were stained with 4 µM FM4-64 for 5 min.The root cell morphology was also observed by a laser scanning confocal microscope after 2 min staining with 10 μg/mL propidium iodide (PI).The excitation and detection wavelengths were 405 nm and 420-480 nm for mTRQ2, 488 nm and 600-700 nm for FM4-64; 561 nm, and 570-670 nm for PI.

Statistical analyses
The JASP software ver.0.16 (JASP team, 2022) was used for statistical analyses.The Cd accumulation data were analyzed by one-way ANOVA, followed by Dunn's test (p < .05) to examine the significance of differences among Col-0 wild-type and the respective transgenic lines.Welch's test (p < .05)was applied to examine the significance of differences between the different medium change frequencies for long-term hydroponic experiments.The effects of Cd treatment on ionomic profiles were also examined by Welch's test (p < .05).

Short-term Cd treatment using agar plates
We first examined Cd accumulation in roots and shoots of three different MerC-SYP121 expressing Arabidopsis using our agar plate-based metal(loid)-assay (Uraguchi et al. 2020).The agar reagent that we used in this study has a lower elemental background and is suggested as a suitable reagent for platebased metal(loid)-assay (Uraguchi et al. 2020).
10-d-old established seedlings were exposed to 10 µM Cd for 1 d before Cd determination.This treatment was to assess the short-term Cd uptake and accumulation ability of the tested plant lines without the severe toxicity that would disturb the entire ion uptake and accumulation and mask the effects of MerC-SYP121 expression.Indeed, the Cd treatment did not induce any visible symptoms (Supplementary Fig. S1).The root cell morphology was also not affected by the Cd treatment, and which did not induce PI-positive cell death (Supplementary Fig. S2).The Cd treatment did not alter the transcript levels of merC in roots, whereas the merC level of the pEpi line was slightly reduced in its shoots (Supplementary Fig. S3).For the pEpi line, plasma-membrane localization of mTRQ2-MerC-SYP121 was observed as reported (Uraguchi et al. 2019a) and was not affected by the Cd treatment (Supplementary Fig. S4).
Under this condition, the MerC-SYP121 over-expressing line p35S and the root-surface specific line pEpi accumulated about 15% higher Cd in shoots, compared to the wild-type Col-0 (Figure 2).In contrast, the endodermal specific MerC-SYP121 expressing line pSCR accumulated about 30% less Cd in shoots.In roots, there were no significant differences in Cd accumulation among the tested lines.

Short-term Cd treatment using hydroponic culture
Cd accumulation in the MerC-SYP121 expressing Arabidopsis plants was then examined under more environmentally relevant Cd concentrations using the hydroponic culture system.We first tested short-term Cd treatment (0.1 µM Cd for 4 d).Under the condition, overall Cd concentrations in plants ranged from 20 to 60 µg g −1 dry wt.(Figure 3), one order of magnitude lower than the plate assay values (Figure 2).Comparing the Cd accumulation of the MerC-SYP121 expression lines and Col-0, p35S and pEpi lines again accumulated about 10% to 15% more Cd in the shoots, whereas the pSCR line accumulated 25% less Cd in the shoots compared to Col-0 (Figure 3).While the shoot Cd accumulation patterns were similar to those of the agar plate assay, the results for root Cd accumulation were different.The root Cd concentrations were approximately 40% and 55% higher in pEpi and pSCR lines compared to Col-0, respectively, however, there was no significant difference between p35S and Col-0 (Figure 3).

Long-term Cd treatment using hydroponic culture
We then planned a long-term Cd treatment using the hydroponic system and first searched for a suitable Cd treatment  condition that would not induce severe Cd toxicity using Col-0 as a model.For that purpose, three environmentally relevant Cd concentrations (0.05, 0.1, and 0.2 µM) were designed based on the Cd levels found in slightly/moderately contaminated soils (Arao et al. 2009;Maejima et al. 2007;Uraguchi et al. 2009).Another testing factor was the frequency of the medium exchange.Because the designed Cd concentrations in the hydroponic solution were very low, there was a concern about the depletion of Cd in the solution during cultivation due to plant root Cd absorption.To avoid complete Cd depletion, changing the medium three times a week was examined (x 3) in addition to our standard protocol of weekly medium exchange (x 1) (Kühnlenz et al. 2014).As a result, 0.2 µM Cd treatment with the three-times-a-week medium exchange protocol (x 3) slightly inhibited the plant growth, although it was not much severe as to induce chlorosis, a typical Cd toxicity symptom (Figure 4a).No visible effects in other treatments, regardless of medium change frequency and Cd concentrations.
Based on the results of the preparatory experiment, 0.1 µM Cd was selected for further experiments, and Cd depletion in the medium was monitored weekly during the 4-week Cd treatment (Figure 4b).In the first and second weeks of the Cd treatment, most of the spiked Cd in the medium remained in the solution regardless of the medium change protocols, probably due to the small plant size.But in week 3, only 25% of the Cd input remained in the solution of the weekly medium change treatment (1 ×), whereas nearly 70% of the Cd input was kept in the threetimes-a-week medium exchange treatment (3 ×).A similar trend was also observed in week 4 of the Cd treatment.This suggested severe Cd depletion in the middle of weeks 3 and 4 of the weekly medium exchange treatment (1 ×).Accordingly, shoot and root Cd concentrations drastically differed between the medium change protocols.Approximately 5-times and 1.7-times higher Cd were accumulated in roots and shoots, respectively, when the medium was changed three times a week (Figure 4c).Thus, 0.1 µM Cd treatment with the three times a week medium change protocol was applied for further experiments.
Under this condition, Cd accumulation in the MerC-SYP121 expressing lines was compared to that of Col-0 (Figure 5).After 4-week treatments of 0.1 µM Cd, the shoot of pEpi line accumulated 16% higher Cd compared to that of Col-0 (P < .05),whereas the p35S line only showed 6% higher Cd concentration (not significant).Shoots of the pSCR line accumulated slightly less cadmium compared to Col-0 (not significant).There was no significant difference in root Cd concentrations among the tested lines.

Ionomic profiles of MerC-SYP121 expressing plants
Nutritional element profiles in the plants harvested from the long-term hydroponic experiment were analyzed by ICP-OES (Figure 6).First, we examined ionomic profiles of the tested plants grown under the control condition (Figure 6a).The clustered heatmap showed that there was no significant difference between Col-0 and the MerC-SYP121 expressing plants.It should be noted that although it was not statistically significant, pEpi and pSCR plants tended to accumulate higher Cu in their roots.
We next examined ionomic changes elicited by the 4 weeks Cd treatment of 0.1 µM (Figure 6b).The clustered heatmap showed similar changes as well as line-specific changes among the lines under the subtoxic Cd treatment.In Col-0 roots, the Mn level significantly decreased by the Cd treatment, and Fe and Zn levels also tended to decrease.pEpi  and pSCR lines both showed a significant decrease in their root Mn, Fe, and Zn levels.These cell-type specific lines also showed a decreased tendency of Cu.Among the tested lines, the roots of the p35S line showed unique ionomic patterns under the Cd treatment: contrary to the other lines, the decrease of Mn, Fe, Cu, and Zn was not that much evident in the p35S line.Macronutrients such as Ca, P, S, and K tended to increase under the Cd treatment.
In shoots, the most notable change was a significant increase of Zn in Col-0, p35S, and pEpi lines.Zn increase in shoots was also observed for the pSCR line, but it was much milder.Col-0, as well as pSCR, showed a significant decrease of Fe in shoots, whereas p35S and pEpi lines did not.pSCR also showed a decreasing tendency of Mn and an increasing tendency of S and Mg, which were not evident in the other lines.

Discussion
Although MerC was initially characterized as a mercury transporter (Inoue, Kusano, and Silver 1996;Kusano et al. 1990;Sahlman, Wong, and Powlowski 1997), the Cd uptake activity of MerC was later suggested in bacteria (Sasaki et al. 2005;Ohshiro et al. 2020).The Cd uptake activity was also functional in plants: the p35S-driven MerC-SYP121 expressing Arabidopsis plant accumulated more Cd than the wild-type (Kiyono et al. 2012), attributed to the plasma-membrane localized MerC expressed in roots.The present study further examined the effects of root cell-type specific MerC-SYP121 expression on Cd accumulation, in comparison to the p35S-ubiquitous expression system (Figure 1).The Cd treatment (1 d exposure to 10 µM Cd) resembling that of the previous study (Kiyono et al. 2012) was first examined using the agar plate assay.Roots and shoots were separately analyzed here, whereas whole seedlings were subjected to the analysis in the previous study (Kiyono et al. 2012).The current assay showed the increased Cd accumulation in the shoots of the p35S line and the root surface-specific MerC-SYP121 expressing pEpi line, but the root endodermis specific MerC-SYP121 expressing pSCR line accumulated less Cd in its shoots, compared to Col-0 (Figure 2).It should be noted that the Cd treatment little affected the plant growth, root morphology, and transgene expression (Supplementary Figs.S1-3).Similarly, the increased Cd in p35S and pEpi shoots and less Cd in pSCR shoots were also observed under the more environmentally relevant Cd concentration using the hydroponic culture (0.1 µM for 1 d, Figure 3).The results of these short-term Cd assays suggest that ubiquitous over-expression of MerC-SYP121 enhances shoot Cd accumulation, and more importantly, the root surface specific expression of MerC-SYP121 sufficiently boosts Cd accumulation of Arabidopsis plants.This response of the pEpi line to Cd treatment is similar to the case of Hg(II) (Uraguchi et al. 2019a): we showed that root surface-specific expression of MerC-SYP121 (pEpi lines) enhanced shoot Hg accumulation as the p35S ubiquitous expression system did.Contrarily, the impaired shoot Cd accumulation of the pSCR line (Figures 2 and 3) observed in this study contradicted the previous result of higher Hg accumulation in the pSCR lines (Uraguchi et al. 2019b).The root-surface-specific expression rather than the endodermis-specific expression of MerC-SYP121 is suggested to be effective in facilitating plant Cd accumulation, unlike the case of Hg(II) treatment.We could not find a mechanistic clue for the reduced shoot Cd accumulation in the pSCR line (Figures 2 and 3).The clustering analysis showed that the ionomic profiles of the pSCR line were similar to that of the pEpi line in roots, but slightly differed from the other tested lines in shoots (Figure 6).In addition, plant growth (Supplementary Fig. S1) and root morphology (Supplementary Fig. S2) of the pSCR line did not show significant differences.These results suggest at least that the entire element transport in roots and to shoots was not impaired in the pSCR line.Interactions between the introduced MerC and endogenous Cd transport systems would be a possible cause for the reduced shoot Cd accumulation in the pSCR line.
We further examined the responses of the root cell-type specific MerC-SYP121 expression systems under prolonged Cd treatment with environmentally relevant Cd concentrations (Figures 4 and 5).We first sought subtoxic Cd concentrations in the hydroponic solutions (0.05-0.2 µM Cd, Figure 4a), which are the Cd levels mimicking the low/moderate Cd polluted soils (Arao et al. 2009;Maejima et al. 2007;Uraguchi et al. 2009).To avoid Cd depletion in the medium during plant growth, we changed the medium three times a week (Figure 4b), and this growth condition drastically increased plant Cd concentrations (Figure 4c).With this protocol, 0.2 µM Cd treatment slightly reduced Col-0 shoot growth (Figure 4a), whereas 0.1 µM Cd treatment did not.Similarly, a previous study reported that long-term 0.1 µM Cd treatment did not induce severe toxicity to Arabidopsis, but the higher Cd treatment elicited apparent phytotoxicity (Belleghem et al. 2007).Thus, to avoid possible toxicity, Cd accumulation of the MerC-SYP121 expressing lines was examined under 0.1 µM Cd treatment (Figure 5).After 4-week Cd treatments, the pEpi line showed a statistically significant increase of the shoot Cd accumulation, whereas the p35S and pSCR lines did not.This further suggests the potential of the root-surface specific expression of MerC-SYP121 for facilitating Cd accumulation under the environmentally relevant Cd treatment.It is also noteworthy that the pEpi expression system is likely more effective than the p35S ubiquitous expression.
The presented results also provide some clues for understanding root elemental transport mechanisms.There are several examples of root cell-type specific expressing endogenous transporters/channels for nutrients.Epidermal cells of Arabidopsis roots are the outermost cell type contacting soils, and plant nutrient transporters expressed in the epidermal cells are crucial for proper and efficient nutrient uptake (Huang et al. 1999;Kiba et al. 2012;Lezhneva et al. 2014;Marquès-Bueno et al. 2016;Mudge et al. 2002;Takano et al. 2010).And in the endodermis, a transporter-mediated symplasmic pathway across the endodermal cells is critical to delivering nutritional elements to the xylem, and eventually to shoots (Alassimone, Naseer, and Geldner 2010;Sasaki et al. 2012;Takano et al. 2010;Ueno et al. 2015), because the Casparian strip diffusion barrier blocks apoplasmic element transport toward inner vascular tissues (Barberon and Geldner 2014;Li et al. 2017;Nakayama et al. 2017).Our data of the pEpi lines suggest that epidermis-expressed ectopic transporters are also crucial for the efficient radial transport of Hg (Uraguchi et al. 2019a) and Cd (this study).Roles of the endodermal expressed MerC-SYP121 May differ by the target metals because pSCR lines showed enhanced Hg accumulation (Uraguchi et al. 2019b) but did not for Cd (this study).A reason for the non-positive effects on Cd accumulation by the endodermal specific expression of MerC-SYP121 is not clear from this study including the ionomic analysis (Figure 6), root morphology (Supplementary Fig. S2), and the transgene expression (Supplementary Fig. S3).However, our results overall suggest a possibility that epidermis rather than endodermis is a physiologically significant cell-type for Cd uptake, and the two cell types are equally important for Hg uptake at least in Arabidopsis roots.Therefore, taking advantage of the enhanced accumulation ability for both Hg and Cd, pEpi lines would be a potential tool to remediate co-contaminated soils with Hg and Cd.Pyramiding pEpi and pSCR expression systems would be a more powerful option at least for boosting plant Hg accumulation than a single expression system.It is noteworthy that the total expression level of merC-SYP121 in the pEpi and pSCR lines are far below that of the p35S line (Supplementary Fig. S3), and the transgene effects are comparable (Uraguchi et al. 2019a(Uraguchi et al. , 2019b)).For Cd, although the pSCR expression system reduced shoot Cd accumulation under short-term treatments (Figures 2 and 3), it did not negatively affect the shoot Cd accumulation under the longer and environmentally relevant Cd treatment (Figure 5).Further studies are needed to examine the potential of the plants with pEpi and pSCR dual expression systems under Hg and Cd co-contaminated conditions.

Figure 2 .
Figure 2. Cd accumulation of the MerC-SYP121 expressing Arabidopsis under short-term Cd exposure using the agar plate assay.Col-0 and MerC-SYP121 expressing lines were grown on the control agar medium for 10 d and then transferred to the medium containing 10 µM CdCl 2 for 1 d.Cd concentrations in root and shoot were separately quantified by ICP-OES.Data represent means with SD from three independent experiments (n > 6).Asterisks indicate a significant difference from Col-0 (*P < .05,Dunn's test).

Figure 3 .
Figure 3. Cd accumulation of the MerC-SYP121 expressing Arabidopsis under short-term hydroponic Cd treatment.The 5-week-old seedlings grown with the control hydroponic solution were subjected to 0.1 µM CdCl 2 treatment for 4 d.Cd concentrations in root and shoot were separately quantified by ICP-OES.Data represent means with SD from two independent experiments (n = 9-13).Asterisks indicate a significant difference from Col-0 (*P < .05,Dunn's test).

Figure 4 .
Figure 4. Effects of different medium exchange frequencies on the growth and Cd accumulation of Col-0 under environmentally relevant Cd conditions.(a) Phenotypes of Col-0 plants hydroponically exposed to Cd for 4 weeks with different medium change frequencies.(b) Residual Cd in the hydroponic solution (% of the spiked Cd) during 4 weeks of Cd treatment with weekly (x 1) and three times a week (x 3) medium change protocols.Data represent means with SD (n = 6).Asterisks indicate a significant difference between the medium change frequencies (*P < 0.05, Welch's test).(c) Cd concentrations in root and shoot after 4 weeks of Cd treatment (0.1 µM) with different medium change frequencies were quantified by ICP-OES.Data represent means with SD (n = 6).Asterisks indicate a significant difference between the medium change frequencies (*P < 0.05, Welch's test).

Figure 5 .
Figure 5. Cd accumulation of the MerC-SYP121 expressing Arabidopsis under long-term hydroponic Cd treatment.The 2-week-old seedlings grown with the control hydroponic solution were subjected to 0.1 µM CdCl 2 treatment for 4 weeks.The Cd concentrations in root and shoot were separately quantified by ICP-OES.Data represent means with SD from two independent experiments (n = 7-10).Asterisks indicate a significant difference from Col-0 (**P < .01,Dunn's test).

Figure 6 .
Figure 6.Ionomic profiles and Cd-induced changes.(a) Hierarchically clustered heatmaps showing ionomic profiles of Col-0 and the MerC-SYP121 expressing plants grown under the control condition.Plants were grown for 6 weeks with the control hydroponic solution and ionomic profiles in shoots and roots were analyzed by ICP-OES.Means obtained from three independent experiments (n = 7-14) were used for the heatmap analysis.The color scale reflects the log 2 transformed nutritional element levels normalized to the values of Col-0.Note that there is no significant difference between Col-0 and each transgenic line.(b) Hierarchically clustered heatmaps showing ionomic profile changes of Col-0 and the MerC-SYP121 expressing plants grown under the 0.1 µM Cd condition for 4 weeks.Nutritional element concentrations in the plant samples used for Cd determination (Figure 5) were analyzed by ICP-OES.The color scale reflects the log 2-fold changes of the nutritional element levels in the Cd treatment compared to each control.Asterisks indicate significant changes induced by the Cd treatment (*P < .05,**P < .01,***P < .001,Welch's test).