Author for correspondence: Leon V. Kochian Tel: +1 607 255 2454 Email: email@example.com
•In this paper, we conducted a detailed analysis of the ZIP family transporter, NcZNT1, in the zinc (Zn)/cadmium (Cd) hyperaccumulating plant species, Noccaea caerulescens, formerly known as Thlaspi caerulescens. NcZNT1 was previously suggested to be the primary root Zn/Cd uptake transporter. Both a characterization of NcZNT1 transport function in planta and in heterologous systems, and an analysis of NcZNT1 gene expression and NcZNT1 protein localization were carried out.
•We show that NcZNT1 is not only expressed in the root epidermis, but also is highly expressed in the root and shoot vasculature, suggesting a role in long-distance metal transport. Also, NcZNT1 was found to be a plasma membrane transporter that mediates Zn but not Cd, iron (Fe), manganese (Mn) or copper (Cu) uptake into plant cells.
•Two novel regions of the NcZNT1 promoter were identified which may be involved in both the hyperexpression of NcZNT1 and its ability to be regulated by plant Zn status.
•In conclusion, we demonstrate here that NcZNT1 plays a role in Zn and not Cd uptake from the soil, and based on its strong expression in the root and shoot vasculature, could be involved in long-distance transport of Zn from the root to the shoot via the xylem.
Noccaea caerulescens is a Zinc (Zn)/Cadmium (Cd) hyperaccumulator, formerly known as Thlaspi caerulescens, which tolerates exposure to very high levels of Zn and Cd both in the soil and the plant. Both metals accumulate to high levels in the shoot and seed when plants are grown on elevated levels of Zn or Cd in soil or nutrient solution (Reeves & Brooks, 1983; Chaney, 1993; Brown et al., 1995a,b). This ability to hyperaccumulate heavy metals has intrigued plant biologists for many years about the possible value N. caerulescens may have in phytoremediation of Zn- and Cd-contaminated soils (Baker et al., 1994; Brown et al., 1995b). However, its slow growth and diminished shoot biomass limits its usefulness for phytoremediation (Ebbs et al., 1997). Exploiting the genetic potential of this species by transferring metal hyperaccumulating traits to higher biomass plants may be an effective approach for generating novel metal accumulator plants useful in phytoremediation (Brown et al., 1995a). To accomplish this goal, a better understanding of the basic molecular and physiological mechanisms responsible for Zn/Cd hyperaccumulation is needed. Because Zn hyperaccumulation in N. caerulescens is related to the elevated or altered expression of a number of Zn-responsive genes, this plant may serve as a useful tool for studying plant mechanisms of Zn sensing and homeostasis (Lasat et al., 2000; Pence et al., 2000).
Comparison of Zn transport in N. caerulescens and the related nonaccumulator species Thlaspi arvense, has demonstrated that Zn transport is altered at several sites along the transport path from the soil to the shoot. For example, root Zn influx is considerably greater in the hyperaccumulator species under a wide range of external Zn concentrations (Lasat et al., 1996). Furthermore, a significant fraction of the Zn transported into the root symplasm remains in the cytoplasm in N. caerulescens, whereas in T. arvense a larger fraction of the absorbed Zn is sequestered in the root vacuole and made unavailable for translocation to the shoot (Lasat et al., 1998). Associated with these two transport alterations, there is a much larger xylem loading of Zn (five-fold greater) in N. caerulescens than in T. arvense (Lasat et al., 1998). Finally, radiotracer (65Zn) uptake experiments conducted with leaf sections and leaf protoplasts from T. arvense and N. caerulescens demonstrated a greater capacity in N. caerulescens to absorb Zn into leaf cells than in the nonaccumulator (Lasat et al., 1998). Collectively, these results show that alterations of Zn transport in both roots and shoots contribute to Zn hyperaccumulation in N. caerulescens.
Noccaea caerulescens also possesses a great ability to accumulate Cd in the shoots, with certain ecotypes able to accumulate up to 10 000 mg kg−1 DW without any indication of toxicity, whereas the typical shoot concentration of Cd in nonaccumulator plants is between 0.1 and 10 mg kg−1 DW (Kabata-Pendias & Pendias, 2000). Significant ecotypic variation in Cd hyperaccumulation exists: ecotypes such as Ganges, from the south of France, are much more effective at accumulating Cd in the shoots, accumulating three-fold more than ecotypes such as Prayon used in this study. This Cd hyperaccumulation phenotype involves increased root uptake and translocation to the aerial portions of the plant coupled with highly efficient mechanisms that provide Cd tolerance at the leaf level, partially due to sequestration in the leaf cell vacuole (Ma et al., 2005).
The first metal transporter identified in a metal hyperaccumulating plant species was the ZIP family member, NcZNT1, from N. caerulescens (Pence et al., 2000). The ZIP family of micronutrient transporters is named for the two founding members: the high-affinity Zn transporter, ZRT1, from yeast, and the high-affinity iron transporter, IRT1, from Arabidopsis leading to the naming of the ZRT-IRT like Protein, or ZIP family of transporters (Eide et al., 1996; Zhao & Eide, 1996). When expressed in yeast, NcZNT1 was shown to mediate high-affinity 65Zn influx as well as low affinity 109Cd uptake (Pence et al., 2000). The authors suggested that NcZNT1 may be involved in uptake of Zn across the root-cell plasma membrane based on similar Zn/Cd uptake kinetics in Noccaea roots and yeast expressing NcZNT1, and the correlation between higher root NcZNT1 expression and root Zn uptake in a comparative study between N. caerulescens and the related nonaccumulator T. arvense (Lasat et al., 2000; Pence et al., 2000).
Initial studies on NcZNT1 in N. caerulescens suggest that it is a root Zn uptake transporter, but a comprehensive characterization of this transporter has not been conducted to date. In order to determine whether NcZNT1 does play a role in Zn and Cd uptake from the soil, a detailed analysis of tissue and cell specific gene expression, protein localization, and transport function in yeast and transgenic Arabidopsis was conducted for NcZNT1 to better understand the role this transporter plays in Zn uptake and hyperaccumulation. In addition, the NcZNT1 promoter was dissected to identify promoter regions that are involved in regulation of ZNT1 expression, in order to better understand the role NcZNT1 plays in micronutrient homeostasis and hyperaccumulation.
Materials and Methods
Yeast (Saccharomyces cerevisiae) strains and nutritional requirements for the five different yeast strains used in these studies include: DY1457 (wild-type), zrt1/zrt2Δ (MATα ade6 can1 his3 leu2 trp1 ura3 zrt1::LEU2 zrt2::HIS3), fet3/fet4Δ (MATa can1 his3 leu2 trp1 ura3 fet3::HIS3 fet4::LEU2), smf1Δ (SLY8; MATa ura3 lys2 ade2 trp1 his3 leu2 smf1::HIS3), and ctr1/ctr3Δ (MATa ctr1::ura3::Kanr ctr3::TRP1 his3 lys2-802 CUP1r). Each strain was maintained on yeast peptone dextrose (YPD) until introduction of either the gene of interest or empty vector. Cultures of each mutant containing one of the genes of interest were grown in liquid synthetic complete-uracil (SC-URA) to an optical density (OD) of 1 and then serially diluted 10, 100, 1000 and 10 000-fold. Each dilution was plated out onto the specific restrictive media for that mutant. For the zrt1/zrt2Δ mutant, the restrictive media contained SC-URA plus 1 mM EDTA and 500 μM ZnCl2. For smf1Δ the media was SC-URA containing 15 mM EGTA. The ctr1/ctr3Δ mutant was assayed for growth on YPE media (10 g yeast extract, 20 g peptone l–1, 10% ethanol). For the fet3/fet4Δ mutant, cells were grown on SC-URA (pH 4.0) containing 80 μM bathophenanthroline disulfonate (BPS). For Zn, Cu and Cd accumulation studies, the cells were grown to OD = 1 in SC-URA medium, before 50 μM of either ZnSO4, Cu SO4 or CdCl2 was added to the culture medium and the cells were allowed to accumulate the metal for 1 h. Cells were then pelleted for 5 min at 3000 g and washed twice with 5 mM CaCl2. Metal accumulation was determined by ICP atomic absorption emission spectroscopy (ICP-AES).
Uptake of Cd as determined by 109Cd
Cadmium uptake was determined as previously reported by Pence et al. (2000), except rather than using a simple uptake solution, cells were resuspended to OD = 1.5 in SC-URA, and an equal volume of 50 μM Cd in SC-URA was added. Counts were taken after 30 s to determine cell wall binding, and 3, 5 and 10 min for actual uptake.
For qRT-PCR, plants of Noccaea caerulescens (J & C Presl.) ecotype Prayon were grown on a modified Johnson’s solution containing 1.2 mM KNO3, 0.8 mM Ca(NO3)2, 0.1 mM NH4H2PO4, 0.2 mM MgSO4, 50 μM KCl, 12.5 μM H3BO3, 1 μM MnSO4, 0.1 μM NiSO4, 1 μM ZnSO4, 0.5 μM CuSO4 and 2 mM MES (pH 5.5). For all metal treatments, plants were grown for 2 wk on this medium and then switched to one of similar composition – Zn, copper (Cu), manganese (Mn) or iron (Fe) was omitted or 5 μM CdSO4 or varying amounts of Zn were added – and plants were grown for an additional 7 d. Plants were harvested, snap frozen in liquid nitrogen, and ground to a fine powder using a mortar and pestle. Total RNA was isolated using the Plant RNeasy RNA mini kit (Qiagen) and cDNA was synthesized using the QuantiTect Reverse Transcription Kit (Qiagen). Transcript levels were measured using GoTaq® qPCR Master Mix (Promega) with the primer pair ATCCTCTGTGATGCTGGCGAATC and AAGGCTTTAGCAGCTACAAAGAGATTTCC. 18S was used as an internal reference, amplified using the primer pair CGCTATTGGAGCTGGAATTACC and AATCCCTTAACGAGGATCCATTG to normalize across treatments. Quantitative real-time reverse transcription polymerase chain reaction (RT-PCR) was performed using an ABI 7500 real-time PCR system and SYBR Green kit (Applied Biosystems, Foster City, CA, USA). PCR conditions were 95°C for 5 min followed by 40 cycles of 95°C for 30 s, 50°C for 30 s and 60°C for 1 min. A dissociation curve was performed after each of the two biological replicates to ensure that only one product was being amplified. NcZNT1 and 18S were amplified using normal PCR conditions to ensure amplification of the target gene and then cloned into pGEM-T Easy (Promega) for target verification. The data were compared using a one-way ANOVA with Tukey’s test for post hoc analysis.
NcZNT1 was cloned into either pSAT6 or pSAT1 eGFPs, in frame on either the C or N terminus of the proteins, and transiently expressed in freshly isolated protoplasts from 3- to 4-wk old Arabidopsis thaliana (L.) Heynh seedlings according to Sheen (2001). Confocal images were taken 15–18 h post-transfection of the Arabidopsis protoplasts. Images were taken on a Leica TCS-SP5 confocal microscope (Leica Microsystems, Exton, PA, USA) with a ×63 magnification, 1.2 numerical aperture, water-immersion objective. The eGFP was excited with the blue argon ion laser (488 nm), and the emitted fluorescence was collected at wavelengths from 505 to 545 nm.
Cloning of the NcZNT1 promoter
Genome walking was performed as described by Liu et al., (1995) by digesting 20 ug of genomic DNA with EcoRI, HindIII and XhoI, and ligating known sequences onto the N. caerulescens DNA. The full promoter (1.1 kb upstream of the start codon) was amplified using the primer pair TGTGCTTGTAATCATGGT and CATCACAGAAGATTTTCGTGGGAG. The promoter DNA sequence is presented in Supporting Information Fig. S7.
For the NcZNT1p::YFP promoter, an 850 bp section upstream of the transcriptional start was isolated from the N. caerulescens genome using genome walking and then cloned into pBAR upstream of an YFP reporter. The pBAR construct was transformed into A. thaliana ecotype Colombia plants via Agrobacterium transformation. Transformants growing in soil were selected by spraying 1–2-wk old plants with 150 μg ml−1 of glufosinate every day for a week. To check for NcZNT1 promoter activation T2 plants were grown for 14–21 d under either replete or Zn limiting conditions (0 μM Zn), and plants were assayed for YFP expression using a Leica TCS-SP5 confocal microscope (Leica Microsystems) with a ×20 objective, 1.2 numerical aperture, water-immersion objective. To better define structures in the roots, the cell wall was stained using 0.1% propidium iodine.
Seedlings of N. caerulescens (ecotype Prayon) were grown in the same modified Johnson’s solution listed above containing 5 μM Zn for 50 d and then transferred to solutions with 0, 5, or 100 thinsp;μM Zn, respectively. After 4 d of growth in these solutions, the roots were excised and used for immunohistological staining. Antibodies against NcZNT1 were obtained by immunizing rabbits with the synthetic peptide C-HGQSHGHVHVHGSHDVENG (positions 201–219 of NcZNT1). The procedures for immunostaining were followed as described previously (Yamaji & Ma, 2007). Fluorescence was observed using a laser-scanning confocal microscope (LSM700; Carl Zeiss).
Overexpression of NcZNT1 in planta
NcZNT1 was cloned into the pBI101 vector downstream of the 35S promoter and then transformed into A. thaliana Col-0 via agrobacterium-mediated transformation using the C58 line. Transformants were selected on half-strength MS with 50 mg ml−1 Kanamycin until true breeding lines could be obtained. Seeds of NcZNT1 overexpression lines were surface sterilized in dilute bleach (0.5% NaOCl) and then with 50% ethanol, washed five times with ultra pure water before being imbibed in 0.1% (w/v) low melting point agarose at 4°C for 5 d. Homozygous lines were then grown on sufficient (1 μM ZnSO4) or high Zn (30 μM ZnSO4), high Cu (10 μM CuSO4), high Fe (250 μM Fe: EDDTA), high Mn (250 μM MnSO4) and Cd (25 μM CdSO4) media for 2 wk and pictures of the roots were taken using a Nikon D200 camera (Nikon, Tokyo, Japan) with a 60 mm lens. Total root length of each plant was determined using Root Reader 2D software (http://www.plantmineralnutrition.net). Relative root length was then calculated by dividing the mean total root length of each high-metal-treated plant by the mean total root length of the control (sufficient) plants. The data were compared using a one-way ANOVA with Tukey’s test for post hoc analysis. Each line was grown three separate times with at least 100 plants counted for each treatment and compared with wild-type Columbia for sensitivity to each metal. To study plant mineral concentration, true breeding overexpression lines were grown for 3 wk on a modified Johnson’s solution containing 5 and 30 μM ZnSO4. Roots were desorbed for 15 min in 5 mM CaCl2 and plants were then separated into roots and shoots for mineral concentration analysis. ICP-AES was used to determine root and shoot mineral concentrations.
5′ deletions of NcZNT1 promoter
Portions of the NcZNT1 promoter ranging from 1.1 kb upstream of the start codon to 213 bp upstream were amplified to add HindIII and XbaI restriction sites to the 5′ and 3′ ends of the promoter during amplification via PCR. PCR products were then cleaned using a PCR clean up kit (Qiagen) and digested overnight with HindIII and XbaI. The digest was again cleaned using the Qiagen PCR clean up kit and ligated in pGPTV. The various constructs were transformed into A. thaliana ecotype Colombia plants via Agrobacterium transformation. Transformants growing in soil were selected by spraying 1–2 wk old plants with 150 μg ml−1 of glufosinate every day for a week. To analyse NcZNT1 expression in transgenic plants, at least five different transgenic lines harbouring each promoter-GUS reporter insertion were grown on hydroponics under replete, Zn, Cu, Mn or Fe limiting conditions or on a range of Zn concentrations up to 50 μM. To stain for GUS activity, plants were incubated in 0.1 M NaPO4 buffer pH 7.0, 1 mM EDTA, 5 mM iron ferricyanide and iron ferrocyanide, 1% Triton-X 100 and 1 mg ml−1 X-Gluc.
Relative expression of NcZNT1
A comparative quantitative real-time PCR analysis was undertaken on NcZNT1 to discern the response of the expression of NcZNT1 to changes in Zn status as well as to various metal treatments (Figs S1, S2). Under nutrient replete conditions, NcZNT1 exhibits high expression in the roots and a lower (c. two-fold lower) expression in the shoots. The highest expression was seen in the roots of low Zn-grown plants; root NcZNT1 expression decreased as plant Zn status increased, although none of the differences in root NcZNT1 expression as a function of changes in plant Zn status were found to be statistically significant (Fig. S1a). Expression of NcZNT1 in the shoots was highest in plants grown on sufficient levels of Zn (1 μM) and then declined as plant Zn status increased. NcZNT1 expression was also quantified in plants grown on limiting levels of Cu, Mn or Fe as well as in response to 5 μM Cd. Under Cu limiting growth conditions, root NcZNT1 transcript abundance was not significantly different compared with expression in roots of control plants. In shoots, a decrease in expression was seen in NcZNT1 expression in response to low Cu status; however, this decrease in expression was again not statistically significant. A sharp downregulation of NcZNT1 was seen under Fe and Mn limiting conditions with transcript levels c. five-fold lower in the roots of low Fe and Mn-grown N. caerulescens plants. A similar but much smaller response was seen in shoots of low Fe or Mn-grown plants. In response to plant Cd exposure, plants grown on nutrient solution containing 5 μM Cd induced strong increases in NcZNT1 transcript levels in the roots relative to roots on replete-grown plants, with NcZNT1 expression increasing c. five-fold in roots of Cd-treated N. caerulescens seedlings. There was no difference in NcZNT1 transcript levels in the shoots when plants were treated with Cd compared with control plants.
NcZNT1 metal transport properties
In order to investigate the transport capabilities of NcZNT1, the open reading frame of NcZNT1 was cloned into the pFL61 yeast expression vector driven by the constitutive phosphoglycerate kinase promoter and this construct was then transformed into one of five different yeast backgrounds. The five different yeast lines consisted of four different metal uptake mutants defective in Zn (zrt1Δ/zrt2Δ), Cu (ctr1Δ/ctr3Δ), Fe (fet3Δ/fet4Δ), or Mn (smf1Δ) uptake, as well as wild-type yeast, which was used to study NcZNT1-mediated Cd transport. Metal transport specificity was assayed by the ability of the transporter to restore growth on low Cu, Fe, Zn or Mn in the appropriate mutant strain, while Cd transport in wild-type yeast expressing NcZNT1 was assayed via Cd accumulation assays using ICP atomic emission spectroscopy (ICP-AES).
Expression of NcZNT1 in the four different mutant uptake yeast strains showed complementation only in the zrt1/zrt2Δ double mutant, indicating that it could mediate Zn transport but not that of Cu, Fe or Mn (Fig. S3).
In addition to the yeast complementation studies, we directly quantified metal accumulation in yeast expressing NcZNT1. We did not conduct Zn accumulation studies as we have previously shown that NcZNT1 protein mediates high-affinity Zn uptake when expressed in yeast (Pence et al., 2000) which supports the Zn complementation data reported here. To gain a better understanding of NcZNT1-mediated Cd transport, we quantified yeast Cd accumulation in cells expressing NcZNT1. We found no evidence that NcZNT1 mediates Cd uptake, as there was no significant difference in Cd accumulation by yeast expressing NcZNT1 compared with wild-type yeast containing the empty vector. Because previous studies had shown NcZNT1 capable of transporting Cd in the zrt1/zrt2Δ yeast mutant, we tested the zrt1/zrt2Δyeast mutant both under 1 h net Cd accumulation and Cd influx using short-term uptake (minutes) assays with radiolabeled Cd. In the 1 h net Zn accumulation experiment, NcZNT1 expressing cells took up considerably less Cd than cells containing the empty vector alone (Fig. 1a; P < 0.01). Radiotracer flux analysis is a more sensitive transport assay than net accumulation measured via ICP spectroscopy, so we performed 109Cd uptake for 3, 5 and 10 min. Again, we found that yeast cells expressing NcZNT1 took up considerably less Cd than cells containing the empty vector alone (Fig. 1b; P < 0.01).
Tissue-specific NcZNT1 localization using promoter::reporter constructs
A 1.1 kb region upstream of the NcZNT1 start codon was cloned in front of a YFP reporter and transformed into transgenic Arabidopsis Col-0 plants. T3 generation homozygous lines were recovered and gene expression was assayed. When the transgenic Arabidopsis seedlings were grown for 3 wk on media lacking Zn, activation of the NcZNT1 promoter::YFP reporter was found to be broadly distributed throughout cells of the root tip, including the root cap, apical meristem, epidermis and stelar region (Fig. 2a). It should be noted that earlier in the time course of Zn deficiency (2 wk on –Zn media); NcZNT1 expression was not seen in the root apex (data not shown). NcZNT1 expression in the mature root of seedlings grown without Zn for 3 wk was found to be localized predominantly in the stele, especially in the pericycle and stelar parenchyma cells, and lower expression was also seen in the root epidermis and cortex (Fig. 2b–e). In the shoot, expression was seen in the vasculature and leaf guard cells as well as the shoot apical meristem (data not shown), an expression pattern very similar to what we previously reported for NcZNT1 in Noccaea shoots using a quantitative in situ hybridization (QISH) technique (Küpper et al., 2007).
To verify that the expression patterns seen in the roots of Arabidopsis were an accurate representation of what is seen in Noccaea roots, an NcZNT1 peptide antibody was generated to a portion of the protein between transmembrane domains 3 and 4. As seen in Fig. 3, immunolocalization of NcZNT1 in Noccaea roots showed that the protein expression pattern was the same as seen for NcZNT1 gene expression in transgenic Arabidopsis presented in Fig. 2. The highest NcZNT1 protein expression was seen in the stele, but NcZNT1 protein was also seen in root hairs and other epidermal cells, as well as the root cortex. The localization of NcZNT1 protein to the cell periphery is consistent with a plasma membrane localization. NcZNT1 protein abundance also showed a strong Zn dependence, with the highest levels of protein seen in roots of –Zn grown Noccaea plants and NcZNT1 protein abundance decreasing with increasing Zn in the media (Fig. 3a–c). Western blot analysis with the root microsomal fraction showed that only one band was detected (Fig. S4), indicating the high specificity of this antibody to the NcZNT1 protein.
Protein localization within the cell
The NcZNT1 protein was fused to eGFP protein on either the N and C terminus with expression driven by a double 35S promoter with a translational enhancer; these constructs were transiently expressed in Arabidopsis leaf protoplasts. As seen in Fig. 4, when protoplasts were imaged using a confocal microscope, eGFP::NcZNT1 was found to be localized to the plasma membrane. No signal could be detected when the C-terminal fusion was transiently expressed in protoplasts.
Overexpression of NcZNT1 in transgenic Arabidopsis
The possible function of NcZNT1 in plant micronutrient nutrition was also examined in transgenic Arabidopsis seedlings overexpressing NcZNT1 under the control of the 35S promoter. When seedlings overexpressing NcZNT1 were grown on high Zn levels in the nutrient solution (30 μM Zn), plants showed a significant increase in susceptibility to Zn toxicity compared with wild-type Col-0 seedlings, based on relative root growth. A total of 16 independent lines overexpressing NcZNT1 were tested for sensitivity to high Zn in the growth media. Root growth in the three most Zn sensitive transgenic lines exhibited a 40–55% inhibition of root growth by high Zn, whereas root growth in Col-0 was inhibited 30% by the same Zn treatment (Fig. 5). This response can also be seen in the photograph of individual Arabidopsis Col-0 and NcZNT1 overexpression lines in response to high Zn (Fig. S5). We also looked at Zn accumulation in the three best-performing transgenic lines using ICP-AES to quantify bulk shoot and root Zn accumulation, but did not have the experimental sensitivity to detect differences in accumulation between wild-type and transgenic lines grown on lower Zn levels (Zn < 5 μM). When the top three performing lines were grown on a higher level of Zn (30 μM) for 7 d, a significant difference in Zn accumulation could be seen in the roots, with roots of the overexpression lines accumulating c. 50% more Zn than the roots of Col-0 plants. A similar trend in Zn accumulation was seen in the shoots of the overexpression lines, which accumulated c. 10% more Zn in the shoots; however, this difference was not statistically different from Zn accumulation in Col-0 plants (Fig. 6).
The Zn fluorophor Zinquin was also used to study the relative accumulation of Zn into roots of the transgenic Arabidopsis overexpression lines and to investigate the increased sensitivity of NcZNT1 overexpression to high Zn. Zinquin is a membrane-permeable Zn fluorophor which binds intercellular Zn. Arabidopsis lines expressing NcZNT1 exhibited a strong Zn fluorescence in both the root tip and mature root regions when grown on 30 μM Zn for 10 d, indicating that high NcZNT1 expression led to increased Zn influx into root cells. Very weak Zn-based Zinquin fluorescence was observed in the roots of Col-0 wild-type plants grown under the same conditions (Fig. 7).
The NcZNT1 overexpression line 1-1, which was the most sensitive transgenic overexpression line to high Zn in the media, was also examined for root tolerance to Cu, Cd, Fe and Mn. When this transgenic Arabidopsis line was grown on high levels of Cu (10 μM), Cd (25 μM), Fe:EDDHA (250 μM) and Mn (250 μM), the line NcZNT1-1 showed no difference in tolerance to high Cu, Cd or Fe relative to Col-0, but increased tolerance to high Mn (Figs 8, S3).
5′ deletions of the NcZNT1 promoter
In order to identify regions of the NcZNT1 promoter that control its expression in plants, sequential 5′ deletions starting from −1045 upstream of the ATG start codon of NcZNT1 were cloned upstream of a GUS reporter and stably transformed into Arabidopsis. We feel Arabidopsis is a reasonable choice for this research as the NcZNT1 expression pattern in transgenic Arabidopsis using the YFP reporter in Fig. 2 was very similar to the pattern of NcZNT1 protein immunolocalization in Noccaea roots (Fig. 3), and the ability to stably transform Arabidopsis is much easier than N. caerulescens, which we can only transform at a very low efficiency. To analyse NcZNT1 expression in transgenic plants, at least five different transgenic lines harbouring each promoter-GUS reporter insertion were grown on hydroponic media under replete, Zn, Cu, Mn or Fe limiting conditions, or on nutrient solution supplemented with a a range of Zn concentrations up to 50 μM. NcZNT1 expression analysis in homozygous T3 progeny allowed us to identify two distinct promoter regions involved in NcZNT1 expression. These two regions are located between −667 and −591 and between −511 and to −469 upstream of the start codon. The latter region is involved in Zn-dependent NcZNT1 expression and has the conserved sequence recently identified by Assuncao et al. (2010) for AtZIP4, RTGTCGACAY, which was found by Assunção and colleagues to be bound by two bZIP transcription factors. Also, as was found for the Arabidopsis ZIP4 promoter, the sequence is repeated in the NcZNT1 promoter and with the repeats, the Zn-dependent region more likely encompasses a region spanning from −569 to −477 rather than −511 to −469 (Fig. 9).
The other identified promoter region is upstream of the two bZIP binding domains seems be involved in the high levels of NcZNT1 expression. The region, which is located from −667 to −591 upstream of the start codon, is activated as early as 6 d after planting (DAP) and the expression it drives is independent of changes in plant status for Zn or other metals (Fig. S6). For example, when plants were grown on nutrient solutions containing between 0 and 50 μM Zn, strong NcZNT1 expression was seen in plants grown on these different Zn levels anywhere between 6 d and 3 wk. This expression also seems to be independent of changes in the plant status for other metals as the strong NcZNT1 expression seen in nutrient replete plants did not change in plants grown under –Cu, –Fe or –Mn conditions (data not shown). Expression of this promoter region did not change the localization of NcZNT1 expression, but it did result in more rapid induction of GUS expression, as well as stronger expression compared with GUS expression driven by the Zn-dependent promoter region described above. For comparison, 5′ deletions of the AtZIP4 promoter driving the GUS reporter were transformed into Arabidopsis seedlings and only regions containing one or both of the bZIP binding domains were able to activate the GUS reporter, and then only after growth for 3 wk on –Zn media.
To try to understand if the mapped region of the NcZNT1 promoter between −667 and −591 can function by itself as an activator of NcZNT1 expression, this promoter region was cloned upstream of an eGFP reporter and transiently expressed in Arabidopsis protoplasts. Transformed protoplasts failed to display any eGFP expression when monitored for 48 h post-transformation.
The full length (c. 1.1 kb) promoter sequence is shown in Fig. S7 with the two Zn-responsive regions identified by Assuncao et al. (2010) highlighted in yellow and the putative hyperexpression region highlighted in green.
In a previous publication, it was suggested that NcZNT1 is involved in the uptake of Zn and Cd from the soil due to the fact that it is very highly expressed in roots of N. caerulescens and its concentration-dependent kinetics for Zn and Cd uptake in yeast were similar to the kinetics for root Zn and Cd uptake in Noccaea (Pence et al., 2000). It was also shown in the previous publication that NcZNT1 is highly expressed in the shoot. Hence, it was suggested that it may be a primary Zn/Cd uptake transporter in root and leaf cells in N. caerulescens. However, to date, no subsequent detailed work on the role of this transporter in metal hyperaccumulation in N. caerulescens has been carried out. The characterization of NcZNT1 presented here supports our previous suggestions (Pence et al., 2000) that NcZNT1 is most likely involved in Zn uptake from the soil. However, the data presented in the current manuscript indicates that NcZNT1 is not involved in root Cd uptake. In our previous study of NcZNT1 uptake in yeast, the uptake media was a low salt solution containing only CaCl2, glucose and MES buffer (Pence et al., 2000). In that study, the data indicated that NcZNT1 was a high-affinity Zn uptake transporter and only absorbed Cd with a low affinity. In the current study we used a full mineral uptake solution supplemented with Cd, so it is likely that Zn and possibly other ions in the uptake solution effectively outcompeted Cd transport by NcZNT1. Another piece of evidence that NcZNT1 does not transport Cd is that our transgenic Arabidopsis lines overexpressing NcZNT1 only showed increased sensitivity to Zn and not to any other metal including Cd. Unlike the closely related family member AtIRT1, which when overexpressed in plants conferred Cd sensitivity (Connolly et al., 2002), NcZNT1 does not. The final piece of evidence is that when NcZNT1 was expressed in two different yeast backgrounds (wild-type and zrt1/zrt2Δ), the cells expressing NcZNT1 absorbed amounts of Cd equal to or less than the cells expressing the empty vector in both yeast backgrounds (Fig. 1). The lower Cd levels seen in the zrt1/zrt2Δ yeast mutant may be explained by an upregulation of other yeast metal transporters that may mediate Cd uptake in the zrt1/zrt2Δ mutant in response to Zn deficiency. One gene controlled by ZAP1, the transcriptional regulator of ZRT1 and ZRT2 in response to changing Zn status, is FET4. This gene is upregulated in zrt1/zrt2Δ cells in response to the Zn deficiency caused by the zrt1/zrt2Δ genotype, when they are expressing the empty vector; thus, FET4 could be mediating the increased Cd uptake (Lyons et al., 2000). However, in zrt1/zrt2Δ cells expressing NcZNT1, enough Zn may be absorbed via NcZNT1 from the uptake media containing both Cd and Zn to downregulate the other transporters induced by Zn deficiency, as NcZNT1 mediates high-affinity Zn uptake but only absorbs Cd at a low affinity (Pence et al., 2000).
The response of NcZNT1 gene expression to changes in plant Zn status, where expression increases as Zn status decreases, is consistent with its role as a Zn root transporter (Fig. S1). It is interesting to note that NcZNT1 expression is also induced by Cd exposure and, to a lesser degree, by Cu deficiency, as NcZNT1 does not appear to transport either metal. The lower levels of root NcZNT1 transcript abundance in response to Mn and Fe limitation is also of interest. Possibly NcZNT1 hyperexpression is controlled by a number of different factors by the plant, but at this time these responses are not well understood and may reflect Zn levels in the plant being too high relative to cellular Fe concentrations.
The eGFP-NcZNT1 expression studies suggest that NcZNT1 is localized in the plasma membrane (Fig. 4). Also, the findings from other experiments presented here support this. For example, the Zn sensitivity data for transgenic Arabidopsis seedlings overexpressing NcZNT1, as well as the imaging of roots containing the Zn fluorescent dye, Zinquin, are all consistent with a plasma membrane localization for NcZNT1 (Figs 5, 7). Although the microscopic resolution of the Zinquin fluorescence does not allow us to determine whether the Zn is in the cytoplasm or is localized to an endomembrane compartment, the general pattern of Zn-dependent Zn fluorescence does support our hypothesis for NcZNT1’s function in Zn uptake. We also cannot determine exactly which tissues are stained with the Zinquin florescent dye; however, it appears that the Zn is outside of the stele and root vasculature, and most likely is located in cells of the root cortex and epidermis. It should be noted that because in the overexpression lines NcZNT1 is under the control of the 35S promoter, and thus expressed to high levels in all cells, this could lead to an alteration in the normal regulation of Zn accumulation in roots of wild-type N. caerulescens plants. But in general, the bright and more generally distributed Zinquin fluorescence in the NcZNT1 overexpression lines is consistent with a role for NcZNT1 in Zn uptake into the root symplasm. These findings also suggest that Zn and Cd hyperaccumulation is not a simple Mendelian trait. It appears that a suite of metal transporters are functioning together to move high levels of Zn and Cd into the shoot including HMA4, MTP1 and HMA3 (Bernard et al., 2004; Papoyan & Kochian, 2004, Desbrosses-Fonrouge et al., 2005; Ueno et al., 2011).
The location of the NcZNT1 protein in N. caerulescens roots supports its role in Zn uptake from the soil, with protein expression seen in both the epidermis and the cortex of Noccaea roots. However, the high stelar expression suggests an additional role for NcZNT1 associated with long-distance Zn transport. Possibly, as Zn moves radially across the root and crosses the endodermis, further radial movement within the stele could involve the sequential uptake and then release of Zn as it moves to the xylem vessels for long-distance transport. This might suggest that NcZNT1 may play a role in the hyperexpression phenotype not only by helping uptake of Zn, but also Zn re-absorption once the Zn is loaded into the xylem and transported to the aerial portions of the plant. Other circumstantial evidence which supports the hypothesis that NcZNT1 plays some role in Zn hyperaccumulation is the finding from reciprocal grafting experiments that roots drive the hyperexpression phenotype in N. caerulescens (de A Guimarães et al., 2009).
The very similar expression patterns for both the immunolocalization of the NcZNT1 protein in Noccaea roots and NcZNT1 gene expression in transgenic Arabidopsis expressing the NcZNT1 promoter::GUS reporter construct, indicated to us that Noccaea promoter::reporter studies could be carried out in transgenic Arabidopsis plants. Hence, as depicted in Fig. 8, we used stably transformed Arabidopsis seedlings as a tool for dissection of the NcZNT1 promoter. Two distinct promoter regions controlling NcZNT1 expression were identified based on these studies. One region is very similar to the recently reported promoter motif of the closely related Arabidopsis gene, AtZIP4 (Assuncao et al., 2010), with only a single base pair difference between the Arabidopsis and Noccaea palindromic sequences. However, a second novel NcZNT1 promoter region was identified in our study that is upstream of the Zn-dependent region common to AtZIP4 and NcZNT1 and might be involved in the NcZNT1 hyperexpression. This region showed GUS activation as soon as 6 DAP and was not influenced by changes in Zn status whereas the putative Zn-dependent promoter region was not activated until at least 2 wk of plant growth. Furthermore, this promoter region functions independently of Noccaea Zn status. It was interesting to note that plants grown under Fe or Mn limiting conditions also did not substantially change the expression of this NcZNT1 promoter region in transgenic Arabidopsis, because we found, as noted in Fig. S2, that the transcript levels of NcZNT1 in N. caerulescens plants are substantially lower when grown under Fe or Mn limiting conditions.
In summary, the findings presented here indicate that some of our hypotheses concerning the function of NcZNT1 presented in Pence et al. (2000) were wrong. We now suggest that NcZNT1 is a root Zn uptake transporter mediating Zn uptake from the soil in Noccaea caerulescens. Furthermore, the higher NcZNT1 gene and protein expression in the root stele leads us to propose a new role for NcZNT1 in long-distance Zn transport from root to shoot, possibly in the reabsorption of Zn into cells associated with the long-distance transport pathway. Transport via NcZNT1 seems to be fairly specific for Zn compared with other essential micronutrients (Cu, Mn, Fe) and the toxic metal, Cd. Finally, here we have identified two distinct regions of the NcZNT1 promoter which regulate its expression, one of which responds to changes in plant Zn status and the other promoter region functioning independent of Zn status and possibly involved in the hyperexpression of NcZNT1.
We would like to thank Dr David Eide for providing us with the fet3/fet4Δ and smf1Δ yeast mutants, Dr Dennis Thiele for providing us with the ctr1/ctr3Δ yeast mutant, Michael Rutzke for running the many ICP samples, and the Plant Cell Imaging Center at Boyce Thompson Institute for the use of their confocal microscope.