Differential expression and regulation of iron-regulated metal transporters in Arabidopsis halleri and Arabidopsis thaliana – the role in zinc tolerance


  • Varanavasiappan Shanmugam,

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    2. Molecular and Biological Agricultural Sciences, Taiwan International Graduate Program, National Chung-Hsing University and Academia Sinica, Taipei, Taiwan, 11529
    3. Graduate Institute of Biotechnology and Department of Life Sciences, National Chung-Hsing University, Taichung, Taiwan, 402
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Jing-Chi Lo,

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    2. Institute of Plant Biology, National Taiwan University, Taipei, Taiwan 10617
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Chia-Lin Wu,

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    Search for more papers by this author
  • Shan-Li Wang,

    1. Department of Soil and Environmental Sciences, National Chung-Hsing University, Taichung, Taiwan 402
    Search for more papers by this author
  • Chong-Cheong Lai,

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    Search for more papers by this author
  • Erin L. Connolly,

    1. Department of Biological Sciences, University of South Carolina, Columbia, SC 29208, USA
    Search for more papers by this author
  • Jing-Ling Huang,

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    Search for more papers by this author
  • Kuo-Chen Yeh

    1. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan, 11529
    2. Molecular and Biological Agricultural Sciences, Taiwan International Graduate Program, National Chung-Hsing University and Academia Sinica, Taipei, Taiwan, 11529
    3. Graduate Institute of Biotechnology and Department of Life Sciences, National Chung-Hsing University, Taichung, Taiwan, 402
    Search for more papers by this author

Author for correspondence:
Kuo-Chen Yeh
Tel: +886 2 2787 2056
Email: kcyeh@gate.sinica.edu.tw


  • To avoid zinc (Zn) toxicity, plants have developed a Zn homeostasis mechanism to cope with Zn excess in the surrounding soil. In this report, we uncovered the difference of a cross-homeostasis system between iron (Fe) and Zn in dealing with Zn excess in the Zn hyperaccumulator Arabidopsis halleri ssp. gemmifera and nonhyperaccumulator Arabidopsis thaliana.
  • Arabidopsis halleri shows low expression of the Fe acquisition and deficiency response-related genes IRT1 and IRT2 compared with A. thaliana. In A. thaliana, lowering the expression of IRT1 and IRT2 through the addition of excess Fe to the medium increases Zn tolerance.
  • Excess Zn induces significant Fe deficiency in A. thaliana and reduces Fe accumulation in shoots. By contrast, the accumulation of Fe in shoots of A. halleri was stable under various Zn treatments. Root ferric chelate reductase (FRO) activity and expression of FIT are low in A. halleri compared with A. thaliana. Overexpressing a ZIP family member IRT3 in irt1-1, rescues the Fe-deficient phenotype.
  • A fine-tuned Fe homeostasis mechanism in A. halleri maintains optimum Fe level by Zn-regulated ZIP transporters and prevents high Zn uptake through Fe-regulated metal transporters, and in part be responsible for Zn tolerance.


In plants, heavy metals such as copper (Cu), zinc (Zn), manganese (Mn), iron (Fe), cobalt (Co) and nickel (Ni) function as micronutrients at minimum quantity and are often added to mineral fertilizers to enhance the growth and development of crop plants. However, they also are considered toxic to plants, particularly when acquired in excess quantities. Thus, for essential metals, the physiological range between deficiency and toxicity is extremely low (Clemens, 2006) and necessitates strict control of metal ion homeostasis. At the cellular level, metals such as Fe and Cu can activate the reduced forms of oxygen produced during the essential cellular oxidation. Such a reaction could generate hydroxyl radicals, which are responsible for alteration of macromolecules and ultimately contribute to cell death (Briat & Lebrun, 1999). Despite the environmental and dietary significance of heavy metals, their importance has so far been under-recognized. Our understanding of the mechanisms that control heavy metal uptake and sequestration in plants is limited. Therefore, gaining knowledge of the mechanisms that control the uptake, accumulation and tolerance of heavy metals in plants is imperative.

Among the essential heavy metals, Zn has been known to be involved in several physiological and metabolic processes that are essential for plant growth and development (Guerinot & Eide, 1999). When Zn is deficient in soils, these processes in plants are severely impaired and the productivity of plants is adversely affected, which often results in poor-quality production (Cakmak, 2002). Zinc toxicity in plants could occur because of soils contaminated by mining, agricultural soils treated with sewage sludge, and urban and semi-urban lands contaminated by effluents and wastes containing excess Zn (Chaney, 1993). Zinc toxicity symptoms in plants usually become visible at a leaf concentration of ≥ 300 mg kg−1 Zn on a DW basis, although some crops may even show symptoms at just over 100 mg kg−1 Zn (Marschner, 1995).

Several organisms have evolved homeostasis mechanisms to cope with surrounding soil Zn levels. When soil Zn levels are high, plants keep very tight control over the internal concentrations of Zn. The Zn homeostasis must therefore be well regulated within plants. However, some plant species (e.g. Zn hyperaccumulators) can accumulate large amounts of Zn, > 10 000 mg kg−1 Zn DW, without any symptoms of toxicity (Baker & Brooks, 1989). More than 500 metal hyperaccumulator species from a wide range of unrelated families exist; among them c. 15 are Zn hyperaccumulators (Broadley et al., 2007; Verbruggen et al., 2009; Kramer, 2010).

Arabidopsis halleri has long been known as a potential Zn hyperaccumulator (Ernst, 1974; Bert et al., 2000; Zhao et al., 2000). It can also accumulate cadmium (Cd) (Kupper et al., 2000; Cosio et al., 2004; Zhao et al., 2006). A detailed study of A. halleri growing both on metal-contaminated and noncontaminated sites revealed that the hyperaccumulation of Zn and Cd is a constitutive trait (Bert et al., 2002). In recent years, A. halleri has emerged as a model species for studies of the mechanism involved in accumulation and tolerance of Zn and Cd because it is closely related to A. thaliana, whose genome is thoroughly explored. Recent studies comparing the expression profiles of A. thaliana and two subspecies of A. halleri identified several candidate genes related to metal tolerance and accumulation (Becher et al., 2004; Weber et al., 2004; Chiang et al., 2006; Talke et al., 2006).

Many physiological and molecular responses to Zn deficiency and toxicity have been described. Two bZIP family transcription factors, bZIP19 and bZIP23, found to be involved in Zn-deficiency response in A. thaliana (Assuncao et al., 2010). Heavy metal ATPase (HMA) transporters of the type 1b P-type ATPases play roles in root-to-shoot translocation of Zn (Hussain et al., 2004; Verret et al., 2004; Barabasz et al., 2010). The C-terminal domain of AtHMA4 is found to be important for Zn translocation to shoots of A. thaliana (Mills et al., 2010). Efficient long-distance Zn transport has been proven to be required for the tolerance of high Zn. Knocking down the expression of A. halleri HMA4, diminishes Zn tolerance (Hanikenne et al., 2008). While A. halleri HMA4 plays a crucial role in Zn tolerance, overexpression of HMA4 in A. thaliana does not increase Zn tolerance (Hanikenne et al., 2008). HMA4 is a required factor for Zn tolerance, but it is not sufficient in nonhyperaccumulators such as A. thaliana. Several other transporters have been reported to confer Zn tolerance and homeostasis. Metal tolerance protein (MTP1) of the cation diffusion facilitator family, is a tonoplast-localized transporter that plays a role in Zn tolerance (van der Zaal et al., 1999; Kobae et al., 2004; Desbrosses-Fonrouge et al., 2005; Gustin et al., 2009; Shahzad et al., 2010). More recently, ZIF1, a novel tonoplast-localized major facilitator superfamily transporter, was found to be a component of Zn homeostasis in plants (Haydon & Cobbett, 2007). Nramp3 and Nramp4 confer Zn and Cd tolerance (Oomen et al., 2009). Although Zn transporters are being widely investigated, other molecules involved in Zn tolerance or homeostasis are little understood. In recent studies, nicotianamine is known to play a role on Zn hyperaccumulation (Weber et al., 2004). Interestingly, the plant heavy metal chelator, phytochelatin, was recently found to play a role in Zn homeostasis. The synthesis of phytochelatin was proven to be essential for the detoxification of excess Zn and contributes to its accumulation (Tennstedt et al., 2009). In addition, plant defensins, a well-known plant innate immune component, also were demonstrated for their role in conferring Zn tolerance (Mirouze et al., 2006).

Limited effort has been invested in understanding cross-homeostasis between heavy metals with similar biological properties. This phenomenon has been emphasized further in a recent observation in Ni hyperaccumulator, Alyssum inflatum, in that Fe and Cu homeostasis are involved in causing Ni toxicity (Ghasemi et al., 2009). Nongrass plants develop the Strategy-I response when facing Fe deficiency (Connolly & Guerinot, 2002; Schmidt, 2003). Gene products in this response include ferric chelate reductase (FRO2) and two metal transporters of ZIP family, IRT1 and its homolog, IRT2. These two transporters are not only regulated by Fe deficiency, but also possess ability to transport Zn (Rogers et al., 2000; Vert et al., 2001; Kim & Guerinot, 2007). The expression of these genes are induced in response to iron deficiency under transcriptional and post-transcriptional regulation of the central transcription factor, FIT (Bauer et al., 2007). In this study, we observed Fe homeostasis contributing to the tolerance of excess Zn.

Materials and Methods

Plant materials and growth conditions

The growth and propagation procedure for A. thaliana and A. halleri ssp. gemmifera was described previously (Chiang et al., 2006). Both A. thaliana and A. halleri were grown on half-strength Murashige and Skoog (MS) medium containing half-strength MS salt, 0.25% phytagel, 500 mg l−1 of 2-morpholinoethanesulfonic acid (MES) and 1% sucrose, at pH 5.7. Plant materials were cultured in the growth condition with light intensity at 70 μmol m−2 s−1 with a 16-h light/8-h dark cycle at 22°C. The tissue culture grown A. halleri cuttings (2 wk after rooting) were used as experimental materials wherever A. halleri mentioned in the paper for further growth on half-strength MS medium for the same periods (age of pretreatment) as A. thaliana seedlings before treatments.

Molecular cloning of genes

To clone genes, total RNA was extracted from A. halleri ssp. gemmifera with use of TRIzol reagent (Invitrogen). Genomic DNA of A. halleri ssp. gemmifera was isolated with use of DNeasy Plant Mini Kit (Qiagen). To characterize the full-length, RNA ligase-mediated rapid amplification of 5′ and 3′ cDNA ends, the GeneRacer kit (Invitrogen) was used. Polymerase chain reaction-based cloning and DNA sequencing were employed to conduct the full-length of AhIRT1 and AhFIT in A. halleri ssp. gemmifera. The construct of AtIRT1 and empty vector (pFL61) were transformed into yeast strains as described previously (Lin et al., 2009). A forward primer, 5′-TGA AAT TAG CGG CCG CAT GGC TTC AAC TTC AGC A-3′, and a reverse primer, 5′-ATA GAC TCG CGG CCG CCT AGA AAG ATC GGT TTT TAG T-3′, were used for AhIRT1 amplification and the PCR product cloned into the NotI site of pFL61. To over-express AtIRT3 in an irt1-1 mutant background, CaMV 35S promoter was used. A PCR fragment of full-length AtIRT3 coding sequence was amplified with a forward primer, 5′-ATG TCG ACA TGT TCT TCG TCG ATG TTC TT-3′ and a reverse primer, 5′-ATG AGC TCC TAA TTA AGC CCA AAT GGC A-3′. This fragment was then cloned into pCAMBIA1390 with SalI and SacI sites. Agrobacterium tumefaciens strain, GV3101, harboring the plasmid 35S::AtIRT3/pCAMBIA1390 was used to transform irt1-1 mutant line by in planta transformation.

Yeast growth conditions

Yeast strains of zrt1zrt2, fet3fet4, smf1 and BY4741 harboring AtIRT1, AhIRT1 or their vector pFL61 only were grown on synthetic defined media lacking uracil supplemented with 2%d-glucose (SD-Ura). Yeast Extract-Peptone-Adenine-Dextrose (YPAD) media; Yeast Extract-Peptone-Dextrose (YPD) supplemented with adenine (40 mg l−1) was used for yeast growth control. For complementation tests, zrt1zrt2 strains were grown on SD-Ura (pH 5.8) supplemented with EDTA 1 mM and 0, 0.5 or 0.75 mM ZnSO4; fet3fet4 strains were selected on SD-Ura (pH 5.5) with 0, 10 or 50 μM FeSO4. Yeast strains of smf1 were grown on SD-Ura medium (pH 6.5) containing 0 or 5 mM ethylene glycol-bis(beta-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA). Yeast wild-type strain, BY4741, was grown on SD-Ura (pH 5.8) containing 0, 5, or 10 μM CdSO4. Yeast cells were resuspended in SD-ura without adding any metal and diluted to the OD600 of 1, 0.1, 0.01 and 0.0001, then 10 μl of each yeast culture was dropped on selective medium. All plates were incubated at 30°C for 1–3 d before photograph.

Real-time quantitative PCR (Q-PCR) and reverse-transcription (RT)-PCR

For RT-PCR, 5-d-old A. thaliana seedlings and A. halleri cuttings were grown on normal (half strength MS) or Fe-deficient (−Fe, lack of Fe with additional 300 μM FerroZine) for another 5 d. Transcript-specific primers were designed for the amplification of a 500- to 620-bp fragment of each gene. Primers for amplifying the AtIRT1 fragment were (forward) 5′-TTT TGC AAT CTC TCC AGC AA-3′ and (reverse) 5′-GCA AGA GCT GTG CAT TTG AC-3′ and for AhIRT1, (forward) 5′-TTT CGC AAT CTC TCC GGC AA-3′ and (reverse) 5′-GCA AGA GTT GTG CAT TTG CG-3′. The same set of primers was used for the detection of AtIRT2 and AhIRT2: forward 5′-TAC ATT TCG TTC CTC CGT CC-3′ and reverse 5′-TCG AAC AAG TGA TGG AAG CA-3′. To detect the control Actin8 expression in both A. thaliana and A. halleri, the forward primer was 5′-CCA CAT GCT ATC CTC CGT CT-3′ and the reverse primer was 5′-CTG GAA AGT GCT GAG GGA AG-3′. A PCR was performed in cycles of 30 s at 94°C, 30 s at 55°C and 60 s at 72°C. To check the expression of AtIRT3 in the transgenic lines, transcript-specific primers of AtIRT3 were used as follows: forward primer, 5′-TCT CTC TCA GCA ACA GAG TCC A-3′ and reverse primer, 5′-CAA TGT GAA TCC CAC CAC TG-3′.

For Q-PCR, 5-d-old A. thaliana seedlings and A. halleri cuttings were treated on normal (half strength MS), Fe-deficient (−Fe) and excess Zn (+Zn, additional 250 μM ZnSO4 for A. thaliana and 1 mM for A. halleri) containing media for 3 d. The Q-PCR involved mixing 0.25 ng cDNA template with SYBR Green PCR Master Mix and reaction in an ABI PRISM 7300 (Applied Biosystems) following the manufacturer’s instructions. Primers were designed to amplify the same region of cDNA based on the sequence of A. halleri IRT1, IRT2 and FIT genes, accession numbers were AJ580312, AY960754 and FJ410331, respectively. The AhIRT1 forward primer was 5′-GGC CAC GAG CCT CTA CAC AA-3′and reverse primer was 5′-ACC ATG TCC ATG AGG CAT GA-3′. The AtIRT1 forward primer was 5′-CAC CAT TCG GAA TAG CGT TAG G-3′ and reverse primer was 5′-CCA GCG GAG CAT GCA TTT A-3, as described by Yang et al. (2010). The forward primer for AhIRT2 detection was 5′-CAT AGG TAT TGG AGC TTG GGA TT-3′ and the reverse primer was 5′-GAT AGT CCA ATG ACC ACG GAA TG-3′. The forward primer for AtIRT2 was 5′-CAT GGT ATT GGA GGT TGG CAT-3′ and the reverse primer was 5′-GAT AGT CCA ATG ACC ACA GAA TGA A-3′. The AhFIT forward primer was 5′-ATG CGG TTT CGT ATG TTC AAG A-3′ and reverse primer was 5′-GTC CCG CGA TAT CGG ATT T-3′. The forward primer for AtFIT was 5′-TCT TCG ACG AAT TGC CTG ACT-3′ and reverse primer was 5′-TTC ACC ACC GGC TCT AAC ACT-3′. Relative quantitative results were normalized to the expression of Actin8 (Chiang et al., 2006).

To ensure similar PCR efficiency for both species, we used the same pair of primers for this Actin8 control to detect high-homology short targets. In several Zn treatments, Actin8 was expressed similarly and was within 1-cycle difference in both species. The Q-PCR involved three biological replicates. The efficiency tests for the use of primers and data calculation followed the manufacturer’s instructions (ABI PRISM 7300; Applied Biosystems). The expression of IRT1, IRT2 and FIT in both A. halleri and A. thaliana grown under different treatments were normalized to the expression of these genes in A. thaliana grown under normal (half strength MS) conditions.

Measurement of Zn and Fe contents

To measure the concentration of Zn and Fe, 5-d-old A. thaliana were grown on the half strength MS medium (containing 50 μM FeSO4 and 15 μM ZnSO4) supplemented with ZnSO4 0–250 μM and combined with ferric citrate 100 or 200 μM for 12 d. The A. halleri cuttings were transferred to the medium containing 0–2 mM ZnSO4 concentrations for another 10-d. Roots and shoots were harvested separately. Tissues were washed three times with 10 mM CaCl2 and H2O. Reagent grade water with a resistance of 18.2 MΩ made by the Milli-Q water purification system (Millipore Co., MA, USA) was used for all experiments. Samples were dried at 70°C for 3 d and weighed, then cleaned by a microwave cleaning procedure before use. Shoot samples (200 mg) and root samples (50 mg) were transferred into a Teflon vessel and digested with 5 ml 65% HNO3 and 2 ml H2O2 (both Suprapur; Merck) in a MarsXpress microwave digestion system (CEM, Matthews, NC, USA). Tomato leaves (SRM-1573a) from the National Institute of Standards and Technology (Gaithersburg, MD, USA), which contained known Zn and Fe concentrations, were used as a standard. The volume was adjusted to 20 ml with H2O and filtered by use of a 0.45-μm membrane filter. The Zn and Fe contents in digested samples were analyzed by use of inductively coupled plasma-optical emission spectrometry (ICP-OES; PerkinElmer OPTIMA 5300) or atomic absorption spectroscopy (PerkinElmer). The calibration solutions were prepared by adequate dilution of multi-element stock standard solutions (ICP Multi Element Standard Certipur VIII; Merck). Digested reagent blank solutions were analysed, and the results were checked before sample analysis. Matrix-induced background interference was corrected by use of reagent blank subtraction. Spectral interferences originating from the biological sample matrix were corrected by selecting an alternative spectrum or by using interfering element corrections to overcome potential interference. Each elemental concentration in the sample was determined by triplicate measurements.

Fe3+ chelate reductase enzymatic assays

For FRO activity assay, 2-wk-old A. thaliana and A. halleri were transferred to half-strength MS (containing 50 μM FeSO4 and 15 μM ZnSO4), Fe-deficient (−Fe) or Zn-excess (+Zn, additional 100 μM for A. thaliana and 1 mM ZnSO4 for A. halleri) for another 3 d or 5 d. The activity of root-associated Fe3+ chelate reductase (FRO, EC was quantified via spectrophotometry (Yi & Guerinot, 1996). Samples from three to five plants were pooled for assaying. The entire root tissue was collected and submerged in reaction solution for the reductase assay. The assay solution consisted of 0.1 mM Fe3+-EDTA and 0.3 mM FerroZine (Sigma-Aldrich) in double-distilled H2O for measurement of Fe3+ reduction. After incubation for 20 min at room temperature in the dark, the formation of a purple-colored Fe2+–FerroZine complex was measured at 562 nm (extinction coefficient of 28.6 mM−1 cm−1). An aliquot of identical assay solution containing no plant material was used as a blank. The formation of Fe2+–FerroZine complex represented FRO activity and was expressed as μmole Fe2+ formation g-fresh weight−1 h−1.

Modeling of Zn and Fe solubility

The minteqa2 program version 4.03 (Allison et al., 1991), capable of calculating equilibrium aqueous speciation and precipitation-dissolution of metals, was used to evaluate the likelihood of Fe and Zn precipitation in the half-strength MS. The input concentration of each component in the media was based on the total content of every component in the half-strength MS and additional Zn and Fe. The model was run with a constant pH at 5.7 and the saturation indices with respect to various Fe and Zn solids were determined. If the saturation index of a solid is > 1, precipitation of that solid could occur in the media.

IRT1 antibody generation and efficiency test

The IRT1 peptide antibody was raised against a synthetic peptide (PANDVTLPIKEDDSSN) that corresponds to amino acids of the IRT1 deduced protein sequence in A. thaliana (LTK BioLaboratories, Taipei, Taiwan), as referred to in previous studies (Connolly et al., 2002; Vert et al., 2002).

As the IRT1 peptide antibody was raised according to A. thaliana deduced amino acid sequence (PANDVTLPIKEDDSSN) and the corresponding amino acid sequence in A. halleri differed in one amino acid (see the Supporting Information Fig. S2), the efficiency of IRT1 peptide antibody in detecting the amino acid sequence from both species was determined. For this purpose, two different peptides (At: PANDVTLPIKEDDSSN, Ah: PANDVTLPIKEDDSAN) were synthesized and purified. The two peptides were then diluted and blotted on PVDF membrane by use of the PR 648 Slot Blot Filtration manifold (Hoefer Scientific Instruments, San Francisco, CA, USA). Membranes were immunoblotted, and peptide bands were visualized and quantified. The efficiency of the antibody to against AtIRT1 by AhIRT1 was calculated with three independent experiments. Efficiency factor (AtIRT1 detection/AhIRT1 detection) was defined as 1.71 for the calculation of relative expression of AhIRT1.

Isolation of protein and immunoblot analysis

Ten-day-old A. thaliana and A. halleri were transferred to 1/2 MS, Fe-deficient (−Fe) and Zn-excess (100 μM and 1 mM ZnSO4 for A. thaliana and A. halleri, respectively) or excess Zn and Fe (Zn+Fe, final 250 μM ZnSO4 and 200 μM ferric citrate) for another 3 or 7 d. Extracts were prepared by grinding root tissue in liquid nitrogen and resuspended in 2× extraction buffer containing 15% glycerol, 6.5% SDS, 0.125 M Tris-HCl, pH 6.8. Samples were centrifuged at 4°C for 15 min at 13 200 g and protein concentration was determined using the bicinchoninic acid protein assay (Pierce, Rockford, IL, USA). Total protein (10 μg) was separated by NuPAGE 4–12% Bis-Tris Gel (Invitrogen) and transferred to PVDF membranes (Immobilin-P; Millipore). For immunodetection, membranes were blocked with 1× phosphate-buffered saline (PBS) containing 5% nonfat dry milk and 2% Tween-20 for 1 h and then incubated in 1 : 3000-diluted IRT1 peptide antibody for 1 h. The membranes were then washed in 1× PBS containing 0.5% Tween 20 four times at 15-min intervals. Membranes were then incubated for 1 h with 1 : 6000-diluted horseradish peroxidase-conjugated antimouse secondary antibody then washed with 1× PBS and 0.5% Tween 20 four times at 15-min intervals. Specific protein bands were visualized by use of Western Blot Chemiluminescence Reagent Plus kit (PerkinElmer) and quantified (Fujifilim Luminescent image analyzer, LAS-1000 Plus).


Fe uptake-related genes are regulated differentially in A. halleri and A. thaliana

In order to obtain the preliminary expression data of Fe related genes, we looked at the previous microarray data (Chiang et al., 2006) and revealed differential expression of Fe homeostasis-associated genes in A. halleri and A. thaliana. Two of the four genes encoding the Fe storage protein ferritin, FER1 (At5g01600) and FER4 (At2g40300), were expressed at higher levels in A. halleri than in A. thaliana, with FER1 (At5g01600) overexpressed by c. 3-fold. By contrast, expression of the IRT2 (At4g19680) Fe transporter gene was sixfold lower in all three A. halleri subspecies than in A. thaliana. The signal for the high-affinity Fe transporter IRT1 (At4g19690) was below the detection threshold in both species. The RT-PCR analysis revealed that expression of IRT1 and IRT2 was lower in A. halleri than in A. thaliana but FER1 was expressed at similar levels in both species (Fig. 1). These data revealed that reduced expression of genes related to Fe uptake under comparable Fe storage in A. halleri. We further characterized in A. thaliana whether this action is related to Zn tolerance.

Figure 1.

 Reverse-transcription polymerase chain reaction (RT-PCR) of IRT1, IRT2 and FER1 in roots of Arabidopsis halleri (Ah) and Arabidopsis thaliana (At). RNA samples were isolated from A. halleri and A. thaliana grown on normal (half-strength Murashige and Skoog (MS)) conditions. Numbers of the amplification cycles indicated were empirically tested to represent nonsaturated amplification. The expression of Actin8 was an internal control.

Excess Fe can rescue A. thaliana from Zn toxicity

To lower the gene expression of the Fe uptake system in A. thaliana, we increased the Fe concentration in the culture medium. At the optimal Fe concentration (50 μM, half-strength MS media), Zn stress produced by excess Zn caused growth reduction and a severe chlorotic phenotype in shoots (Fig. 2a). However, this phenotype was suppressed by the addition of 200 μM Fe to the medium (Fig. 2a). At the optimal Fe concentration, the shoot mass was decreased to c. 75% and 25% of the weight of plants grown under normal conditions with the addition of 50 μM Zn and 250 μM Zn, respectively (Fig. 2b). But at the same concentration of Zn (250 μM) with excess Fe (additional 200 μM), the shoot biomass was reduced to only c. 60% (Fig. 2b). The anion salt (citrate) does not play a role in the rescue (data not shown). These results clearly demonstrated that excess Fe could rescue Zn stress in A. thaliana.

Figure 2.

 Excess iron (Fe) rescues Arabidopsis thaliana from excess-zinc (Zn) stress. Five-day-old A. thaliana seedlings were treated under various combinations of Zn and Fe concentrations as indicated for 12 d. Additional Zn (0, 50, 100, 250 μM) and Fe (0, 100 and 200 μM) in the half-strength Murashige and Skoog (MS) medium are indicated (a). Quantification of shoot DW is shown in (b). Fe 0 μM, black bars; Fe 100 μM, light grey bars; Fe 200 μM, dark grey bars. The relative DW values were normalized to that in half-strength MS Zn concentration. Additional Fe (0, 100 and 200 μM) in the half-strength MS medium is indicated. Eighty-four plants were pooled for each experiment, and results are means of duplicate experiments. Bars indicate standard error of the mean.

We also conducted chemical modeling for the equilibrium speciation and potential precipitation of metals using the minteqa2 program. Under the experimental conditions, the possibility of Fe and Zn precipitation was minimal and Fe and Zn predominantly existed as free ions or complexes in the media. Therefore, the phenomena observed could not be caused by the effect of interaction between additional Zn and Fe in the media.

Reduced Fe accumulation in shoots is associated with Zn sensitivity

To further investigate Zn accumulation in A. thaliana in response to different Fe concentrations, we measured Zn content in roots and shoots. The Zn accumulation in shoots was not affected by excess Fe (200 μM; Fig. 3 upper panel), ranging from 0.1 to 1.3 mg g−1 DW for optimal to excess Zn levels, respectively, regardless of Fe concentration in the media. By contrast, Zn accumulation in roots showed a decreasing trend with increases in Fe concentration (Fig. 3, lower panel). However, Zn was accumulated to a higher level in roots with elevated Zn concentrations under optimal Fe concentration (Fig. 3, lower panel), with a threefold increase in root Zn accumulation under optimal Fe compared with excess Fe concentration (additional 200 μM) at Zn levels from 50 to 250 μM.

Figure 3.

 Effect of iron (Fe) on zinc (Zn) accumulation in Arabidopsis thaliana shoots and roots. The Zn concentration in the shoots and roots of A. thaliana treated with the various Zn and Fe combinations shown in Fig. 2 were measured with inductively coupled plasma-optical emission spectrometry. Additional Fe (Fe 0 μM, black bars; Fe 100 μM, light grey bars; Fe 200 μM, dark grey bars) in the half-strength Murashige and Skoog (MS) media is indicated. For roots, 84 samples were pooled for each experiment, and results are means of duplicate experiments. For shoots, c. 40 samples were pooled for each experiment, and results are means of four experiments.

In A. halleri, the shoot Fe content was not much affected by elevated Zn (Fig. 4a). However, in A. thaliana shoots, Fe concentrations decrease with increasing Zn in the medium. The Fe concentration was reduced, from 0.09 to 0.04 mg g−1 DW, with increasing Zn concentration from 0 to 250 μM (Fig. 4b). The root Fe accumulation pattern was not affected significantly by elevated Zn conditions in both species (Fig. S1). Thus, excess Zn can reduce the Fe content in shoots of A. thaliana but not A. halleri. Additional Fe, which rescued the excess Zn stress, also restores the Fe content in the shoot (Fig. 4b). These data suggest that reduced Fe content may, in part, explain the differences in Zn toxicity observed between the two species.

Figure 4.

 Effect of excess zinc (Zn) on iron (Fe) accumulation in Arabidopsis halleri and Arabidopsis thaliana. Iron concentration in shoots of A. halleri (a) and A. thaliana (b) treated with various concentrations of Zn were measured by inductively coupled plasma-optical emission spectrometry [correction added after online publication 18 January 2011: the x-axis label for part (a) of this figure was corrected from Additional Zn (μM) to Additional Zn (mM); additionally within part (b) of this figure, the x-axis label was corrected from Additional Zn (mM) to Additional Zn (μM)]. For A. halleri, two shoot samples were pooled for experiments performed in duplicate. For A. thaliana, c. 40 shoot samples were pooled for experiments performed in duplicate. Bar represents standard error.

Excess Zn elevates the expression of Fe-regulated transporters IRT1 and IRT2 in A. thaliana but not A. halleri

To examine the expression of genes involved in Fe acquisition under excess Zn in both Arabidopsis species, we cloned AhIRT1 gene and its cDNA (GeneBank accession FJ410332), and a partial cDNA encoding AhIRT2. The essential Fe transporter IRT1 in A. halleri is 345 amino acids in length, whereas A. thaliana IRT1 is 347 amino acids. Alignment of the deduced amino acid sequence with AtIRT1 indicates that they are highly conserved (97% identity; Fig. S2). This sequence information was used to design primers for Q-PCR. We found that the expression of IRT1 and IRT2 was consistently higher in A. thaliana than in A. halleri, and induced by tenfold and fivefold under Fe deficiency, respectively. Conversely, excess Zn causes fivefold induction in both IRT1 and IRT2 in A. thaliana. By contrast, the expression of IRT1 and IRT2 in A. halleri showed little induction or was unchanged with 1 mM Zn treatment, although the induction of IRT1 and IRT2 was observed in A. halleri under Fe-deficient conditions (Fig. 5).

Figure 5.

 Quantitative PCR of IRT1 (a) and IRT2 (b) mRNA in Arabidopsis halleri and Arabidopsis thaliana. Root total RNA was isolated from A. halleri (Ah) and A. thaliana (At) seedlings grown in half-strength Murashige and Skoog (MS) (CK, black bars) or iron (Fe)-deficient (−Fe, light grey bars) or zinc (Zn)-excess (+Zn, dark grey bars) half-strength MS medium for 3 d. Relative expression fold was calculated by comparison with AtIRT1 (a) and AtIRT2 (b) expression in A. thaliana roots. Means and SE were calculated from three replicates and plotted.

The accumulation of IRT1 protein in A. halleri and the function of AhIRT1

Previously, IRT1 was found to be regulated post-transcriptionally in response to Fe (Connolly et al., 2002). It is important to see the protein level change in IRT1 expression in both Arabidopsis species. Therefore, to examine IRT1 protein accumulation under the conditions used in this study, we generated a peptide antibody specific for AtIRT1, as described previously (Connolly et al., 2002; Vert et al., 2002) and detected IRT1 accumulation by immunoblotting.

First, we examined the efficiency with which the AtIRT1 antibody detects IRT1 protein in both A. thaliana and A. halleri (Fig. S3). The antibody was 1.71-fold more efficient on detecting AtIRT1 than detecting AhIRT1. We found that AhIRT1 is about similar in size but slightly smaller than AtIRT1. This data supports the prediction of amino acid sequence.

The accumulation of IRT1 proteins in A. thaliana and A. halleri is consistent with RNA levels. Overall, the expression of AhIRT1 is low, although it also can be induced by 7 d of growth on Fe-deficient or excess Zn medium. The accumulation of AtIRT1 is much greater than that of AhIRT1. Under a 7-d Fe-deficiency treatment, A. halleri showed a slightly chlorosic phenotype in the new leaves, but this phenotype was not observed under excess Zn (data not shown). Like previous studies, AtIRT1 accumulation is induced by Fe deficiency (Eide et al., 1996; Connolly et al., 2003). It is also induced by excess Zn (250 μM). This induction is repressed by additional Fe (Fig. 6). In order to investigate the function of AhIRT1, we examined its metal transport ability in a yeast system. AhIRT1 possesses similar biological abilities to transport Fe, Zn, Mn and Cd as AtIRT1 does in yeast complementation experiments (Fig. S4). These data suggest that AhIRT1 is a multiple metal transporter, similar to AtIRT1, but its expression is controlled at low level in A. halleri.

Figure 6.

 IRT1 protein accumulation in Arabidopsis halleri and Arabidopsis thaliana. Western blot analysis of 10 μg total proteins of A. halleri and A. thaliana treated with different zinc (Zn) and iron (Fe) conditions (half-strength Murashige and Skoog (MS) (C), Fe-deficient (−Fe), Zn-excess (+Zn) or excess on Zn and Fe (Zn + Fe)) for 3 d or 7 d. The numbers represent relative accumulation of IRT1 proteins (see the Materials and Methods section for details).

FRO activity and the expression of FIT in A. thaliana and A. halleri

To investigate other genes in the system, we analyzed the expression of FRO and FIT. FRO functions in Strategy-I Fe uptake system and can be upregulated under Fe deficiency and excess Zn (Robinson et al., 1999; Hell & Stephan, 2003; Schmidt, 2003). Ferric chelate reductase assays revealed that the FRO activity in A. halleri roots was about one-fifth of that in A. thaliana. The FRO activity increased in A. thaliana after root transferred to Fe-deficient and Zn-excess conditions for 3 d (Fig. 7a). Surprisingly, the low FRO activity in A. halleri remained unchanged with 1 mM Zn, even with a 5-d period, upon transfer to Fe-deficient or Zn-excess media (Fig. 7a). A twofold induction of FRO activity was seen after 7-d Fe deficiency (data not shown).

Figure 7.

 Expression of FRO (a) and FIT (b) in Arabidopsis halleri and Arabidopsis thaliana. (a) To examine FRO activity, A. halleri (Ah) and A. thaliana (At) seedlings were transferred to half-strength Murashige and Skoog (MS) (CK, black bars) or Fe-deficient (−Fe, light grey bars) or Zn-excess (+Zn, dark grey bars) half-strength MS medium for 3 d (3D) or 5 d (5D). Experiments were performed in duplicate. Bar represents standard error of means of three independent experiments. (b) FIT expression in A. halleri and A. thaliana treated with half-strength MS (CK), Fe-deficient (−Fe) and excess Zn (+Zn) conditions for 3 d were detected by quantitative-PCR and Actin8 was used as the control [correction added after online publication 18 January 2011: within the x-axis label for part (b) of this figure, the abbreviations Ah and At were corrected to At and Ah, respectively].

The expression of the central transcription factor for the regulation of Fe-deficiency responses, FIT, was also examined (Colangelo & Guerinot, 2004). The genomic and coding sequences of the FIT gene from A. halleri were cloned and characterized (GeneBank accession no. FJ410331). The AhFIT is highly conserved compared with AtFIT. The deduced amino acid sequence of AhFIT shows 96% identity with that of AtFIT (Fig. S5). The expression of AhFIT is slightly lower than that of AtFIT when grown under normal condition and not regulated with the Fe-deficient or excess Zn treatments. By contrast, AtFIT was induced under Fe deficiency and excess Zn conditions (Fig. 7b). These data suggest that Zn hyperaccumulation in A. halleri is facilitated by concomitant alterations in Fe homeostasis system.

Overexpressing IRT3 in irt1-1 rescues its growth defect and low shoot Fe content under Fe-limited conditions

Despite the low-level expression of Fe-deficiency response genes under high levels of Zn, A. halleri can still maintain unaltered shoot Fe content (Fig. 4b). Recently it was reported that IRT3, a ZIP family transporter, associated with one major quantitative trait locus cosegregated with high Fe accumulation phenotype in A. halleri (Willems et al., 2010). This condition raises the possibility of involvement of other transporters particularly ZIP family members in maintaining Fe level in A. halleri. IRT3 has the ability to transport Fe and Zn in yeast and high expression of IRT3 was observed in A. halleri while IRT3 expression is only elevated under Zn deficiency in A. thaliana (Fig. S6; Chiang et al., 2006; Lin et al., 2009). The irt1-1 mutant in A. thaliana had severe growth defects when grown under Fe-limited conditions and impaired in maintaining shoot Fe concentrations (Vert et al., 2002). To test whether IRT3 has any role in Fe uptake or accumulation in planta, we performed a complementation experiment in which we overexpressed IRT3 under the control of CaMV 35S promoter in irt1-1 mutant background and this system also represented A. halleri IRT3 expression level. Three transgenic lines (35S::IRT3/irt1-1) have higher level of expression of IRT3 and these lines were selected for further analysis (Fig. 8a). Under Fe-deficient conditions, the transgenic plants can accumulate higher amounts of Fe in the shoot tissues compared with irt1-1 (Fig. 8b). The 35S::IRT3/irt1-1 lines rescue the growth defect phenotype of irt1-1 when grown on soil under Fe-limited conditions (Fig. 8c). These results might explain the high expression of Zn responsive/regulated transporters and their role in Fe uptake and accumulation in A. halleri.

Figure 8.

 Overexpression of IRT3 in irt1-1 mutant background. (a) To detect the expression of AtIRT3 in 35S::IRT3/irt1-1 transgenic lines, 2-wk-old seedlings grown on half-strength Murashige and Skoog (MS) medium were used for RNA isolation and cDNA synthesis. Reverse-transcription PCR (24 cycles) was performed to detect the expression. Expression of Actin8 was used as an internal control. (b) Seedlings were grown on half-strength MS for 10 d and transferred to iron (Fe)-free medium for further 5 d before sampling for measuring the concentration of Fe. (c) Phenotype of 4-wk-old Ws, irt1-1, and 35S::IRT3/irt1-1 transgenic lines grown on soil watered twice a week.


Both Zn and Fe are considered essential micronutrients for plant growth and development. Zinc deficiency can severely limit crop yields because it plays an essential role in several physiological and metabolic processes at the cellular level (Ramesh et al., 2004). Under Fe deficiency, most of the nongraminaceous plants activate a ferric reduction-based Strategy-I response to take up Fe from the soil (Schmidt, 2003). In our study, we observed differential expression of the Strategy-I Fe acquisition system in A. halleri, a Zn/Cd hyperaccumulator, compared with A. thaliana, with Fe uptake-related genes downregulated in A. halleri. Although IRT1 and IRT2 can be induced in A. halleri under Fe-deficient conditions, their expression is much lower in A. halleri than in A. thaliana (Figs 5, 6). In addition, these genes are expressed at low levels under excess Zn in A. halleri. In a previous study, low expression of IRT1 in Thlaspi caerulescens was reported (van de Mortel et al., 2006). IRT1 protein levels are lower in A. halleri than those in A. thaliana under excess Zn- and Fe-deficiency conditions (Fig. 6). Interestingly, ferric reductase activity in A. halleri was constitutively low under Fe-deficient and Zn-excess conditions whereas ferric reductase activity is strongly induced by these conditions in A. thaliana (Fig. 7). Short term treatment of elevated Zn concentrations in A. thaliana resulted in no change in shoot Fe content, but this still can induce IRT1 and ferric reductase activity (Becher et al., 2004).

In the present study, Fe content was reduced in shoots but not roots of A. thaliana with elevated Zn concentrations (Fig. 4), whereas shoot Zn accumulation is associated with media Zn concentration and not affected by Fe concentration. The addition of Fe not only reversed the reduced level of Fe but also reduced the accumulation of Zn in roots. These observations imply that in the competition between Zn and Fe, Zn is a priority in root-to-shoot transport and Fe is a priority in root uptake or accumulation that repulses Zn uptake in A. thaliana. By contrast, A. halleri shoot Fe content did not change under elevated Zn concentrations, with little change in the root Fe accumulation (Fig. 4; Fig. S1). This clearly indicates that Fe uptake and translocation is not altered in A. halleri upon Zn excess.

The unaltered shoot Fe content might be a reason for the reduced level of expression of Strategy-I Fe acquisition system in A. halleri. Conversely, the reduced shoot Fe concentrations under excess Zn might trigger the Fe-deficient signal that increases the expression of IRT1 and IRT2 transporters in the root of A. thaliana. Both IRT1 and IRT2 expression and FRO activity can be induced to high levels under Fe-deficient or Zn-excess conditions in A. thaliana. Their low expression under Fe-sufficient or Fe-excess conditions and high expression under Fe-deficient conditions were observed previously (Vert et al., 2001, 2009; Connolly et al., 2002, 2003). By contrast, in A. halleri, the Strategy-I Fe uptake system was expressed at a very low levels under both Fe-deficient and Zn-excess conditions, although IRT1 and IRT2 are inducible under Fe deficiency. Thus, in A. thaliana, the reduced Fe concentration in shoots under excess Zn is accompanied by upregulation of the Strategy-I Fe acquisition system. Fe deficiency could be sensed by plants, and the induction of the Fe acquisition system by excess Zn in A. thaliana could be responsible for Zn sensitivity because IRT1 and IRT2 can also transport Zn (Rogers et al., 2000; Vert et al., 2001, 2009; Kim & Guerinot, 2007; Yang et al., 2010). Knocking out the expression of IRT1 leads to a significant reduction in root Zn content under Fe-deficient conditions in A. thaliana (Vert et al., 2002). A higher level of Zn (100 μM) leading to decreased expression of IRT1 in A. thaliana was reported in a previous study (Connolly et al., 2002), but it was observed only under Fe-deficient conditions. Plants grown on Fe-sufficient medium with 500 μM Zn had neither AtIRT1 mRNA nor protein (Connolly et al., 2002). This might result from toxicity at the higher levels of Zn-leading to decreased expression of IRT1. This low AtIRT1 expression was also observed at high Zn concentrations in our experimental conditions (data not shown).

In A. halleri, expression of the Strategy-I Fe acquisition system remains low even under Fe-deficient conditions, which might help to control the Zn uptake at certain level. Though AhIRT1 possess similar metal uptake property (Fig. S4) as similar to AtIRT1, the differential expression of this transporter in two different species emphasizes that their regulation could be different under elevated Zn concentrations. Thus, A. halleri maintains highly efficient Fe usage, with low expression of the Fe acquisition system and an Fe concentration unaffected by excess Zn in shoots. This could bebecause of the elevated expression of Zn-regulated ZIP transporters in A. halleri specifying the metal uptake, and this could be a better adoptive mechanism under metal-rich environments. In our recent study, we found that overexpressing IRT3, a member of ZIP family, elevated Fe accumulation in A. thaliana (Lin et al., 2009). In this study, overexpressing IRT3 could complement the Fe-deficiency phenotype and restore Fe content in the shoot tissues of the irt1-1 mutant (Fig. 8). This system substantially represents A. halleri in which the expression of IRT3 is higher under normal and high concentrations of Zn (Chiang et al., 2006). High expression of IRT3 in root stele was observed in A. thaliana under Zn-deficient treatment (Lin et al., 2009). The high IRT3 expression in root stele could be responsible for better Fe translocation in A. halleri when considering the constitutively high expression of IRT3 in A. halleri (Fig. S6). Recently, Verbruggen’s group reported that A. halleri accumulated twofold higher Fe concentrations in the shoot tissues when compared with nonhyperaccumulator A. lyrara. In addition, it was found that two major quantitative trait loci (QTL) are linked with the Fe accumulation phenotype. IRT3, along with another two metal homeostasis genes, is associated with one of QTL (Willems et al., 2010). Thus, the high expression of ZIP transporters such as IRT3, ZIP3, ZIP6, ZIP9 and ZIP12 (Chiang et al., 2006) may play a function for Fe availability in A. halleri.

In general, Fe homeostasis in plants is well regulated through Fe uptake, trafficking and intracellular storage mechanisms (Curie & Briat, 2003; Hell & Stephan, 2003; Schmidt, 2003). Tobacco plants ectopically expressing ferritin exhibit an Fe-deficient phenotype with high FRO activity (Van Wuytswinkel et al., 1999). However, our observations suggest that A. halleri does not exhibit or delay Fe deficiency in this regard. The major Strategy-I Fe uptake transporter IRT1 was expressed at a lower level in A. halleri than in A. thaliana, but the Fe distribution/storage-associated genes nicotianamine synthase (NAS) and FER were more highly expressed in A. halleri than in A. thaliana (Talke et al., 2006). Together with the steadily low activity of FRO and FIT (Fig. 7), at least under our experimental conditions, the Strategy-I Fe-uptake machinery in A. halleri behaves as if the plant has sufficient Fe. Alternatively, the A. halleri Strategy-I Fe uptake system could possess high specific activity toward Fe and thus be sufficient to sustain Fe acquisition. It is worth mentioning here that overexpressing an AtIRT1 mutant defective in Fe-induced turnover accumulates higher concentrations of Fe in the roots without increasing FRO activity in A. thaliana, as found in AtIRT3 overexpressing lines (Kerkeb et al., 2008; Lin et al., 2009). This suggests that FRO activity is not the only limiting factor for Fe uptake in Strategy-I plants and that higher expression of ZIP transporters also could contribute to the high efficiency of Fe uptake.

It is also intriguing that in A. halleri, the expression of FIT maintains similar levels of expression under all our experimental conditions (Fig. 7). In A. thaliana, FIT targets many iron regulated genes including FRO2, IRT1 and IRT2 (Colangelo & Guerinot, 2004), and these genes have a low expression in A. halleri (Figs 5, 7). Conversely, few FIT regulated genes, including HMA3 and ZIP9, are expressed at high levels in A. halleri (Becher et al., 2004; Weber et al., 2004; Chiang et al., 2006). This raise the possibility of the presence of additional regulatory networks in response to metal deficiency or excess conditions in A. halleri. In a recent observation, it was found that FIT, together with bHLH38 or bHLH39, members from bHLH family transcription factors, were responsible for the induction of IRT1 and FRO2 (Yuan et al., 2008). The possible involvement of other transcription factors, including bHLH family in A. halleri under Fe-deficiency conditions, is not clear at this stage.

A distinct regulation of Fe homeostasis could exist in A. halleri. This mechanism could be accompanied or aided by Zn/Cd tolerance and accumulation in A. halleri. The elevated expression of dual-function ZIP and Nramp transporters could contribute, at least in part, to the Fe homeostasis in two subspecies of A. halleri (Curie et al., 2000; Thomine et al., 2000; Becher et al., 2004; Weber et al., 2004; Chiang et al., 2006; Cailliatte et al., 2010). In particular, mutation of Nramp3 confers sensitivity to Zn and Cd (Oomen et al., 2009). Overexpression of Nramp3 in A. thaliana did result in downregulation of IRT1 and FRO2 (Thomine et al., 2003). Therefore, the high expression of Nramp3 in A. halleri (Weber et al., 2004; Chiang et al., 2006) might play a role in maintaining Fe homeostasis and further add to Zn/Cd tolerance. Zinc accumulates greatly in the aerial regions of plants; however, Fe accumulation in A. halleri remains at relatively normal levels and occurs mainly in roots (Kubota & Takenaka, 2003; Becher et al., 2004). The maintenance of Fe content in plants may be critical to avoid the oxidative damage caused by the Fenton reaction through excess Fe (Hell & Stephan, 2003). Therefore, gating the uptake of Fe by downregulating the expression of IRT1 and IRT2 should play an important role in high expression of ZIP and Nramp transporters. This hypothesis is worthy of further exploration.

In A. thaliana with high expression of Fe-regulated metal transporters, excess Zn can be taken up and even overly when it causes Fe deficiency. However, the major uptake of Zn through Zn-regulated ZIP transporters in roots and efficient translocation to shoot and detoxification might help to maintain sufficient shoot Fe concentrations in A. halleri, eventually leading to lower expression of the Fe acquisition system to prevent losing control of Fe-regulated multimetal transporters under excess Zn. In summary, we report a differential expression and regulation of the Strategy-I Fe uptake system in the Zn hyperaccumulator A. halleri compared with the nonhyperaccumulator A. thaliana. For engineering heavy metal tolerance in non-hyperaccumulators, repression of Fe-regulated multimetal transporters may be worthy of consideration.


The work was supported by grants from the National Science Council (94-2311-B-001-002 and 95-2311-B-001-050-MY2) and Academia Sinica. ELC is grateful to the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service (grant number 2004-35100-14934) for financial support. We thank Drs Shu-Hsing Wu and Wolfgang Schmidt for helpful discussion and critical reading of the manuscript.