ZINC TOLERANCE INDUCED BY IRON 1 reveals the importance of glutathione in the cross-homeostasis between zinc and iron in Arabidopsis thaliana

Authors

  • Varanavasiappan Shanmugam,

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

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

    Corresponding author
    1. Molecular and Biological Agricultural Sciences, Taiwan International Graduate Program, National Chung-Hsing University and Academia Sinica, Taipei, Taiwan
    2. Graduate Institute of Biotechnology and Department of Life Sciences, National Chung-Hsing University, Taichung, Taiwan
    3. Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan
      (fax +886 2 2651 5600; e-mail kcyeh@gate.sinica.edu.tw).
    Search for more papers by this author

(fax +886 2 2651 5600; e-mail kcyeh@gate.sinica.edu.tw).

Summary

Zinc is an essential micronutrient for plants, but it is toxic in excess concentrations. In Arabidopsis, additional iron (Fe) can increase Zn tolerance. We isolated a mutant, zinc tolerance induced by iron 1, designated zir1, with a defect in Fe-mediated Zn tolerance. Using map-based cloning and genetic complementation, we identified that zir1 has a mutation of glutamate to lysine at position 385 on γ-glutamylcysteine synthetase (GSH1), the enzyme involved in glutathione biosynthesis. The zir1 mutant contains only 15% of the wild-type glutathione level. Blocking glutathione biosynthesis in wild-type plants by a specific inhibitor of GSH1, buthionine sulfoximine, resulted in loss of Fe-mediated Zn tolerance, which provides further evidence that glutathione plays an essential role in Fe-mediated Zn tolerance. Two glutathione-deficient mutant alleles of GSH1, pad2-1 and cad2-1, which contain 22% and 39%, respectively, of the wild-type glutathione level, revealed that a minimal glutathione level between 22 and 39% of the wild-type level is required for Fe-mediated Zn tolerance. Under excess Zn and Fe, the recovery of shoot Fe contents in pad2-1 and cad2-1 was lower than that of the wild type. However, the phytochelatin-deficient mutant cad1-3 showed normal Fe-mediated Zn tolerance. These results indicate a specific role of glutathione in Fe-mediated Zn tolerance. The induced accumulation of glutathione in response to excess Zn and Fe suggests that glutathione plays a specific role in Fe-mediated Zn tolerance in Arabidopsis. We conclude that glutathione is required for the cross-homeostasis between Zn and Fe in Arabidopsis.

Introduction

Plants require optimal concentrations of a number of essential nutrients, including heavy metals such as iron (Fe), copper (Cu), manganese (Mn), zinc (Zn), boron (B), and molybdenum (Mo), to effectively complete their life cycle. For normal growth and metabolism of plants, these metal ions are required in trace quantities. However, when these essential metals accumulate in excess amounts in plant tissue, they become extremely toxic and cause physiological alterations and growth inhibition. Thus, plants must tightly regulate metal homeostasis to cope with adverse environmental conditions. The metal homeostasis network in plants operates from mobilization to storage to acquire an optimal concentration of each micronutrient (Hansch and Mendel, 2009).

Zinc is an essential micronutrient for all organisms and is the second most abundant metal in living organisms after Fe (Andreini et al., 2006; Broadley et al., 2007). It is necessary for several physiological and metabolic processes of plants (Ramesh et al., 2004). Zinc acts as a cofactor for more than 300 enzymes (Guerinot and Eide, 1999); Zn-dependent enzymes are functional in all major organelles of the cell and are involved in DNA replication, transcription, and translation, and photosynthesis (Luciano et al., 1998; Kramer, 2005a,b). The optimal leaf Zn concentration required for normal growth in most crops ranges from 15 to 20 μg g−1 dry weight (DW) (Marschner, 1995). Zinc deficiency results in leaf chlorosis, stunted growth, shortened internodes, and small leaves (Guerinot and Salt, 2001; Cakmak, 2002). Conversely, excess amounts of Zn lead to symptoms of toxicity.

Zinc toxicity occurs in soils contaminated by mining and smelting activities (Chaney, 1993). Symptoms of toxicity are seen in plants when the Zn concentration ranges from 100 to 300 μg g−1 leaf DW depending on the plant species (Chaney, 1993; Marschner, 1995). Because the physiological range between optimal and toxic concentrations of Zn is narrow, plants have a strict homeostasis mechanism to cope with surrounding soil Zn levels. The mechanisms plants use to overcome Zn toxicity include uptake and translocation of the metal from roots to shoot, detoxification through chelation and sequestration, homeostasis of other nutrients, and a stress protection process (Kramer, 2005a,b; Kramer et al., 2007; Verbruggen et al., 2009).

Many genes involved in Zn tolerance have been identified and thoroughly investigated. Heavy metal ATPase (HMA) transporters of the type 1b P-type ATPases play roles in the root-to-shoot translocation of Zn (Hussain et al., 2004; Verret et al., 2004; Hanikenne et al., 2008; Kim et al., 2009; Barabasz et al., 2010). In hyperaccumulators such as Arabidopsis halleri, HMA4 is involved in the translocation of Zn to the aerial parts and is responsible for Zn tolerance (Hanikenne et al., 2008). Metal tolerance protein 1 of the cation diffusion facilitator family, is a tonoplast-localized transporter that plays a role in Zn tolerance (Kobae et al., 2004; Gustin et al., 2009; Kawachi et al., 2009; Shahzad et al., 2010). A novel tonoplast-localized major facilitator superfamily transporter, ZIF1, was found as a component of Zn homeostasis in plants (Haydon and Cobbett, 2007). Also, synthesis of phytochelatin (PC) is essential for detoxification of excessive Zn and contributes to its accumulation (Tennstedt et al., 2009). PLANT CADMIUM RESISTANCE 2 (PCR2), a membrane protein of the PCR family, is involved in Zn extrusion and long-distance Zn transport from root to shoot (Song et al., 2010). In addition to Zn transporters and metal tolerance proteins, other signaling molecules are also associated with Zn tolerance. One recent example is nitric oxide production in Solanum nigrum which is associated with long-term Zn tolerance by modulating the root architecture (Xu et al., 2010).

Excess or deficiency status of a particular metal or mineral elements can influence the uptake or translocation process of other metals and affect overall plant growth and development (Ghasemi et al., 2009; Bose et al., 2011; Chou et al., 2011; Fukao et al., 2011; Sarvari et al., 2011). In the nickel (Ni) hyperaccumulator Alyssum inflatum, Fe and Cu homeostasis are involved in causing Ni toxicity (Ghasemi et al., 2009). In many plant species, excess application of magnesium (Mg) alleviates aluminum (Al) toxicity (Kinraide, 2003; Kinraide et al., 2004) and is relevant to various mechanisms proposed, including protection against oxidative damage and better carbon partitioning from shoots to roots (Bose et al., 2011). In Arabidopsis thaliana, low Mg status improved tolerance to increased cadmium (Cd) exposure, together with increased shoot Fe concentration (Hermans et al., 2011), which further suggests the influences of Mg level on Cd tolerance and shoot Fe accumulation. The increased Fe concentration is attributed to the protection of the photosynthetic apparatus under excess Cd stress (Hermans et al., 2011). In addition, excess Cd induced symptoms of Fe deficiency in a few plant species (Solti et al., 2008; Han et al., 2011; Sarvari et al., 2011). Although these observations reveal the occurrence and importance of cross-metal homeostasis, the detailed molecular mechanism and key molecules involved in this processes remain obscure.

We previously observed that excess Zn in A. thaliana can reduce the accumulation of Fe in shoots and thus induce significant Fe deficiency. Interestingly, excess Fe could rescue the Zn stress even in plants with toxic levels of Zn in the shoot (Shanmugam et al., 2011). To better understand the genes involved in this Fe-induced Zn stress tolerance, we used genetic screening to identify mutants defective in this process with an ethyl methanesulfonate (EMS) mutant pool. We found a defective mutant that cannot tolerate high levels of Zn in the presence of additional Fe. We used a map-based cloning approach to identify the genetic locus responsible for the Fe-mediated Zn tolerance phenotype. ZIR1 was found to encode γ-glutamylcysteine synthetase (γ-ECS, γ-glutamylcysteine ligase, GSH1), the enzyme that catalyzes the first step of glutathione biosynthesis. In plants, glutathione plays an important role in biotic and abiotic stress tolerance. Glutathione is a key enzyme in the glutathione–ascorbate cycle that scavenges hydrogen peroxide (Foyer and Halliwell, 1976; Noctor and Foyer, 1998; Noctor et al., 1998; Foyer and Noctor, 2005). It is the precursor of PCs, glutathione oligomers that chelate heavy metal such as Cd (Ha et al., 1999). Our identification of ZIR1 as GSH1 suggests that adequate levels of glutathione are important for metal tolerance, in particular Fe-mediated Zn tolerance, in Arabidopsis.

Results

Identification of a zir1 mutant defective in Fe-mediated Zn tolerance

In previous study, we observed that wild-type seedlings can tolerate toxic levels of Zn in the presence of excess Fe in the medium (Shanmugam et al., 2011). To identify the genes involved in Fe-mediated Zn tolerance, we used genetic screening for mutants defective in this process. EMS-mutated seeds were screened to select mutants lacking Zn tolerance under excess Fe conditions. Approximately 14 900 seeds were screened, and 359 mutants were selected for further analysis. Some of the mutants showed severe growth defects at later stages and were unable to produce seeds. The remaining mutants were further characterized. We obtained a defective mutant, designated zinc tolerance induced by iron 1, zir1. This zir1 mutant cannot tolerate a high level of Zn in the presence of additional Fe, as compared with the wild type (Figure 1). The zir1 seedlings were smaller in size than the wild type, even under control (1/2 MS) conditions. They were more sensitive to excess Zn and showed no Fe-mediated Zn tolerance.

Figure 1.

 The phenotype of wild-type and zir1 seedlings under excess Zn or excess Zn and Fe.
Five-day-old Col-0 and zir1 seedlings grown on ½ MS were treated with 250 μm Zn (Zn) or 250 μm Zn and 200 μm Fe (Zn + Fe) for 8 days. As a control, seedlings were grown on ½ MS.

Map-based cloning identifies ZIR1 as a γ-glutamylcysteine ligase

To identify the nature of the zir1 mutation, zir1 was crossed with the wild type (Col-0), and the F1 progeny obtained were analyzed under conditions of excess Zn and Fe. All F1 seedlings were able to rescue the Zn toxicity (data not shown). F2 progenies were obtained by selfing F1 seedlings and were tested under conditions of excess Zn and Fe. The obtained population was segregated into wild-type to mutant phenotypes at a ratio of 3:1; thus, the mutation is in a single nuclear gene and is recessive (Table S1 in Supporting Information).

Next, we used a map-based cloning approach to identify the gene responsible for the zir1 phenotype. The zir1 mutant was crossed with another A. thaliana ecotype, Landsberg erecta (Ler), which is similar to Col-0 in Fe-mediated Zn tolerance (data not shown). From the F2 progenies, 70 seedlings showing a hypersensitive phenotype were used for map-based cloning of the ZIR1 gene. Insertion–deletion polymorphism (INDEL) markers were used in the first phase of mapping, and the zir1 phenotype co-segregated with markers on the lower arm of chromosome 4. To refine the position of ZIR1, we further increased the population by 500. The markers CER432705 and CER428541 flanking the zir1 mutation region showed very low recombination frequencies. These markers defined a region of about 30 kb containing eight annotated genes on the lower arm of chromosome 4 of Arabidopsis.

We PCR-amplified and sequenced the genomic DNA (gDNA) sequences of these eight candidate genes from the zir1 mutant. We identified a single G-to-A nucleotide transition at position 2421 from the start codon of the gene At4g23100 (Figure 2) as compared with the wild type. This mutation replaced the amino acid glutamate with a lysine residue at position 385 of a 522 amino-acid-long protein (Figure S1). At4g23100 encodes a γ-glutamylcysteine synthetase (also designated GSH1) that catalyzes the first step in glutathione biosynthesis (May and Leaver, 1994).

Figure 2.

 Map-based cloning of zir1.
The position of ZIR1 is indicated in the genetic map of the lower arm of chromosome 4 of Arabidopsis. Insertion–deletion polymorphism (INDEL) and single nucleotide polymorphism (SNP) markers were used for fine mapping to locate the region of ZIR1. RF, recombination frequency. Markers T12H17 and T22A6 flanked the mutation region of about 230 kb that had a lower recombination frequency. Scores for Col-0 or Ler (Col-0/Ler) are indicated in brackets. The SNP markers CER432705 and CER428541 flanked the mutation region of about 30 kb that contains eight functional genes between At4g23070 and At4g23130 (http://www.arabidopsis.org). Gene sequencing identified the point mutation (G to A transition) in At4g23100, which encodes a γ-glutamylcysteine synthetase (γ-ECS/GSH1) with glutamate (E) to lysine (K) substitution at amino acid 385.

Point mutation in zir1 alters thiol accumulation

Because GSH1 is involved in the biosynthesis of glutathione, we determined the glutathione level in both the wild type and zir1 by HPLC using internal standards as reference and for quantification. Only about 15% (27.3 ± 5.7 nmol g−1 FW) of glutathione was present in zir1 as compared with the wild type (180 ± 8.2 nmol g−1 FW; Figure 3a,b). In addition, we analyzed other thiol compounds (cysteine and γ-glutamylcysteine, γ-EC) in the glutathione biosynthesis pathway. The cysteine level was 10 times higher in zir1 than in the wild type. γ-Glutamylcysteine levels were similar for the wild type and zir1 (16.3 ± 2.9 versus 15.7 ± 2.9 nmol g−1 FW, SD). Therefore, accumulation of both glutathione and cysteine is impaired in the zir1 mutant. Semi-quantitative PCR of transcript levels of GSH1 and GSH2 showed no dramatic change in the level of these genes (Figure 3c). Furthermore, we measured GSH1 enzyme activity in wild-type and zir1 seedlings, along with cad2-1 (cadmium-sensitive 2-1), a previously identified mutant allele of GSH1 (Cobbett et al., 1998). The GSH1 enzyme activities were 50% for zir1 (8.9 ± 0.9 nmol min−1 mg−1 protein) and 57% for cad2-1 (10.2 ± 0.9 nmol min−1 mg−1 protein) that of the wild-type level (17.8 ± 2.3 nmol min−1 mg−1 protein, Table S2). Therefore, the point mutation (G-to-A transition) in zir1 does not alter the expression level of GSH1 but reduces its activity, and the reduced glutathione content in zir1 is independent of the expression levels of GSH1 or GSH2.

Figure 3.

 Analysis of GSH1 and GSH2 transcripts and determination of thiol contents in wild-type and zir1 seedlings.
(a) Chromatogram of thiols extracted from shoot samples of 5-week-old, soil-grown plants of Col-0 (left) and zir1 (right). The extracts were labeled with monobromobimane, separated by reverse phase HPLC, and detected by fluorescence. In both chromatograms, arrows indicate, from left to right, the compounds identified as cysteine, γ-glutamylcysteine (γ-EC), and glutathione.
(b) Contents of glutathione and cysteine in Col-0 and zir1 seedlings as identified by HPLC and quantified by using peak areas of individual compounds against the corresponding standard curves. The data represent the mean ± SD of samples (Col-0, = 4; zir1, = 5) collected in two independent experiments. FW, fresh weight.
(c) Reverse transcriptase-PCR of GSH1 and GSH2 transcripts in shoots of 17-day-old Col-0 and zir1 seedlings grown on ½ MS. The number of amplification cycles is indicated. Actin 8 was an internal control.

Genetic complementation restores the wild-type level of glutathione in zir1 and Fe-mediated Zn tolerance

To confirm that the phenotype of zir1 was caused by the identified point mutation in GSH1, we used two genetic complementation approaches. First, the full-length gDNA of the wild-type copy of GSH1 including its 1270-bp promoter region was transformed into the zir1 mutant. As a second approach, the wild-type copy of the GSH1 coding sequence (CDS) was expressed under the control of the CaMV 35S promoter in the zir1 mutant. We selected four independent transgenic lines from each type of construct for quantification of thiol contents. Figure 4a shows two representative lines from each type of construct: each showed glutathione accumulated to the wild-type level. The cysteine levels in the complementation lines were lower than that of zir1 level (Figure S2).

Figure 4.

 Genetic complementation of zir1.
(a) Glutathione content in shoot samples of 2-week-old seedlings grown on ½ MS for Col-0, zir1, and complementation lines. Lines containing a genomic copy of GSH1 are marked as GSH1gDNA, and those overexpressing the coding sequence of GSH1 are labeled as 35S:GSH1CDS. Data represent the mean ± SD from at least three independent experiments. FW, fresh weight.
(b) The Fe-mediated Zn tolerance phenotype of complementation lines. Seven-day-old seedlings of Col-0, zir1, and complementation lines were treated with 250 μm Zn (Zn) or 250 μm Zn and 200 μm Fe (Zn + Fe) for 7 days. As a control, seedlings were grown on ½ MS. Seedlings of one representative line from each type of construct, 15-4 for GSH1gDNA and 15-3 for 35S:GSH1CDS, are shown. The bar represents 1 cm.

To associate the reduced glutathione level and loss of the Fe-mediated Zn tolerance phenotype with the mutation in GSH1, we tested the complemented lines for rescuing the Zn stress phenotype in the presence of additional Fe. Two-week-old seedlings of the complemented lines in each approach completely restored the Fe-mediated Zn tolerance (Figure 4b). Thus, the genetic complementation of the wild-type copy of GSH1 in zir1 seedlings could rescue the reduced glutathione level and the Fe-mediated Zn tolerance phenotype in the zir1 mutant. These data confirm that ZIR1 encodes for GSH1 and that the loss of glutathione production and Fe-mediated Zn tolerance phenotypes of zir1 seedlings are caused by the point mutation identified in GSH1.

Blocking glutathione biosynthesis in the wild type causes increased Zn sensitivity

To test whether the reduced glutathione accumulation in the wild type leads to loss of Fe-mediated Zn tolerance in Arabidopsis, we added the glutathione biosynthesis inhibitor buthionine sulfoximine (BSO) to the medium. Buthionine sulfoximine inhibits GSH1 and results in depletion of the cellular GSH level (Griffith and Meister, 1979; May and Leaver, 1993). As shown in Figure 5, wild-type seedlings lost the Fe-mediated Zn tolerance in the presence of BSO (100 μm). However, at the same concentration of BSO, zir1 did not show any significant difference in the Zn stress phenotype (Figure S3). The addition of cysteine (1 mm) in the medium did not affect the Fe-mediated Zn tolerance in the wild type (data not shown). Therefore, a substantial reduction in glutathione level can cause loss of Fe-mediated Zn tolerance.

Figure 5.

 Blocking the glutathione biosynthesis in Col-0 leads to a defect in Fe-mediated Zn tolerance.
Seven-day-old Col-0 seedlings were treated with excess Zn or excess Zn and Fe in the presence or absence of 100 μm of buthionine sulfoximine (BSO), a specific inhibitor of GSH1, for 7 days. The bar represents 1 cm.

Phytochelatin is not responsible for Fe-mediated Zn tolerance

Glutathione is the precursor of PCs. Phytochelatins were recently found to be involved in the Zn detoxification process in Arabidopsis (Tennstedt et al., 2009). To examine whether the Fe-mediated Zn tolerance effect is specifically due to glutathione or PCs, we tested the phenotype in the PC-deficient mutant cad1-3. Compared with the wild type and zir1, cad1-3 showed a slight reduction in shoot biomass under excess Zn but still possessed Fe-mediated Zn tolerance ability (Figure 6). Therefore, glutathione but not PC is largely responsible for the Fe-mediated Zn tolerance phenotype.

Figure 6.

 The phytochelatin-deficient mutant possesses Fe-mediated Zn tolerance.
(a) Seven-day-old Col-0, cad1-3, and zir1 seedlings were treated with excess Zn or excess Zn and Fe for 7 d and photographed. The bar represents 1 cm.
(b) The relative shoot fresh weight (FW) of individual lines described in (a). Data represent the mean ± SD from three independent experiments (> 12).

Analysis of other mutant alleles of GSH1

Previous studies reported four mutant alleles of GSH1 in Arabidopsis: cad2-1; (Cobbett et al., 1998), root meristemless1 (rml1; Vernoux et al., 2000), regulator of APX2 1-1 (rax 1-1; Ball et al., 2004), and phytoalexin-deficient mutant 2-1 (pad 2-1; Parisy et al., 2007). The amino acid changes that result in a point mutation in all five alleles (including zir1) of the GSH1 protein are shown in Figure S1.

Among these mutant alleles, cad2-1, the first identified allele of GSH1, was reported as Cd sensitive; the study demonstrated the role of glutathione in metal tolerance in Arabidopsis (Cobbett et al., 1998). The mutant allele rml1 shows severe growth defects and is unable to establish an active post-embryonic meristem, even under normal growth conditions, because it contains only 3% of the wild-type level of glutathione (Vernoux et al., 2000). This allele reveals the importance of glutathione in the general developmental processes in Arabidopsis. The mutant alleles rax1-1 and pad2-1, both reported later, were found to be sensitive to photooxidative stress and biotic stress, respectively (Ball et al., 2004; Parisy et al., 2007). Among the four alleles, rax1-1 contains the highest level of glutathione (40–50% of the wild-type level), then cad2-1 (40%), pad2-1 (22%), and rml1 (2.7%) (Vernoux et al., 2000; Ball et al., 2004; Parisy et al., 2007).

To correlate the phenotype with glutathione levels, we compared the glutathione levels and degree of the Fe-mediated Zn tolerance in two additional alleles, cad2-1 and pad2-1. We measured glutathione content in leaves of 5-week-old, soil-grown cad2-1 and pad2-1 mutant, wild-type, and zir1 mutant plants. Among the tested mutants, zir1 had the lowest glutathione level (15% of wild-type level), then pad2-1 (22%) and cad2-1 (39%) (Figure 7a). To test the degree of Fe-mediated Zn tolerance, 7-day-old GSH1 mutant and wild-type seedlings were treated with excess Zn alone or with excess Zn and Fe for 7 days before phenotype observation. The shoot biomass and root growth of both pad2-1 and cad2-1 mutants were comparable to that of the wild type under control and excess Zn conditions. The cad2-1 mutant still possessed the Fe-mediated Zn tolerance phenotype of the wild type. However, the pad2-1 mutant lost this phenotype, with analogy to zir1 (Figure 7b,c). Therefore, internal glutathione level is related to the Fe-mediated Zn tolerance phenotype, and a minimal level, between 22 and 39%, of wild-type glutathione may be required for Fe-mediated Zn tolerance.

Figure 7.

 Analysis of other GSH1 alleles for the Fe-mediated Zn tolerance.
(a) Glutathione content in the shoot samples of 5-week-old soil-grown plants of Col-0, zir1, pad2-1 and cad2-1. Data represent the mean ± SD of three to four samples collected in two separate experiments.
(b) Phenotype of Col-0, zir1, pad2-1, and cad2-1 seedlings grown under excess Zn or excess Zn and Fe for 7 days. The bar represents 1 cm.
(c) The relative shoot fresh weight (FW, upper panel) and root growth (lower panel) of Col-0, zir1, pad2-1, and cad2-1 seedlings were treated as described in (b). For shoots, data represent the mean ± SD from four independent experiments (= 7). For roots, data represent the mean ± SD from four to seven independent experiments (= 7).

Accumulation of glutathione under excess Zn and Fe conditions

Because the reduced level of glutathione impaired the Fe-mediated Zn tolerance, as observed in the zir1 and pad2-1 mutants (Figure 7b,c), glutathione may play an important role in the homeostasis of Zn or in the cross-homeostasis of Zn and Fe. To test whether the accumulation of glutathione is regulated by these conditions, we measured shoot glutathione content after treatment with high concentrations of Zn or Zn and Fe. The accumulation of glutathione in wild-type seedlings was increased by 2- and 1.6-fold with excess Zn and excess Zn and Fe conditions, respectively. However, glutathione synthesis was not induced by these treatments in pad2-1 or cad2-1 seedlings, although the glutathione levels were higher in the shoot of cad2-1 than pad2-1 seedlings (Figure 8a).

Figure 8.

 Relationship between the glutathione level and recovery of shoot Fe content.
(a) Seven-day-old seedlings of Col-0, pad2-1, and cad2-1 were treated with excess Zn or excess Zn and Fe for 3 days, and shoot samples were collected for glutathione measurement. Data represent mean ± SD from three independent experiments (= 10).
(b) Seven-day-old seedlings underwent treatments as indicated for 9 days, and shoot samples were harvested for Fe content analysis by inductively coupled plasma-optical emission spectrometry. Data represent the mean ± SD from three independent experiments. DW, dry weight; FW, fresh weight.

Recovery of shoot Fe contents in GSH1 mutants

From our previous study, we know that Zn stress reduces the shoot Fe levels in wild-type seedlings and that supplying excess Fe in the medium can restore these levels. To examine the recovery of Fe levels, we analyzed the shoot Fe concentration in pad2-1 and cad2-1 seedlings grown under conditions of excess Zn and Fe. The pattern of Fe accumulation in shoots of pad2-1 and cad2-1 mutants was similar to that of the wild type under control and excess Zn treatments (Figure 8b). Under excess Fe and excess Zn, the shoot Fe content was not restored to wild-type levels in pad2-1 and cad2-1 mutants. However, the recovery was related to the level of glutathione, which was higher in the cad2-1 than the pad2-1 mutant (Figure 8b). This finding agrees well with the Fe-mediated Zn tolerance phenotype (Figure 7). We analyzed Zn and Fe contents in zir1 seedlings grown only under control treatments because the mutant showed a severe phenotype with excess Zn and Fe. Under control treatments, shoot Fe content (55%) but not shoot Zn content was greatly reduced as compared with the wild type (Figure S4).

Discussion

zir1 is a mutant allele of GSH1

In this study, we isolated an EMS mutant, zir1, with an abolished Fe-mediated Zn tolerance phenotype (Figure 1). Map-based cloning and complementation studies indicated that ZIR1 encodes GSH1, which catalyzes the first step of glutathione biosynthesis (Figures 2–4). The mutation causes reduced GSH1 activity (Table S2). These data indicated that the single mutation causing an amino acid substitution of glutamate to lysine at position 385 reduced GSH1 activity.

In plants, glutathione plays an important role in biotic and abiotic stress tolerance. It is a key compound in the glutathione–ascorbate cycle that scavenges hydrogen peroxide (Foyer and Halliwell, 1976; Noctor and Foyer, 1998). It is also the precursor of PCs, glutathione oligomers that chelate non-essential heavy metals such as Cd (Ha et al., 1999). Overproduction of glutathione in Arabidopsis resulted in a slightly enhanced tolerance to various abiotic stresses, including metal stress such as exposure to Cd (Xiang et al., 2001). Glutathione biosynthesis is mediated through two ATP-dependent processes, one in which GSH1 (EC.6.3.2.2) catalyzes the formation of a peptide bond between the carboxyl group of glutamate and the amino group of cysteine, thus resulting in the formation of γ-EC. This step is the rate-limiting step in glutathione biosynthesis under conditions of increased demand for glutathione (Noctor and Foyer, 1998; Noctor et al., 1998). The second step comprises the ligation of a glycine residue to γ-EC by glutathione synthetase (GSH2, EC.6.3.2.3). Both GSH1 and GSH2 are encoded by single genes in Arabidopsis (GSH1, At4g23100; GSH2, At5g27380) (May and Leaver, 1994; Rawlins et al., 1995; Ullmann et al., 1996). GSH1 is exclusively located in plastids, whereas GSH2 is dually targeted to plastids and the cytosol (Wachter et al., 2005). Both glutathione biosynthesis enzymes are essential in plants. Knockout of GSH1 and GSH2 is lethal to embryos and seedlings, respectively (Cairns et al., 2006; Pasternak et al., 2008).

Mutants of four other previously reported GSH1 alleles vary in glutathione content, between 3% and 50% of the wild-type level. The allele with the lowest glutathione level, rml1, exhibited the most severe developmental defects (Vernoux et al., 2000). Two mutants with higher glutathione content, pad2-1 (22%) and cad2-1 (40%), grew normally under control conditions but were more sensitive to biotic and Cd stress, respectively (Cobbett et al., 1998; Parisy et al., 2007). The rax1-1 mutant had about 40–50% of the wild-type glutathione level and was identified by its constitutive expression of ascorbate peroxidase 2 (Ball et al., 2004). All these mutants had altered amino acid residues near the GSH1 active site, which could explain their decreased glutathione content (Hothorn et al., 2006). However, except for the rml1 mutant, the other mutants with glutathione deficiency were not affected in overall growth under normal conditions.

We found that the mutant allele of GSH1, zir1, contains only about 15% of the wild-type glutathione level under normal conditions. The zir1 mutant plants could flower normally and set seeds but grew slightly slower and were smaller than the wild type. Antisense lines of GSH1, which contain about 2–7% of the wild-type level of glutathione, were found to be smaller but could flower and set seeds normally (Xiang et al., 2001). The pad2-1 mutant has only a slightly higher glutathione content than zir1 but grows similar to the wild type under normal conditions (data not shown). Therefore, different lesions could have different effects on the regulatory and catalytic function of GSH1 and result in shuttle phenotypes between mutant alleles.

Fe-mediated Zn tolerance is associated with glutathione levels

Wild-type seedlings have an adequate level of glutathione and can overcome Zn stress in the presence of additional Fe. However, zir1, with a low glutathione level, cannot be rescued by Fe (Figure 1). Elevating the glutathione levels in the zir1 mutant by complementation with wild-type gDNA copy or overexpressing the CDS of GSH1 restored the Fe-mediated Zn tolerance phenotype (Figure 4). Therefore, the Zn sensitivity in the presence of Fe, as observed in zir1, is due to the reduced level of glutathione. Blocking glutathione biosynthesis by using BSO in the wild type led to loss of Fe-mediated Zn tolerance, which provides further evidence that the glutathione level is important for this phenotype (Figure 5).

To further support this notion, we examined the Fe-mediated Zn tolerance phenotype in the pad2-1 and cad2-1 mutants of GSH1, which contained 22% and 39% of the wild-type glutathione level, respectively (Figure 7a). The sensitivity to excess Zn in pad2-1 and cad2-1 mutants was similar to that of the wild type. This finding is consistent with previous observations in the cad2-1 mutant and supports that a moderate level of glutathione is sufficient for plants to cope with excess Zn stress (Hugouvieux et al., 2009). However, the pad2-1 mutant containing less glutathione loses the Fe-mediated ability to tolerate Zn, whereas the cad2-1 mutant still possesses the Fe-mediated Zn tolerance phenotype (Figure 7b,c). A moderate level of glutathione (22% of wild type) may be enough to cope with Zn stress similar to the wild type, but a higher glutathione level may be required for Fe-mediated Zn tolerance similar to the wild type (Figure 7b,c). Thus, glutathione is an essential factor for Fe-mediated Zn tolerance. An adequate level of GSH is required for Fe-mediated Zn tolerance. The degree of Fe-mediated tolerance is directly associated with the amount of glutathione.

Phytochelatins and GSH have distinct roles in metal tolerance

Glutathione is as a precursor in the synthesis of PCs mediated by PC synthase 1 (PCS1). Phytochelatins are required for tolerance to heavy metals, including Cd and As, in plants (Grill et al., 1987, 1989; Cobbett and Goldsbrough, 2002; Rea et al., 2004; Sung et al., 2009). The essential role of PCS1 (At5g44070) in heavy-metal homeostasis, particularly in response to Cd, was demonstrated by its identification as the gene affected by the cad1 mutation (Clemens et al., 1999; Ha et al., 1999; Vatamaniuk et al., 1999). Recently, PCs were found to be involved in the detoxification and accumulation of Zn in Arabidopsis (Tennstedt et al., 2009). A mutant deficient in PCs, cad1-3, treated with excess Zn and Fe still revealed the Fe-mediated Zn tolerance phenotype (Figure 6). Therefore, PCs are not responsible for the phenotype of Fe-mediated Zn tolerance. This result reveals the importance of glutathione specifically in Zn and Fe cross-homeostasis and Fe-mediated Zn tolerance in Arabidopsis.

Regulation and role of GSH in Fe-mediated Zn tolerance

The most documented function of glutathione is as an antioxidant participating in the ascorbate–glutathione cycle (Foyer and Halliwell, 1976; Noctor and Foyer, 1998; Noctor et al., 1998, 2002). Glutathione also plays an important role in redox signaling, particularly under biotic and abiotic stresses in plants. Among the various metal stresses tested, glutathione levels were found to be increased under Cd but not Cu or Se (IV) stress conditions in Arabidopsis (Hugouvieux et al., 2009; Sung et al., 2009). We found elevated accumulation of glutathione under excess Zn conditions (Figure 8). Zinc stress may be associated with induced glutathione production. The relationship between metal toxicity and cellular redox imbalance has been reported in many plant species. Excess Zn can induce oxidative damage of cellular components and change the status of antioxidative systems (Cuypers et al., 2001; Wojcik et al., 2006; Morina et al., 2010). Glutathione may have a protective role in preventing cellular oxidative damage by quenching the level of reactive oxygen species (ROS) in plants (Yadav, 2010). Metal hyperaccumulators, especially Zn, Pb, and Ni hyperaccumulators (e.g. Thlaspi caerulescens and Sedum alfredii), contain elevated levels of glutathione (Freeman et al., 2004; Sun et al., 2005; Kramer, 2010). Removal and reduction of ROS in oxidative damage, together with metal chelating function, were proposed to be specific characteristics of glutathione-mediated metal tolerance in heavy metal hyperaccumulators (Freeman et al., 2004; Freeman and Salt, 2007).

In A. thaliana, excess Zn can cause reduced shoot Fe content and induce genes related to Fe acquisition (Fukao et al., 2011; Shanmugam et al., 2011). Alternatively, excess Fe in the medium can elevate the normal shoot Fe level and rescue the Zn stress (Shanmugam et al., 2011). Excess Zn had a similar effect on reducing shoot Fe accumulation in the wild type and glutathione-deficient pad2-1 and cad2-1 mutants (Figure 8b). Interestingly, glutathione-deficient mutants are defective in restoring shoot Fe contents with excess Fe and excess Zn. In particular, pad2-1, which has lower glutathione levels than the cad2-1 mutant, was severely impaired in restoring Fe content in the shoot. Therefore, Fe-mediated Zn tolerance is clearly associated with restoring shoot Fe content. Moreover, this action is glutathione dependent.

Studies in yeast and mammalian cells reveal the importance of glutathione in Fe metabolism, including maturation of Fe–S protein and transport of dinitrosyl Fe complex (Sipos et al., 2002; Richardson and Lok, 2008; Kumar et al., 2011). Interestingly, a recent study in yeast suggested that rather than being involved in thiol-redox control, glutathione plays an important role in maturation of Fe–S cluster and Fe homeostasis (Kumar et al., 2011). In Arabidopsis, although the relationship between glutathione levels and Fe accumulation or homeostasis remains obscure, the reinstatement of Fe content in the shoot is related to glutathione levels (Figure 8). Fe-mediated Zn tolerance may be mainly protected by facilitating Fe shortage caused by Zn stress through many aspects (Shanmugam et al., 2011). In this process, the presence of glutathione is an essential factor to facilitate reinstatement of Fe in this cross protection.

Glutathione and Fe accumulation

The induction of glutathione under excess Zn could be direct or indirect. A strategy-I Fe deficiency response induced by excess Zn, including the induction of FIT and FRO2 expression and accumulation of IRT1 (Walker and Connolly, 2008), has been reported recently (Fukao et al., 2011; Shanmugam et al., 2011). Of note, under Fe-limited conditions, zir1 showed strategy-I Fe deficiency responses similar to those of the wild type, but the accumulation of Fe in shoot tissues was reduced (K.C. Yeh, unpublished data). This finding implies that glutathione is mainly required for the accumulation of Fe in shoot tissues.

The induction of putative Fe deficiency signals, NO and S-nitrosoglutathione (GSNO), under conditions of excess Zn and Fe deficiency was reported (Graziano et al., 2002; Xu et al., 2010). Glutathione is a key component in the formation of GSNO and dinitrosyl–Fe complexes (DNIC) (Richardson and Lok, 2008; Ramirez et al., 2010). A DNIC was proposed to have a role in translocation of Fe from roots to shoots and may improve the availability of Fe in monocot plants (Graziano et al., 2002; Ramirez et al., 2008, 2011). These lines of research are worth further exploration in Arabidopsis or dicot plants.

In summary, the identification of zir1 as GSH1 and the regulation of glutathione under conditions of excess Zn and Fe reveal the importance of glutathione in Fe-mediated Zn tolerance in Arabidopsis. Glutathione may have several roles in heavy-metal responses in plants. Our data suggest that glutathione has a more specific role than PCs in Arabidopsis in Zn and Fe cross-homeostasis.

Experimental procedures

Plant materials and growth conditions

Wild-type A. thaliana (ecotype Columbia 0), GSH1 mutant alleles (pad2-1, cad2-1) and PC-deficient mutant (cad1-3) were used in this study. The EMS mutated seeds were obtained from Lehle Seeds (http://www.lehleseeds.com). Seeds were surface-sterilized with 70% ethanol for 5 min, and then treated with 1.2% bleach containing 0.02% SDS for 15 min, rinsed five times with sterilized water, and kept in the dark at 4°C for 3–5 days for seed stratification. Washed seeds were grown on half-strength MS medium (1/2 MS salt; Sigma-Aldrich, http://www.sigmaaldrich.com/) containing 1% sucrose (J. T. Baker, http://www.jtbaker.com), 0.5 g L−1 of 2-morpholinoethanesulfonic acid (MES; J. T. Baker), and 0.7% agar Type A (Sigma-Aldrich) at pH 5.7 for 5–7 days. Subsequently, seedlings were transferred to excess Zn (ZnSO4) and excess Fe (Fe citrate) treatments as indicated in the figure legends. Soil-grown plants were obtained by sowing seeds in pots containing a mixture of organic substrate, vermiculite, and mica shoot (9:1:1). Both medium- and soil-grown plants were subjected to light intensity 70 μmol photons m−2 sec−1 under a 16-h light/8-h dark cycle at 22°C.

Map-based cloning of zir1

A previously described methodology (Jander et al., 2002) was adopted for mapping of zir1. The F2 population obtained from a cross between Arabidopsis ecotype Ler and zir1 was screened for the zir1 phenotype (lack of Fe-mediated Zn tolerance). Seedlings possessing zir1 phenotypes were analyzed for the segregation of polymorphisms by using INDEL markers (Berendzen et al., 2005) for first phase of mapping. For fine mapping, single nucleotide polymorphism (SNP) markers were used with the databases for the Monsanto Arabidopsis Polymorphism and Ler Sequence Collections (http://www.arabidopsis.org). Amplification and sequence determination of full-length gDNA of eight annotated genes flanked by markers CER432705 and CER428541 (Figure 2) from zir1 involved specific primers derived from the wild-type sequence. Primer design, PCR amplification, and sequencing were conducted by Genomics BioSci & Tech (http://www.genomics.com.tw). GSH1 mutant alleles including zir1 were analyzed for point mutations by sequencing PCR products obtained with the primer pair 5′-CCA ACT ATC TAC GGA TGT CAC AGG-3′ and 5′-TCC AGT CAG CTG TCA GAT CC-3′.

Isolation of RNA and RT-PCR

Total RNA was extracted from wild-type and zir1 seedlings by use of the RNeasy Plant Mini Kit (Qiagen, http://www.qiagen.com/). Subsequently, 2 μg RNA was treated with RQ1 RNase-Free DNase (Promega, http://www.promega.com/), and the reaction buffer was replaced with 5× First-strand RT Buffer (Invitrogen, http://www.invitrogen.com/). The complementary DNA (cDNA) was synthesized by use of SuperScript® III Reverse Transcriptase (Invitrogen). Transcript-specific primers were designed and synthesized for amplification of 1500–1600 bp of GSH1 and GSH2 genes. Primers for the GSH1 fragment were 5′-CTC CGT CAA GCT TGA CGA ATT TCA-3′ (forward) and 5′-CAG ATA GCA TAA ACT CAC ACC CAA-3′ (reverse), while for GSH2 we used 5′-ATG GGC AGT GGC TGC TCT TC-3′ (forward) and 5′-GCT GTC CAA GAC TCC AAA ACC-3′ (reverse). To detect Actin8 expression (internal control), we used the primer pair 5′-CCA CAT GCT ATC CTC CGT CT-3′ (forward) and 5′-CTG GAA AGT GCT GAG GGA AG-3′ (reverse). Amplification of both GSH1 and GSH2 involved the conditions 30 sec at 94°C, 30 sec at 55°C, and 90 sec at 72°C. Amplification of Actin8 was similar, except for an extension time of 60 sec.

Genetic complementation of zir1

Both the GSH1 genomic DNA (gDNA), including its promoter region (1270 bp), and overexpression CDS of wild-type GSH1 were transformed into the zir1 mutant background. For PCR amplification of gDNA, we used the primers forward, BamH1-5′-ATT TGG ATC CAC ATA CAA CAT AGA AGT GCA CAA-3′ and reverse Cla1-5′-ATT TAT CGA TGT GTG TGG ACT GTG GAG ACT-3′ cloned into pCAMBIA1305.1. For PCR amplification of CDS, we used the primers forward NcoI-5′-AAA CCA TGG CGC TCT TGT CTC AAG-3′ and reverse Pml-5′-CAC GTG TTA GTA CAG CAG CTC TTC GAA-3′ cloned into pCAMBIA 1305.1. Agrobacterium tumefaciens strain GV3101, which harbored 35S:AtGSH1/pCAMBIA1305.1 and AtGSH1gDNA/pCAMBIA1305.1, was used for in planta transformation.

Extraction and determination of thiols

Total thiol levels in leaves were determined by use of reverse-phase HPLC coupled with fluorescence detection after reduction with sodium borohydride (NaBH4) and derivatization with monobromobimane (mBBr) as described (Arisi et al., 1997; Cobbett et al., 1998) with minor modification. Approximately 100 mg of frozen ground tissue was extracted with 1 ml ice-cold extraction buffer [100 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)-HCl containing 10 mm MgCl2 and 1 mm EDTA, pH 7.5] and centrifuged at 16,000g, at 4°C for 30 min. Thiols in supernatant were reduced with 40 mm NaBH4 in 0.2 mmN-cyclohexyl-2-aminoethanesulfonic acid (CHES), in a pH 9.3 environment. Derivatization further involved 15 mm mBBr and was stopped by adding 5% acetic acid. Samples were then centrifuged (at 16 000 g) and diluted two-fold with 2.5% acetic acid and microfiltered before injection into HPLC. The HPLC analysis involved the Agilent 1100 Series system (Agilent Technologies, http://www.agilent.com/) with a C18 column (Eclipse XDB, 5 μm × 150 mm × 4.6 mm; Agilent Technologies) and elution buffers: buffer A, 10% (v/v) methanol, 0.25% (v/v) acetic acid, pH 4.3, and buffer B, 90% (v/v) methanol, 0.25% (v/v) acetic acid, pH 4.3. The linear gradient of buffer A from 100 to 90% was run for 10 min at a flow rate of 1.5 ml min−1. The mBBr conjugates were detected at excitation/emission wavelengths of 380 and 480 nm, respectively. The injection volume was 100 μl. Standard mixtures (cysteine, γ-EC and GSH) treated exactly the same as sample supernatants were used for peak identification and thiol quantifications. The peak areas from eight different concentrations of each standard solution were used for calibration curves, and the r2 values were 0.99 or higher for all thiol compounds. Thiols in plant samples were quantified by use of peak areas against the standard curve.

Analysis of GSH1 activity

The activity of GSH1 enzyme was determined as described (Cobbett et al., 1998). Briefly, 10-day-old seedlings were frozen with liquid nitrogen and ground in extraction buffer containing 100 mm TRIS-HCl, 10 mm MgCl2, 5 mm EDTA, pH 8, at a tissue to buffer ratio of 1:4 (w/v). Tissue extracts were centrifuged at the top speed of the microcentrifuge (16 000 g) at 4°C for 10 min. The supernatant was desalted by use of a Sephadex G-25 Column (Illustra® MicroSpin® G-25, GE, http://www.gehealthcare.com/illustra). To measure the GSH1 enzyme activity, 180 μl of desalted extract was incubated at 37°C for 45 min in a total volume of 360 μl reaction mixture containing 100 mm HEPES (pH 8), 40 mm MgCl2, 30 mm glutamate, 0.8 mm cysteine, 0.4 mm DTT, 5 mm ATP, 5 mm phosphoenol pyruvate and 5 U pyruvate kinase. At the end of the reaction, 50 μl of the reaction mixture was added to 200 μl of 50 mm CHES buffer (pH 9.2) and immediately underwent derivatization for 15 min with 20 μl of 15 mm mBBr and stopped by adding 400 μl of 10% acetic acid. Samples were centrifuged at 16 000 g for 10 min, and supernatant was analyzed in HPLC; the γ-EC amount was quantified as described in the previous section. Protein concentration was measured by use of the BCA protein assay kit (Bio-Rad, http://www.bio-rad.com/).

Measurement of shoot Fe content

The Fe content was analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES; OPTIMA 5300; Perkin-Elmer, http://www.perkinelmer.com/) as described (Shanmugam et al., 2011). Briefly, shoot samples were washed with CaCl2 and deionized water and dried at 70°C for 3 days before acid digestion by use of the MarsXpress microwave digestion system (CEM, http://www.cem.com/).

Statistical analysis

Student’s t-test was used to analyze differences between two groups. < 0.05 was considered statistically significant.

Acknowledgements

This work was supported by grants from DPIAB 098S0050067-AA and Academia Sinica. We thank Dr Christopher Cobbett for providing cad1-3 and cad2-1 seeds. Technical support from Chong-Cheong Lai and I-Chien Tang is gratefully acknowledged. We thank Dr Heiko Kuhn for manuscript editing.

Ancillary