Metallothionein deficiency impacts copper accumulation and redistribution in leaves and seeds of Arabidopsis

Authors


Summary

  • Most angiosperm genomes contain several genes encoding metallothionein (MT) proteins that can bind metals including copper (Cu) and zinc (Zn). Metallothionein genes are highly expressed under various conditions but there is limited information about their function. We have studied Arabidopsis mutants that are deficient in multiple MTs to learn about the functions of MTs in plants.
  • T-DNA insertions were identified in four of the five Arabidopsis MT genes expressed in vegetative tissues. These were crossed to produce plants deficient in four MTs (mt1a/mt2a/mt2b/mt3).
  • The concentration of Cu was lower in seeds but higher in old leaves of the quad-MT mutant compared to wild-type plants. Experiments with stable isotopes showed that Cu in seeds came from two sources: directly from roots and via remobilization from other organs. Mobilization of Cu out of senescing leaves was disrupted in MT-deficient plants. Tolerance to Cu, Zn and paraquat was unaffected by MT deficiency but these plants were slightly more sensitive to cadmium (Cd). The quad-MT mutant showed no change in resistance to a number of microbial pathogens, or in the progression of leaf senescence.
  • Although these MTs are not required to complete the plant's life cycle, MTs are important for Cu homeostasis and distribution in Arabidopsis.

Introduction

Transition metals such as copper (Cu), iron (Fe), manganese (Mn) and zinc (Zn) are essential for all organisms because they provide critical functionality to a variety of proteins. Cu is required for many essential processes in plants including photosynthesis, respiration, ethylene perception, metabolism of reactive oxygen species and cell walls (Penarrubia et al., 2009). Some of the proteins that require Cu include: the electron carrier plastocyanin; cytochrome c oxidase in mitochondria; cytosolic, stromal and peroxisomal Cu/Zn superoxide dismutases; ethylene receptors; apoplastic oxidases; and laccases (Burkhead et al., 2009).

The availability of Cu in the environment, requirement for Cu in plant tissues during development and reactivity of Cu in cells, have led to the evolution of a regulated system that controls Cu uptake and distribution in plants. Components of this system include membrane transporters, transcription factors and miRNAs that modulate gene expression, and chaperones that deliver Cu to specific targets (Burkhead et al., 2009; Penarrubia et al., 2009). However, little is known about the function of the metal binding proteins known as metallothioneins (MTs) in plants.

MTs are low molecular weight (4–8 kDa), cysteine-rich proteins that bind metals such as Zn, Cu or Cd (Leszczyszyn et al., 2013). Most angiosperm genomes encode a small number of MT proteins that can be divided into four types on the basis of conserved cysteine residues (Freisinger, 2011; Leszczyszyn et al., 2013). Arabidopsis has a seven-member gene family encoding all four plant MT types (Supporting Information Fig. S1; Cobbett & Goldsbrough, 2002). A study of Arabidopsis MT gene expression, including GUS reporter genes driven by AtMT promoters, has provided detailed information about tissue-specific expression (Guo et al., 2003). The promoters of AtMT1a (At1g07600) and AtMT2b (At5g02380) are active in the phloem and are copper-inducible; the promoters of AtMT2a (At3g09390) and AtMT3 (At3g15353) are active predominantly in mesophyll cells and are also induced by Cu in young leaves and root tips. Arabidopsis MT1a, MT2a, MT2b and MT3 are also highly expressed during leaf senescence and induced by Cu in trichomes.

Numerous studies have described the expression of plant MT genes in specific tissues and in response to various biotic and abiotic factors. However, the functions of plant MTs are still poorly understood. The conserved arrangements of cysteine residues in the four plant MT types and their tissue-specific expression suggest that MTs may have a number of distinct functions (Freisinger, 2011; Leszczyszyn et al., 2013). MTs have been proposed to participate in a variety of processes including metal ion homeostasis and tolerance (Zhou & Goldsbrough, 1994; Murphy & Taiz, 1995a,b; van Hoof et al., 2001; Cobbett & Goldsbrough, 2002; Mengoni et al., 2003; Roosens et al., 2004; Zimeri et al., 2005); oxidative stress protection (Akashi et al., 2004; Wong et al., 2004); root development and seed germination (Yuan et al., 2008); pathogen defense signaling (Wong et al., 2004); and the senescence program (Butt et al., 1998; Garcia-Hernandez et al., 1998; Mira et al., 2001; Gepstein et al., 2003; Guo et al., 2003). A number of studies suggest that plant MTs may participate in metal ion homeostasis, especially for Cu, during both vegetative growth and senescence. In Arabidopsis, rice and the metal hyperaccumulator Noccaea caerulescens MT RNA expression was strongly induced by Cu treatment, and to a lesser degree by Cd and Zn (Zhou & Goldsbrough, 1994; Hsieh et al., 1995; Guo et al., 2003; Sancenon et al., 2004). Copper tolerance of some Arabidopsis ecotypes correlated with RNA expression of a type 2 MT gene in roots (Murphy & Taiz, 1995a,b). Copper-tolerant populations of two Silene species showed increased expression of a type 2 MT as a result of gene duplication (van Hoof et al., 2001; Mengoni et al., 2003). Guo et al. (2008) showed that Arabidopsis mutants lacking MT1a accumulated less Cu in roots compared to wild-type plants when exposed to high concentrations of CuSO4.

During senescence some nutrients are transferred from older tissues to growing organs of the plant. Nutrients can be mobilized out of Arabidopsis leaves during senescence (Himelblau & Amasino, 2001; Colangelo & Guerinot, 2006; Waters & Grusak, 2008). RNA expression of MT genes increases dramatically during leaf senescence in a number of species (Hsieh et al., 1995; Foley et al., 1997; Chen et al., 2003; Kohler et al., 2004). In Arabidopsis, RNA levels of all MT genes expressed in vegetative tissues are elevated in senescing leaves (Butt et al., 1998; Guo et al., 2003). These observations suggest a role for MTs in the senescence program.

MTs have also been suggested to protect cells against oxidative stress. Recombinant watermelon MT protein can detoxify hydroxyl radicals and suppress hydroxyl radical-catalyzed degradation of DNA in vitro (Akashi et al., 2004). Wong et al. (2004) proposed that MTs participate in pathogen defense signaling as scavengers of reactive oxygen species (ROS). Recombinant OsMT2b protein was shown to scavenge superoxide- and hydroxyl-radicals in vitro. Transgenic rice plants overexpressing OsMT2b showed increased susceptibility to bacterial blight and rice blast diseases, and reduced production of elicitor-induced hydrogen peroxide compared to wild-type plants.

In order to examine the functions of MT genes in plants, mutants containing T-DNA insertions in several Arabidopsis MT genes were identified. Arabidopsis plants carrying insertions in four of the five MT genes expressed in vegetative tissues, referred to as the quad-MT mutant (mt1a-2/mt2a-1/mt2b-1/mt3-1), were used in a variety of experiments. MT-deficient plants developed normally under standard conditions and showed few changes in tolerance to most biotic and abiotic stresses tested. However, the concentration of Cu was higher in leaves and lower in seeds of quad-MT mutants compared to wild-type plants. We conclude that MTs play an important role in Cu homeostasis in Arabidopsis, specifically the remobilization of Cu from senescing leaves.

Materials and Methods

Plant material

Arabidopsis lines with T-DNA insertions in MT1a (mt1a-2), MT2a (mt2a-1) and MT2b (mt2b-1) were identified using the SIGNAL database (http://signal.salk.edu/cgi-bin/tdnaexpress) and obtained from ABRC (Table 1). The T-DNA insertion for MT3 was produced in the Wassilewskija (Ws) ecotype, identified in the FLAGdb/FST database (http://genoplante-info.infobiogen.fr) and ordered from INRA (Versailles, France). Fig. 1(a) shows the T-DNA insertion in each MT gene. Plants homozygous for T-DNA insertions were identified by PCR using primer pairs targeting both the wild-type (WT) and T-DNA insertion alleles. Primer sequences are listed in Supporting Information Table S1. Crosses performed to develop the MT-deficient mutants are given in Table S2. To remove additional T-DNA sequences in the genome, single mutants were backcrossed to the Col-0 wild-type (Table 1). The quad-MT mutant (mt1a-2/mt2a-1/mt2b-1/mt3-1) was backcrossed three additional times to the WT (Col-0).

Table 1. Arabidopsis T-DNA insertion lines used in this study
GeneLocusAlleleAccession numberInsertion siteaBCb
  1. a

    The insertion site is given relative to the translational start site: −, upstream, in the promoter; +, downstream from the start codon.

  2. b

    BC, number of backcrosses made to the wild-type (Col-0).

MT1a At1g07600 mt1a-2 Salk_76355+193
MT2a At3g09390 mt2a-1 Salk_59712−1382
MT2b At5g02380 mt2b-1 Salk_37601+183
MT3 At3g15353 mt3-1 FLAG_354B05+3423
Figure 1.

Analysis of RNA expression from insertion mutant alleles of Arabidopsis metallothionein (MT) genes. (a) The diagram shows the positions of T-DNA insertions in Arabidopsis MT genes. The insertions are indicated relative to the translational start site. The black and gray boxes indicate regions that are transcribed into mRNA; black boxes, indicate the protein coding sequence, gray boxes represent untranslated regions. (b) RT-PCR analysis of MT gene expression in the T-DNA insertion lines. The number of PCR cycles were as follows: 25 cycles for MT1; 36 cycles for MT1ac, MT1c, MT2b, MT3, and ACT2/8; 31 cycles for MT2a. The second panel in (b), labeled MT1ac-BglII, shows the MT1a and MT1c amplification products after digestion with BglII. The MT1c RT-PCR product contains a site for BglII resulting in a DNA fragment that is 32 bp smaller than the MT1a product. ACT2/8 expression was used as positive control. (c) Quantitative RT-PCR analysis of MT1a and MT1c RNA expression. cDNA made from leaf RNA was amplified with primers that targeted both MT1a and MT1c (AtMT1acF, AtMT1acR) or with primers that amplified only MT1c RNA (AtMT1cF, AtMT1cR). Transcript levels were normalized to ACT2 and are expressed as a percentage of MT1a+MT1c in wild-type plants. WT, wild-type, closed bars; quad, quad-MT mutant, open bars.

Growth of Arabidopsis plants

Arabidopsis seeds were surface sterilized with 30% (v/v) household bleach containing a drop of Tween 20 for 8 min followed by three washes with sterile water. Seeds were stratified at 4°C for 2–3 d in the dark and germinated on 100 × 100 × 15 mm square Petri dishes (Simport, Beloeil, Canada) containing half-strength MS salts supplemented with 0.05% (w/v) MES and 1% (w/v) sucrose, solidified with 1% (w/v) agar. Plants were grown under fluorescent lights with a 16 h photoperiod regime (light intensity 60–80 μmol m−2 s−1) at 24 ± 2°C. To measure tolerance to various stress conditions, seeds were germinated vertically on agar medium containing varying concentrations of CuSO4, ZnSO4, CdSO4 or paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride; Sigma). Plants were grown for 8 d and plates were imaged using a scanner at a resolution of 300 dpi. Root length was determined using ImageJ software (http://rsbweb.nih.gov/ij/). To grow Arabidopsis plants in a hydroponic system, seeds were sown on top of a rockwool (Grodan, Milton, ON, Canada) plug inserted in a 1.5-ml microcentrifuge tube soaked in hydroponic solution. The bottom of the tube was cut off so the rockwool was in contact with the solution. The hydroponic medium contained: 625 μM Ca(NO3)2·4H2O, 312.5 μM K2SO4, 250 μM MgSO4·7H2O, 62.5 μM KH2PO4, 32.5 nM KCl, 562.5 nM MnSO4·H2O, 3.125 μM H3BO3, 475 nM ZnSO4·7H2O, 62.5 nM CuSO4·5H2O, 25 nM Na2MoO4·2H2O, 6.875 μg l−1 Fe-Sequestrene (Sprint® 330; Becker Underwood, Ames, IA, USA). The microcentrifuge tubes were inserted through holes in a 1.2 cm-thick Styrofoam raft floating in a plastic box containing 500 ml of hydroponic solution. Two weeks after germination seedlings were transferred to a larger plastic box (35 × 25 × 9 cm) containing 3 l of solution that was aerated using an aquarium pump. The solution was replaced weekly. Plants were grown at 24 ± 2°C. Photoperiod and light intensity varied between experiments and details are provided in the figure legends.

RNA expression analysis

Total RNA was extracted using the TRIzol® Reagent (Invitrogen) and treated with DNAse using a Turbo DNA-free™ Kit (Applied Biosystems/Ambion). The concentration, integrity and purity of DNAse-treated total RNA were determined with a NanoDrop™ 1000 spectrophotometer (Thermo Scientific, Waltham, MA, USA) and by formaldehyde agarose gel electrophoresis. RNA samples were reverse transcribed using 1 μg of DNAse-treated total RNA and 0.5 μl of AMV-reverse transcriptase (Promega) according to the manufacturer's instructions. The cDNA was diluted 10 times and 2 μl was used as template in a 15-μl PCR reaction with the following thermal profile: 3 min at 95°C, 36 cycles (30 s at 95°C, 30 s at 54°C, 60 s at 72°C), and 10 min at 72°C. The numbers of PCR cycles used are described below or in Fig. 1. Table S1 provides information on the primers used in this study. The level of expression of ACT2/8 (At3g18780/At1g49240) was used to normalize RNA expression across different samples.

For quantitative real-time PCR, the cDNA was diluted 50 times and 2 μl was used as a template in a 15 μl real-time PCR reaction. qRT-PCR reactions were performed in a Mx3000P™ qPCR System (Agilent Technologies, Santa Clara, CA, USA) using SYBR Green to monitor cDNA amplification (Brilliant SYBR Green QPCR Master Mix). The following thermal profile was used: 10 min at 95°C for 1 cycle, 40 cycles (30 s at 95°C, 30 s at 60°C, 30 s at 72°C), and 1 cycle (1 min at 95°C, 30 s at 60°C, 30 s at 95°C). To distinguish between the transcripts of MT1a and MT1c, primers that amplify both MT1a and MT1c (AtMT1acF, AtMT1acR) and specific primers that target only MT1c (AtMT1cF, AtMT1cR) were employed. The transcript levels of MT1a+MT1c and MT1c were normalized with ACT2 (ATC2F, ATC2R).

Experiments with stable Cu isotopes in hydroponics

65CuSO4·5H2O (97.9% pure) was obtained from Isoflex (San Francisco, CA, USA). Wild-type and MT-deficient Arabidopsis plants were germinated in a hydroponic solution where most of the Cu was provided in the form of 65CuSO4. Plants were grown in 65Cu-enriched medium until they started to bolt, typically c. 4 wk. At this stage the solution was replaced with normal medium containing 69% 63Cu and 31% 65Cu. Plants were grown until they completed their life cycle (approx. eight more weeks) when seeds and leaves were harvested.

Metal ion analysis of plant material

Plant tissues were harvested and dried at 65°C for 2 d. Before drying, roots were washed for 10 min each in deionized water, citrate buffer (20 mM sodium citrate, 1 mM EDTA, pH 4.2), 25 mM CaCl2 (pH 5.0) and a final wash in deionized water. Tissue samples were digested in 1 ml of concentrated nitric acid (AR Select; Avantor Performance Materials, Center Valley, PA, USA) at 115°C for 12 h. Digested samples were diluted to 4 ml with deionized water and analyzed by graphite furnace-atomic absorption spectrometry (GF-AAS) (AA-6800; Shimadzu, Kyoto, Japan) or inductively coupled plasma-mass spectrometry (ICP-MS) (ELAN DRC-e; PerkinElmer, Waltham, MA, USA).

In order to determine the concentrations of Cu isotopes in samples, counts of 63Cu and 65Cu were obtained from the ICP-MS instrument. The concentrations of 63Cu and 65Cu in the samples were then calculated using standard curves prepared for 63Cu and 65Cu using a standard that contained the natural abundance of these Cu isotopes.

Induction of leaf senescence

Arabidopsis plants were grown for 4 wk in a hydroponic solution as described. Fully expanded green leaves of similar size were excised using a sharp razor blade and placed on top of a moist filter paper in a 100 × 15 mm Petri dish. Leaves were incubated in the dark at 24 ± 2°C for 5 d. Each day, leaves were weighed, frozen in liquid nitrogen and stored at −80°C. One millilitre of 80% (v/v) acetone was added to a microcentrifuge tube containing a single detached leaf and incubated overnight at room temperature in the dark. The total chlorophyll content was determined by measuring absorbance at 663.2 and 646.8 nm using a Shimadzu UV-1650PC spectrophotometer (Shimadzu, Kyoto, Japan). Chlorophyll concentration was calculated as described by Lichtenthaler (1987).

Photooxidative damage assay

A detached leaf assay was performed to determine the sensitivity of MT-deficient mutant plants to photooxidative damage induced by paraquat (Mengiste et al., 2003). For this assay, WT (Col-0) and MT-deficient mutant Arabidopsis plants were grown in a potting mix (Super Fine Germinating Mix; Conrad Fafard, Agawam, MA, USA) in a growth chamber with fluorescent light on a 12 h photoperiod regime (200 μmol m−2 s−1) at 22 ± 4°C with 60% RH. Plants were watered by subirrigation. Leaves were removed and placed on half-strength MS agar medium without sucrose. A single 2-μl drop of 100 μM paraquat or water was placed on each leaf. Leaves were incubated under a transparent cover to maintain high humidity for 2 d before photographs were taken.

Plant disease assays

Wild-type (Col-0) and MT-deficient mutant Arabidopsis plants were grown as described above. Botrytis cinerea strain BO5-10 was grown as described previously (AbuQamar et al., 2006). The Botrytis spore suspension was adjusted to 3 × 105 spores ml−1 and spray-inoculated onto Arabidopsis plants using a Preval sprayer (Valve Corp., Yonkers, NY, USA). Control plants were sprayed with 1% Sabouraud maltose broth buffer. Plants were kept under a transparent cover to maintain high humidity in a growth chamber with 21°C : 18°C, day : night temperatures and a 12 h photoperiod regime (200 μmol m−2 s−1). Alternaria brassicicola strain MUCL 20297 was grown as described previously (Mengiste et al., 2003). Single leaf inoculations were performed by placing a 5-μl drop of 5 × 105 spores ml−1 spore suspension of A. brassicicola on individual leaves of soil-grown plants. After inoculation, plants were kept under a transparent cover to maintain high humidity and transferred to a growth chamber with 24°C temperature and a 12 h photoperiod regime (200 μmol m−2 s−1). Photographs of representative plants were taken 6 d after inoculation (DAI).

Bacterial disease assays were performed as described previously (Mengiste et al., 2003). Fully expanded leaves of 4 wk-old soil-grown Arabidopsis plants were infiltrated with a suspension (OD600 = 0.001 in 10 mM MgCl2) of Pseudomonas syringae pv tomato DC3000. Photographs of representative plants were taken 6 DAI.

Statistical analysis

Statistical analyses of the data were performed using the function Student's t and/or ANOVA – Tukey's HSD (Honestly Significant Difference) tests from the statistical analysis software KaleidaGraph® – Data Analysis and Graphic Presentation for Business, Science and Engineering; v3.6 (Synergy Software, Paramus, NJ, USA).

Results

MT gene expression in T-DNA insertion mutants

Arabidopsis mutants with T-DNA insertions in MT1a (mt1a-2), MT2a (mt2a-1), MT2b (mt2b-1), and MT3 (mt3-1) were used to study the functions of MTs in plants (Table 1, Fig. 1a). No dramatic phenotypes were observed in mutants that lacked a single MT. A series of crosses was performed (Table S2) to develop plants deficient in the expression of four MT genes that are highly expressed in vegetative tissues, referred to as the quad-MT mutant (mt1a-2/mt2a-1/mt2b-1/mt3-1).

The RNA expression of MT genes in single gene mutants and the quad-MT mutant was analyzed by RT-PCR (Fig. 1b). RNA transcripts were not detected from the mt2b-1 and mt3-1 alleles. A low level of RNA expression was observed for mt2a-1, likely because the T-DNA insertion is in the promoter of this gene. In both the mt1a-2 mutant and the quad-MT mutant a PCR product was amplified using primers for the MT1a gene. Further experiments showed that this came from MT1c RNA, which shares 94% nucleotide sequence identity with MT1a. The 3′ UTR of MT1c contains a BglII restriction site that could distinguish between MT1a and MT1c (Fig. S2). After digestion with BglII the RT-PCR product in the mt1a-2 and quad-MT mutants was smaller than in the wild-type (WT) or the other mutants (Fig. 1b). These results indicate that essentially all of the MT1 RNA in the mt1a-2 and quad-MT mutants is from MT1c and that the mt1a-2 allele is silent. Additional quantitative real-time PCR experiments were performed to examine the expression of MT1a and MT1c RNAs. Using primers that amplified both transcripts, the combined abundance of both RNAs in the quad-MT mutant was only 9% of the level in WT, and all this RNA came from MT1c (Fig. 1c). The expression of MT1c RNA was c. four-fold higher in the quad-MT mutant compared to WT, perhaps to compensate for lack of other MTs. Overall, these results demonstrate that mt1a-2, mt2b-1 and mt3-1 are null alleles that produce no detectable RNA. However, there is a low level of RNA expression from MT2a and MT1c in the quad-MT mutant.

Tolerance of MT-deficient plants to abiotic stresses

MTs can bind metal ions such as Cu, Zn and Cd, and can also scavenge reactive oxygen species. Several studies have proposed that MTs in plants contribute to both metal tolerance and oxidative stress tolerance. To examine whether MT deficiency affects metal tolerance in Arabidopsis, WT and quad-MT mutants were germinated on medium containing increasing concentrations of CuSO4, CdSO4 or ZnSO4. Root growth of quad-MT mutant seedlings was no more sensitive to CuSO4 or ZnSO4 than in WT seedlings (Fig. 2a,b). Both genotypes were equally inhibited by higher concentrations of Cu or Zn. Root growth of the quad-MT mutant was, however, somewhat more sensitive than WT to Cd (Fig. 2c).

Figure 2.

Impact of metallothionein (MT)-deficiency on seedling growth in the presence of copper (Cu), zinc (Zn) or cadmium (Cd). Wild-type (closed circles) and quad-MT mutant (open circles) Arabidopsis seedlings were grown for 8 d on MS medium containing increasing concentrations of CuSO4 (a), ZnSO4 (b) or CdSO4 (c). The control treatment was standard MS medium containing 0.05 μM CuSO4 and 15 μM ZnSO4. Root length was measured after 8 d; results are the means ± SE of 50 seedlings from each genotype. Significant differences from the wild-type were determined by Student's t-test: **, < 0.01.

MT-deficient mutants were subjected to paraquat-induced oxidative stress. The quad-MT mutant seedlings were slightly less sensitive to paraquat than WT seedlings (Fig. S3a). A paraquat-induced photo-oxidative damage assay was also performed where leaves excised from plants grown in potting mix were treated with 100 μM paraquat. Only the mt3-1 mutant showed a slight increase in sensitivity to paraquat (Fig. S3b). However, the mt3-1 leaves also exhibited loss of chlorophyll under control conditions. Increased sensitivity of the mt3-1 mutant to paraquat may come from the Ws background in which this mutation was originally produced. The quad-MT mutant, which contains the same mt3-1 allele backcrossed into the Col-0 genetic background, responds to paraquat in a manner similar to the Col-0 control.

Senescence and chlorophyll degradation in detached leaves of MT-deficient mutants

Gene expression studies have shown that Arabidopsis MT genes expressed in vegetative tissues are upregulated during leaf senescence (Guo et al., 2003). To test the hypothesis that MTs play a role in leaf senescence, the effect of MT deficiency on the leaf senescence program was examined. Senescence was induced by placing detached leaves in the dark and chlorophyll content was measured over 5 d (Fig. S4). There were small differences in chlorophyll content between WT and quad-MT mutant leaves on the third and fourth days of dark treatment, with the quad-MT mutant having slightly less chlorophyll. However, quad-MT mutant leaves had slightly lower amounts of total chlorophyll compared to the WT leaves throughout the experiment. No differences in leaf senescence were observed when the MT-deficient mutant was grown to maturity under various standard conditions (data not shown). The results suggest that these MTs are not involved in the progression of visible symptoms of leaf senescence.

Fungal and bacterial disease resistance

The effect of MT deficiency on resistance to fungal and bacterial diseases was examined. Wild-type and MT-deficient Arabidopsis plants were inoculated with the fungal pathogens Alternaria brassicicola and Botrytis cinerea, and the bacterial pathogen Pseudomonas syringae. Figs S5, S6 and S7 show results from 6 DAI for A. brassicicola and B. cinerea, and 3 DAI for P. syringae, respectively. Disease symptoms from Alternaria brassicicola, Botrytis cinerea or Pseudomonas syringae infection were similar overall between MT-deficient mutants and WT plants. These results indicate that these Arabidopsis MTs do not make a significant contribution to resistance against these fungal and bacterial pathogens.

Cu accumulation in MT-deficient plants under Cu-excess conditions

Guo et al. (2008) showed that Arabidopsis mutants lacking MT1a (mt1a-2) accumulated 30% less Cu in roots than wild-type when plants were exposed to 30 μM CuSO4. To examine whether deficiency of multiple MTs expressed in vegetative tissues would have a greater impact on Cu accumulation, WT and quad-MT mutant plants were grown in a hydroponic system for 5 wk and exposed to 5 μM CuSO4 for 5 d. Shoots and roots of the quad-MT mutant accumulated c. 45% and 30% less Cu, respectively, compared to the WT (Fig. 3). These observations suggest that, in addition to MT1a, other MTs play a role in the accumulation of Cu in both roots and shoots when plants are exposed to excess Cu.

Figure 3.

Reduced accumulation of copper (Cu) in metallothionein (MT)-deficient plants exposed to excess Cu. Wild-type (closed bars) and quad-MT mutant (open bars) Arabidopsis plants were grown in a hydroponic system for 5 wk and then treated with 5 μM CuSO4 for 5 d. Control plants were grown continuously in the standard hydroponic solution containing 0.06 μM CuSO4. Shoot (a) and root (b) tissues were collected and analyzed for Cu concentration by GF-AAS. Results are the means ± SE of between four and six replicates. Significant differences from the wild-type were determined by Student's t-test: *, < 0.05; **, < 0.01.

Altered accumulation of Cu in leaves and seeds of MT-deficient plants

We next examined the impact of MT deficiency on Cu accumulation during vegetative and reproductive stages of development. Wild-type and quad-MT mutant plants were grown in a hydroponic system. Rosette leaves were harvested after 4 wk, just before plants started to flower. Seeds and leaves were also collected after 12 wk, when most leaves had fully senesced. There was no difference between WT and MT-deficient plants in the concentration of Cu in leaves of 4-wk-old plants (Fig. 4). However, the concentration of Cu increased in leaves of 12-wk-old MT-deficient plants but decreased in WT leaves. At 12 wk, leaves of the quad-MT mutant had more than twice the concentration of Cu compared to WT. By contrast, the concentration of Cu in seeds of MT-deficient plants was less than half that in WT seeds.

Figure 4.

Changes in concentrations of copper (Cu) in leaves, and seeds of metallothionein (MT)-deficient Arabidopsis plants. Wild-type (closed bars) and quad-MT mutants (open bars) were grown in a hydroponic system. Leaves were harvested at 4 and 12 wk; seeds were collected at 12 wk. Cu concentration was measured by ICP-MS. Results are the means ± SE of six replicates. Significant differences from the wild-type were determined by Student's t-test: *, < 0.05; **, < 0.01.

In order to examine Cu accumulation in MT-deficient plants in more detail, another experiment was performed where concentrations of the stable isotopes of Cu in the hydroponic medium were modified. Seeds were germinated in medium containing 68% 65Cu and grown for 4 wk until plants started to flower. Plants were transferred to medium containing the natural ratio of Cu isotopes (69% 63Cu, 31% 65Cu) and grown to maturity under these conditions. Leaves were sampled at 2-wk intervals and seeds were collected at 10 and 12 wk. The concentration of Cu and the abundance of Cu isotopes in these samples were measured by ICP-MS.

The concentration of Cu declined in WT leaves as they matured and senesced (Fig. 5a). By contrast, the concentration of Cu gradually increased in leaves of the quad-MT mutant so that at 12 wk the quad-MT mutant leaves had more than double the Cu concentration in WT leaves. By contrast, the concentration of Cu in seeds of the quad-MT mutant was less than half that of WT seeds (Fig. 5a), similar to results from the first experiment.

Figure 5.

Accumulation of copper (Cu) in leaves and seeds during growth of Arabidopsis plants. Wild-type (closed bars) and quad-metallothionein (MT) mutants (open bars) were grown hydroponically in 65Cu-enriched medium (68% 65Cu, 32% 63Cu) for 4 wk, then transferred to medium containing the natural Cu isotopic ratio (69% 63Cu, 31% 65Cu) and grown for another 8 wk. Leaves were harvested after 4, 6, 8, 10 and 12 wk; seeds were collected at 10 and 12 wk. Cu concentration (a) and isotopic abundance (b) were measured by ICP-MS. Results are the means ± SE of four replicates. Significant differences from the wild-type were determined by Student's t-test: *, P < 0.05; **, P < 0.01.

The percentage of each Cu isotope in these samples provides information about whether Cu in specific organs is acquired directly from the medium or redistributed from other sources within the plant. For example, 63Cu in leaves increased from 35% at week 4 to 63% at week 6 (Fig. 5b). This increase in the percentage of 63Cu indicates that leaves acquired a large amount of Cu from the hydroponic medium during the 2 wk after plants were transferred to medium containing natural Cu. There was also a substantial increase in the biomass of rosette leaves (from 8.7 mg to 76 mg dry weight) during this time. Although there was a two-fold difference in the concentration of Cu in seeds between WT and quad-MT mutant plants, they both contained 69% 63Cu (Fig. 5b). This result indicates that in this experiment almost all the Cu in seeds came from Cu taken up from the nutrient solution after plants were transferred to medium with the natural Cu isotope abundance.

In the experiment described in Fig. 5 the 4-wk-old plants that were transferred from 65Cu-enriched medium to normal medium were small. Consequently, the amount of 65Cu that could be remobilized from within the plant to developing seeds was limited. In a second experiment growth conditions were modified to increase the biomass of the rosette and the pool of 65Cu that could be remobilized from leaves to seeds (Fig. 6). This was accomplished by increasing light intensity and removing the first inflorescence when it emerged, which allowed the rosette of leaves to continue to grow and accumulate more 65Cu. When the second inflorescence emerged c. 1 wk later plants were transferred from 65Cu-enriched medium to normal medium for the remainder of their life cycle.

Figure 6.

Accumulation of copper (Cu) in leaves and seeds during growth of Arabidopsis plants. Wild-type (closed bars) and quad-metallothionein (MT) mutants (open bars) were grown hydroponically in 65Cu-enriched medium (97% 65Cu, 3% 63Cu) for 4 wk. The primary inflorescences were removed to promote vegetative growth. When the first flowers opened on secondary inflorescences, plants were switched to medium containing natural Cu (69% 63Cu, 31% 65Cu) and grown for another 8 wk. Leaves were harvested at 4 wk; leaves and seeds were collected at 12 wk. Cu concentration (a) and isotopic abundance (b) were measured by ICP-MS. Results are the means ± SE of six replicates. Significant differences from the wild-type were determined by Student's t-test: **, < 0.01.

Cu accumulation in leaves and seeds under these conditions was similar to other experiments. In MT-deficient plants the Cu concentration was higher in old leaves and lower in seeds compared to WT (Fig. 6a). Increasing the amount of biomass produced while plants were growing in 65Cu-enriched medium reduced the percentage of 63Cu in the seeds (Fig. 6b). In this experiment, 63Cu comprised between 57% and 59% of the Cu in seeds. This result demonstrates that a substantial fraction of the Cu in these seeds came from the pool of Cu taken up by plants when they were grown in 65Cu-enriched medium and subsequently redistributed to developing seeds.

Another experiment was performed using hydroponics to examine the accumulation of Cu in Arabidopsis seeds. Plants were grown for the first 6 wk in a hydroponic medium containing 0.12 μM CuSO4. A short photoperiod was used to delay flowering, increase the amount of biomass and allow leaves to accumulate as much Cu as possible. After 6 wk, plants were transferred to medium containing no added CuSO4 and photoperiod was increased to promote flowering. Control plants were grown in medium containing 0.12 μM CuSO4 from germination to maturity. In this experiment the Cu concentration was higher in leaves and seeds of WT plants compared to other experiments, likely because of these changes in growth conditions.

Removing Cu from the medium during reproductive development had no effect on WT plants in terms of growth of the inflorescence or the weight and appearance of seeds, but the seeds contained a lower concentration of Cu than those from control plants (Fig. 7c). In the quad-MT mutant, however, seeds from plants grown in medium lacking Cu during reproductive development were significantly smaller, yellow rather than brown in color, and contained a very low concentration of Cu (Fig. 7). The combined effect of Cu-deficient growth conditions and lack of MTs had a significant impact on seed development in the quad-MT mutant. The germination rate of these seeds was below 5% compared to more than 95% for seeds from MT-deficient plants grown with sufficient Cu (Table S3). In WT plants the concentration of Cu in leaves declined after plants were transferred to medium without added Cu, indicating that Cu was mobilized out of leaves during senescence. By contrast, when MT-deficient plants were transferred to medium without Cu, the concentration of Cu in leaves did not change as they senesced (Fig. 7c). These results provide further evidence that MTs are required to mobilize Cu out of senescing leaves. When the supply of Cu is limited, the ability to mobilize Cu from internal sources is important for seed development.

Figure 7.

Impact of copper (Cu) deficiency during reproductive growth on seed development and Cu accumulation in Arabidopsis plants. Wild-type (WT, closed bars in b, c) and quad-metallothionein (MT) mutant (open bars in b, c) Arabidopsis plants were grown in a hydroponic system with 0.12 μM CuSO4 and a 12-h light photoperiod for 6 wk until the first flower opened. The photoperiod was changed to 16-h light and plants were grown under either Cu-sufficient conditions (control) or transferred to medium without any CuSO4 (−Cu). Leaves were collected at 6 and 12 wk, seeds were collected at 12 wk. Seed morphology (a), seed weight (b), and Cu concentration in leaves and seeds (c) under control and Cu deficient (−Cu) conditions are presented. Bars, 1 mm. The data are means ± SE of six replicates. Significant differences from WT were determined by Student's t-test: **, < 0.01.

Discussion

We used a reverse genetics approach to examine the functions of MT genes in plants. Because mutants deficient in a single MT showed no dramatic phenotypes we developed plants that were deficient in the expression of the four MT genes that are most highly expressed in vegetative tissues. Three of the T-DNA insertions in the quad-MT mutant abolish RNA expression but the mt2a-1 allele retains a low level of expression. The only other MT gene that is expressed in vegetative tissues is MT1c. qRT-PCR analysis demonstrated that even in the quad-MT mutant MT1c RNA is expressed at less than 10% of the level of MT1a RNA. Because the MT1a and MT1c genes are < 4 kb apart, it would be challenging to combine independent insertions in these two genes. Alternative approaches will be needed to develop Arabidopsis plants that lack all MTs. Nevertheless the quad-MT mutant is a useful tool to examine the functions of MTs in plants.

There were no obvious differences between WT and quad-MT mutant plants under a variety of typical growth conditions (in potting mix, on agar medium, in hydroponics). This suggests that these MTs are not essential for plants to complete their life cycle under ‘standard’ conditions. Lack of MTs had no effect under most of the stress conditions we tested, including pathogen infection, oxidative stress and induced senescence. We did not, however, test the full range of environmental factors that can affect plant growth. The quad-MT mutant also retains a low level of RNA expression from MT1c and the mt2a-1 allele, which may be sufficient to provide the functions of MTs under some of these conditions.

Evidence that MTs contribute to metal tolerance in plants is limited. No differences in tolerance to Cu and Zn were observed between MT-deficient and WT seedlings (Fig. 2) and similar results were obtained when tolerance to Cu and Zn was compared under other conditions. It is difficult to reconcile our results with those of Murphy & Taiz (1995a,b) who reported a striking correlation between expression of MT2a RNA and Cu tolerance in a group of Arabidopsis ecotypes. Our results do not rule out a role for MTs in Cu tolerance in some hyperaccumulators (van Hoof et al., 2001; Mengoni et al., 2003). Other mechanisms, including metal transport and compartmentation, are likely to be more important for Cu and Zn tolerance in most plants (Burkhead et al., 2009; Palmer & Guerinot, 2009; Penarrubia et al., 2009; Pilon et al., 2009; Puig & Penarrubia, 2009). MT-deficient plants were, however, more sensitive to Cd than WT (Fig. 2c). Guo et al. (2008) showed that plants deficient in MT1a, MT2b and phytochelatins (mt1a-2/mt2b-1/cad1-3) were more sensitive to Cd than plants lacking just phytochelatins. It would be interesting to examine the combined effect of deficiency of all MTs and phytochelatins on metal tolerance, especially for Cd.

The concentration of Cu in leaves of 12-wk-old plants was higher in the quad-MT mutant than in WT. There are two possible explanations for the higher concentration of Cu in leaves of MT-deficient plants: either the quad-MT mutant takes up more Cu into leaves or WT plants transfer more Cu out of their leaves. Experiments using stable isotopes of Cu can discriminate between these possibilities. If the quad-MT mutant took up more Cu than WT after plants were transferred to medium containing 69% 63Cu, the MT-deficient leaves would contain a higher percentage of 63Cu. In our experiments, however, there was no consistent difference between WT and quad-MT mutant in the percentage of 63Cu in senesced leaves (Figs 5, 6). These results indicate that leaves of WT and MT-deficient plants took up similar amounts of Cu. If there is no difference in Cu uptake, the lower concentration of Cu in WT plants can be explained by greater remobilization of Cu out of leaves to other organs. MTs must therefore play an important role in the mobilization of Cu from senescing leaves. The lack of MTs reduces export of Cu from senescing leaves resulting in the higher concentration of Cu in old leaves of the quad-MT mutant.

There are a number of differences between the two stable isotope experiments presented here. The percentage of 63Cu in all tissues was higher in the first experiment (compare Figs 5 and 6). Although the 65CuSO4 used to prepare the medium contained 98% 65Cu, the nutrient solution in the first experiment contained only 68% 65Cu. This was likely due to contamination from natural Cu (69% 63Cu, 31% 65Cu) in the reagents used to prepare the medium. In the second experiment contamination was reduced so that the medium contained 97% 65Cu and the percentage of 63Cu measured in plant tissues was lower in all cases. One objective of the second stable isotope experiment (Fig. 6) was to increase leaf biomass and provide a larger source of Cu that could potentially be remobilized from leaves to seeds. This was accomplished by growing plants under higher light intensity and by removing the first inflorescence so that more leaf tissue developed before plants were transferred to medium containing natural Cu. These modifications were successful as the percentage of 63Cu in seeds was lower (c. 58% 63Cu) in the second experiment (Fig. 6).

Further evidence that MTs are important for mobilizing Cu out of senescing leaves was obtained from the experiment where CuSO4 was withheld from the nutrient solution when plants started to flower and produce seeds (Fig. 7). Without a supply of Cu from the medium, the only Cu available for reproductive tissues was that already present within the plant. The concentration of Cu in leaves of WT plants declined dramatically under these conditions indicating that Cu was transferred out of leaves as they senesced. In the MT-deficient mutant the Cu concentration did not change as these leaves aged demonstrating that MTs are required for redistribution of Cu from leaves.

The second striking phenotype in the quad-MT mutant is the low concentration of Cu in seeds. It is simple to suggest a causal relationship between increased Cu in leaves and reduced Cu in seeds of the quad-MT mutant: accumulation of more Cu in leaves reduces the amount of Cu available for seeds. However, this does not fully consider the complex dynamics of growth, uptake, accumulation and redistribution of nutrients in different plant tissues. Waters & Grusak (2008) analyzed the accumulation of mineral nutrients in Arabidopsis organs and concluded that seeds were more dependent on direct supply of nutrients from roots rather than mobilization of nutrients from other organs such as leaves. However, the methods they used could not unequivocally distinguish between uptake and remobilization as the source of specific minerals in seeds.

Experiments with stable isotopes and withholding CuSO4 from the nutrient medium demonstrated that Cu in Arabidopsis seeds comes from two sources: Cu can be transported directly from roots to seeds and also remobilized from other tissues to seeds. When Cu is not available from the roots, Cu must be redistributed from internal sources. Without Cu in the nutrient solution during seed development, WT seeds developed normally but had a low Cu concentration (Fig. 7). The Cu in these seeds must have been remobilized from other organs within the plant. Withdrawing Cu from the medium during seed development had a dramatic effect on the quad-MT mutant, resulting in seeds with a very low Cu concentration, reduced seed weight, abnormal appearance and low germination rate. Stable isotope experiments established that the quad-MT mutant is defective at mobilizing Cu out of leaves during senescence. Lack of Cu from the growth medium combined with a limited supply of Cu from internal sources in MT-deficient plants resulted in production of abnormal seeds with a very low Cu concentration. These results do not rule out the possibility that MTs may participate in other processes that affect the delivery of Cu into developing seeds. In these experiments we did not attempt a complete mass balance analysis to account for all of the Cu in the plants. This could provide valuable information on the relative importance of remobilization of Cu from internal sources compared to supply of Cu from the growth medium.

The model presented in Fig. 8 shows that Cu taken up by roots can be transported to leaves and inflorescences in WT and MT-deficient plants. Cu is also remobilized from senescing leaves to developing seeds in WT plants. In the absence of MTs, redistribution of Cu from leaves is disrupted and this accounts for the high concentration of Cu in leaves of the quad-MT mutant. Loss of this source of Cu from senescing leaves may also contribute to the low concentration of Cu in seeds of MT-deficient plants.

Figure 8.

Proposed model for altered accumulation of copper (Cu) in metallothionein (MT)-deficient Arabidopsis plants. In wild-type plants (left panel) Cu that accumulates in seeds comes from two sources: directly from the roots as well as from senescing leaves. In the quad-MT mutant (right panel) Cu cannot be remobilized from senescing leaves because of the lack of MTs. The inability to mobilize Cu from leaves results in the high concentration of Cu in senesced leaves and may contribute to the low concentration of Cu in seeds of MT-deficient plants.

Cu accumulation has been studied in a number of Arabidopsis mutants that affect Cu homeostasis including those lacking COPT1 and HMA5 (Sancenon et al., 2004; Andres-Colas et al., 2006). None of these have the same phenotype of high Cu in leaves and low Cu in seeds observed in the quad-MT mutant. However, there are striking similarities between the quad-MT mutant and the Arabidopsis ysl1/ysl3 mutant with regard to Cu accumulation (Waters et al., 2006; Chu et al., 2010). YSL proteins are metal-nicotianamine transporters that are encoded by small gene families in plants. Research on YSL proteins has focused on Fe homeostasis and the Arabidopsis ysl1/ysl3 double mutant exhibits symptoms of Fe deficiency. However, this mutant also accumulates more Cu in leaves and less Cu in seeds compared to WT, and is defective in exporting Cu from leaves during senescence (Waters et al., 2006), similar to the quad-MT mutant. The ysl1/ysl3 mutant also shows altered homeostasis of Fe, Zn and Mn, which are not altered in MT-deficient plants. Knockout of the YSL16 gene in rice also affects Cu distribution in leaves and seeds (Zheng et al., 2012).

MTs may function to deliver Cu to YSL transporters in senescing leaves. The senescence program involves the catabolism of leaf proteins resulting in release of metal ion cofactors. Many MT genes are highly expressed during leaf senescence in Arabidopsis and other plants (Hsieh et al., 1995; Foley et al., 1997; Butt et al., 1998; Chen et al., 2003; Guo et al., 2003; Kohler et al., 2004). MTs may provide protection against metal ion toxicity under these conditions. However, the similarities between MT-deficient plants and those lacking YSL transporters suggest that MTs and YSLs may interact. MTs may function as chaperones to deliver the Cu released during protein degradation to the transport system so that Cu can be exported from senescing cells and remobilized to other organs. Although MT deficiency does not affect the overall senescence program, it specifically inhibits the export of Cu from leaves.

These results show that the Arabidopsis MTs studied here are not involved in adaptation to several biotic and abiotic stresses, but are important for the accumulation and distribution of Cu in leaves and seeds. Further experiments will be required to understand the underlying molecular mechanisms that lead to altered accumulation of Cu in different organs of MT-deficient mutants. It will also be necessary to determine the contribution of individual MT genes to these phenotypes.

Acknowledgements

We thank Jenny Olszewski, Kyle Mohler and Kimberly Chapman for help identifying quad-MT mutant plants. We acknowledge the Purdue Ionomics Center, Dr David Salt, Brett Lahner and Stephen Sassman for ICP-MS analysis, and Dr Paul Schwab for GF-AAS analysis. This research was supported in part by a grant from the USDA-NRI Plant Responses to the Environment Program (no. 01-35100-10613).

Ancillary