Copper speciation and isotopic fractionation in plants: uptake and translocation mechanisms


Author for correspondence:

Brooke Ryan

Tel: +61 8 83036787



  • The fractionation of stable copper (Cu) isotopes during uptake into plant roots and translocation to shoots can provide information on Cu acquisition mechanisms.
  • Isotope fractionation (65Cu/63Cu) and intact tissue speciation techniques (X-ray absorption spectroscopy, XAS) were used to examine the uptake, translocation and speciation of Cu in strategy I (tomato–Solanum lycopersicum) and strategy II (oat–Avena sativa) plant species. Plants were grown in controlled solution cultures, under varied iron (Fe) conditions, to test whether the stimulation of Fe-acquiring mechanisms can affect Cu uptake in plants.
  • Isotopically light Cu was preferentially incorporated into tomatoes (Δ65Cuwhole plant-solution = c. −1‰), whereas oats showed minimal isotopic fractionation, with no effect of Fe supply in either species. The heavier isotope was preferentially translocated to shoots in tomato, whereas oat plants showed no significant fractionation during translocation. The majority of Cu in the roots and leaves of both species existed as sulfur-coordinated Cu(I) species resembling glutathione/cysteine-rich proteins.
  • The presence of isotopically light Cu in tomatoes is attributed to a reductive uptake mechanism, and the isotopic shifts within various tissues are attributed to redox cycling during translocation. The lack of isotopic discrimination in oat plants suggests that Cu uptake and translocation are not redox selective.


Copper (Cu) is an essential micronutrient for plant growth, playing a key role in the function of many proteins. Cu is involved in electron transport during photosynthesis, lignin formation and cell wall metabolism (Burkhead et al., 2009). Cu-deficient plants (< 5 mg kg−1 in vegetative tissue) exhibit reduced growth and development, with reproductive organs and youngest leaves displaying the most severe symptoms (Burkhead et al., 2009; Broadley et al., 2012). At above optimal concentrations (usually > 10 μM in solution), Cu can act as a toxin to plants, causing nutrient loss and oxidative stress (Martins & Mourato, 2006; Reichman et al., 2006). Under toxic conditions, damaging hydroxyl radicals are formed that attack the cell structure, and inhibit photosynthesis (Fernandes & Henriques, 1991; Yruela, 2009).

The mechanisms of Cu mobilization and uptake by roots from soil solutions remain unclear. Cu in soils is strongly associated with organic matter, as well as iron (Fe) and aluminium (Al) oxides (Flemming & Trevors, 1989). Plants have developed two uptake mechanisms for the acquisition of highly insoluble Fe(III) in soils: strategy I plants (dicotyledons and nongraminaceous monocotyledons) mobilize Fe through acidification of the rhizosphere, causing the dissolution of Fe/Al oxides, and the upregulation of Fe reductases, which reduce Fe(III) to Fe(II), whereas strategy II plants (graminaceous monocotyledons) can complex Fe(III) through the release of high-affinity Fe(III) root exudates, phytosiderophores (Guerinot & Yi, 1994; Marschner & Romheld, 1994). These Fe reduction and complexation mechanisms have been shown to affect Cu speciation in controlled laboratory environments. For instance, ferric reductase oxidase 2 (FRO2) is capable of reducing Cu(II) and FRO3 is upregulated during Cu deficiency (Robinson et al., 1999; Burkhead et al., 2009; Palmer & Guerinot, 2009; Bernal et al., 2012). In addition, phytosiderophores, or similar exudates, are capable of Cu complexation (Welch et al., 1993; Georgatsou et al., 1997; Schmidt et al., 1997; Yruela, 2009). The importance of these mechanisms in Cu mobilization and uptake in the soil environment remains unclear. Although no studies have confirmed the mobilization and uptake mechanism of Cu, there appears to be a strong overlap between Fe and Cu uptake, with Cu shown to competitively inhibit Fe uptake (Schmidt et al., 1997; Michaud et al., 2008; Yruela, 2009).

Stable isotope fractionation can be an effective method of monitoring Cu biogeochemical cycling, in particular, nutrient uptake. Bigalke et al. (2010a) found that organic top soil horizons of three hydromorphic soils were isotopically enriched in light Cu, relative to mineral fractions, suggesting a biologically induced enrichment from soil–plant Cu cycling. A study of two soil-grown monocots (rye, Elymus virginicus, and hairy leaved sedge, Carex hirsutella) found preferential incorporation of the light Cu isotope into vegetative tissues, relative to the soil, of between −0.94‰ and −0.33‰ (Weinstein et al., 2011). Similarly, a recent study of hydroponically grown tomato and wheat plants showed enrichment of the light Cu isotope in roots relative to the growth medium (Δ65Curoot-solution) of between −0.84‰ and −0.47‰ for strategy I species and −0.48‰ and −0.11‰ for strategy II species (Jouvin et al., 2012). The authors attributed this isotopic shift to the reduction of Cu(II) to Cu(I) before uptake into the root symplasm. Their study used chelator-buffered solutions which complicate the interpretation of isotope data, as fractionation during Cu-chelate dissociation and uptake of chelated Cu through the apoplastic pathway need to be considered. A study in which the speciation of Cu provided to the plant is held constant is required to gain a more complete understanding of the mechanisms involved during uptake and translocation.

Although the magnitude and direction of isotope fractionation will be influenced by the binding environment and redox state of Cu, it is not possible to confirm the speciation and bonding environment without other techniques. X-Ray absorption spectroscopy (XAS) provides an in situ, nondestructive method of assessing Cu speciation in plant tissues. Previous work has found that Cu exists as both Cu(I) and Cu(II) in the vegetative tissue of Larrea tridentate (creosote bush), despite the fact that Cu absorbed from the soil is Cu(II), indicating redox changes within the plant or during uptake (Polette et al., 2000). The Cu accumulator Crassula helmsii was found to bind Cu almost exclusively through oxygen (O)-rich ligands (Kuepper et al., 2009), whereas Thlaspi caerulescens, a hyperaccumulator for Cd and Zn but not Cu, stored excess Cu with strong ligands rich in sulfur (S), possibly metallothioneins (Mijovilovich et al., 2009).

The aim of this study was to use highly precise isotope fractionation and intact tissue speciation techniques to examine the uptake and translocation of Cu in plants. Tomato (Solanum lycopersicum), a strategy I plant species, and oat (Avena sativa), a strategy II species, were grown hydroponically for 30 d, with an Fe deficiency treatment imposed during the last week of growth, to test the hypothesis that the stimulation of Fe-acquiring mechanisms can affect Cu uptake and translocation in plants. The speciation of Cu in the nutrient solution was closely controlled without using chelator buffering, so that Cu(II) activities in Fe-sufficient and Fe-deficient nutrient solutions were similar and isotope signatures in plants were not affected by isotopic fractionation in the solution or the possible uptake of intact Cu chelates from the hydroponic solutions.

Materials and Methods

Hydroponic plant growth

Six tomato (Solanum lycopersicum L.) and six oat (Avena sativa L.) plants were grown for a total of 30 d under controlled conditions in a nutritionally complete solution at pH 4.5. To maintain Fe in a soluble form for plant growth, Fe-specific chelators were used: N,N′-bis(2-hydroxybenzyl)ethylenediamine-N,N′-dipropionic acid (HBED) (12 μM) for tomatoes and ethylenediamine-N,N′-bis(2-hydroxyphenylacetic acid) (EDDHA) (12 μM) for oats. The EDDHA chelate was used for oats, given the inefficiency of strategy II plants to extract Fe from the HBED complex (Parker & Norvell, 1999). These chelates were selected on the basis of GEOCHEM modelling of nutrient solutions at pH 4.5, as they yielded the highest possible ‘free’ Cu(II) concentrations, with HBED having no Cu complexation and EDDHA only complexing 6% of the total Cu (Supporting Information Table S1 for GEOCHEM modelling). The Cu concentration in the solution was 1 μM and the nutrient solution was kept at pH 4.5 in order to avoid significant complexation, maintaining > 85% of Cu in solution as ‘free’ Cu2+ (Table S1); this was confirmed using a Cu ion-selective electrode. The Cu concentration selected for this hydroponic study was chosen to enable sufficient concentrations in plant tissues to allow for accurate and precise isotope ratio and speciation analysis. The maintenance of all Cu as free Cu2+ at lower concentrations would have been extremely difficult. On day 24, Fe deficiency was induced in three replicates from each species by placing the plants in a nutrient solution that contained no Fe or chelator, but was otherwise nutritionally identical. Nutrient solutions were changed every 4 d for the first week, every 2 d for the second week and daily for the final 2 wk of plant growth, to avoid significant Cu isotope fractionation as a result of the build up of Cu-complexing exudates. This allowed the assumption of a constant Cu isotope ratio. The frequency of solution changes in conjunction with pH buffering from 2-(N-morpholino)ethanesulfonate (MES; 2.5 mM) prevented pH changes in solution (Methods S1).

All plants were harvested on day 30 and separated into roots, stems and leaves. The leaves for Fe-deficient treatments were further separated into green and chlorotic leaves. The roots were washed to remove any apoplastically bound Cu by submerging in ice-cold ultrapure deionized water (Millipore) for 5 min followed by 1 mM LaCl3 and 0.05 mM CaCl2 for 5 min (Weiss et al., 2005). Plant tissues (roots, stems and leaves) were frozen and freeze–dried for 7 d to a constant mass (ModulyoD; Thermo Scientific, Waltham, MA, USA). The plant tissues were homogenized using stainless steel scissors.

Digestion of plant tissues

Approximately 0.05–0.1 g of the freeze–dried plant tissues (roots, stems, leaves) were cold digested in 3 ml of nitric acid (HNO3) and 1 ml of hydrogen peroxide (H2O2) in polytetrafluoroethylene (PTFE) vials, followed by a 2-h hotplate reflux at 140°C. Samples were then evaporated to dryness and redissolved in 7 ml HNO3 and 3 ml H2O2, and transferred to closed PTFE vessels. The samples were microwave digested in sealed vessels (Ethos E; Milestone, Shelton, CT, USA; Methods S1). The digest extracts were then dried and redissolved in 7 M HCl for Cu purification, and Cu, macro- and micronutrients were measured on inductively coupled plasma mass spectrometry (ICP-MS) or ICP-optical emission spectroscopy (ICP-OES; Table S4).

The accuracy of the digestion and analysis procedures for the determination of total Cu concentrations in plants was assessed using National Institute of Standards and Technology (NIST) 1573a tomato leaves standard reference material. The total Cu concentrations determined in the tomato leaves standard reference material was in close agreement with the certified value (4.6 ± 0.2 mg kg−1 (mean ± SD) compared with a reference value of 4.7 ± 0.1 mg kg−1; Table S2).

Cu purification

Cu for isotope analysis was purified using AG-MP-1 anion exchange resin (BioRad), following the procedure described by Marechal et al. (1999) (Table S3). High concentrations of matrix elements in the plant tissues required the column purification to be run twice. For stem and leaf tissues, a third column purification was needed to ensure that alkali metals were below the instrumental detection limits. The column purification procedure was optimized using NIST 1573a tomato leaves (Fig. S1, Methods S1). Digest and column purified Cu extracts were checked for Cu recoveries, and samples with 100 ± 8% were found to show no significant difference in δ65Cu values.

The Cu fraction from the columns was collected in PTFE vials and evaporated to dryness on a hotplate. The samples were resuspended in 0.5 ml HNO3 and refluxed for 2 h at 140°C to digest any organic compounds that may have been released from the column (Bigalke et al., 2010a,b), before being evaporated to dryness again. Samples were redissolved in 2% HNO3, and 0.1-ml aliquots of the sample solutions were taken for total Cu analysis to determine Cu recoveries from the column purification procedure.

Cu isotope measurements

Cu isotope ratios (65Cu/63Cu) of purified samples were determined using a multicollector-inductively coupled plasma-mass spectrometer (MC-ICP-MS; Thermo Scientific). The Cu isotope ratios were expressed as δ65Cu values relative to NIST 976 Cu isotopic reference material according to Eqn 1:

display math(eqn 1)

Cu isotope ratios were measured by methods outlined previously by Ehrlich et al. (2004) and Bigalke et al. (2010a,b), with internal mass bias correction with NIST 986 Ni, and are briefly described in Methods S1. The precision of each individual sample is constrained by multiple analyses of each sample, and reported as twice the standard deviation in Table 1. In addition, long-term external reproducibility for the entire method (i.e. digestion, purification and isotope analysis) is defined by the processing and analysis of NIST 1573a tomato leaves (= 5), with each replicate sample measured at least three times. This standard has δ65Cu = 0.63 ± 0.16‰ (mean ± 2SD). The error afforded to each sample as a result of sample preparation is therefore assumed not to exceed ± 0.16‰. Treatments have their own variations, reported in Table 1 as 2SD, generated from the inherent variability between replicate plants.

Table 1. Biomass (DW), copper (Cu) concentrations and isotopic measurements for tomato (Solanum lycopersicum) and oat (Avena sativa) plants grown in iron (Fe)-sufficient (+Fe) and Fe-deficient (−Fe) hydroponic conditions
SpeciesTreatmentTissue n Biomass (DW) (g)SDCu (mg kg−1)SDΔ65Cutissue-solution2SD
Green leaves30.0980.0355011−0.560.51
Chlorotic leaves30.0530.014328.3−0.260.71
Green leaves30.0680.069212.20.000.13
Chlorotic leaves30.0700.016202.5−0.270.20
Sample n Cu (mg kg−1)SDδ65Cu2SD
  1. Data for National Institute of Standards and Technology (NIST) 1573a tomato leaf samples used for method development, as well as δ65Cu value for hydroponic nutrient solution, are included in this table.

NIST 1573a tomato leaves54.60.230.630.16
Nutrient solution4  0.250.04

The samples and standards were measured in 2% HNO3 at concentrations of 300 μg Cu l−1 (63Cu = c. 7–10 V) and 1000 μg Ni l−1 (60Ni = c. 10–12 V) in low-resolution mode. The samples were analysed using wet plasma conditions with a 100-μl PFA nebulizer and platinum sampler and skimmer cones.

All measured δ65Cu values were normalized to the δ65Cu value for the nutrient solution (Δ65Cutissue-solution) by subtracting the δ65Cu value of the solution Cu (+0.25‰). Similarly, fractionation during Cu translocation, which was simplified to include only a Cu-supplying tissue (or solution) (a) and a Cu-receiving tissue (b), is defined as Δ65Cub-a (‰) according to the following equation:

display math(Eqn 2)

The whole-plant Δ65Cuwhole plant-solution (‰) values were calculated from a mass balance equation (Eqn 3) to determine the uptake-induced fractionation:

display math(Eqn 3)

(Fi, fraction of Cu in a given tissue i (e.g. root, stem or leaves); δ65Cui (‰), isotope ratio of Cu in tissue i).

Similarly, to examine the fractionation between roots and shoots (Δ65Cushoot–root = δ65Cushoot − δ65Curoot), a mass balance calculation was used to determine the δ65Cu of shoots (‰) (i.e. total of stems and leaves), from the Cu amount (mg), calculated from the Cu concentration and total dry weight, and the δ65Cu (‰) value of stems and leaves:

display math(Eqn 4)

X-Ray absorption spectroscopy

Cu K-edge XAS of frozen roots and leaves of replicate Fe-sufficient and Fe-deficient tomato and oat plants were recorded on the XAS beamline at the Australian Synchrotron (AS), Melbourne, Australia. Extended X-ray absorption fine structure (EXAFS) scans were run on one root tissue replicate from each of the Fe-sufficient and Fe-deficient oat plants, but only on the Fe-deficient tomato plant roots, because of the Cu concentration limitations in the Fe-sufficient roots. Only X-ray absorption near-edge structure (XANES) scans were run on the corresponding leaf tissues, given the significantly lower Cu concentrations. The Cu concentrations in stem tissues were below the detection limit of the beamline, precluding the collection of XAS data.

The X-ray beam was monochromated by diffraction from a pair of Si(111) crystals. Plant tissues were cut to c. 1 × 1 cm2 in size, secured between Kapton tape and stored in a −70°C freezer and cooled to c. 10 K in a Cryo Industries (Manchester, NH, USA) cryostat for analysis. Spectra were recorded in fluorescence mode on a 100-pixel Ge detector array at 90° to the incident beam, with Ni absorption filters and Soller slits placed between the sample and fluorescence detector to improve the fluorescence to background ratio. The energy ranges used for XANES data collection were as follows: pre-edge region, 8779–8959 eV (10-eV steps); XANES region, 8959–9029 eV (0.25-eV steps); post-edge region, 9029–9167 eV (0.2-Å−1 steps in k space). EXAFS spectra for roots were collected at the following energy ranges: pre-edge region, 8779–8959 eV (10-eV steps); XANES region, 8959–9029 eV (0.25-eV steps); EXAFS region, 9029–9623 eV (0.035-Å−1 steps in k space to 13 Å−1). A Cu foil standard was recorded simultaneously in transmission mode downstream of the sample to calibrate the energy scale to the first peak of the first derivative of the Cu edge (8980.3 eV).

Data reduction and analysis, including calibration, averaging and background subtraction of all spectra and principal component analysis (PCA), target transformation and linear regression analyses of XANES spectra, were performed using the EXAFSPAK software package (G. N. George, Stanford Synchrotron Radiation Lightsource, Menlo Park, CA, USA). Linear combination fits of XANES spectra were performed over the region 8950–9070 eV. Spectra of model Cu compounds for target and linear regression analyses were collected during the same beamtime, employing the same conditions as outlined above (Fig. S2; additional information on XAS analysis and model Cu compounds can be found in Methods S2).

Statistical analysis

Cu concentration and isotope data were analysed using ANOVA (GenStat, 14th edn; VSN International Ltd, Hemel Hempstead, UK) to determine significant differences between treatments (i.e. Fe-sufficient and Fe-deficient treatments), plant species and plant tissues (roots, stems, leaves). A 5% level of significance was used to compare treatment means.


Plant biomass

Total concentrations of micro- and macro-elements in tomato and oat plants were found to be above critical deficiency concentrations suggested to limit plant growth and survival (Huett et al., 1997; Reuter et al., 1997). Fe deficiency in this study was visually observed in Fe-deficient treatments as chlorosis in leaves and measured by ICP-OES. Deficiency of Fe was found to have no significant effect on the dry biomass of roots, stems or leaves for either test species (Tables 1, S4).

Cu concentrations in roots of Fe-sufficient tomatoes (448 ± 56 mg kg−1) were significantly lower than those in Fe-deficient roots (766 ± 170 mg kg−1; Table 1). Similarly, oat plants showed significantly less Cu in Fe-sufficient roots (611 ± 80 mg kg−1) than in Fe-deficient roots (1093 ± 189 mg kg−1; Table 1). Increased Cu uptake in Fe-deficient plants had no observable effect on root growth (visually and root dry weight).

The absence of a significant difference in total Cu concentrations in stem and leaves between Fe treatments of tomato and oat plants, despite the very different root Cu concentrations (Table 1), aligns with the known regulation of Cu translocation between roots and shoots (e.g. Alaoui-Sossé et al., 2004; Chen et al., 2011). Concentrations in the range 2–50 mg kg−1 (DW) are considered to be sufficient for tomatoes (Mills & Jones, 1996). The Cu concentrations in tomato leaves of both treatments were slightly above this range (Table 1); however, no visible signs of toxicity were observed.

Cu isotope fractionation during uptake

The whole-plant δ65Cu for tomatoes and oats indicated a preferential incorporation of the light isotope into the plants from solution, but the magnitude of enrichment was much greater for tomatoes than for oats (Fig. 1). There was no significant difference between Fe treatments for tomatoes, with Δ65Cuwhole plant-solution values of −1.05 ± 0.36‰ and −0.99 ± 0.37‰ for Fe-sufficient and Fe-deficient plants, respectively (Fig. 1). Such a large negative fractionation indicates the presence of a dominant reductive uptake mechanism, that is, the reduction of free Cu(II) to Cu(I) during uptake into the symplasm, as Cu reduction is known to induce such significant light Cu isotopic enrichment (Criss, 1999; Zhu et al., 2002; Kavner et al., 2008; Jouvin et al., 2012).

Figure 1.

Average Δ65Cu(tissue-nutrient solution) (mean ± 2SD, = 3) of roots, stems and leaves for tomato (Solanum lycopersicum) and oat (Avena sativa) plants after 30 d of growth in either iron (Fe)-sufficient (+Fe) or Fe-deficient (−Fe) nutrient solutions.

Oat plants showed a small enrichment in the light Cu isotope during uptake and no significant difference between Fe treatments, with Δ65Cuwhole plant-solution values of −0.20 ± 0.11‰ and −0.11 ± 0.03‰ for Fe-sufficient and Fe-deficient plants, respectively (Fig. 1). This fractionation is too small to imply a dominant reductive uptake mechanism, suggesting that another uptake mechanism is at play. Cu-phytosiderophore complexation and uptake may have contributed to uptake, which may be expected to induce slight heavy Cu isotope enrichment in the plant, given the small preference for heavy isotopes in organic complexes (Guelke & Von Blanckenburg, 2007; Bigalke et al., 2010a,b).

Cu isotope fractionation during translocation

Tomato plants showed a large fractionation of Cu isotopes during translocation between different tissues, whereas oat plants maintained a fairly constant Cu isotope signature throughout all tissues (Fig. 1).

The δ65Cu value of the whole plant results from the fractionation during uptake. δ65Cu in plant tissues depends on both δ65Cu of the whole plant and the isotopic fractionation during translocation. For instance, preferential translocation of a heavier isotope from a plant tissue will cause the Cu remaining in that tissue to be enriched with the light isotope. An overview of isotopic composition within the whole plant and the different tissues is given in Table 2.

Table 2. Schematization of copper (Cu) isotope fractionation as Cu is translocated from roots to shoots, and then between stem and leaves, for iron (Fe)-sufficient (+Fe) and Fe-deficient (−Fe) conditions in tomato (Solanum lycopersicum) and oat (Avena sativa)
SpeciesTreatmentΔ65Cuwhole plant-solution (‰)Root to shoot translocationStem to leaf translocation
Tissue F Δ65Cutissue-solution (‰)Tissue F Δ65Cutissue-solution (‰)
  1. The Δ65Cu values are reported normalized to the nutrient solution value, and F is the fraction of total plant Cu contained within the given tissue. Mass balance calculated values shown in italics.

Shoot 0.37 0.40Stem0.03−0.08
Shoot 0.14 0.43Stem0.02−0.22
Green leaves0.09−0.56
Chlorotic leaves0.03−0.26
Shoot 0.07 0.24Stem0.02−0.20
Shoot 0.04 0.08Stem0.010.00
Green leaves0.030.00
Chlorotic leaves0.01−0.27

For tomato, the value of Δ65Cuwhole plant-solution is c. −1.0‰, suggesting a reductive uptake mechanism, as discussed above. The Cu translocated to shoots is more enriched in the heavy isotope, whereas Cu retained in roots is more enriched in the light isotope, compared with the whole plant (Table 2), indicating preferential translocation of the heavy Cu isotope to above-ground tissues from roots. Thus, the root will become more enriched in the light isotope as more Cu is translocated to the shoot (Rayleigh fractionation, see Notes S1). This agrees with the more negative Δ65Curoot-solution for the Fe-sufficient than for the Fe-deficient plants, as relatively more Cu is translocated to shoots in the Fe-sufficient plants (Notes S1). These data point towards a conserved, regulated mechanism of root to shoot Cu transport, which leads to a positive delta shift that is consistent across both Fe-sufficient and Fe-deficient treatments. The distribution of Cu in shoots between stems and leaves leads to heavier Cu isotope enrichment in the stems of both treatments, suggesting a preference for the light Cu isotope to be translocated to the leaves. The values of Δ65Culeaves-solution and Δ65Cushoot-solution are similar, because 91% and 84% of vegetative tissue Cu exists in the leaves in the Fe-sufficient and Fe-deficient plants, respectively. In contrast with the nonchlorotic leaves, the chlorotic leaves showed similar isotopic composition to the stems (Δ65Cuchlorotic leaves-stem = −0.04‰; Table 2). The large Cu isotope fractionations between plant tissues throughout the tomato plants of both treatments suggest that redox-selective mechanisms are occurring during translocation.

The oats showed a constant Cu isotope signature (within error) throughout all tissues in the Fe-sufficient and Fe-deficient treatments, with the exception of the chlorotic leaves of the Fe-deficient treatment (Δ65Cuchlorotic leaves − Δ65Custem = −0.27‰; Table 2). The lack of any significant Cu isotope fractionation during translocation within the oat plants indicates that redox-selective transport is not a dominant translocation mechanism for Cu in this plant species.

XAS analysis

The Cu K-edge XANES spectra of root and leaf samples of both species showed that Cu speciation was dominated by Cu(I), as identified by the low rising edge inflection energy of Cu(I) species at c. 8984 eV (Kau et al., 1987; Fig. 2) and linear combination fitting of model compounds using EXAFSPAK (Fig. 3a,b; Table S5). PCA indicated that the samples had three primary Cu species present and, as such, the three model compounds with the lowest target transformation residuals were fitted to the sample spectra: Cu(I)-glutathione (GSH), Cu(I)-cysteine and Cu(II)-histidine (Figs S3, S4).

Figure 2.

X-Ray absorption near-edge structure (XANES) spectra of root and leaf samples of iron (Fe)-sufficient (+Fe) and Fe-deficient (−Fe) tomato (Solanum lycopersicum) and oat (Avena sativa) plants, with standard spectra used for linear combination fitting. Similar spectra have been grouped together as follows: tomato −Fe roots, tomato −Fe leaves, oat +Fe roots and oat −Fe roots (a–d); oat +Fe leaves and oat −Fe leaves (e–f); tomato +Fe roots (g); tomato +Fe leaves (h); Cu(I)-GSH (i); Cu(I)-cysteine (j); Cu(II)-histidine (k).

Figure 3.

Proportion of copper (Cu) species present in root tissues (a) and leaf tissues (b) for iron (Fe)-sufficient (+Fe) and Fe-deficient (−Fe) tomato (Solanum lycopersicum) and oat (Avena sativa) plants, as determined by linear combination fitting of X-ray absorption near-edge structure (XANES) Cu K-edge spectra (see Supporting Information Table S4). Cu(II)-histidine, grey bars; Cu(I)-cysteine, white bars; Cu(I)-glutathione, black bars. Note the different concentration scales on the two graphs.

XANES fitting indicated that Fe-deficient tomato roots contained 86% Cu(I), whereas Fe-sufficient tomato roots contained 58% Cu(I) (Fig. 3a, Table S5). However, in the tomato plants grown in Fe-sufficient solution, the three model compounds accounted for only 85–88% of the spectral fit for roots and leaves, suggesting that a fourth, unidentified component may have been present (Table S5). Although having significantly different total Cu concentrations, tomato roots showed very similar Cu(II) concentrations of 120 and 108 mg kg−1 for Fe-sufficient and Fe-deficient plants, respectively (Fig. 3a), suggesting a highly controlled re-oxidation of the stored Cu(I) to Cu(II). Cu(I) dominated the speciation of tomato leaves in both treatments. Oat roots and leaves showed similar Cu speciation in both treatments, with the speciation strongly dominated by Cu(I) species (Fig. 3a,b). It should be noted that differentiating between Cu(I) and Cu(II) species in photosynthetic tissues/proteins can be difficult (Lee et al., 2009) and, as such, the reported Cu(I) and Cu(II) proportions in the leaves should be taken as approximate.

EXAFS fitting of the tomato and oat root samples indicated that Cu was bound by S-donor ligands in a three-coordinate complex (e.g. Fig. 5; explaining why so much of the XANES data can be fitted with Cu(I)-GSH; Table S5). The fitting error in the EXAFS spectra was reduced significantly by including a Cu–Cu interaction at 2.66–2.71 Å (Tables 3, S6). This indicates that Cu in root tissues is being stored in small Cu clusters bridged by S ligands. The EXAFS spectra and corresponding Fourier transform plots of the observed and fitted data can be seen in Figs 5 and S5–S7. The most realistic Cu coordination number in the S-bridged cluster is three, forming a trigonal S-bridged poly-Cu cluster, as identified for the transcription factor proteins Mac1 and Ace1 (Brown et al., 2002). However, the best fit of the EXAFS data varies with each sample from 2.5 to 3.5 S atoms (Table 3). This is probably the result of secondary Cu···Cu interaction which sits out of phase to the Cu···Cu interaction occurring at c. 2.7 Å (Brown et al. (2002). The speciation observed in these plants may have been quite different if plants had been grown at a lower free Cu concentration, as lower tissue Cu concentrations would see a decrease in excess root Cu, and hence less requirement for detoxification with metallothionein-type proteins.

Table 3. Best-fit parameters for extended X-ray absorption fine structure (EXAFS curve fitting of tomato (Solanum lycopersicum) and oat (Avena sativa) plant roots, grown in iron (Fe)-sufficient (+Fe) and Fe-deficient (−Fe) conditions, as determined by fitting error, physical reasonableness of the parameters and visual inspection of the fitted spectra
SampleShellsInteraction N R (Å)σ2ΔEo (eV)Error
  1. A full set of trialled fitted parameters can be found in Supporting Information Table S5. (The k range was 1–13 Å−1 and a scale factor (inline image) of 0.9 was used for all fits. ΔE0 = E0−9000 (eV), where E0 is the threshold energy. Values in parentheses are the estimated standard deviations derived from the diagonal elements of the covariance matrix and are a measure of precision. The fitting error is defined as [Σk6exp − χcalc)2k6χexp2]1/2.)

Tomato −Fe2





2.254 (3)

2.706 (3)

0.0071 (2)

0.0022 (2)

−14.2 (5)

−14.2 (5)

Oat +Fe2





2.261 (4)

2.686 (5)

0.0046 (2)

0.0051 (5)

−13.5 (8)

−13.5 (8)

Oat −Fe2





2.251 (3)

2.661 (8)

0.0066 (2)

0.0062 (8)

−15. 8 (6)0.420


Tomatoes (strategy I)

Cu uptake

A large enrichment of the light Cu isotope, relative to the growth medium, is seen in both Fe-sufficient and Fe-deficient tomato plants (Fig. 1). The uptake of Cu into tomato plants induces a fractionation of c. −1.00‰ in both treatments, in line with previously observed biologically induced Cu(II) to Cu(I) reduction during uptake (Δ65Cubacteria-solution = −1.2‰ to −4.4‰; Navarrete et al., 2011).

Jouvin et al. (2012) also found light Cu isotope enrichment in the roots of hydroponically grown strategy I plants (−0.84‰ to −0.63‰). These values were Δ65Curoot-solution values, for unwashed roots, relative to solution, and probably contained apoplastically sorbed, isotopically heavy Cu. Adsorbed, heavy Cu would have masked some of the light isotope enrichment present in the root symplasm, perhaps explaining why the light isotope enrichment found by Jouvin et al. (2012) was less than that observed in this study. Moreover, Jouvin et al. (2012) used chelate buffering with nitrilotriacetic acid (NTA), ethylenediaminetetraacetic acid (EDTA) or N-hydroxyethylenediaminetriacetic acid (HEDTA) as Cu-complexing ligands. There may have been direct, passive uptake of the Cu complexes from solution (Iwasaki & Takahashi, 1989; Laurie et al., 1991; Collins et al., 2002), which might also explain the less negative fractionation found by Jouvin et al. (2012). Regardless of these experimental differences, the large light Cu isotope enrichment in both studies is evident, and suggests that Cu(II) in hydroponic solutions is reduced to Cu(I) at the root membrane before it is taken up.

Reductive uptake of Cu in strategy I plants is reasonable, given the prevalence of the high-affinity Cu transporter protein (COPT) family in roots, which is a selective Cu(I) transporter (Sancenon et al., 2003; Yruela, 2009). It is likely that Cu is being reduced by the Fe reductases FRO2 and FRO3, previously reported to be capable of reducing Cu(II) (Bernal et al., 2012). The upregulation of Fe reductases may explain the significantly higher total Cu concentration in tomato roots in the Fe-deficient treatment (Fig. 3a), because of increased Cu reduction and uptake via COPT transporters.

The Cu EXAFS data for Fe-deficient roots indicate that the majority of Cu is being stored in small S-bridged Cu clusters (Fig. 5). These observed structures are probably reflections of strong metal complexes, such as metallothioneins and phytochelatins, both rich in cysteine, and glutathione as a cofactor, which are upregulated under excess Cu to buffer cell Cu concentrations (Cobbett & Goldsbrough, 2002). The root Cu EXAFS data in Fig. 5 (and Figs S5, S6) appear to be analogous to those presented for Cu-phytochelatin complexes by Polette et al. (2000), including the Cu–Cu interaction at 2.7 Å. The increase in Cu concentrations in roots coincides with an increase from 0% Cu(I)-cysteine species in Fe-sufficient plants to 20% in Fe-deficient plants (Fig. 3a, Table S5). This may indicate the stimulation of metallothionein/phytochelatin synthesis.

XAS analysis of Cu in cowpea roots after 24 h of exposure to 1.5 μM Cu found Cu to be bound predominantly to polygalacturonic acid, an O-rich ligand found in the cell wall of the rhizodermis and outer cortex (Kopittke et al., 2011). Given that the cell wall is the site of apoplastic binding and transport, and the root washing procedure employed in the current study was designed to remove adsorbed apoplastic Cu, it is not surprising that little Cu binding to O-rich ligands was observed in the current study. Kuepper et al. (2009) noted that the cell wall and vacuoles, which are thought to be the site of Cu storage in Cu hyperaccumulator plants, lack strong S-rich ligands, such as metallothioneins, leaving only weak O-rich ligands. The current findings showed a limited amount of O-rich Cu binding, suggesting a limited role for the vacuole in Cu storage, and cell wall-bound Cu should have been removed during root washing. Co-precipitates of Fe and oxidized Cu may have formed on the root surface, known as Fe plaques, but, again, the lack of Cu(II) observed by XAS led to the conclusion that this was not a dominant Cu storage mechanism. The XAS data support the hypothesis that non-Cu accumulator plants store excess Cu in S-rich metallothionein-type structures, as suggested by Mijovilovich et al. (2009).

Cu translocation

The Δ65Cushoot-solution values were c. −0.40‰, compared with the whole-plant value of c. −1.0‰, for both the Fe-sufficient and Fe-deficient treatments (Table 2). The similar Cu(II) concentrations and Δ65Cushoot-solution values across both treatments indicate a highly regulated mechanism of Cu transport between roots and stems, and, possibly, a highly regulated re-oxidation of Cu(I) to Cu(II), given the consistent amount of Cu(II) in roots, despite the very different total Cu concentrations (Fig. 3).

Root to shoot transport of Cu in tomato plants is dependent on the mugineic acid-derived metal chelate, nicotianamine (NA; Pich & Scholz, 1996; Curie et al., 2009). The chloronerva tomato mutant is incapable of synthesizing NA and was found to accumulate Cu in its roots, whereas vegetative tissues suffered Cu deficiency, as long-range Cu transport was severely impaired (Pich & Scholz, 1996). Liao et al. (2000) found that < 0.05% of Cu existed in tomato xylem as free Cu(II), with essentially all Cu present as Cu-NA. Given that NA strongly complexes Cu(II), oxidation of Cu(I) in roots would need to occur for NA complexation and subsequent translocation. Both oxidation and complexation products tend to be enriched in the heavier isotope, under equilibrium conditions, given the increased thermodynamic stability achieved with a heavier mass; however, oxidation induces much stronger heavier isotope enrichment than does complexation (Δ65CuCu(II)-Cu(I) = +1.9 to +5.3‰ vs Δ65Cucomplex-free = +0.27‰; Mathur et al., 2005; Asael et al., 2007; Balistrieri et al., 2008; Bigalke et al., 2010b). Isotopically heavy Cu found in the shoots of this study may suggest selective transport of oxidized Cu(II) or Cu(II)-NA to the xylem for translocation. Further investigations are required to confirm this proposed mechanism of transport.

In contrast with our study, Jouvin et al. (2012) found isotopically lighter Cu (Δ65Cushoot − Δ65Cushoot-root = −0.37 to −0.10‰) in shoots compared with roots. As apoplastic Cu was not desorbed in their study, isotopically heavy Cu on the roots could have masked the extent of light Cu isotope enrichment in the roots. Investigations are currently underway to quantify the degree of heavy isotope enrichment caused by root adsorption of Cu. In addition, the translocation mechanisms used by the plant may be concentration specific and, given that our plants were exposed to, and contained, higher concentrations of Cu, this may have affected the translocation mechanism employed. However, both studies indicate the storage of Cu in roots and the strong regulation of Cu translocation to shoots.

The Δ65Cu values of (green) leaves were more negative than those of the stem, indicating enrichment of the light isotope as Cu is translocated from stem to leaves (Table 2; Δ65Culeaf – Δ65Culeaf-stem = −0.53‰ and −0.34‰ for Fe-sufficient and Fe-deficient plants, respectively). This is supported by the presence of the Cu(I) COPT transporter protein in leaf tissues (Burkhead et al., 2009) and the XANES data, which suggest that c. 80% of all Cu present in the leaves of both treatments is Cu(I) (Fig. 3b). In addition, dissociation of the Cu(II)-NA complex for Cu transport into leaf tissues would also favour the lighter isotope.

Translocation fractionation between green and chlorotic leaves appears to be different, with no apparent isotopic discrimination observed between chlorotic leaves and stems (Δ65Cuchlorotic leaves − Δ65Cuchlorotic leaves-stem = −0.04‰; Table 2). This indicates that there may be two different mechanisms of Cu transport at play in green and chlorotic leaves. It is not possible to determine whether this difference between chlorotic (young) and green (old) leaves is a result of the imposed Fe deficiency or simply a consequence of harvesting young and old leaves separately, as old and new leaves were not separated from the Fe-sufficient plants for comparison. Differences in isotopic signature may be related to the relative fraction of Cu contained in the given tissues. Green leaves are older and contain Cu that was translocated early in growth, when relatively more Cu was translocated from roots to shoots. As growth progresses, the roots begin to take up an excess of Cu which is stored in the roots, whereas Cu in the shoots continues to be regulated. This results in Fshoot becoming smaller, and hence δ65Cushoot becoming more positive. This could potentially cause a lighter isotope enrichment in older leaves, relative to young leaves, as predicted by a Rayleigh curve (Notes S1).

Conceptual model of Cu uptake and translocation

The uptake and translocation of Cu in strategy I tomato plants appears to be driven by redox-selective processes, as evidenced by the strong Cu isotope fractionations between the nutrient solution and different plant tissues. XAS data indicated a dominance of Cu(I) species throughout roots and leaves, and suggested that excess root Cu is being stored in detoxifying small Cu clusters, bridged by S ligands. A conceptual model of the uptake and translocation processes for tomato plants based on the findings of this study, combined with the known importance of Cu-NA for Cu transport through vascular tissue from the literature (Pich & Scholz, 1996; Curie et al., 2009), is presented in Fig. 4(a). We propose that the uptake of Cu occurs predominantly as Cu(I), with oxidation to Cu(II) required for xylem loading, followed by complexation, most probably with NA, for translocation.

Figure 4.

(a) A conceptual model of the uptake and translocation mechanisms for Cu in strategy I tomato (Solanum lycopersicum) plants based on the findings in this study. (b) A conceptual model of the uptake and translocation mechanisms for Cu in strategy II oat (Avena sativa) plants based on the findings in this study.

Oat plants (strategy II)

Cu uptake

Oat plants showed an overall small enrichment in the light Cu isotope, relative to the nutrient solution, with no significant difference between Fe treatments (Fig. 1). The observed Cu isotope fractionations for oat plants are similar to the Δ65Curoot-solution values of wheat and rice reported by Jouvin et al. (2012) (−0.48‰ to −0.11‰). The small fractionation of Cu isotopes on uptake into oat plants is too small to be attributed to a reductive uptake mechanism, which induces larger fractionation (Navarrete et al., 2011; Jouvin et al., 2012). ZIP divalent cation transporters have been suggested as potential Cu(II) transporters (Wintz et al., 2003; Schaaf et al., 2004). They are mostly nonspecific for Cu, and hence do not involve specific Cu binding. Uptake into oat roots may also be affected by the exclusive strategy II phytosiderophore complexation mechanism. Heavy Cu isotopes would be expected to preferentially form the Cu(II)-phytosiderophore complex (Guelke & Von Blanckenburg, 2007; Bigalke et al., 2010a,b); however, absorption, possibly via the Fe-phytosiderophore membrane yellow-stripe-like (YSL) transporter, should not induce any significant Cu isotope fractionation, given that the complex is too large for mass-dependent fractionation (Schaaf et al., 2004; Palmer & Guerinot, 2009). The production of phytosiderophore is enhanced under Fe deficiency, given that it is predominantly an Fe-acquiring mechanism; hence, the increased Cu uptake under Fe deficiency may have resulted from an increase in Cu-phytosiderophore absorption. A combination of these mechanisms may be contributing to Cu uptake in oats.

XANES Cu speciation analysis found that the majority of Cu in oat roots for both Fe treatments was present as Cu(I) species, as was the case for tomato roots (Fig. 3a). However, unlike tomatoes, no evidence for reduced Cu uptake was found. EXAFS data indicate that the root Cu is bound with S-containing ligands in small Cu clusters, as found for the Fe-deficient tomato plants (Figs S5, S6). This suggests that Cu(II) was taken up into roots and, once absorbed, was reduced and stored as Cu(I) metallothionein/phytochelatin-type complexes as a method of detoxification, given the excess amount of Cu taken up into the roots. This suggests that, although the uptake mechanisms for strategy I and II plants differ, the method of storing and detoxifying excess Cu is similar in these nonaccumulator plants.

Figure 5.

Example extended X-ray absorption fine structure (EXAFS) spectra (left) and Fourier transform (right) of tomato (Solanum lycopersicum) roots grown in an iron (Fe)-deficient nutrient solution. Black line, experimental data; green line, fitted data. Fitted parameters can be found in Table 3.

Cu translocation

The oat plants studied showed a constant Cu isotope ratio, within error, throughout the plant, indicating that no significant fractionation of Cu isotopes occurred during translocation (Δ65Cushoot – Δ65Cushoot-root = −0.04‰ and +0.04‰ for Fe-sufficient and Fe-deficient plants, respectively; Table 2). Similar results have been found for Fe in other strategy II plants (wheat) with a constant isotopic signature throughout the plant, whereas strategy I plants showed significant fractionation (Guelke & Von Blanckenburg, 2007; Guelke-Stelling & von Blanckenburg, 2012). These authors suggested that complexation-driven processes that lacked redox cycling, or that involved only quantitative redox changes, were responsible for the relatively conserved Fe isotope signature. Given the similar lack of redox cycling observed for Cu, it is suggested that, in order to conserve the Cu isotope signature between roots and stems, translocation of Cu(I) present in roots occurs, or Cu(I) is quantitatively oxidized to Cu(II) after moving into the xylem; that is, no selection for oxidized species occurs at the root–stem boundary, and hence no fractionation is observed. Further investigations are required to confirm this proposed mechanism for Cu transport.

The translocation of Cu from stem to leaves (green), as determined by Δ65Culeaves – Δ65Culeaves-stem, showed no Cu isotope fractionation (Δ65Culeaves − Δ65Culeaves-stem = −0.02‰ and 0‰ for Fe-sufficient and Fe-deficient plants, respectively; Table 2). The XANES data for oat leaves indicate that 22% of the Cu in Fe-sufficient leaves and 14% of the Cu in Fe-deficient leaves is Cu(II), with the majority of Cu present as Cu(I) (Fig. 3b). It is unclear what mechanisms are responsible for the translocation of Cu from stems to leaves in the oat plants, but the small Cu isotope fractionation observed suggests that it is not a redox-selective process and, possibly, more of a complexation-driven mechanism.

As observed with strategy I tomato plants, Cu isotope fractionation following the translocation of Cu from the stems was different when moving Cu to green leaves vs chlorotic leaves (0‰ and −0.27‰ for green and chlorotic leaves, respectively; Table 2). Young, chlorotic leaves contained isotopically light Cu, as opposed to the lack of isotopic discrimination in chlorotic tomato leaves. The reason behind this difference between the two species is unclear at this stage. Evidently, translocation between roots and shoots is significantly different between the two plant species studied; hence, it is reasonable to conclude that the translocation mechanisms between stems and leaves would also differ.

Conceptual model of Cu uptake and translocation

The uptake and translocation of Cu in strategy II oat plants does not appear to be driven by redox-selective processes, but by complexation-driven processes. As is the case with tomatoes, XAS found Cu(I) to be the dominant Cu species present in the roots and leaves of oat plants. A conceptual model of the uptake and translocation processes for Cu in oats, based on the findings in this study, is presented in Fig. 4(b). More research is required to elucidate the mechanisms driving Cu translocation in strategy II plants.

Concluding remarks

The results obtained here align well with findings reported previously for Fe uptake and translocation in strategy I and II plants (Guelke & Von Blanckenburg, 2007). The Cu concentrations used in this study were high to enable all Cu to remain as free Cu2+ for clear interpretation of the isotope data, and to generate plants with sufficiently high Cu concentrations to allow XAS analysis. Although further research is needed to elucidate the role of Cu concentration on the uptake and translocation of Cu in soil–plant systems, it is likely that the uptake mechanisms suggested in this study hold for monocots and dicots at various Cu concentrations. This has been shown with Fe, where dicots continually have significantly lighter Fe isotopic signatures than monocots, although the monocot δ56Fe appears to become slightly more negative during high Fe supply, because of the use of the strategy I mechanism under Fe-sufficient conditions (Guelke & Von Blanckenburg, 2007; Kiczka et al., 2010; Guelke-Stelling & von Blanckenburg, 2012). Although uptake mechanisms are not expected to change drastically with Cu supply, Cu speciation at lower Cu concentrations may. Cu-S species would probably be less dominant, given that less immobilization of Cu would be required.

The combination of isotopic and intact tissue speciation techniques has provided useful, new information on the behaviour of Cu in plant systems. Both species studied store and immobilize Cu using similar Cu-bridged S complexes that resemble metallothioneins. However, the uptake and translocation mechanisms employed by the strategy I and II species studied are evidently different, with tomatoes relying heavily on redox-selective transport, and with oats showing no sign of dominant redox selection in either uptake or translocation. To the authors' knowledge, this is the first identification that different Cu uptake and translocation mechanisms may exist between monocot and dicot plant species. Further studies using molecular techniques will be required to confirm the suggested mechanisms.


We acknowledge travel funding provided by the International Synchrotron Access Program (ISAP) managed by the Australian Synchrotron and funded by the Australian Government. Technical support from Claire Wright and Claire Weekley has been indispensible and is greatly appreciated.