Metal ions such as iron (Fe), zinc (Zn), manganese (Mn) and copper (Cu) are essential for healthy plant growth, being required for structural and catalytic roles in proteins involved in metabolism and development. However, an excess of free metal ions is toxic to cells because of the generation of reactive oxygen species or through displacement of other metal ions from sites within metalloproteins rendering them nonfunctional. Therefore plants have mechanisms for metal homeostasis that allow uptake and distribution of metals to tissues while maintaining metals within cells or subcellular compartments below levels that cause toxic symptoms (Clemens, 2001). Over the past decade much progress has been made in identifying the components of these mechanisms, in particular the transporters of free ions and, to a lesser extent, the factors that regulate these transporters. These have been reviewed recently (Grotz & Guerinot, 2006).
Very little cellular metal ion content is expected to exist as free ions. For example, in Escherichia coli, free Zn2+ was calculated in the femtomolar range, despite a millimolar total concentration in whole cells (Outten & O’Halloran, 2001). Metal ions not occupying sites in proteins are expected to be bound to low molecular-weight metal ligands. Metal ligands can have intracellular roles as chelators for sequestering metal ions in the cytosol or in subcellular compartments or, for metals with low solubility such as Fe, they may be used either for mobilization from the soil or for translocation within the plant. In order to fulfil these roles, transport mechanisms are required for the secretion or uptake of metal ion ligands by cells or for the movement of ligands between subcellular compartments. Research over the past 5 yr has made progress in the identification and characterization of some of these transporters and has provided insight into the role of a number of metal ion ligands in metal homeostasis. The focus of this review is to describe recent advances in elucidating the role of small, organic molecules as metal-binding ligands and the identity of proteins implicated in their transport. In particular, the contribution of mugineic acid (MA), nicotianamine (NA), organic acids, histidine and phytate (Fig. 1) to metal homeostasis and the proteins implicated in their transport as free ligands or as metal–ligand complexes are described. In the absence of a candidate for transport of the ligand the evidence for its existence is discussed. The enzymatically synthesized metal-binding peptide phytochelatin and its transport are not discussed here as little evidence supports a role for this ligand in essential metal homeostasis. For a review see Cobbett & Goldsbrough (2002).
Mugineic acid and nicotianamine
Fe3+ is relatively insoluble in alkaline soils and two alternative mechanisms for Fe acquisition have evolved in plants. Nongraminaceous monocots and dicots, the ‘strategy I plants’, acidify the rhizosphere (presumably via an H+-ATPase) to increase Fe solubility and use a ferric-reductase to reduce Fe3+ to Fe2+ which is transported into roots via an Fe2+ transporter. Graminaceous species utilize strategy II whereby a metal-binding ligand, MA, is synthesized enzymatically (Fig. 2) from three molecules of S-adenosyl methionine and is secreted from roots to bind Fe3+ in the rhizosphere. The Fe(III)–MA then enters the roots via a specific transporter (Curie & Briat, 2003). In maize, this transporter was identified using the yellow-stripe1 (ys1) mutant, which is deficient in uptake of Fe–MA and has an Fe-deficiency phenotype. YS1 was the first transporter of a metal-ion ligand identified in plants (Curie et al., 2001) and is a member of the oligopeptide-transporter family, members of which transport tri-, tetra-, penta- and hexapeptides (Yen et al., 2001). The ys1 phenotype indicates a primary defect in Fe transport and heterologous expression in Xenopus oocytes demonstrated that YS1 functions as an Fe(III)–MA/H+ symporter (Schaaf et al., 2004). Roberts et al. (2004) also demonstrated Fe(III)–MA transport mediated by YS1 by complementation of an Fe uptake-deficient yeast strain. YS1 may also play a role in the homeostasis of other metals as it could also transport Zn, Cu or Ni as MA complexes (Schaaf et al., 2004). Although competition assays for Fe-uptake in yeast failed to detect transport of Zn-, Ni- or Cu–MA by YS1, Co–MA was identified as a possible substrate (Roberts et al., 2004). Upregulation of transcripts encoding enzymes involved in MA synthesis (Fig. 2) in response to both Fe and Zn deficiency has been demonstrated in barley roots and correlated with increased secretion of MA into the rhizosphere. Furthermore, under these conditions, Zn was more readily absorbed when supplied as Zn–MA than as Zn2+ (Suzuki et al., 2006).
Strategy I plants, such as Arabidopsis and tobacco, do not synthesize MA but do synthesize NA, which is a precursor for MA (Fig. 2). Nicotianamine is present in all plants and, like MA, is believed to play a primary role in Fe homeostasis but is not secreted and is believed to function within the plant to maintain the solubility of Fe (Curie & Briat, 2003). There is also increasing evidence for a role for NA in Zn and Cu homeostasis. Nicotianamine synthase (NAS) genes are upregulated in roots and shoots of plants grown under Zn or Cu deficiency, as well as under Fe deficiency (Wintz et al., 2003). In addition, NAS transcripts were more highly expressed in roots and shoots of the Zn-hyperaccumulating species Arabidopsis halleri compared with Arabidopsis thaliana, and heterologous expression of AhNAS2 or AhNAS3 in Schizosaccharomyces pombe and Saccharomyces cerevisiae, respectively, conferred increased tolerance to Zn, suggesting a direct role for NA in Zn tolerance in vivo (Becher et al., 2004; Weber et al., 2004). Comparative microarray experiments between A. thaliana and A. halleri also identified S-adenosyl methionine synthetase as being more highly expressed in A. halleri, further suggesting upregulation of NA biosynthesis in the Zn hyperaccumulator (Talke et al., 2006).
Recent evidence for the role of NA in planta comes from depletion of NA in tobacco by transgenic overexpression of nicotianamine aminotransferase (NAAT) from barley. The levels of Fe, Cu and Zn decreased in leaves and floral organs of transgenic plants, suggesting a role for NA in long-distance translocation of these metals. Leaves of transgenic lines exhibited interveinal chlorosis and synchrotron radiation-induced X-ray fluorescence spectrometry indicated impaired translocation of Fe, but not Zn, into interveinal regions of leaves, suggesting that NA is required for unloading Fe from vascular tissues. In a reciprocal experiment, overexpression of barley NAS in transgenic tobacco led to increased Fe, Zn and Cu content in leaves and flowers and enhanced the Fe and Zn content of pollen and seeds, further supporting a role for NA in transport of these metals (Takahashi et al., 2003). This is consistent with calculations for the likely ligands in vascular tissues, based on the metal and ligand composition of xylem and phloem, which suggested that NA should complex all Fe, Zn and Cu in the phloem and all Zn and Cu in the xylem (von Wiren et al., 1999). Nicotianamine may also be important for Ni tolerance with a number of recent studies showing that transgenic overexpression of NAS in Arabidopsis or tobacco confers increased tolerance to Ni (Douchkov et al., 2005; Kim et al., 2005; Pianelli et al., 2005). Furthermore, using high-performance liquid chromatography coupled to inductively coupled plasma mass spectrometry (HPLC–ICPMS) and electrospray MS/MS, Ni–NA complexes have been detected in Ni-exposed roots of the Zn-hyperaccumulator Thlaspi caerulescens (Vacchina et al., 2003).
YS1 can transport Fe–NA and Ni–NA complexes in addition to metal–MA complexes (Roberts et al., 2004; Schaaf et al., 2004). There are eight YELLOW STRIPE-LIKE (YSL) genes in A. thaliana, even though MA is not synthesized by this species, and it was suggested that the Arabidopsis orthologues may transport metal–NA complexes. Consistent with this AtYSL2 could complement yeast Fe- and Cu-uptake, but not Zn-uptake mutants, when metals were supplied as NA complexes (DiDonato et al., 2004). By contrast, Schaaf et al. (2005) were unable to detect transport by AtYSL2 of any metal–NA or metal–MA complex in Xenopus oocytes or in yeast. Nevertheless, YSL2 expression was downregulated when availability of Fe or Zn was low or when Cu concentration was high, consistent with a role in metal homeostasis. YSL2 promoter–GUS fusions showed highest expression in vascular parenchyma, in the pericycle and in endodermal cells in roots and in leaf veins. YSL2-GFP localized only to lateral plasma membranes, suggesting a possible role in movement of metal–NA complexes into or out of vasculature (DiDonato et al., 2004; Schaaf et al., 2005). A ysl2 T-DNA mutant presented no visible phenotype and was not affected in metal accumulation (DiDonato et al., 2004), suggesting that the function of YSL2 in Arabidopsis is redundant under the conditions tested. In rice, OsYSL2 is expressed in phloem cells in roots and leaves and in developing embryos, and is upregulated in Fe-deficient plants. Fe–NA and Mn–NA, but not other metal–NA or metal–MA complexes, generated currents in Xenopus oocytes expressing OsYSL2 suggesting a role for YSL2 in phloem transport and seed loading of Fe and Mn as NA complexes (Koike et al., 2004).
AtYSL1 is expressed predominantly in vascular tissues of leaves and flowers, with low expression in roots, and is upregulated in the presence of Fe. Expression was also observed in the vasculature of siliques, in pollen grains and in developing seeds. The seeds of ysl1 mutants had decreased levels of Fe and NA, but not of Cu, Mn or Zn. In addition, ysl1 seeds had slower germination rates on medium with low Fe availability (Le Jean et al., 2005). YSL3 is also expressed in vascular tissues of roots, leaves and flowers, with high expression in pollen, and is expressed more highly in plants grown in the presence of Fe. A ysl1 ysl3 double mutant showed decreased Fe content in roots and leaves and exhibited interveinal chlorosis when grown on soil, and this phenotype was suppressed by supplementation with Fe, indicating that the phenotype is caused by an Fe deficiency. In addition, fertility of ysl1 ysl3 was impaired and seed yield was low. Fe, Zn and Cu seed content were lower in the double mutant and seed yield could be improved by Fe supplementation, although not to wild-type levels. YSL1 and YSL3 are also upregulated during leaf senescence and mobilization of Cu and Zn (although not Fe) was decreased in ysl1 ysl3 senescing leaves (Waters et al., 2006). Experimental evidence for the subcellular localization or substrate specificity of YSL1 or YSL3 has not been reported. However, given that many YSL transporters confer uptake of metal–NA complexes in heterologous systems, it seems likely that these transporters also mediate uptake of metal–NA across the plasma membrane. The evidence summarized above suggests that these proteins are involved in the unloading of metal–NA from vasculature into developing tissues, in mobilization of metal–NA from senescent leaves, and in efficient loading of metal–NA into seeds.
Three YSL transporters have been identified in the Zn-hyperaccumulator T. caerulescens. TcYSL3 and TcYSL5 showed constitutively elevated expression in both roots and shoots compared with expression of the orthologues in A. thaliana while TcYSL7 showed high expression only in shoot tissues of the hyperaccumulator. TcYSL3 and TcYSL7 were detected in xylem parenchyma and phloem by in situ hybridization, which is consistent with a role in vascular loading or unloading. Exposure to Ni, but not Zn or Cd, led to decreased expression of TcYSL5 and TcYSL7 in roots and a transient increase in TcYSL7 expression in shoots although regulation of these transcripts by Fe was not tested. Heterologous expression of TcYSL3, but not TcYSL5 or TcYSL7, enhanced growth of an Fe-uptake yeast mutant when Fe was supplied as Fe–NA and yeast expressing TcYSL3 showed increased uptake of Fe or Ni when supplied as metal–NA (Gendre et al., 2006). Together, these data suggest that YSL transporters in T. caerulescens may play a role in metal hyperaccumulation or tolerance in this species by participating in vascular transport of metal–NA complexes.
Nicotianamine may also be important for sequestration of Fe in vacuoles. Immunohistochemical detection of NA in pea and tomato suggested that under excess Fe supply, NA concentrations increase and accumulate in vacuoles (Pich et al., 2001). This suggests a requirement for NA or Fe–NA transport into vacuoles and YSL transporters may also fulfil these roles. It should be noted, however, that this could not occur by proton symport, which is the mechanism of transport mediated by YS1. Therefore, if YSL transporters localize to internal membranes, it would seem more likely that they would function to mobilize metal–NA from these subcellular compartments.
Organic acids have been associated with metal hyperaccumulation and tolerance in a range of plant species and have been proposed as important cellular ligands for Zn, Cd and Ni (Rauser, 1999). Analysis of tissues from metal-hyperaccumulator species using X-ray absorption techniques has identified organic acids as the predominant ligands. By X-ray absorption spectrometry (XAS) and extended X-ray absorption fine structure (EXAFS) analysis, citrate was identified as the predominant ligand for Zn in leaves of T. caerulescens (Salt et al., 1999; Kupper et al., 2004) while Zn-malate was the major Zn species in aerial tissues of A. halleri (Sarret et al., 2002). Similarly, Ni-citrate accounted for one quarter of the Ni species in leaves of the Ni-hyperaccumulator Thlaspi goesingense and in the related nonaccumulator Thlaspi arvense (Kramer et al., 2000). The identification of the vacuole as the major subcellular compartment for Zn, Cd and Ni (Kramer et al., 2000; Ma et al., 2005) and favouring of the formation of metal–organic acid complexes in the acidic environment of the vacuolar lumen suggest that citrate and malate are probably relevant only as ligands for these metals within vacuoles. In addition, secretion of malate or citrate from root apices is a well established mechanism for tolerance to Al in a range of plant species (Delhaize & Ryan, 1995) and citrate is believed to be the predominant ligand for Fe in xylem (von Wiren et al., 1999).
Although organic acids have been identified as metal ligands in vacuoles, it is not clear how these ligands are sequestered. AttDT (tonoplast dicarboxylate transporter) was identified by similarity to the Na+/dicarboxylate symporters from mammals and was shown to localize to the tonoplast. T-DNA mutants in AttDT contained lower malate content in bulk leaf tissue and had decreased malate transport across the tonoplast and accumulated less malate in isolated intact vacuoles compared with the wild type but were unaffected in transport of citrate into vacuoles (Emmerlich et al., 2003; Hurth et al., 2005). If malate is an important ligand for Zn or other metals in vacuoles of A. thaliana these mutants might be affected in tolerance to these metals. Interestingly, AttDT was among a number of transcripts identified as an Al-induced gene located within a quantitative trait locus (QTL) for Al tolerance in A. thaliana (Hoekenga et al., 2003).
An alternative candidate for a vacuolar metal ligand transporter in Arabidopsis is ZINC-INDUCED FACILITATOR1 (ZIF1). zif1 mutants are hypersensitive to Zn and ZIF1 localized to the tonoplast, suggesting a possible role in vacuolar sequestration of Zn. ZIF1 is expressed throughout developing tissues and is induced in plants exposed to Zn. ZIF1 is a member of the major facilitator superfamily (MFS), which transport a wide range of small, organic molecules, and is most similar to the drug/H+ antiporters from bacteria (Haydon & Cobbett, 2007). Although the substrate has not been identified, ZIF1 may transport a Zn ligand, or a Zn-ligand complex, across the tonoplast, and one possibility is an organic acid.
A well characterized malate transporter in plants is Al-activated malate transporter1 (ALMT1) which was first cloned from wheat (Hordeum vulgare) and mediates Al-activated efflux of malate in Xenopus oocytes, cultured tobacco cells and transgenic rice. High expression of ALMT1 in root apices of Al-tolerant wheat strains cosegregated with tolerance in F2 and F3 populations of crosses between a tolerant and nontolerant line (Sasaki et al., 2004). Although no identifiable homologues exist in bacterial or mammalian genomes, 14 similar proteins exist in Arabidopsis. AtALMT1 is specifically expressed in root epidermis and heterologous expression in Xenopus oocytes demonstrated Al-enhanced efflux of malate by AtALMT1. An Atalmt1 mutant failed to excrete malate from roots in the presence of Al and was more sensitive to Al compared with the wild type. Although AtALMT1 is located close to a QTL for Al tolerance in Arabidopsis, complementation experiments between the Al-sensitive Landsberg ecotype and almt1 in the Al-tolerant Columbia background indicated that it was not the gene responsible for the QTL (Hoekenga et al., 2006). The other 13 ALMT proteins in Arabidopsis are yet to be characterized but may be involved in transport of organic acids in various tissues with implications for metal homeostasis.
Ferric reductase defective3 (frd3) mutants are allelic to manganese accumulator1 (man1) and exhibit leaf chlorosis and constitutive activation of components of strategy I Fe-acquisition, including rhizosphere acidification and expression of ferric reductase and the Fe-uptake protein IRT1, consistent with Fe deficiency. The constitutive activation of Fe uptake leads to overaccumulation of Mn and Zn in aerial tissues, but not Fe, which accumulates to high concentrations in roots (Delhaize, 1996; Rogers & Guerinot, 2002). FRD3 encodes a member of the multidrug and toxin efflux (MATE) family of transporters and is expressed in root pericycle and xylem parenchyma. In roots of the frd3 mutant, Fe is localized predominantly to the central vascular cylinder, indicating a failure either to mobilize Fe to aerial tissues or to unload Fe from vasculature. In addition, the proportion of shoot Fe in the symplasm is lower in frd3 compared with wild type, suggesting that the Fe was delivered apoplastically to shoot tissue rather than being unloaded into the symplasm from the xylem (Green & Rogers, 2004). Together, these observations suggest that FRD3 is necessary for unloading Fe from vascular tissue, notwithstanding that FRD3 is expressed only in roots. As MATE proteins transport small, organic molecules, the FRD3 substrate may be an Fe ligand, or Fe-ligand complex, that is required for efficient unloading of Fe from the vasculature into leaf tissue. More recent experiments have demonstrated that xylem exudate of frd3 contains lower levels of citrate. Citrate also generated currents in Xenopus oocytes expressing FRD3 and overexpression of FRD3 in transgenic Arabidopsis confers tolerance to Al. These data suggest a role for FRD3 as a citrate exporter (E. E. Rogers and co-workers, unpublished data), in agreement with the suggestion that citrate is the major ligand for Fe in xylem (von Wiren et al., 1999).
Histidine has a high affinity for binding metals both as the free amino acid and as metal-coordination residues in proteins. In particular, free His has been implicated as an important ligand in Ni hyperaccumulation. A linear, Ni-induced increase in His content was detected in xylem exudate of the hyperaccumulator Alyssum lesbiacum and, in plants exposed to Ni, was 36-fold higher in the xylem sap of A. lesbiacum compared with the nonaccumulator Alyssum montanum. XAS indicated coordination of His to Ni, and foliar application of His to A. montanum conferred enhanced Ni tolerance (Kramer et al., 1996). The elevated His content of A. lesbiacum has been associated with constitutively high expression of enzymes for His biosynthesis, in particular ATP-PRT1, which encodes the first committed step of the biosynthetic pathway. Overexpression of ATP-PRT from A. lesbiacum in transgenic A. thaliana conferred enhanced Ni tolerance, although not increased Ni accumulation (Ingle et al., 2005). Further studies using A. lesbiacum and a nonaccumulator, Brassica juncea, showed that Ni exposure led to increased concentrations of Ni and His in xylem exudate in both species. However, this did not correlate with increased total His in roots, suggesting a redistribution of existing pools of His in roots, rather than increased synthesis. Although it is possible that His acts to increase mobility of Ni in root symplasm, Ni–His may be the form that is loaded into xylem via a specific transporter. Indeed, addition of an inhibitor of translation, cycloheximide, greatly decreased xylem loading of both Ni and His, indicating that loading from root symplasm into xylem requires de novo synthesis of Ni-induced proteins, and this may include an as-yet unidentified Ni–His transporter (Kerkeb & Kramer, 2003).
Although organic acid complexes have been suggested to account for the majority of Zn in aerial tissues, X-ray absorption studies using T. caerulescens indicated that Zn–His was the second most abundant ligand in mature leaves and that His is complexed to 70% of Zn in roots (Salt et al., 1999). In addition, studies of the Zn species in a range of aerial tissues of T. caerulescens at various developmental stages indicated that the speciation of Zn is dynamic, and suggested that His may be more important in chelating Zn in developing and older tissues. This may be related to a shift in the cellular or subcellular distribution of Zn at different developmental stages. For example, a higher proportion of Zn might be localized to the cytosol in developing and senescent tissues as compartments are loaded and dismantled, respectively, requiring a strong ligand, such as His, to chelate free Zn ions at higher, cytosolic pH, rather than the weak interactions with organic acids at lower pH in vacuoles (Kupper et al., 2004).
Phytate myo-inositol hexakisphosphate (phytate) is primarily a storage molecule for phosphorus (P) but has also been implicated in binding and storage of metals, particularly Zn, in roots of a number of Zn-tolerant species (Rauser, 1999). There is a positive correlation between Zn and P concentrations in roots of A. halleri and Arabidopsis lyrata and EXAFS studies suggested that Zn is coordinated to P as Zn–phytate in roots of both species and in leaves of A. lyrata (Sarret et al., 2002). Measurement of phytate in vacuoles and protoplasts of cultured Catharanthus roseus cells by ion chromatography indicated that phytate is synthesized in the cytosol and accumulated in vacuoles. Phytate production was increased following the addition of inorganic phosphate to the culture medium and was further induced when Mg, K or Zn was also supplied. However, the Zn-induced phytate did not accumulate in vacuoles but formed insoluble aggregates in nonvacuolar compartments, presumably the cytosol (Mitsuhashi et al., 2005).
Energy-dispersive X-ray spectroscopy (EDXS) analysis has also implicated metal–phytate complexes as a transient storage form of Mn and Zn in developing seeds of Arabidopsis. Mn–phytate was identified in the endoplasmic reticulum (ER) and Zn–phytate in vacuolar compartments of the chalazal endosperm which is believed to play a role in loading of minerals into the developing embryo. The minerals are mobilized at different stages of embryo development with Zn–phytate crystals disappearing from the endosperm early, between the globular and heart stages, while Mn–phytate was mobilized from the endosperm later, during bent-cotyledon stages. The disappearance of Mn–phytate coincided with increased Mn content and the expression of Mn-requiring proteins in embryos (Otegui et al., 2002). As phytate is believed to be synthesized in the cytosol, the compartmentalization observed in the ER and vacuoles suggests a mechanism for phytate transport into these organelles. INOSITOL TRANSPORTER4 encodes a plasma membrane-localized, high-affinity H+/myo-inositol symporter that is one of four related MFS proteins from Arabidopsis (Schneider et al., 2006) but no plant protein has been identified that mediates subcellular sequestration of phytate.