|Home | About | Journals | Submit | Contact Us | Français|
Fe is essential for plant growth. At the same time, Fe is highly reactive and toxic via the Fenton reaction. Consequently, plants tightly control Fe homeostasis and react to Fe deficiency as well as Fe overload. The ability of plants to respond to Fe availability ultimately affects human nutrition, both in terms of crop yield and the Fe concentration of edible tissues. Thus, elucidating the mechanisms of Fe uptake and transport is essential for the breeding of crops that are more nutrient rich and more tolerant of Fe-limited soils.
This review covers Fe transport and homeostasis in plants, focusing on the research published in the past five years. Because Fe transporters often have a broad range of substrates, we also examine the relationship between Fe and the toxic metals that often accompany Fe uptake, namely Cd, Co, and Ni. We begin by discussing Fe uptake into the root, then long-distance transport to the shoot, and finally, the loading of Fe into seeds. And, as Fe is essential to the metabolism of the mitochondria and chloroplast, we also look at the recent discoveries in Fe transport and homeostasis at the intracellular level. We do not cover the regulation of these transporters as this topic has been recently reviewed.1
Plants mainly acquire Fe from the rhizosphere. Although Fe is one of the most abundant metals in the earth's crust, its availability to plant roots is very low. Fe availability is dictated by the soil redox potential and pH. In soils that are aerobic or of higher pH, Fe is readily oxidized, and is predominately in the form of insoluble ferric oxides. At lower pH, the ferric Fe is freed from the oxide, and becomes more available for uptake by roots. Because 30% of the world's cropland is too alkaline for optimal plant growth,2 and some staple crops, like rice, are especially susceptible to Fe deficiency,3 much research has focused on how plants cope with Fe limitation.
The responses to Fe deficiency include changes in root morphology,2 and up-regulation of genes involved in Fe uptake. 4, 5 In fact, in Arabidopsis thaliana, up to 85% of the genes expressed in particular regions of the root are differentially regulated by Fe.4 This transcriptome analysis was made possible by the isolation, via fluorescence activated cell sorting analysis, of cells from specific root layers that were expressing GFP under the control of cell-specific promoters.6 The transcript levels within each layer were then measured via microarray analysis. This allows detection of differential expression profiles among specific cell types that cannot be seen when the root as a whole is examined. Large transcriptional differences between layers in response to Fe deficiency were identified, indicating layer-specific roles (Figure 1). The expression of genes related to metal transport and chelation was increased in the epidermis, while genes related to root hair morphogenesis were downregulated; in the stele, genes associated with signaling and stress responses were upregulated. These results suggest that sensing of Fe levels and control of the Fe deficiency response occurs in the vasculature, while regulation of Fe levels in the root is facilitated by modulating uptake in the epidermis.
When these Fe deficiency-induced changes were compared with the response to salt stress, it was found that the vast majority of the transcriptome is altered by environmental stress, and that these changes are most dramatic in the root epidermis. Interestingly, there is also a small set of genes unaffected by stress; this core may define the essential features of each cell type, and mediate the appropriate transcriptional responses to environmental stresses. Of the changes in the epidermis, two specific strategies of Fe uptake have been identified in plants. Non-graminaceous plants reduce Fe3+ via a membrane-bound reductase to make it accessible for uptake by a Fe2+ transporter, while grasses secrete phytosiderophores (PS) that readily bind Fe3+, and the Fe-PS complexes are then transported back into the roots.
Components of the reduction strategy have been described in many non-graminaceous species,7–10 but it is best characterized in Arabidopsis (Figure 2). In response to Fe deficiency, protons are released into the rhizosphere, by AHA H+-ATPases expressed in the epidermis.11, 12 This lowers the soil pH, making Fe more soluble. While AHA1, AHA2, and AHA7 are all upregulated in the root epidermis in response to Fe deficiency,4, 5 AHA2 is the primary root H+-ATPase in the Fe deficiency response. 13 The expression level of AHA2 is highest among the three, and only the loss of AHA2 was found to reduce rhizosphere acidification during Fe deficiency.13
The NADPH-dependent ferric chelate reductase, AtFRO2, then reduces Fe3+ to Fe2+. Electrons are transferred from NADH+ across four heme groups, to Fe in the rhizosphere.14 This appears to be the rate-limiting step in Fe uptake in Arabidopsis.15 In fact, the transgenic overexpression of ferric chelate reductases in the roots of rice, tobacco and soybeans has been successful in increasing tolerance to Fe-limiting conditions.16–18
Once reduced, Fe (II) can then be transported into the root epidermal cells by the divalent metal transporter AtIRT1.19–21 AtIRT1 also transports Zn, Mn, Cd, Co,22, 23 and Ni.24 Additional root epidermal transporters for these metals have not yet been identified; but in the irt1-1 loss of function mutant, shoot accumulation of Fe, Mn, Zn, and Co decreases significantly, and the plants become Cd tolerant,20, 25 suggesting that AtIRT1 is a primary transporter for these metals under Fe deficiency. A similarly broad range of metals was found to be transported by the tomato orthologs LeIRT1 and LeIRT2, which complement yeast mutants defective in the uptake of Fe, Zn, Mn, and Cu.7 Thus, the Fe deficiency response also leads to the uptake of metals other than Fe, all of which are potentially toxic.
The presence of Cd and Co have been shown to exacerbate Fe deficiency. Cd interferes with Fe movement from root to shoot, as Cd treatment of Brassica napus produces a dramatic increase in Fe accumulation in the roots, while the shoots become Fe starved.26 The level of Fe in the xylem and phloem saps also decreased significantly, while the level of Cd increased. This suggests that the presence of Cd impairs Fe access to the phloem, rather than uptake, at least in B. napus. A similar phenotype was seen in mung bean seedlings treated with Co: Fe uptake increased, but Fe was unable to move from the root to the shoot.27 Additionally, the Fe deficiency response in Arabidopsis is upregulated in Co-treated plants (Morrissey and Guerinot, unpublished data).
The uptake of Cd during Fe deficiency is of special interest, as Cd is considered one of the most toxic crop contaminants. Cd has the opposite bioavailability profile compared to Fe: in aerobic conditions where Fe is oxidized and insoluble, Cd becomes more soluble.28 Thus, the soil conditions that trigger IRT1 expression also enhance Cd availability for uptake by IRT1. Because ingestion of plants is the primary route of Cd exposure for non-smokers,29 efforts have been made to understand and modify the selectivity of IRT1. Expression of mutagenized IRT1 in yeast found that amino acid substitutions in the first third of the protein, especially the first extracellular loop, could modulate selectivity of Fe, Zn, Mn, and Cd.30 Mutations that destroyed IRT1 function were almost exclusively found in the two alpha helices that compose the fourth and fifth transmembrane domains; it is believed that this region forms a metal binding pocket that facilitates ion movement across the membrane. Ultimately, this demonstrates that the broad substrate range of Fe transporters can be adjusted, producing variants optimized to biofortify crops while excluding toxic contaminants.
The influx of toxic metals via IRT1 is counteracted in part by the expression of the metal effluxer FPN2 (IREG2) during Fe deficiency. FPN2 is localized to the vacuolar membrane24 in the two outermost root layers of Arabidopsis, and appears to sequester metals in the vacuole. When expressed in yeast, FPN2 confers tolerance to Ni24 and Co (Morrissey and Guerinot, unpublished data). The Fe-regulated expression pattern and root localization of FPN2 suggests that it serves as an adaptation to the influx of Ni and Co during Fe deficiency; accordingly, the loss of FPN2 results in increased sensitivity to Ni and Co (Morrissey and Guerinot, unpublished data). A similar role has been described for MTP3, which is also Fe regulated, localizes to the vacuolar membrane and is thought to sequester Zn in the vacuole during Fe deficiency.31
Fe itself is highly reactive and potentially toxic. It is unclear which ligands bind Fe after transport, but buffering Fe uptake is clearly important. Indeed, a phenotype long attributed to phosphorous deficiency was found to be caused by Fe toxicity.32 Inhibition of root elongation in the Arabidopsis ecotype Col-0 was known to occur during phosphorous limitation, and was believed to be caused by a phosphorous-sensing regulatory pathway. Instead, it was recently shown that root growth is restored during phosphorous deficiency by simply removing Fe from the growth medium. The inhibition was actually caused by the toxic effects of Fe that was likely no longer complexed with phosphate, greatly increasing its bioavailability. The influx of Fe via IRT1 may be buffered in the outer root layers by the expression of FPN2, which transports Fe (unpublished data), presumably sequestering excess Fe in the vacuole. A similar buffering role has been proposed for IRT2, which transports Fe and Zn 33 and has been localized to vesicles in the epidermis.34
Another aspect of the Fe deficiency response in non-graminaceous plants is the secretion of phenolic compounds into the rhizosphere2 and the uptake of apoplastic Fe. It has been observed that as much as 75% of Fe in the roots is attached to the apoplast,35 as the negatively charged carboxyl groups of the cell walls serve as a cation sink.2 This pool decreases when a plant becomes Fe deficient, suggesting mobilization into the symplast.36 How this Fe is taken up is unclear, but it was recently found that phenolics exuded by the root in response to Fe deficiency facilitate the utilization of apoplastic Fe, and the recovery from Fe deficiency.37 Phenolics secreted by red clover roots were shown to efficiently strip Fe from purified cell walls. To determine if this is an essential component of the Fe deficiency response, phenolics were filtered from the liquid growth medium by constant recirculation through a resin column. Under normal growth conditions, the level of apoplastic Fe was found to decrease in response to Fe limitation, and although initially chlorotic, the leaves began to re-green. The filtering of phenolics, however, resulted in no decrease in apoplastic Fe, while the Fe concentration in the shoot became lower. This produced plants that were much more chlorotic, and could not recover from Fe deficiency. Ferric chelate reductase activity and proton extrusion also increased, but this alone was not able to counteract the severe chlorosis. Thus, phenolic-mediated mobilization of apoplastic Fe is an integral part of the Fe deficiency response (at least in red clover), although it is unclear how phenolics facilitate uptake. Perhaps phenolics mediate extraction of Fe from the negatively charged cell walls, allowing transport into the root symplast. Whether a Fe-phenolic complex is directly transported into the root is unclear, as potential Fe chelate transporters have not yet been characterized in the root epidermis of non-grasses. A candidate would be the Fe-nicotianamine (NA) transporter, AtYSL3, that is up-regulated in the root epidermis under Fe deficiency,4 although it has not been tested for Fe-phenolic transport.
Grasses depend on the uptake of Fe chelated by soluble siderophores with a high affinity for Fe3+.2 In response to Fe deficiency, mugineic acid (MA) family PS are synthesized from L-methionine and released from the root epidermis (Figure 3), perhaps via anionic channels or vesicles.38 In barley, the genes required for sulfur uptake, methionine synthesis, and PS synthesis are dramatically up-regulated in the first 24 hours of Fe deficiency.39 In rice, expression of the OsIRO2 transcription factor increases dramatically over the course of the first five days of Fe starvation, and is believed to activate the expression of genes related to PS synthesis and Fe uptake.40
The resulting Fe (III)-PS complexes are readily transported into the root epidermis via a high-affinity uptake system.41 The chelation strategy is less sensitive to pH than the reduction strategy, and there is a strong correlation between the volume of PS released and resistance to Fe limiting soils. For instance, barley, which is adapted to alkaline soils, releases a much greater volume of PS than most rice species,39 which are adapted for growing in anaerobic soils where Fe is more soluble. Indeed, in Oryza sativa var. japonica, which grows poorly on calcareous soils, the overexpression of enzymes in the barley PS synthesis pathway greatly increased PS secretion.3 This resulted in a four-fold increase in grain yield by rice grown on Fe limited soil.
The most well characterized PS-Fe transporter is the maize oligopeptide transporter (OPT) family member, ZmYS1. ZmYS1 is expressed in roots in response to Fe deficiency, and its loss results in decreased Fe uptake, and a constitutive Fe deficiency response; in the leaf, the decrease in Fe containing proteins impairs chlorophyll synthesis, resulting in a yellowing between the veins, or interveinal chlorosis.42, 43 The transport of Fe-PS by ZmYS1 appears to be well adapted to high pH solutions - the type of environment that is Fe limiting for plants. While reduction of Fe3+ becomes more difficult with increasing soil pH due to the pH optimum of the reductase, the expression of ZmYS1 in oocytes showed transport of Fe-PS is still efficient at very high pH.44
The maize PS deoxymugineic acid (DMA) also readily chelates other metals, and ZmYS1 has been shown to transport PS complexed with Zn, Cu, and Ni at the same rate as Fe-PS. 44 ZmYS1 also transported Ni, Fe(II), and Fe(III) complexed with the PS precursor NA. Thus, like IRT1 in non-graminaceous species, ZmYS1 also serves as gateway for a broad range of metals, including those toxic to plants and humans. Interestingly, HvYS1, which is 95% similar to ZmYS1, only transports Fe-PS 45. Domain swapping between the two transporters showed that the extracellular loop between the sixth and seventh transmembrane regions provided the selectivity.46 When the loops were synthesized in vitro, the HvYS1 peptide formed an alpha helix in solution, while the ZmYS1 peptide remained flexible, suggesting that this structural difference dictates substrate specificity.
Maize DMA also appears to bind Cd in soil, and while Cd disrupts Fe homeostasis in maize, the Cd-PS complex is not readily transported by ZmYS1.47 However, the presence of Cd in growth media up-regulates the Fe deficiency response in maize, and results in reduced Fe levels in the xylem sap. At the same time, the level of Cd uptake in maize was similar in wild-type and ys1 mutants, suggesting Cd primarily enters the roots through another transporter,47 perhaps a Ca transporter or channel, or a divalent metal transporter similar to IRT1. An IRT1 ortholog was recently identified in barley, and heterologous expression in yeast indicates that it can transport Fe, Mn, Zn, and Cd.48 Like AtIRT1, HvIRT1 is up-regulated in response to Fe deficiency, but the tissue specific localization has yet to be determined. Thus, it is premature to say whether barley transports free Fe into the root epidermis like non-grasses. HvIRT1 is also up-regulated under Mn deficiency and higher expression of HvIRT1 correlated with increased Mn uptake by a Mn efficient genotype of barley. 48
Another graminaceous species, rice, combines components of the reduction strategy seen in non-graminaceous plants with Fe-PS uptake. Of the 18 yellow-stripe like (YSL) genes in rice, OsYSL15 is the primary transporter responsible for uptake of Fe-PS from the rhizosphere.49, 50 OsYSL15 is up-regulated in response to Fe deficiency and is expressed on the plasma membrane in the root epidermis, in addition to the stele, flowers, and developing seeds. Two osysl15 insertional mutants exhibited chlorotic phenotypes under Fe deficiency and had reduced Fe concentrations in their shoots, roots and seeds.50 Reducing OsYSL15 expression with RNAi resulted in severe germination defects, indicating an important role in Fe homeostasis, although these could relate more to Fe loading of seeds than Fe-PS uptake by roots. 49
But, as mentioned above, rice produces much less PS than maize and barley, making it less tolerant of calcareous soils. Rice compensates by expressing the divalent metal transporters OsIRT1 and OsIRT2 in the root epidermis in response to Fe deficiency. Both are similar to AtIRT1, and transport Fe when expressed in yeast.51 In fact, in rice mutants without the ability to synthesize PS, these Fe(II) transporters were found to be dramatically up-regulated in the roots: 30-fold for OsIRT1, and 64-fold for OsIRT2.52 This compensated in terms of Fe uptake in waterlogged soil where Fe2+ is readily available; surprisingly, the mutant plants even accumulated more Fe in roots and shoots than wild-type under these conditions. In aerobic soils (where Fe is limiting) and in hydroponic solution where only Fe3+ was supplied, these mutant plants died; thus, the reduction strategy alone is inadequate in Fe2+-limited conditions.
The elevated expression of OsIRT1 and OsIRT2 in rice mutants lacking PS also resulted in increased concentrations of Zn, Cu, Mn and Cd in the shoot,52 while ectopic expression of OsIRT1 increased Fe, Zn, and Cd accumulation.53 This suggests these transporters have a similar range of substrates as AtIRT1. When expressed in yeast, both OsIRT1 and OsIRT2 transport Cd, and Fe limitation in rice increases Cd uptake and translocation,28 as in Arabidopsis.
Despite its use of Fe2+ transporters, rice roots have very low ferric chelate reductase activity.16 This is likely because many rice varieties are adapted to the anaerobic paddy environment where Fe2+ is readily available. To recreate the reductase strategy system found in non-graminaceous plants, rice was transformed with a ferric chelate reductase.16 The gene encoding the yeast ferric chelate reductase FRE1, which has optimal activity in acidic conditions, was mutagenized and selected for elevated activity in high pH environments. A common mutation in the alleles tolerant of high pH was a substitution of methionine with arginine at position 312, near one of the four heme-coordinating sites. It would be interesting to determine the structural and functional significance of this change. The resulting ferric chelate reductase coding sequence was fused with the OsIRT1 promoter, to ensure that expression of the modified gene was Fe regulated and that it would be expressed in the same tissue as the Fe2+ transporter. This resulted in plants that thrived on high pH soil compared to wild type, and produced 7.9 times greater grain yield. Interestingly, the concentration of shoot Fe increased only slightly, while the seed levels did not increase at all. This indicates a tight regulation of Fe homeostasis in rice, and that uptake is immediately down-regulated once the necessary amount of Fe has been taken up.16 This also demonstrates that components of the reduction strategy can be incorporated into grass species to augment Fe uptake, improving crops.
Similarly, incorporating the components of the chelation strategy could increase Fe uptake in non-grass crops, although this has not yet been successfully demonstrated. All plants synthesize the PS precursor NA, and constitutive overexpression of the PS synthesis enzyme nicotianamine aminotransferase (NAAT) in tobacco has been shown to consume NA, leading to intervenial chlorosis and sterility.54 This suggests that introduction of PS synthesis into non-grasses is feasible, but we still do not understand how PS are secreted from roots, so this may represent another step that will have to be engineered.
After entering the epidermis, Fe is likely bound by unknown chelators or chaperones, due to its potential reactivity. Fe moves symplastically through the interconnected cytoplasm of the root, perhaps diffusing along the concentration gradient.2 At the pericycle, Fe is effluxed into the xylem, and moves towards the shoot through the transpiration stream. Although Cu chaperones have been identified in many organisms, 55 including plants,56 the existence of a cytosolic Fe chaperone in plants is unproven. Interestingly, the first cytosolic Fe chaperone was recently identified in humans: PCBP1, a ubiquitously expressed RNA binding protein that also facilitates Fe loading of ferritin.57 When human ferritin is expressed in yeast, it loads very little Fe, indicating the requirement of a chaperone; when PCBP1 is co-expressed, the ferritin fills with Fe. PCBP1 is also found complexed to ferritin in vivo, and is able to bind Fe in vitro; additionally, the knockdown of PCBP1 increases cytosolic Fe levels in cultured cells. Thus, PCBP1 likely delivers cytosolic Fe to ferritin, facilitating Fe loading. In the Arabidopsis genome, the genes most similar to HsPCBP1 have been characterized for their role in binding viral RNA (BTR1) and regulation of flower development (HEN4 and FLK). However, PCBP1 was also first characterized as an RNA binding protein,58 and its Fe function was only identified in 2008.
In the vasculature, Fe is likely chelated to prevent precipitation. The chelators believed to bind Fe have properties appropriate for their respective environments: citrate readily binds Fe at the xylem pH of 5.5, while NA prevents precipitation of Fe at the pH of phloem sap, 7.5.59 The exchange of Fe from citrate to NA has been predicted to occur at pH 5.5.60
When Fe enters the xylem, it is believed to complex with citrate.59 In Arabidopsis, citrate is effluxed into the xylem via FRD3, which is expressed in the root vasculature.61 FRD3 is a member of the multidrug and toxic compound extrusion (MATE) family, of which several other members also efflux citrate to mitigate aluminum toxicity;62–64 indeed, overexpression of FRD3-GFP increases aluminum tolerance in Arabidopsis.65 While FRD3 mRNA is detected under Fe sufficiency, it is up-regulated two-fold in response to Fe deficiency. The loss of FRD3 results in severe chlorosis and a constitutive Fe deficiency response.66 Xylem exudate collected from the top of frd3 mutant roots contained nearly 50% less translocated Fe, while Perls staining showed significant Fe3+ accumulation in the root vasculature.65 The shoots of frd3 plants have been shown to accumulate slightly less Fe than wild type.61, 65, 67 Without citrate, Fe does not efficiently move through the xylem and is not utilized by the shoot; instead it likely precipitates on the apoplast walls. Accordingly, adding citrate to the growth media re-greened the plants, abolished Fe3+ accumulation in the root vasculature, and reduced the Fe deficiency response to wild-type levels.65 Despite the severe phenotype, there is only a 40% decrease in citrate in the xylem, suggesting a role for other citrate effluxers.
The constitutive Fe deficiency response of the frd3 mutant results in increased IRT1 expression and Fe uptake. But when plants increase IRT1 expression, the uptake of Zn, Mn, and Co, and Cd via IRT1 also goes up. In the shoot, growth is tied to the availability of Fe; consequently, the concentration of Fe remains relatively constant under limitation, as growth is retarded. The shoot concentrations of Zn, Mn, Co, and Cd, however, keep increasing with increasing IRT1 expression. At the same time, Mo concentrations decrease, as the acidification of the rhizosphere reduces its availability in soil. This unique pattern in Fe deficient plants was further substantiated by analyzing the shoot metal profiles of over 70,000 Arabidopsis plants grown with different levels of Fe supplementation in Purdue Ionomics Information Management System (PiiMS)68. It has been described as an ionomic signature for Fe deficiency,69 and could be utilized as a biomarker to identify Fe deficient plants.
The ortholog of FRD3 was recently identified in rice. While not Fe regulated, OsFRDL1 was found to transport citrate when expressed in Xenopus oocytes, and the loss of OsFRDL1 results in chlorotic plants with Fe precipitation in the xylem.70 And, like atfrd3, the osfrdl1 loss of function insertion mutant has increased OsIRT1 expression, and accumulates more Zn and Mn in the shoot. Similarly, the elevated expression of OsIRT1 and OsIRT2 in MA-free rice mutants also resulted in increased concentrations of Zn, Cu, Mn and Cd in the shoot.52 This suggests that the Arabidopsis Fe deficiency signature could be adapted to rice, as the substrate specificity of uptake and translocation to the shoot does not diverge between these two species, or perhaps even non-grasses to grasses. Interestingly, the loss of OsFRDL1 reduced the concentration of Fe3+ in the xylem sap, but not Fe2+.70 This suggests that there is an additional chelator besides citrate involved in Fe movement in the xylem.
The Arabidopsis orthologs of ZmYS1 and HvYS1 do not transport Fe-PS from soil, but play a significant role in the distribution of Fe, most likely via the phloem. The eight Arabidopsis Yellow Stripe Like (YSL) transporters are proposed to transport Fe chelated by the PS precursor NA in and out of the phloem.59 The expression pattern of the rice YSLs also suggest a role in the long distance transport of Fe complexes, including delivery to the seeds.49 OsYSL15 and OsYSL2 are both upregulated in response to Fe deficiency, and may coordinate long distance Fe transport from root to shoot to seed, via the phloem: OsYSL15 in the root vasculature, flower, and developing seed; and OsYSL2 in the phloem companion cells of the shoot.49, 71 Interestingly, expression in oocytes showed that OsYSL2 transports Fe-NA but not Fe-PS, 71 while OsYSL15 transports Fe-PS but not Fe-NA.49 OsYSL18, like OsYSL15, also transports Fe-PS but does not appear to be involved in uptake from the rhizosphere. Rather, based on its expression pattern, it may be involved in DMA-mediated Fe distribution in reproductive organs, lamina joints and phloem cells at the base of the sheath.72
NA is a non-proteogenic amino acid ubiquitous in higher plants, synthesized by the condensation of 3 molecules of S-adenosylmethionine in a reaction catalyzed by nicotianamine synthase (NAS). NA complexes with Fe2+ and Fe3+; it has a higher affinity for Fe3+, but forms a more stable complex with Fe2+.73 NA also readily binds Cu2+, Ni2+, Co2+, Zn2+, and Mn2+, in decreasing order of affinity.59
The immunolabeling of NA in tomato root and shoot sections showed NA increases within cells in response to increasing Fe levels.74 This is in contrast to barley, where NAS1 expression in roots is up-regulated by Fe deficiency,75 as NA is used as a precursor to MA. Similarly, all three rice NAS genes were upregulated in the roots in response to Fe deficiency, especially in the vasculature.76 However, while barley roots accumulate (and secrete) higher levels of MA, rice accumulates much more NA in its roots, under both Fe sufficiency and Fe deficiency. This suggests that the NAS expression in rice is used to produce NA predominately for long distance Fe transport. NA and MA levels were also very high in Fe deficient rice leaves, while only trace amounts were detected in barley.75 This further indicates a more significant role for NA in long distance Fe transport in rice compared to barley, and also a possible role for MA in Fe translocation. The low levels of NA in barley raises the question of what chelators barley uses for long distance Fe transport. It is interesting that there is such a divergence between graminaceous species, and that rice appears to utilize components found in non-graminaceous species for both Fe uptake and long distance transport.
In Arabidopsis, there are four NAS genes. During Fe deficiency, NAS2 and NAS4 were upregulated in the root,77 suggesting a role in Fe translocation to the shoot. NAS3 expression increased four-fold after the transition from vegetative to reproductive growth, suggesting NA also mediates Fe movement to the flowers. Despite the varied patterns and Fe regulation, all the single mutants had wild type NA levels, indicating functional redundancy, presumably because NA is mobile. In fact, interveinal chlorosis and sterility were observed only when the quadruple mutant was created. 77
NA levels have a significant affect on metal homeostasis. The overexpression of NAS in tobacco and Arabidopsis increases NA levels, resulting in the increased accumulation of Fe, Zn, Mn, and Ni in the shoots.54, 78, 79 It is unclear if these changes are the result of greater root to shoot translocation facilitated by NA, or increased metal uptake in the roots driven by the creation of new Fe sinks in the shoots. Similarly, increasing NA levels in rice and Arabidopsis alters the localization of Fe. Rice plants with disrupted PS synthesis accumulated up to 43 times more NA than wild type in Fe-starved roots; this dramatic change in NA increased long-distance Fe transport, resulting in seeds accumulating significantly more Fe than wild type.52 The increased movement of Fe-NA to the seeds most likely involved OsYSL2, which was significantly up-regulated. In Arabidopsis, an NAS overexpressing line accumulated 100 times the NA, resulting in decreased root Fe, and constitutive IRT1 and FRO2 expression.80 These plants accumulated significantly more Fe in both the roots and shoots when exposed to Fe deficient conditions, yet remained chlorotic compared to wild type. This indicates that ubiquitous NA increases Fe translocation, but impairs the efficient utilization of Fe in the shoot.
Conversely, the loss of NA, either by loss of NAS function in tomato and Arabidopsis, or the depletion of NA by the overexpression of the NA-consuming NAAT, leads to symptoms of Fe deficiency like interveinal chlorosis, reduced growth, and sterility.54, 77, 81 When NA was depleted in tobacco by NAAT overexpression, Fe accumulated only in the veins of the leaf; the addition of exogenous NA, however, resulted in Fe movement throughout the entire leaf.54 This suggests that NA is essential for Fe mobilization from the vasculature into the interveinal tissues. Based on what is known about Fe-NA transport in Arabidopsis, the YSLs likely play a role in this mobilization.59
NA also plays a significant role in Ni tolerance and localization. In Arabidopsis, exposure to Ni induces the expression of all four NAS genes,79 and overexpression of NAS confers resistance to Ni. 79,82 Conversely, the NAS quadruple knock-down mutant was found to be highly sensitive to Ni. 77 In the Ni hyperaccumulator Thlaspi caerulescens, the high expression level of NAS genes in the shoots (relative to Arabidopsis thaliana) appears to be a key component of Ni tolerance and long distance transport.83 When treated with Ni, NAS expression was only detected in the shoots, yet NA began accumulating in the roots. At the same time, Ni-NA was detected in the xylem and Ni rapidly accumulated in the shoots. This suggests that in T. caerulescens, NA is translocated from the shoot to the root to bind Ni, and the Ni-NA is then translocated back, at least in part, via the xylem sap. Additionally, three TcYSL family members have much higher expression than their Arabidopsis orthologs and, although not regulated by Ni, TcYSL3 has been demonstrated to readily transport Ni-NA when expressed in yeast.84 Thus, there may be a sharing of transporters and translocation pathways by Ni-NA and Fe-NA.
Of the eight members of the Arabidopsis YSL family, three members - YSL1, YSL2, and YSL3 - have been characterized. All but YSL3 transport Fe-NA when expressed in yeast, and all were found to be expressed in a broad range of tissues, especially the vasculature.41, 85–88 It is often proposed that the YSLs serve to translocate Fe from the xylem into the phloem, so that it can move to young, growing tissues. The YSLs are also believed to load Fe from senescent leaves for long distance transport to the flower, for loading into the developing seed. Curie et al. 59 have recently reviewed the YSLs.
Of particular interest is the ysl1 ysl3 double mutant, which had lower Fe levels in both leaves and seeds.88 Flower and seed set were especially affected, which is discussed in the seed section, below (Section 4.1). The double mutant also displayed what was described as interveinal chlorosis, somewhat similar to the chlorosis seen in the NA-free tomato mutant chloronerva.81 Despite the chlorosis, the Fe deficiency response was not altered,88 unlike the NA-free tomato mutant. It is interesting that Fe starvation in the interveinal areas of the leaf is not enough to trigger the Fe deficiency response. This suggests a tissue-specific component of Fe sensing in Arabidopsis, in which the veins may be more important than the interveinal tissues. A similar role was proposed for the root vasculature, based on the stele-specific up-regulation of signaling genes during Fe deficiency.4
In rice, OsIRT1-GUS expression was detected in the phloem of roots and shoots.51 Expression was up-regulated in response to Fe deficiency, especially in companion cells. It is proposed that OsIRT1 transports Fe(II) into the phloem, where it is then chelated by NA. This role does not appear to apply to AtIRT1, as its expression within the root has only been detected in the epidermis.20
In addition to NA, an Fe binding protein has been identified in the phloem sap of 7 day old castor bean shoots.89 The Iron Transport Protein, or ITP, is a dehydrin, expressed in the shoots of both seedlings and adult plants. When radiolabeled Fe was applied to the cotyledons, nearly all was recovered in the phloem sap associated with the 17 kD ITP protein, indicating that Fe moves quickly to the phloem and nearly all is bound by ITP. The purified ITP protein was found to preferentially bind Fe3+ but not Fe2+. Unfortunately, obtaining large amounts of phloem sap from plant model organisms is difficult, and ITP remains reported only in castor beans. The most similar genes in Arabidopsis have annotations related to stress, and several are highly up-regulated in response to Fe deficiency in the root (BTI1, BTI2, At1G54410, At2G44060), although none are specific to the stele.4 But, working under the assumption that an ITP exists in other plant species, it has been proposed that NA serves as a shuttle, facilitating Fe movement in and out of the phloem (via the YSLs), while the actual movement of Fe within the phloem occurs via ITP.
In barley, an Fe discrimination center (DC) in the basal shoot was identified by monitoring the dynamics of Fe uptake and translocation with a Positron Emitting Tracer Imaging System (PETIS).90 PETIS allows the non-destructive visualization of metal movement in live plants, in real time. In both Fe sufficient and Fe deficient barley plants, 52Fe was found to first accumulate at a central location in the basal part of the shoot (Figure 5), yet the Fe was only translocated to the leaves in the Fe deficient plants, suggesting that this region regulates Fe distribution in barley.90 They also found that damaging the phloem impaired Fe movement to young leaves, but not old leaves. This provides more evidence that young leaves receive Fe primarily from the phloem, while older leaves receive Fe from the xylem. PETIS has also been used to visualize and quantify the uptake and translocation of radiolabeled 52Fe and 62Zn in rice plants.16, 91
Fe moves to the seeds, most likely via the phloem, as the flow of the xylem is driven by transpiration, and seeds do not transpire 92. Developing seeds receive Fe from the roots and from senescent leaves. The level of remobilization from shoot to seed varies by species: rice transports only 4% of shoot Fe to the seeds,93 versus wheat, that transports 77% of shoot Fe to the seeds.94 The timing and regulation of senescence has been shown to have a significant effect on Fe accumulation in the seeds. In wheat, the knockdown of the NAC family of transcription factors with RNAi was found to delay senescence by over three weeks, and to decrease seed Fe by over 30%.95 How developmental changes, photosystem deconstruction, and Fe remobilization interact is still unclear. It should be noted that crop breeding has often selected for improved grain maturation time, but ignored nutrient accumulation in the grain as a desirable trait. Consequently, many staple crops are agronomically productive, but have low levels of nutrients like Fe in the seed.
Cereal seeds provide more than 50% of the world's energy intake,96 and are a large part of the diet in many developing countries. As the plant based diet offers relatively low amounts of bioavailable Fe, large portions of the developing world suffers from Fe deficiency, including over 60% of all children in Africa and Southeast Asia.97 In response to this, research has focused on understanding how nutrients are transported to seeds, and how this can be increased. Overexpression of Fe related genes, however, often creates sinks in the leaves, rather than the seed.98 This shows the importance of determining how Fe levels are sensed at the tissue and intracellular level, and how this ultimately affects Fe allocation to the seed.
Fe-NA is essential for flower and seed development. The loss or depletion of NA results in deformed flowers and sterility, as well as significant decreases in floral Fe accumulation.54, 81 This fits well with the observed NAS expression pattern in tobacco, with highest expression being seen in flowers, especially in anthers and pollen.54 Interestingly, the grafting of NA-depleted tobacco shoots onto NAS overexpressing shoots restored Fe mobilization in leaves and flower development, but could not completely rescue the impaired seed set. 54 Thus, it would appear that the Fe-NA requirement for normal seed development is especially high. The Arabidopsis NAS quadruple knock-down mutant also becomes chlorotic when reproductive growth begins, and accumulates significantly more Fe in the leaves during flowering (+216%).77 At the same time, the level of seed Fe only decreased 46%, while IRT1 expression in the flower increased, suggesting compensation. When a second quadruple mutant with NA synthesis completely abolished was created, the result was sterility. It should also be noted that IRT1 expression in the flower is also exclusively in the anther,20 indicating a role for both Fe-NA and Fe(II) in flower development.
Thus, as Fe-NA is critical in seed development, the YSLs play an important, if not essential, role in Fe-NA delivery to the developing seed. In Arabidopsis, YSL1 expression was found in and around leaf veins, especially in senescent leaves, in addition to expression the flower, pollen, young siliques and embryo.87 This suggests a role in Fe loading from senescent leaves for transport to developing seed. Indeed, the seeds of the ysl1 loss of function mutant lines contained 30–65% less Fe, and germinated more slowly on Fe deficient medium. Watering plants with exogenous Fe could not restore Fe accumulation in the seeds, indicating that YSL1 plays a role in seed loading that can not be compensated for by other transporters or chelators. The expression pattern of YSL3 is somewhat similar to YSL1: in the vasculature of shoots, and in pollen and anthers.88 When the two loss of function mutants were crossed, the resulting double mutant was chlorotic, and most flowers did not produce siliques. The few resulting seeds were small and irregular, and 80% less likely to germinate than wild type. These phenotypes are similar to the floral deformity and sterility seen in plants lacking NA,54, 81 indicating that seed development requires not just the availability of NA, but also specific Fe-NA transporters.
OPT3, a member of the family that includes ZmYS1 and the AtYSLs, plays an essential role in Fe loading of the seed. OPT3 is expressed in pollen, the silique vasculature, and the developing embryo; additionally, expression in the root and shoot vasculature is up-regulated in response to Fe deficiency.99, 100 Unlike the ysl mutants, the opt3 null mutant is embryo lethal, indicating an essential role for AtOPT3 in seed development. An opt3 knock down line, opt3-2, allowed embryo formation in seeds, but these accumulated significantly less Fe.101 The opt3-2 plants also exhibit constitutive expression of genes involved in the root Fe deficiency response, regardless of exogenous Fe supply. This leads to the accumulation of very high levels of Fe in leaves, resulting in brown necrotic spots, especially during the seed-filling stage. The substrate of OPT3 is unknown, but its phenotypes and relation to the YSLs suggests it likely transports chelated Fe or a Fe chelator. There are eight other members of the Arabidopsis OPT subfamily, and many are expressed in the vasculature and reproductive organs; none, however, have reported phenotypes, most likely due to functional redundancy.100
In Arabidopsis, it has been observed that developing seeds store Mn and Zn complexed with phytate in the vacuoles of the embryo and endosperm, and transiently in the ER.102 The storage state of Fe in Arabidopsis seeds was unknown, although it was long assumed to be stored in ferritin in the plastid. This was based on earlier experiments in legumes that found as much as 90% of Fe in ferritin.103 Recent work in Arabidopsis has found that there is very little ferritin in seeds,104 raising the possibility that in Arabidopsis most seed Fe is bound by phytate or some other chelator in the vacuole.
In Arabidopsis, VIT1 transports Fe2+ into the vacuole, and is expressed in the vasculature, especially during embryo and seed development.105 While the loss and overexpression of VIT1 does not affect total Fe levels in seeds, Fe is severely mislocalized in the loss of function mutant. Visualization of Fe distribution by synchrotron X-ray fluorescence microtomography106 showed Fe concentrated in provascular strands of the embryo in wild-type seeds, while in the vit1 mutant, Fe was not associated with the vascular system but rather was seen throughout the hypocotyl and radicle and was concentrated in a layer of cells just inside the abaxial epidermis of the cotyledons (Figure 6). 105 This mislocalization of Fe resulted in decreased seedling viability on Fe limited soil. Thus, vacuolar Fe loading via VIT1 is essential for proper Fe distribution in the embryo, which in turn determines seedling viability under low Fe conditions. The Fe stored in the vacuoles of the vasculature may be in the Fe3+ form, as Perls staining of Fe in embryos101 strongly resembles the vascular localization of Fe demonstrated by SXRF. 105
In Arabidopsis, NRAMP3 and NRAMP4 also localize to the vacuoles in the vasculature, but transport Fe out of the vacuole. Like vit1, the nramp3 nramp4 double mutant seeds contain the same level of Fe as wild-type, but produce seedlings that grow poorly on Fe limited soil.107 Visualization of wild-type seeds by electron microscopy showed what were likely Fe-phytate globoids in the vacuole, which disappeared as germination progressed. But, in the double nramp3 nramp4 mutant, the globules remained during germination, indicating that Fe was not being mobilized from the vacuole, causing the germination defects seen on Fe limited media. As the Fe uptake transporter IRT1 is not expressed until the third day of germination, the first two days of growth rely on mobilization of vacuolar Fe stores via NRAMP3 and NRAMP4; hence the germination phenotypes of vit1 and nramp3 nramp4 mutants. Interestingly, this also demonstrated that the primary storage form of Fe in Arabidopsis seeds is not ferritin, as had been assumed based on earlier work in legume seeds. Instead, the vacuolar globoids, which are ions complexed with phytate,108 appear to be a primary Fe storage form in Arabidopsis seeds. It would thus be interesting to determine the Fe storage form in the reduced phytate Arabidopsis mutant.109 Perhaps ferritin levels increase to compensate, or another chelator (e.g. NA) is able to bind Fe in the vacuole.
Thus, both loading of Fe into the seed vacuole via VIT1, and its release during germination via NRAMP3 and NRAMP4 are essential for seed viability under Fe limited conditions, and implicate the vacuole as an integral component of Fe storage in seeds.
In addition to vacuoles, Fe is found in the plastids bound with ferritin. Arabidopsis has four ferritin genes, of which only FER2 is expressed in developing seeds, and during germination. 110 Accordingly, FER2 is the only ferritin up-regulated in response to the plant hormone abscisic acid (ABA). The fer2 loss of function mutant does not affect Fe accumulation or seed viability under normal conditions; in fact, ferritin was estimated to account for only 5% of total seed Fe.104 When the other three ferritin genes were knocked out, the flowers accumulated more Fe, and were highly sensitive to Fe supplementation. This resulted in deformed, less functional flowers, and increased oxidative stress. Thus, ferritin likely serves more as a Fe buffer, sequestering free Fe to prevent oxidative stress.
The seeds of staple crops contain large amounts of the antinutrient phytate in the maize embryo and the aleurone cells of wheat, rice, and barley 111. Phytate is composed of a phosphorylated myo-inositol ring, and strongly chelates metal cations, including Fe, Zn, and Mn 112; these salts accumulate as globules in the vacuole, as mentioned above. As phytate represents around 1–2% of the dry weight of cereal seeds 113, this poses a serious impediment to dietary Fe uptake. In the developing world, the prevalence of phytate in the plant-based diet is believed to contribute to the high rate of Fe deficiency and anemia.114 In fact, high fiber diets have been shown to induce Fe deficiency in healthy women,115 as phytate is presumably binding Fe from other foods in the intestine, making it unavailable for uptake. Conversely, Fe stored within ferritin is believed to be safe and have high bioavailability.116 Thus, several strategies have been employed to reduce the amount of phytate in seeds, while increasing the amount of ferritin.
One obvious approach is to disrupt phytate biosynthesis. Early attempts to reduce phytic acid across the whole plant successfully reduced accumulation in seeds, but these plants often germinated poorly and were more susceptible to stress.117 Recently in Arabidopsis, the disruption of the inositol polyphosphate kinases required for the later steps of phytate synthesis, AtIPK1 and AtIPK2, was found to produce seeds with 93% less phytate.109 While these mutations did not affect seed yield or germination, the loss of the phytate precursors did alter phosphate sensing. The authors noted that this could be overcome by using promoters specific for the seeds. Accordingly, low phytate maize and soybean seeds were generated by the seed-specific silencing of an ABC transporter.118 Although it is unclear which membrane the transporter localizes to, or even what it transports, its loss prevents phytate from accumulating in the seed without compromising seed viability. How the reduction of phytate in seeds affects Fe homeostasis has not been examined, but it would be interesting to look at the interplay between vacuolar and plastid Fe pools in these mutants.
A second approach is to overexpress ferritin in seeds. Although the mechanism of dietary ferritin uptake in the human gut is unknown, it is believed that the Fe complexed in ferritin is readily absorbed, and a highly accessible source of Fe.116 Consequently, ferritin has been viewed as a means of increasing bioavailable Fe in staple crops. Indeed, the overexpression of soybean and bean ferritins in rice seed resulted in two to threefold increases in seed Fe content; 119–121 rats fed ferritin overexpressing rice recovered from Fe deficiency, indicating that the Fe is bioavailable.122 Of course, overexpression of ferritin does have consequences for the plant. Overexpression of soybean ferritin in tobacco resulted in a constitutive Fe deficiency response, causing greater Fe uptake and accumulation, but also a 2-fold increase in Cd when grown on contaminated soil; at the same time, the increase in sequestered Fe produced improved resistance to oxidative stress.98, 123 Overexpression of alfalfa ferritin in tobacco produced an increased resistance to Fe overload, oxidative stress, and pathogen invasion.124 Ferritin overexpression with more powerful promoters produced the same fold increase in Fe as transgenics with weaker promoter constructs, suggesting further increasing Fe accumulation is limited by Fe uptake and transport and not by ferritin levels.125
Finally, a combination of the two approaches has been undertaken. Transgenic maize plants were generated to ectopically express Aspergillus phytase and soybean ferritin in the endosperm.126 This increased total Fe content in seeds by 20–70%, and resulted in the degradation of nearly all endogenous phytate. When paste from the resulting seeds was fed to cultured human cells, Fe uptake was significantly higher compared to those fed wild-type seed paste. Thus, attempts to increase bioavailable Fe in seeds are becoming more successful.
It is believed that chloroplasts hold nearly 90% of the Fe within a leaf.127 Indeed, Fe is required for photosynthesis, heme biosynthesis, and Fe-S cluster assembly, all of which take place in the chloroplast yet very little is known about how Fe is transported in and out of this organelle. Transporters likely serves as the gateway for Fe, regulating its levels within the plastid in response to cellular demand outside the plastid.
The expression of the ferric reductase FRO7 on the chloroplast membrane indicates that some Fe is trafficked to the chloroplast in the ferric form, and must be reduced to enter the chloroplast.128 Previous experiments with purified pea chloroplasts showed Fe was likely transported across the inner envelope in the ferrous state.129, 130 Accordingly, the chloroplasts of the fro7 mutant contain 33% less Fe, resulting in decreased photosynthetic efficiency and fewer healthy photosystems128. Additionally, the FRO7-facilitated import of ferrous Fe into the plastid is essential for seedling growth under Fe-limited conditions. While the nramp3 nramp4 double mutant grows poorly when germinated on Fe limited soil,107 the fro7 loss of function mutant dies. Thus, both Fe mobilization from the vacuole and Fe import into the plastid are essential for seedling development when Fe is limiting. This implies that the plastid Fe pool in seeds is insubstantial, which correlates with the low levels of ferritin found in Arabidopsis seed.104 This further implicates the vacuole, rather than the plastid, as the primary site for Fe storage in seeds, and subsequent site for mobilization during germination.
PIC1 (Tic21) was originally identified as a chloroplast translocon component, as it immunoprecipitates with the major components of the Toc and Tic translocon.131 However, it has also been proposed that developmental defects seen in the loss of function mutant are related to impaired Fe homeostasis within the chloroplast, rather than protein translocation. PIC1 localizes to the inner envelope of the chloroplast131, 132 and is essential for Fe homeostasis within the plastid and plant as a whole.132 The heterologous expression of the plastid-localized transporter in yeast suggested that PIC1 transports Fe and Cu across the membrane. Although overall Fe levels in the leaf do not change in the pic1 mutant, the plants are dwarfed and chlorotic, with impaired chloroplast development. These plastids were also found to have elevated levels of ferritin and lacked thylakoids; this suggests Fe was no longer being utilized properly in the plastid, and instead accumulates in ferritin. This mislocalization of Fe also changes the expression of non-plastid, Fe-regulated genes in the shoot cells, and the expression of the root Fe uptake transporter IRT1 was repressed. This indicates that the chloroplast is integral to the Fe sensing mechanism, as the Fe status of the chloroplast affects the Fe homeostasis of the entire cell, in addition to the expression of Fe deficiency response genes in the root.
In Arabidopsis FER1, FER2, and FER3 are predicted to localize to the plastid; FER4 is predicted to localize to the mitochondria, or be dual targeted to both organelles.110 The roles of the ferritin paralogs is differentiated by localization and regulation: FER2 is only expressed in the seeds, while the other three ferritins are expressed in the shoots and flowers, in addition to FER1 expression in the roots.110 Additionally, the expression of the three non-seed ferritins increases in response to high Fe levels, whereas the seed FER2 is expressed in response to the plant hormone ABA .110 Ferritins appear to buffer Fe levels, and sequester excess free Fe to prevent oxidative stress.104 When the three genes encoding non-seed ferritins were knocked out, the triple mutant showed a shift in Fe accumulation from stem to flower when supplemented with Fe, resulting in increased oxidative stress and deformed flowers. This supports the hypothesis that chloroplasts are an important Fe sink, and that ferritins may sequester some Fe in the leaf plastids. This prevents excess Fe movement to the flower, although it is unclear whether this is by physically sequestering Fe in the shoot, or if the Fe status of the plastid regulates long distance Fe transport to the flower. The triple mutant also showed no decrease in photosynthesis,104 indicating that ferritins are not essential for chloroplast development or function. Instead, the ferritins prevent excess free Fe from accumulating in the flower, where it causes damage.
Plant mitochondria require Fe for respiration, heme biosynthesis and the synthesis of Fe-S clusters,133 but the combination of electrons and free Fe is highly toxic. Thus, proper Fe homeostasis in the mitochondria is vital, and both transporters and Fe sequestering proteins have been found to be essential for mitochondria function. Flower development, especially microsporogenesis, is highly dependent on energy from the mitochondria.134 Maintaining mitochondrial Fe levels is thus of high importance, as Fe deficiency produces deformed mitochondria in rice pollen, and reduces seed yield.135 Appropriately, many Fe related genes are highly expressed in the anthers (the portion of the male organ of the flower containing pollen), such as NtNAS, AtOPT3, AtYSL1, AtYSL3 and AtIRT1.20, 54, 88, 100
Recently, the Fe binding proteins ferritin and frataxin have been localized to the mitochondria in several organisms,136–138 including Arabidopsis.139 They appear to play a very important role in metal homeostasis not only in the mitochondria, but also in the whole cell.
Very little research has been done on mitochondrial ferritins in plants, other than confirming their presence in purified mitochondria from pea and Arabidopsis.139 Based on its putative transit peptide, AtFER4 is the most likely to localize to the mitochondria, although it may also target to the chloroplast (Aramemnon Plant Membrane Database). The fer4 loss of function mutant does not have a phenotype, perhaps because one or more of its paralogs are also targeted to the mitochondria, or frataxin is able to compensate. More likely, FER4 is not essential for mitochondria function under normal conditions. While it is expressed in response to Fe overload, it is downregulated in response to oxidative stress.110 Like the mitochondrial ferritin in humans and fruit flies,136, 137 FER4 appears to play an important role in the mitochondria-rich reproductive organs, as FER4 expression was highest in the flowers and floral stalk. It would be interesting to see the effects of FER4 overexpression in plants, as the overexpression of mitochondrial ferritin in human cells has been linked to cytosolic Fe depletion.140, 141
Like FER4, frataxin is expressed in the mitochondria of the flowers,142 in addition to the developing embryo.143 Unlike mitochondrial ferritin, frataxin is not Fe regulated and its loss is embryo lethal. 143,144 Frataxin is essential to growth because it has functions beyond mitochondrial Fe sequestration. In addition to sequestering Fe, frataxin is believed to serve as a chaperone, mediating Fe delivery to the Fe-S cluster assembly scaffold.145 The knock-down of frataxin in Arabidopsis is not lethal, but results in increased ROS and decreased vegetative growth and seed set. 144 The knock-down line also accumulates more Fe in the root, and the expression of FER1 and FER4 is increased, presumably to prevent oxidative stress.146 Interestingly, Fe-S cluster related genes in the mitochondria are upregulated in the mutant; however, the resultant Fe-S containing proteins (like aconitase) have reduced activity. 144 This indicates that Arabidopsis frataxin is essential for functional Fe-S clusters, and that even decreased expression of frataxin has serious phenotypic consequences in terms of sequestration of free Fe and Fe-S cluster assembly.
In Arabidopsis, the only identified mitochondrial Fe transporters are the ATMs - half-molecule ABC proteins that are orthologs of ScATM1. ScATM1 localizes to the mitochondrial inner membrane and is believed to efflux Fe-S clusters from the matrix.147 The Arabidopsis ATMs were first identified by the chlorotic, dwarf phenotype of the atm3 loss of function mutant (or sta1).148 Like Δatm1 yeast, mitochondria of these plants accumulated more non-heme, non-protein Fe than wild type, resulting in increased oxidative stress. When expressed in yeast, the Arabidopsis ATMs localized to the mitochondria, but only AtATM3 was able to rescue the Δatm1 yeast.148, 149 These results suggest that ATM3, if not the other ATMs, could perform a similar function of Fe-S cluster export from the Arabidopsis mitochondria.
ATM3 also appears to play a role in Cd detoxification in Arabidopsis.150 ATM3 is up-regulated in roots treated with Cd or Pb, and the atm3 dwarfs were more sensitive to Cd than wild type, while ATM3 overexpression enhanced Cd resistance. As ATM3 is closely related to the fission yeast SpHMT1, a vacuolar phytochelatin-Cd transporter, it has been suggested that ATM3 may also function to export chelated Cd complexes out the mitochondria, in addition to its role in transporting Fe-S clusters. Plants overexpressing ATM3 also accumulate more Cd in roots and shoot, but the concentration of Fe and other metals were not reported. It would be very interesting to investigate how the constitutive export of Fe-S clusters from the mitochondria affects the Fe deficiency response. The increase in shoot Cd suggests that IRT1 expression may be up-regulated.
As described earlier, the vacuole serves as the most important Fe store in Arabidopsis seeds. The vacuole also plays a role in the roots and shoots, storing and releasing Fe in response to changes in cytosolic Fe levels. Interestingly, it was recently demonstrated that phosphorous availability controls the subcellular localization of Fe in Arabidopsis leaves.151 Using X-probe microscopy, leaf sections from plants grown under phosphorous sufficiency showed Fe and phosphate in globules in the vacuole. Interestingly, these vacuolar Fe globules were only seen in the cells surround the vasculature, and not in other tissue layers. However, under phosphorous deficiency, the localization of Fe shifts to the chloroplast and FER1 expression increases, suggesting ferritin is now binding Fe.
NA is found in the vacuole74 and is likely chelated to Fe there. If Fe-NA complexes are indeed transported in and out of the vacuole, YSL4 and YSL6 are candidates, as they were found in the vacuole proteome of Arabidopsis suspension cells.152 In the shoots of tomatoes and peas, NA was found primarily in the cytoplasm during Fe deficiency and Fe sufficiency, but shown to concentrate in the cytoplasm and vacuole during Fe overload.74 It has been proposed that NA functions as a Fe(II) scavenger, protecting the cell from oxidative stress,73, 74 and this excess Fe-NA may then be sequestered in the vacuole.
When NA is depleted in tobacco, electron microscopy of chlorotic leaf sections showed the appearance of electron dense globules in the vacuole.54 Although these were not analyzed, similar vacuolar globules were identified as containing Fe and phosphate.151 Thus, it is worth speculating whether the depletion of NA shifts Fe sequestration within the vacuole into phytate complexes.
AtNRAMP3 and AtNRAMP4 are both expressed on the vacuolar membrane in the vasculature of the roots and shoots in response to Fe deficiency. VIT1 most likely loads Fe into the vacuole,105 and like NRAMP3 and NRAMP4, VIT1 is expressed in the vasculature. Thus, the role of VIT1 may be filling the vacuole with Fe, which is then released into the cells of the vasculature during Fe deficiency by NRAMP3 and NRAMP4.
Much progress has been made in the past few years in studying Fe homeostasis in Arabidopsis, especially in clarifying ferritin function, and in identifying the vacuole's role in Fe storage and mobilization during seed set and germination. Additionally, the recent large-scale characterization of transcriptional changes in specific root layers of Arabidopsis has proven an invaluable resource, increasing the resolution of our understanding of the Fe deficiency response. Similarly, the depositing of ICP-MS data from thousands of mutant Arabidopsis lines into the PiiMS database 68 has allowed the identification of the Fe deficiency signature, and will likely yield more unexpected discoveries in the future. Some of the discoveries in Arabidopsis can be generalized to all plants but, as studies of rice and other grasses have shown, there are species-specific aspects of Fe metabolism as well. It must also be noted that many aspects of Fe homeostasis and transport in plants remain unclear. Foremost, a mechanism of Fe sensing has not been discovered in plants. Additionally, it is still unknown what chelates Fe once it is transported into the root epidermis, and very little is known about transport in and out of the mitochondria and chloroplast. Current research aims to answer these questions.
We thank members of the Guerinot lab for helpful discussions. Work in our laboratory is supported by grants from the National Science Foundation (IBN-0344305; IBN-0419695; DBI-0606193), the National Institutes of Health (RO1 GM 078536), the Department of Energy (DEFG-2-06ER15809) and the National Institute of Environmental Health Sciences (5 P42 ES007373).
Mary Lou Guerinot is the Ronald and Deborah Harris Professor in the Sciences at Dartmouth College. She earned her bachelor’s degree in biology at Cornell University and her Ph.D. in biology from Dalhousie University. After completing postdoctoral studies at the University of Maryland and at the DOE–MSU Plant Research Laboratory, she joined the Department of Biological Sciences at Dartmouth as an assistant professor in 1985. She was promoted to an associate professor with tenure in 1991 and to full professor in 1997. She was chair of the Department of Biological Sciences from 1994 to 1998, served as the associate dean of the Faculty for the Sciences from 1998 to 2001 and as Vice Provost from 2001 to 2004. Guerinot is a molecular geneticist whose principal expertise and research interests are in the area of metal transport and regulation of gene expression by metals. In particular, she has been systematically dissecting how iron gets from the "soil to the seed" and has identified many key genes. Guerinot is a fellow of the American Association for the Advancement of Science and has served as President of the American Society of Plant Biologists, a 5000-member organization devoted to the advancement of plant science. She has served on the Advisory Committee for the Biological Sciences Directorate at NSF, is a member of the Scientific Advisory Board for the Donald Danforth Plant Science Center and serves on the Board of Directors for TAIR (The Arabidopsis Information Resource). She is currently an associate editor of Plant Cell and Environment as well as a member of the editorial board of Applied and Environmental Microbiology. She teaches Microbiology and Molecular Genetics and has mentored numerous undergraduates, graduates students and postdocs.
Joe Morrissey received his undergraduate education at the University of Minnesota, and performed research in John Ward's laboratory. He entered the biology PhD program at Dartmouth College in 2004, and joined the lab of Mary Lou Guerinot. His graduate studies are focused on characterizing the function of the ferroportin genes in Arabidopsis, and using ionomic screens to identify genes involved in metal homeostasis.