Report

Localization of Iron in Arabidopsis Seed Requires the Vacuolar Membrane Transporter VIT1

See allHide authors and affiliations

Science  24 Nov 2006:
Vol. 314, Issue 5803, pp. 1295-1298
DOI: 10.1126/science.1132563

Abstract

Iron deficiency is a major human nutritional problem wherever plant-based diets are common. Using synchrotron x-ray fluorescence microtomography to directly visualize iron in Arabidopsis seeds, we show that iron is localized primarily to the provascular strands of the embryo. This localization is completely abolished when the vacuolar iron uptake transporter VIT1 is disrupted. Vacuolar iron storage is also critical for seedling development because vit1-1 seedlings grow poorly when iron is limiting. We have uncovered a fundamental aspect of seed biology that will ultimately aid the development of nutrient-rich seed, benefiting both human health and agricultural productivity.

Iron is the most important yet problematic of the essential elements required by plants. It is needed for life-sustaining processes from photosynthesis to respiration, yet it can be toxic at high levels due to its propensity to form hydroxyl radicals that can damage cellular constituents. Like animal cells, plant cells can safely store iron in ferritin (1). However, unlike animal cells, plant cells also have vacuoles in which iron and other potentially toxic metals can be sequestered. Most efforts to date at increasing the iron content of staple foods have been focused on increasing seed ferritin levels (24), but the contribution of the vacuole to seed iron storage has remained largely unexplored.

In yeast, the vacuole serves as the main intracellular storage compartment for iron (57). The yeast CCC1 (Ca2+-sensitive cross-complementer 1) gene encodes an iron/manganese transporter that mediates the accumulation of these metals in the vacuole (8). We have characterized the Arabidopsis ortholog of yeast CCC1, VIT1 (vacuolar iron transporter 1; At2g01770), in order to address the role of the vacuole in iron homeostasis. VIT1 shows 62% amino acid similarity to the yeast CCC1 protein, and secondary-structure analysis programs predict five possible transmembrane domains, consistent with the model previously proposed for yeast CCC1. VIT1-like proteins can be found throughout the plant kingdom, with a distinct clustering of dicot and monocot VIT1-like sequences (Fig. 1A).

Fig. 1.

(A) Phylogenetic tree of plant VIT1 transporters. The deduced amino acid sequences of selected plant VIT1 orthologs were aligned with ClustalW. The tree and bootstrap analyses were performed with MEGA version 2.0 (25). Values indicate the number of times (in percent) that each branch topology was found during bootstrap analysis. At, Arabidopsis; Le, tomato; Gm, soybean; Bn, rapeseed; St, potato; Vv, grape; Mc, common ice plant; Os, rice; Zm, corn; Ta, wheat; and Hv, barley. The scale bar represents 0.1 substitutions per site. (B) VIT1 complements the sensitivity of ccc1 yeast to extracellular iron. Wild-type (WT) and Δccc1 cells were transformed with either an empty vector or a plasmid containing a MET25 promoter–regulated cDNA of tomato (LeVIT1) or Arabidopsis (VIT1). Cells were grown in CM-Ura for 16 hours, washed, and spotted onto CM-Ura-Met plates in the presence of different amounts of iron. The plates were incubated at 30°C for 2 days and photographed. (C to E) Overexpression of VIT1 increases the accumulation of vacuolar iron and manganese and leads to increased iron uptake in Δccc1. Δccc1 cells were transformed with either an empty vector or a plasmid containing a MET25-regulated VIT1. The cells were grown overnight in methionine-free medium, vacuoles were isolated, and the Fe (C) and Mn (D) content of the isolated vacuoles was determined by ICP-MS. The metal content was normalized to vacuolar protein levels. (E) Cells were grown for 16 hours in CM-Ura-Met. Cells were washed and incubated with 0.5 μM 59Fe for 15 min, and the amount of cell-associated radioactivity and cell number were determined.

To determine if VIT1 is a true ortholog of CCC1, we expressed VIT1 in ccc1 mutant yeast that are sensitive to high amounts of extracellular iron and thus fail to grow on media containing elevated levels of iron. This sensitivity is due to the inability of the ccc1 mutant to store iron in the vacuole, leading to increased accumulation of cytosolic iron (8). Expression of VIT1 sustained the growth of the ccc1 mutant yeast on high-iron medium (Fig. 1B). When VIT1 was expressed in ccc1 mutant yeast, vacuolar iron was increased threefold compared to control cells (Fig. 1C). Vacuolar manganese was also increased in yeast cells expressing VIT1 (Fig. 1D). The increases seen are similar to those conferred by expression of the CCC1 gene (6). No increases were seen in Zn or Cd. We also examined the effect of VIT1 expression on iron uptake. Overexpression of CCC1 in yeast cells decreases cytosolic iron levels, leading to increased expression of high-affinity iron transporters in the plasma membrane (9). The iron uptake rate of yeast cells overexpressing VIT1 was markedly increased relative to ccc1 cells (Fig. 1E). This result, together with the increased metal content of the vacuole, provides functional proof that VIT1 mediates iron sequestration into vacuoles.

We next investigated the localization of VIT1 using a green fluorescent protein (GFP)–tagged version of the VIT1 protein. The GFP-VIT1 fusion protein complements the ccc1 mutant phenotype (fig. S1), indicating that GFP tagging does not disrupt the biochemical function or the localization of VIT1. In yeast, GFP-VIT1 colocalizes to the vacuolar membrane with the FM4-64 marker (fig. S1). In transgenic Arabidopsis plants that stably express a GFP-VIT1 gene driven by the 35S promoter, the GFP fluorescence is localized to the vacuolar membrane (Fig. 2, A to D). In Fig. 2, A and C, the VIT1-GFP staining is only seen on the side of the nucleus facing the interior of the cell, that is, distinguishing it from staining of the plasma membrane, which would follow the cell periphery.

Fig. 2.

Subcellular localization, expression, and tissue distribution of VIT1. (A to D) GFP-tagged VIT1 localizes to the vacuolar membrane in plant cells. The GFP fluorescence in a root epidermal cell (A) of transgenic Arabidopsis stably expressing a 35S::GFP-VIT1 construct was visualized with confocal microscopy. (B) The cell walls and nuclei were stained with propidium iodide, shown here by red fluorescence. (C) Overlay of green fluorescence and red fluorescence. (D) Differential interference contrast image of the observed cell. Scale bar: (A) 10 μm. (E) Developmental expression of VIT1. mRNA levels for VIT1 were obtained from AtGenExpress (9). The linearized gcRMA values were plotted on a logarithmic scale. The peak value corresponds to seeds at developmental stage 6. (F to N) The VIT1 gene is expressed along the vasculature in developing seeds and young seedlings. The uidA gene was expressed under the control of the VIT1 promoter (1.0 kb upstream of VIT1 coding sequence). β-Glucuronidase (GUS) assays were performed with either X-GLUC (5-bromo-4-chloro-3-indolyl β-d-glucuronide cyclohexylamine salt) for histochemical staining (F, G, I, J, L to N) or ImaGene Green C12FDGlcU for fluorescent imaging (H and K). [(F to H) and (L)] GUS staining of the developing seeds from transgenic plants. (I to K) GUS staining of wild-type Col-0. (M and N) GUS staining in the seedlings at 0.5, 1, 2, 4, and 6 days after germination. Scale bars: (F to H) 100 μm, (L to N) 1 mm.

VIT1 is expressed at a low level throughout the plant, but there is a large peak in steady-state levels of VIT1 mRNA in the developing seed [Fig. 2E (10)]. Notably, the peak in VIT1 expression coincides with vacuole formation in the developing embryo (11). VIT1 expression is not affected by iron availability, unlike other proteins that have been implicated in iron metabolism such as IRT1, FRO2, FRD3, and FIT1 (1215). To further assess the tissue localization of VIT1, we generated transgenic plants that carry a β-glucuronidase (GUS) reporter whose expression is driven by the VIT1 promoter. Histochemical analysis of GUS activity showed GUS reporter gene expression in the developing embryo and seed (Fig. 2, F to L). Gus staining was also detected in young seedlings, predominantly associated with the vasculature (Fig. 2, M and N).

Because VIT1 is highly expressed in the developing seed and because seeds are an important food source, we examined the iron content of seeds using inductively coupled plasma mass spectrometry (ICP-MS). However, we found no difference in the iron content of seeds or shoots of vit1-1 plants compared to wild-type plants (table S1). We next investigated whether there was a change in the distribution of iron in a vit1-1 mutant, given that VIT1 is expressed in the vascular system and that the vascular system is responsible for the delivery of iron. Previously, metals have been localized in Arabidopsis seed by means of electron beam energy-dispersive x-ray spectroscopy, but this technique requires fixation and embedding of samples for electron microscopy [e.g., (16, 17)]. Such preparation can alter cellular materials and the location of key associated elements. Synchrotron x-ray fluorescence microtomography requires no sample pretreatment, allows noninvasive examination of living materials, and can detect elemental abundances in the sub–microgram per gram range with a resolution of 10 μm or less (18). Seeds are ideal samples for x-ray analysis because of their low moisture content and stability over long periods of data collection. Three-dimensional (3D) images and virtual cross sections of the x-ray attenuation or individual elemental fluorescence within the seed can be rendered, allowing visualization of either variability in density or elemental distribution (19). Three-dimensional tomographic reconstructions can be manipulated by computer analysis (in silico), allowing the investigator to look inside the seed, overlay elements of interest, and investigate elemental co-associations. We collected x-ray fluorescence microtomography data from three seeds each of wild-type Col-0 and the vit1-1 mutant (Fig. 3 and fig. S2). Total x-ray absorption allows one to visualize the cellular structureofthe seed (Fig. 3,B,C,E, and F) and demonstrates the high resolving ability of the technique such that individual cells in the cortex region can be distinguished. The most notable difference between the two seed types is in the distribution of Fe, which in wild-type seeds is strongly localized to the provascular strands of the hypocotyl, radicle, and cotyledons (Fig. 3, D and G). In vit1-1 seeds, Fe is completely absent from these cells and is instead located more diffusely in the hypocotyl and radicle, and in the epidermal cells of the cotyledons, in particular, the abaxial (lower) epidermis (Fig. 3, D and H). There is no difference in the pattern observed for Mn between wild-type and vit1-1 mutant seeds, although in vit1-1 mutants, the distribution of Fe is similar to that for Mn, as is apparent in the overlay of the signals for Fe, Mn, and Zn (Fig. 3D). Zinc is found throughout the seed and shows a similar distribution pattern in wild-type and vit1-1 seed (Fig. 3D). Multiple tomograms can be assembled to show the 3D distribution of various metals; Fig. 3G clearly shows that, in wild-type seed, Fe is associated with the provascular system throughout the embryo and additionally shows a region of high Fe concentration associated with the micropylar region (see also Movie S1).

Fig. 3.

X-ray fluorescence microtomography of Arabidopsis seed. (A) Light micrograph cross section of a mature Arabidopsis seed [modified from (26) with permission]; bar: 62 μm. (B and C) Total x-ray absorption tomographic slices of Col-0 and vit1-1 seeds; bar: 100 μm. (D) X-ray fluorescence tomographic slices of Fe Kα (blue), Mn Kα (green), and Zn Kα (red) fluorescence lines collected from Col-0 and vit1-1 with metal abundances indicated in mg kg–1 (smaller images), and composite images of Fe, Mn, and Zn abundance of Col-0 and vit1-1 (larger images). (E) Three-dimensional rendering of total x-ray absorption of a wild-type Arabidopsis seed. (F) In silico–sectioned (y axis, upper 50% removed) rendering of total x-ray absorption shown in (E). (G and H) Three-dimensional rendering of Fe Kα x-ray fluorescence in Col-0 and vit1-1, respectively, with both seeds identically oriented.

The marked change in metal distribution observed in the vit1-1 mutant implies that most of the iron in the wild-type embryo is stored in the vacuoles of provascular cells. Previous examination of minerals present in electron-dense globoids within protein storage vacuoles from nine different regions of Arabidopsis seed showed the highest Fe levels in the procambium regions of both the hypocotyl-radicle axis and the cotyledons (16), in agreement with our finding that Fe is concentrated in the provasculature. Such a location might allow rapid access to this pool of iron during growth of the germinating seedling. To test whether such localization affects seedling growth, we germinated wild-type and vit1-1 mutants on alkaline (pH 7.9) soil to limit iron availability. vit1-1 seedlings grew poorly compared to wild-type plants (Fig. 4). When plants were grown at pH 5.6, vit1-1 seedlings were indistinguishable from wild-type plants. This finding is very similar to the recently reported phenotype for a nramp3 nramp4 double mutant of Arabidopsis that cannot mobilize iron from the vacuole (20).

Fig. 4.

VIT1 is required for growth on alkaline soil. Wild-type and vit1-1 seedlings were grown for 15 days on either acidic (pH 5.6) or alkaline (pH 7.9) soil. No obvious difference was observed among the plants grown on the acidic soil. Wild-type seedlings developed weak chlorosis on the high-pH soil, consistent with limited iron availability. The growth of vit1-1 was markedly reduced and the leaves of vit1-1 showed severe chlorosis when vit1-1 was grown at pH 7.9.

Taken together, our results demonstrate that proper localization of iron, as well as an ability to access this store, plays important roles in iron homeostasis. It is important to note that Arabidopsis seed contains a single layer of endosperm and an embryo with two cotyledons and a radicle-shoot axis. The cotyledons serve as the main storage organ, similar to other nonendospermic seeds such as soybean, peanuts, and most Brassica species, in which the endosperm is degraded during the seed development and the cotyledons become the primary storage tissue. This is in contrast to grains like rice and wheat that store materials required for germination of the seedling in a multilayer endosperm. However, most of the iron in rice seed, for example, is associated with the embryo and the aleurone layer, not the endosperm, suggesting that VIT1-mediated iron storage in the embryo may play the same role in developing endospermic plants as that described here for Arabidopsis. Furthermore, unlike other Fe transporters characterized to date such as IRT1, which can transport Cd as well as Fe (21), VIT1 does not appear to transport Cd. Cd levels in seeds from lines overexpressing VIT1 were low (< 0.1 part per million), with no significant difference compared to wild-type seeds (P < 0.05). Therefore, any potential biotechnological applications of VIT1 will not have to consider unwanted accumulation of this toxic heavy metal.

Our study demonstrates the power of combining mutant analysis with a technique that can both image and determine the elemental composition of living plant material. Although 2D imaging with x-ray fluorescence has been used before to image the distribution of metals in plant tissues (22, 23), including Arabidopsis seed (24), our ability to render 3D images at high resolution allowed us to determine that Fe was associated with the provascular system throughout the seed and should prompt more studies on spatial distribution of metals in biological samples. Our study also highlights the role of the vacuole in seed iron storage and suggests that the vacuole offers another avenue for increasing the iron content of plant-based diets.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1132563/DC1

Materials and Methods

Figs. S1 and S2

Table S1

References

Movie S1

References and Notes

View Abstract

Navigate This Article