Vacuolar-Iron-Transporter1-Like Proteins Mediate Iron Homeostasis in Arabidopsis

Iron deficiency is a nutritional problem in plants and reduces crop productivity, quality and yield. With the goal of improving the iron (Fe) storage properties of plants, we have investigated the function of three Arabidopsis proteins with homology to Vacuolar Iron Transporter1 (AtVIT1). Heterologous expression of Vacuolar Iron Transporter-Like1 (AtVTL1; At1g21140), AtVTL2 (At1g76800) or AtVTL5 (At3g25190) in the yeast vacuolar Fe transport mutant, Δccc1, restored growth in the presence of 4 mM Fe. Isolated vacuoles from yeast expressing either of the VTL genes in the Δccc1 background had a three- to four-fold increase in Fe concentration compared to vacuoles isolated from the untransformed mutant. Transiently expressed GFP-tagged AtVTL1 was localized exclusively and AtVTL2 was localized primarily to the vacuolar membrane of onion epidermis cells. Seedling root growth of the Arabidopsis nramp3/nramp4 and vit1-1 mutants was decreased compared to the wild type when seedlings were grown under Fe deficiency. When expressed under the 35S promoter in the nramp3/nramp4 or vit1-1 backgrounds, AtVTL1, AtVTL2 or AtVTL5 restored root growth in both mutants. The seed Fe concentration in the nramp3/nramp4 mutant overexpressing AtVTL1, AtVTL2 or AtVTL5 was between 50 and 60% higher than in non-transformed double mutants or wild-type plants. We conclude that the VTL proteins catalyze Fe transport into vacuoles and thus contribute to the regulation of Fe homeostasis in planta.


Introduction
Regulation of the cellular Fe concentration poses the problem of balancing Fe deficiency against toxicity [1]. Equilibrium is maintained by strict regulation of Fe uptake and storage in the cell. Uptake of soil Fe in Arabidopsis is catalyzed by the Fe 2+ transporter AtIRT1 [2,3]. Transcription of AtIRT1 is greatly increased under conditions of Fe deficiency [2], and AtIRT1 is also post-translationally modified to regulate its partitioning between the trans-Golgi network and plasma membranes [4]. Prior to transport, soil Fe 3+ -chelates are reduced by the plasma membrane ferric reductase, AtFRO2 [5]. The coupling of uptake to reduction is a hallmark of strategy I plants, and an analogous mechanism is also found in Chlamydomonas [6] and Saccharomyces [7] among other organisms. The reduction-based acquisition of Fe is aided by secretion of Fe-binding compounds and H + -ATPase-mediated acidification of the rhizosphere [8,9,10,11].
Once in the cell, Fe 2+ can be incorporated into proteins or stored in cellular compartments. In chloroplast and presumably mitochondria, several thousand Fe atoms are stored per ferritin molecule [12]. However, in Arabidopsis only 5% of Fe is stored in ferritin, and vacuoles appear to be the major compartment for seed Fe storage [13]. Vacuolar Fe uptake was shown to be catalyzed by the ferroportin homologue, AtFPN2 [14,15], and by the CCC1-like protein AtVIT1 (16); however, the primary substrates of AtFPN2 have been reported to be Ni and Co in addition to Fe [14,15]. AtVIT1 has a specific function in the vacuolar transport of Fe into xylem parenchyma of developing embryos, and the vit1-1 mutant shows misdistribution of Fe in seeds; although, vit1-1 mutant seeds have unchanged Fe concentration compared with the wild type [16]. Efflux of Fe from the vacuole is catalyzed by two NRAMP proteins, NRAMP3 and NRAMP4. The double mutant nramp3/nramp4 shows decreased Fe mobilization from the vacuole in germinating seeds [17].
Two rice orthologs of AtVIT1, OsVIT1 and OsVIT2 [18], and one in tulip, TgVIT1 [19], have been shown to catalyze vacuolar Fe transport. TgVIT1 catalyzes the transport of Fe into the proximal perianth cell vacuole, which was shown to be essential for blue color development in tulips. Both OsVIT1 and OsVIT2 complemented the yeast Dccc1 mutant, and vacuoles that were isolated from complemented cells had increased Fe and Mn concentrations. In addition, both rice genes complemented the Zn transport mutant, Dzrc1, and vacuoles isolated from these cells also had an increased Zn concentration [18]. The transcript abundance of OsVIT2 increased within hours in roots and shoots grown under conditions of excess Fe (4 mM) but decreased in roots and shoots when rice was grown under Fe-deficient conditions. In contrast, OsVIT1 expression responded only weakly to changes in the Fe status. OsVIT1 and OsVIT2 were expressed in flag leaf blades and sheaths, respectively. Consistent with the localization of expression, the osvit1-1 and osvit2-1 T-DNA knockout mutants had decreased Fe and Zn content in flag leaves with no change in Mn. Seeds of osvit1-1 and osvit2-1 had correspondingly increased Fe and Zn content. OsVIT1 and OsVIT2 were shown to regulate the partitioning of Fe and Zn in rice between source and sink tissues [18].
In an analysis of the transcriptional response of Arabidopsis roots to Fe deficiency, we and others have identified three genes whose mRNA abundance decreased in Fe-deficient roots and whose putative amino acid sequences showed significant homology to AtVIT1 and consequently also to yeast Ccc1p [20,21,22]. These genes belong to a small, five-membered family that has been annotated as nodulin or nodulin-like in databases. These genes will be subsequently referred to as Vacuolar Iron Transporter-Like (VTL). The VTL family was found both in mono-and dicotyledon plants, as well as Chamydomonas and Physcomitrella. Promoter-bglucuronidase (GUS) assays showed expression of AtVTL1 in roots, hypocotyls, and cotyledons of seedlings with the greatest activity associated with the vascular bundle and the root stele [21]. The promoter activity was greatly reduced in Fe-deficient seedlings, consistent with the transcriptional analysis. In the present report, we show that AtVTL1 (At1g21140) and AtVTL2 (At1g76800) are localized to the vacuolar membrane in plants, and that AtVTL1, AtVTL2 and AtVTL5 (At3g25190) complement the Dccc1 mutation in Saccharomyces. Over-expression of AtVTL1, AtVTL2 and AtVTL5 also complemented the Fe deficiencydependent root growth phenotype in the nramp3/nramp4 double mutation and the vit1-1 mutation in Arabidopsis seedlings. These results indicate that the three members of the VTL family are involved in regulation of cellular Fe homeostasis, likely by acting as vacuolar Fe transporters.

Analysis of the Fe Concentration in Arabidopsis Seeds and in Yeast Vacuoles
The Fe content in seeds was determined by the BPDS method [23]. Dried and powdered samples (4-8 mg) were mixed with glass beads (,20; ø 425-600 mm) in 2 ml Eppendorf tubes and incubated at 95uC in 75 ml nitric acid (65%) for 6 h to digest the plant material. Fifty ml of H 2 O 2 (30%) was added and the solution incubated at 56uC for 2 h. The volume was adjusted to 200 ml with water. Twenty ml of this solution were diluted in 980 ml of BPDS buffer (1 mM bathophenanthroline disulfonic acid, 0.6 M sodium acetate and 0.48 M hydroxylammonium chloride). The concentration of Fe-BPDS was determined photometrically at 535 nm. A standard calibration curve was prepared by dilution of a stock FeSO 4 solution dissolved in 0.1 N HCl. Samples were measured in triplicate and the experiments were conducted at least three times.
Vacuoles were isolated as described by Li et al. [25]. Briefly, cells harvest at OD 600 = 0.4 were suspended in 30 ml of 0.1 M Tris-HCl (pH 9.3) buffer containing 10 mM dithiothreitol and incubated for 10 min at 30uC. The cells were washed once with spheroplast buffer (1.2 M sorbitol, 20 mM K-phosphate, pH 7.4) and incubated with 20 mg/ml lyticase for 30 min at 30uC. Spheroplasts were collected by centrifugation at 3,500 g for 5 min., and the pellet was resuspended in 10 ml of 15% Ficoll buffer (15% Ficoll (Sigma) in 0.2 M sorbitol and 10 mM PIPES-KOH, pH 6.8). Fifty mg/ml DEAE-Dextran (Amersham Pharmacia Biotech) were added to the spheroplasts, and the sample was incubated for 3 min. on ice and for an additional 5 min. at 30uC. The lysate (10 ml) was transferred to SW28 centrifuge tubes (Beckman Instruments) and overlaid with 10 ml of 8% Ficoll, 10 ml of 4% Ficoll, and 10 ml of 0% Ficoll. The tubes were centrifuged at 110,000 g for 3 h. The vacuolar fraction was collected from the 0%/4% interphase. Fe concentration of yeast vacuoles was determined by the BPDS method [23].

Transformation and Transient Expression of GFP::VTL
Arabidopsis wild-type Col-0 was transformed using the floral dip method of Clough and Bent [26]. Transgenic plants were selected for BASTA resistance on potting soil.
Onion epidermal cells were co-bombarded with 2.5 mg of plasmids bearing free GFP or GFP-tagged VTL genes (35S:GFP::AtVTL1 or 35S:GFP::AtVTL2) and the vacuole marker plasmid vac-rk CD3 975 [27]. Plasmids were coated on 1 mm gold particles and delivered into onion epidermal cells at a pressure of 900 psi by a PDS 1000/He particle delivery system (BioRad, U.S.A). After bombardment, onion slices (2 cm 2 ) were placed in a Petri dish containing Murashige and Skoog (MS) salts, 30 g/l sucrose and 1.5% agar (pH = 5.7). Following a minimum of 24 h, the epidermis cells were observed under a confocal laser scanning microscope (Zeiss LSM510 Meta) using a 663 water objective. Vac-rk CD3 975 images were captured in the 560 to 615 nm range after excitation at 543 nm with a HeNe laser beam. The GFP images were captured in the 505 to 530 nm range after excitation at 488 nm with an argon laser beam. Image overlay was carried out by Z-stack analysis at 0.8 mm intervals and further processed with the projection function under LSM510-Expert Mode software.

qRT-PCR
Total RNA was isolated using TRIsure (Bioline) and treated with DNase using DNAse Kit (Fermentas) as suggested by the manufacturer. cDNA was synthesized using DNA-free RNA with oligo-dT(18) primer and RevertAid reverse transcriptase (Thermo scientific). The cDNA was used as a PCR template in a 10 ml reaction system using the SensiMix SYBR No-ROX Kit (Bioline) with programs recommended by the manufacturer in a CFX96 Realtime system (BioRAD). Three biological replicates were performed for each sample. The DDC T method was used to determine the relative transcript abundance

Results
Complementation of the yeast mutant Dccc1 by AtVTL1, AtVTL2 and AtVTL5 As in Arabidopsis, the yeast vacuole is the major compartment for Fe storage but also protects the cell from Fe toxicity. The yeast gene CCC1 encodes a vacuolar Fe 2+ /Mn 2+ transporter that catalyzes Fe 2+ uptake into the vacuole. The Dccc1 mutant is hypersensitive to Fe toxicity, not growing in the presence of greater than ca. 3 mM Fe [25]. Because of the homology of the VTL family member to yeast Ccc1p and Arabidopsis AtVIT1 and because of the transcriptional repression of AtVTL1, AtVTL2 and AtVTL5 under conditions of Fe deficiency, the ability of the three AtVTL genes to complement the Dccc1 yeast mutant was investigated. When expressed in yeast, the VTL proteins were effective in complementing of the Dccc1 phenotype in the order AtVTL1. AtVTL2. AtVTL5 (Fig. 1). These results are consistent with an AtVTL-dependent decreased the cytosolic Fe in the Dccc1 mutant, presumably by transport of Fe into the vacuole.
To test this hypothesis, the Fe concentration in isolated yeast vacuoles was investigated. Intact yeast vacuoles were isolated by floatation through a Ficoll step gradient as described by Li et al. [25]. We confirmed the composition of the vacuolar fraction by recentrifugation the fraction on a continuous sucrose gradient followed by marker enzyme analysis (Fig. S1). As expected, the vacuole marker, bafilomycin-sensitive ATPase, co-localized with membrane proteins and no detectable plasma membrane (vanadate-sensitive ATPase) or endoplasmic reticulum (NADPH cytochrome c reductase) activities were detected in the gradient. Vacuoles, isolated from log-phase Dccc1 cells or from Dccc1 cells transformed with the control vector grown in liquid SD medium containing 1 mM FeSO 4 , had approximately 2 nmoles Fe per mg protein, whereas Dccc1 cells expressing AtVTL1, AtVTL2 or AtVTL5 had a greater than 3-fold higher Fe content ( Fig. 2A).
Localization of AtVTL1 was demonstrated by transforming Dccc1 cells with AtVTL1 fused to a 39 His tag (AtVTL1-H 6 ). Dccc1 cells that were transformed with AtVTL1-H 6 complemented the Fe-sensitive phenotype (Fig. S2). Using a His-tag antibody, AtVTL1-H 6 protein was localized to the vacuolar fraction at the 0/4% Ficoll interface (Fig. 2B). Iron was also concentrated in this fraction compared to fractions in other regions of the Ficoll gradient. The immunological signals in the pellet and 15%/8% Ficoll interface were most like the result of vacuolar membranes from ruptured vacuoles that that did not float through the gradient. We demonstrated that AtVTL1 was localized on the yeast vacuolar membrane, where it presumably catalyzed transport of Fe into the yeast vacuole.

Localization of AtVTL1 and AtVTL2 in planta
The localization of AtVTL1 and AtVTL2 was investigated in onion epidermis cells. Attempts to localize AtVTL5 in onion or tobacco leaf cells have not been successful. Onion cells were transformed by particle bombardment with 35S:GFP::VTL1 or 35S:GFP::VTL2 reporter gene constructs. In the case of AtVTL1, the GFP fluorescence co-localized with the vacuolar marker vac-rk CD3-975 (Fig. 3A, [27]). When cells were plasmolyzed in 0.8 M mannitol, the GFP signal remained associated with the vacuole (Fig. 3B). The localization of AtVTL2 in epidermis cells was predominately associated with the vacuolar membrane (Fig. 3A); however, GFP fluorescence was also found to be associated with the plasma membrane and the cytoplasm upon plasmolysis (Fig. 3B). The localization of the two AtVTL proteins on the onion vacuolar membrane was consistent with the complementation results in Saccharomyces and with the increased Fe concentration in vacuoles isolated from VTL-expressing yeast cells. Over-expression of AtVTL1 in Arabidopsis did not greatly alter Fe content of shoots and roots [21], nor were changes observed in the kinetics of the response to Fe deficiency, as determined by the root plasma membrane Fe 3+ -chelate reductase and chlorophyll content (Figs. S3A and B). In an attempt to characterize their functions, we have ectopically expressed the AtVTL genes in the Arabidopsis nramp3/nramp4 double mutant background (17). Over-expression was confirmed by semi-quantitative RT-PCR (Fig. S4). Both NRAMP3 and NRAMP4 have been shown to be Fe vacuolar efflux carriers. The double mutant had no obvious phenotype when grown on soil; however, growth of mutant seedlings was retarded shortly after germination on an Fe-deficient substrate [17]. The phenotype was not persistent and disappeared after a few days of growth or with supplemental Fe in the media. Under Fe deficiency root growth of the nramp3/nramp4 double mutant and of the double mutant transformed with the empty vector (GL1), was greatly inhibited compared to the wild-type control at 5 d following germination (Fig. 4 and Fig. S5; p,0.001) [17]. Importantly, over-expression of AtVTL1, AtVTL2 or AtVTL5 in nramp3/nramp4 complemented this mutant growth phenotype. The early seedling root length was significantly increased (Fig. 4 and Fig. S5; p,0.001) compared to the double mutant; however, the roots were still significantly shorter than those of the wild type (p,0.001). Thus, over-expression of AtVTL1, AtVTL2 or AtVTL5 partially restored the wild-type phenotype in the nramp3/nramp4 double mutant.
The proposed function of NRAMP3 and NRAMP4 was to mobilize Fe from the vacuole of young seedlings [17]. We hypothesized that over-expression of the AtVTLs might have increased total seed Fe, and thereby increased the Fe supply to the germinating seedling. To test this, we analyzed the Fe concentration in nramp3/nramp4 seeds that over-expressed the AtVTL genes. The results of these analyses demonstrated that the AtVTL1, AtVTL2 or AtVTL5 over-expressing lines had between 40 and 60% increased Fe concentration compared to either the wild type or the nramp3/nramp4 double mutant ( Fig. 5; p,0.01). These results supported the concept that an increased content of Fe in the seeds could have supplied the embryo with sufficient Fe even when efflux from the vacuole through the NRAMP transporters was compromised.
As mentioned above, the CCC1-like protein, AtVIT1, is localized on the vacuolar membrane and functions in the transport of Fe into the parenchyma cell of the provascular strands in developing embryos. The vit1-1 mutant showed misdistribution of Fe; although, vit1-1 mutant seeds had unchanged Fe content compared to the wild type (Fig. 5) [16]. vit1-1 had a severely chlorotic phenotype when grown on alkaline soil [16]. In addition, we observed that vit1-1 displayed a short-root phenotype similar to the nramp3/nramp4 mutant when grown on Fe-deficient media in the presence of the Fe 2+ chelator Ferrozine ( Fig. 6 and Fig. S6). Over-expression of AtVTL1, AtVTL2 or AtVTL5 restored root growth to greater than the wild-type length (p.0.001), thus complementing the vit1-1 mutation.

Regulation of AtVTL1, AtVTL2 and AtVTL5 Gene Expression
Finally, we investigated the response of AtVTL1, AtVTL2 and AtVTL5 to nutrient supply using quantitative qRT-PCR. Previously, we reported that the transcriptional activity of these genes positively correlated with the Fe supply [20]. Quantitative RT-PCR analyses confirmed these results (Fig. 7). Compared to Fesufficient controls (40 mM Fe), the transcript abundance of AtVTL1, AtVTL2 and AtVTL5 was decreased under Fe deficiency and increased in plants grown on media containing 120 mM Fe. Similarly, transcript abundance also decreased under Zn deficiency for AtVTL1, AtVTL2 and AtVTL5 (Fig. 7). However, when the Zn concentration was increased to 5 mM, the transcript abundance for AtVTL1, AtVTL2 and AtVTL5 was not significantly different from that of control plants grown on standard media (Fig. 7). The VTL genes were unable to restore growth in the Zn sensitive Dzrc1   (Fig. S7). Thus, a function of AtVTL1, AtVTL2 or AtVTL5 in vacuolar Zn transport is unlikely. Since AtVTL1, AtVTL2 and AtVTL5 responded positively to Fe supply and since their over-expression complemented vacuolar Fe transport mutations in yeast and Arabidopsis mutants, we conclude that VTL genes encode protein involved in Fe homeostasis, presumably as Fe vacuolar transporters.

Discussion
Heterologous expression of AtVIT1, OsVIT1, OsVIT2 and TgVIT1 in the Dccc1 mutant has been shown to complement the Dccc1 mutation by restoring Fe uptake into the vacuole, and thus, to protect the cell from the deleterious effects of high cytosolic Fe ( [6,18,19]. Vacuolar Fe transporter activity has been reported for FPN2, NRAMP3 and NRAMP4. Whereas AtVIT1 and FPN2 transported Fe into vacuoles, the two NRAMPs were shown to be efflux carriers. Previously, we have identified five genes in Arabidopsis, which encoded proteins with significant sequence homology to AtVIT1 [20]. The transcript abundance of AtVTL1, AtVTL2 and AtVTL5 was rapidly decreased in roots under conditions of Fe deficiency [21,28]. We have shown in this report, that heterologous expression of AtVTL1, AtVTL2 and AtVTL5 in yeast Dccc1 cells restored their ability to grow in the presence of 4 mM Fe. Since the heterologous expression of these genes in yeast correlated with increased Fe in the vacuolar fraction, we proposed that these three proteins increased Fe transport into the vacuole and thereby reduced toxic Fe in the cytoplasm of Dccc1 cells.

VTLs are Localized to the Arabidopsis Vacuolar Membrane
Transient expression of GFP::VTL1 and GFP::VTL2 chimeric proteins in onion epidermis cells resulted in a specific colocalization of GFP::VTL1 with the vacuolar marker vac-rk CD3-975. In the case of GFP::VTL2 transient expression, a prominent signal was co-localized with the vacuole marker, but GFP fluorescence was also found in the cytoplasm and associated with the plasma membrane, indicating possible localization of AtVTL2 on other membranes in addition to the vacuole. Attempts at localizing AtVTL5 have not been successful.
Evidence for a function of AtVTL1, 2 and 5 in Fe homeostasis was obtained by complementation of the nramp3/nramp4 and vit1-1 mutations. The NRAMP3 and NRAMP4 proteins act redundantly as vacuolar Fe efflux carriers. The abundance of both NRAMPs was increased during Fe deficiency, and in the double mutant, greening of cotyledons and early root growth of seedlings were delayed compared to controls when seeds were germinated under Fe deficiency [17]. This inhibition of root growth could be largely eliminated by supplying Fe to the growth medium. The root growth phenotype was complemented by over-expression of AtVTL1, 2 or 5 in the nramp3/nramp4 mutant. In addition the total Fe content of seeds from over-expressing AtVTL1, 2 or 5 plants was increased compared to the wild type and the mutant. It might have seemed paradoxical that a mutation in a vacuole efflux transport system could be complemented by a putative vacuolar uptake transporter; however, an increased vacuolar Fe concentration might have been sufficient to compensate for a decreased efflux caused by the mutations in NRAMP3 and NRAMP4.
The vit1-1 mutant also displayed a Fe deficiency-dependent inhibition of seedling root growth that is likely the result of a mislocalization of Fe in the seed radicle. Over-expression of AtVTL1, 2 or 5 might have restored the Fe supply to the growing root. In summary and based on the complementation activity in yeast, the increased vacuolar Fe concentration in yeast cells expressing AtVTL1, 2 or 5, the localization of AtVTL1 and 2 on the vacuolar membrane in onion epidermis cells and the restoration of seedling root growth in the nramp3/nramp4 and vit1-1 mutants, a critical role for the VTL proteins in the subcellular distribution of Fe is apparent.

AtVTLs and Metal Homeostasis
Low substrate specificity is commonly found in cation transports. For example, AtIRT1 is the predominant plasma membrane Fe 2+ transporter and is essential for Fe uptake from the soil [3]. In the absence of Fe or in the presence of excess divalent cations, AtIRT1 can also catalyze the transport of Zn, Mn and Cd [3,29]. Both NRAMP3 and NRAMP4 have been shown also to be vacuolar Mn exporters [30]; although, transcription was correlated with the Fe and not the Mn nutrient status. Lanquar et al. [30] envision a passive function of NRAMP3 and NRAMP4 in cycling of Mn from the vacuole to the plastid in mesophyll cells. The nramp3/nramp4 double mutant is hypersensitive to Zn and Cd, and as Sinclair and Krä mer [31] suggested, the cause of the hypersensitivity appeared to be an indirect effect related to   impaired Fe and Mn transport. Although the transcript data do support a role for the VTLs in Fe homeostasis, a role in homeostasis of other divalent cations analogous to the roles of NRAMP3 and NRAMP4 cannot be excluded.
Five Zn transport genes are among the Arabidopsis genes that response to Fe deficiency [32]. Four of these genes that transport Zn into the vacuole are strongly induced under Fe deficiency. The remaining gene, AtZIP3, was repressed and shown to catalyzed Zn uptake into cells. Zn has been shown to accumulate in roots grown under Fe-deficient conditions, likely the result of the Zn transport activity of IRT1 [3,31]. It is noteworthy that the expression of AtVTL1, AtVTL2 and AtVTL5 was greatly decreased under Zn deficiency but unchanged when grown under Zn excess (Fig. 7). Expression of AtVTL1, AtVTL2 and AtVTL5 in the yeast Dzrc1, a vacuolar Zn transport mutant, did not complement this mutation (Fig. S7).
Although a direct role of the VTL proteins in Zn transport is lacking, an involvement of Zn in Fe homeostasis has been clearly demonstrated. Transcriptional regulation in strategy I plants to Fe deficiency is mediated in part by the bHLH transcription factors FIT (bHLH029) [33,34,35] and POPEYE (bHLH047, PYE) [36]. The central components of PYE regulation include PYE (At3g47640), IAA-LEUCINE-RESISTANT3 (ILR3, bHLH105, At5g54680) and BRUTUS (BTS, At3g47640). PYE is weakly and BTS is strongly induced under Fe deficiency [20,36], and PYE and BTS both interact but with ILR3 in a yeast two-hybrid assay but not with each other [36]. Either a ternary complex of PYE-ILR3-BTS or the competition between PYE and BTS for ILR3 has been proposed as the regulatory switch in the PYE network [36]. PYE directly regulated genes that were associated with the response to Fe deficiency including: FRO3, NAS4 and ZIF1. The transcription of AtVTL1, 2 and 5 was increased 1.9, 16.9 and 2.9-fold, respectively, in the pye-1 knockout mutant [36], and roots of the pye-1 mutant had an elevated Fe content compared with controls. These observations are consistent with the positive correlation between the root Fe concentration and transcript abundance of AtVTL1, 2 and 5 (Fig. 7) [21].
In the ilr3-1 gain-of-function mutant, the transcript abundance of AtVTL1, 2 and 5 was decreased between 3-and 4-fold; an observation that can be explained by a gain-of-function in the PYE-ILR3-BTS repression [22]. In general, BTS and the rice homologs OsHRZs negatively regulate the Fe deficiency response in Arabidopsis and rice, and thus plants with decreased expression of these genes showed tolerance to Fe deficiency [36,37]. Recombinant BTS has been shown to bind both Fe and Zn, primarily at the hemerythrin domain of the protein [36]. In a recent model, Kobayshi and Nishizawa [38] have proposed a mechanism for transcriptional regulation through the BTS/ OsHRZ proteins based on binding of Zn and Fe to the hemerythrin domain of these proteins. In their model, Fe binding to BTS/OsHRZ in the absence of Zn would lead to repression PYE-regulated gene expression as seen in Fig. 7. The implications of their model are germane to the repression of VTL transcription under Zn deficiency observed in our experiments. Taken together, we have identified a group of putative Fe transporters that participate in the Fe homeostasis in Arabidopsis. As vacuolar Fe transporters these proteins may have importance in biofortification efforts in the future. Figure S1 Sucrose density gradient isolation of membrane in the yeast vacuolar fraction. The vacuolar fraction was isolated as described in the Materials and Methods, and the vacuoles ruptured by repeated pipetting. The membranes were layered onto a continuous, 10 to 60% sucrose gradient and centrifuged at 110,000xg in a swing-out rotor over-night. The gradient was fractionated into 1 ml fractions and marker enzymes for the vacuole (bafilomycin-sensitive ATPase), plasma membrane (vanadate-sensitive ATPase) and endoplasmic reticulum (cytochrome c reductase) were determined by the method of Luster and Buckhout (Plant Physiol. 1989