Arabidopsis IRT2 cooperates with the high-aynity iron uptake system to maintain iron homeostasis in root epidermal cells

Planta (2009) 229:1171 1179 DOI 10.1007/s00425-009-0904-8 ORIGINAL ARTICLE Arabidopsis IRT2 cooperates with the high-aynity iron uptake system to maintain iron homeostasis in root epidermal cells Grégory Vert Marie Barberon Enric Zelazny Mathilde Séguéla Jean-François Briat Catherine Curie Received: 12 December 2008 / Accepted: 6 February 2009 / Published online: 28 February 2009 Springer-Verlag 2009 Abstract Iron is an essential nutrient for all organisms but toxic when present in excess. Consequently, plants carefully regulate their iron uptake, dependent on the FRO2 ferric reductase and the IRT1 transporter, to control its homeostasis. Arabidopsis IRT2 gene, whose expression is induced in root epidermis upon iron deprivation, was shown to encode a functional iron/zinc transporter in yeast, and proposed to function in iron acquisition from the soil. In this study, we demonstrate that, unlike its close homolog IRT1, IRT2 is not involved in iron absorption from the soil since overexpression of IRT2 does not rescue the iron uptake defect of irt1-1 mutant and since a null irt2 mutant shows no chlorosis in low iron. Consistently, an IRT2- green Xuorescent fusion protein, transiently expressed in culture cells, localizes to intracellular vesicles. However, IRT2 appears strictly co-regulated with FRO2 and IRT1, supporting the view that IRT2 is an integral component of the root response to iron dewciency in root epidermal cells. We propose a model where IRT2 likely prevents toxicity from IRT1-dependent iron Xuxes in epidermal cells, through compartmentalization. Keywords Arabidopsis IRT2 Iron uptake Metal Transport Root G. Vert M. Barberon E. Zelazny M. Séguéla J.-F. Briat C. Curie (&) Biochimie et Physiologie Moléculaire des Plantes, Institut de Biologie Intégrative des Plantes, CNRS UMR 5004, 2 place Viala, 34060 Montpellier Cedex 1, France e-mail: curie@supagro.inra.fr Introduction Iron is an essential micronutrient for virtually all organisms, but it is not readily available due to its low solubility. In plants, highly specialized systems have evolved in order to acquire iron from the soil. Recent advances have identi- Wed molecular components of the iron uptake system, both in graminaceous plants (Curie et al. 2001) and in nongraminaceous plants (Eide et al. 1996; Henriques et al. 2002; Robinson et al. 1999; Varotto et al. 2002; Vert et al. 2002). In Arabidopsis thaliana, upon iron dewciency, expression of a set of genes encoding a proton ATPase, probably of the AHA family (Palmgren 2001), the ferric-chelate reductase FRO2 (Connolly et al. 2003; Robinson et al. 1999) and the metal transporter IRT1 (Eide et al. 1996; Vert et al. 2001; Vert et al. 2002, 2003) is induced in root epidermal cells. These proteins are responsible for the solubilization of iron through rhizosphere acidiwcation, its reduction from Fe 3+ to Fe 2+ and subsequent Fe 2+ uptake into the cytoplasm, respectively. Several families of putative iron transporters, including NRAMP, ZIP and YSL have been described in Arabidopsis (Curie and Briat 2003). Although the physiological role of YSL and NRAMP members is associated with distribution, partitioning and remobilization of iron in the plant, the ZIP founding member IRT1 is the only transporter involved in iron uptake (Colangelo and Guerinot 2004). IRT1 was Wrst identiwed through functional complementation of a yeast mutant defective in iron uptake (Eide et al. 1996). Characterization of IRT1 in Arabidopsis showed that this gene is highly induced by iron starvation in roots and speciwcally expressed in the epidermis and cortex (Eide et al. 1996; Vert et al. 2002). Moreover, the identiwcation of a knock-out mutant in the IRT1 gene, irt1-1, demonstrated that in low iron conditions, IRT1

1172 Planta (2009) 229:1171 1179 is the main route of iron entry into the plant and mediates the accumulation of additional metal ions (Vert et al. 2002). IRT2, a close homolog of IRT1 in Arabidopsis, was previously shown by functional complementation of yeast metal transport mutants to be a transporter of iron and zinc (Vert et al. 2001). Like IRT1, IRT2 is expressed in the outer layers of the root and upregulated, although to a lesser extent than IRT1, by iron dewciency (Vert et al. 2001). Even though IRT2 was suggested to be involved in iron uptake from the soil (Vert et al. 2001), little is known about its biological function. A loss of function En transposon mutant, irt2-1, was isolated and shown to display a slightly reduced stature but no chlorotic phenotype (Varotto et al. 2002). However, transposon insertion in irt2-1 led to concomitant downregulation of the neighboring IRT1 gene, playing itself a major role in iron uptake (Vert et al. 2002), that may result from the loss of an enhancer function acting on both IRT genes (Varotto et al. 2002). Therefore, the analysis of irt2-1 greatly limits the conclusion on the biological role of IRT2. The present study shows that IRT2 is an intracellular membrane protein whose role in the cellular response to iron dewciency is diverent from the major root iron transporter IRT1. We propose that IRT1 may compartmentalize iron into vesicles, to prevent toxicity by excess free iron in the cytosol as a result of extracellular iron acquisition by the FRO2/IRT1 system. Experimental procedures Plant growth conditions The Arabidopsis thaliana ecotypes used in this study were Columbia (Col-0) and Wassileskija (Ws). The irt1-1 and Wt- 1 mutants used are in the Ws background (Colangelo and Guerinot 2004; Vert et al. 2002). The iron dewciency treatments and split-root experiments were performed exactly as previously described (Seguela et al. 2008; Vert et al. 2003). BrieXy, the localized supply treatments were initiated by transferring one part of the root system to the Fe-free solution after washing the roots with 10 mm sodium dithionite and 1.5 mm 2,2-bipyridyl for 5 min (Bienfait et al. 1985), the other part remaining supplied with 100 μm Fe-EDTA. At this stage, special care was taken to avoid cross-contamination between compartments. Plants were grown for an additional 3 days before harvesting. Plasmid constructions For the IRT2-GFP fusion, IRT2 cdna sequence (At4g19680) was ampliwed by PCR using the following primers: IRT2-GFP F 5 -gctctagataatggctactacc AAGCTCGTC-3, and IRT2-GFP R 5 -cgggatcctttt AGCCCACACGGCGACG-3. Care was taken to substitute the stop codon TAA of IRT2 by AAA, encoding a lysine residue, to fuse IRT2 in frame with the GFP sequence. IRT2 PCR product was cloned in the BamHI and XbaI site of the pbi221 vector (Clontech, Palo Alto, CA, USA) to fuse the GFP sequence in 3 of the IRT2 sequence, under the control of the caulixower mosaic virus CaMV35S promoter. To generate IRT2 overexpressing plants, the IRT2 coding sequence was ampliwed by PCR using the following primers : IRT2ox F 5 -cgggatccataatggctactacc AAGCTCG-3 and IRT2ox R 5 -ggggtaccgtttaagc CCACACGGCGAC-3. The corresponding product was cloned between the BamHI and KpnI restriction sites of the pbib HYG vector (Becker 1990), modiwed to carry the CaMV 35S promoter. The MP90 strain of Agrobacterium tumefaciens was used to transform Arabidopsis following the Xoral dip protocol (Clough and Bent 1998). Seeds obtained from the primary transformants were germinated on hygromycincontaining plates and analyses were carried out on the resistant plants (T2). Metal content The iron concentration in leaves was determined for soilgrown plants as described (Lobreaux and Briat 1991). For heavy metal content (Zn, Mn, Co, Cu, Cd), samples were washed for 5 min in a solution containing 5 mm CaSO 4,10 mm EDTA, dried overnight at 70 C, weighed, then 1 g of dry tissue was completely digested in 70% HNO 3 at 120 C. Trace elements were analyzed by ICP-MS (inductively coupled plasma-mass spectrometry) at the University of Montpellier II and measurements were performed using a PQ II Turbo+ quadrupole ICP mass spectrometer (VG Elemental, England). Prior to analysis, solutions were diluted by a factor of ca. 100, and In and Bi were added to aliquots of the solutions as internal standards for drift correction. Gene expression analyses Total RNA was extracted from plants grown axenically. Northern blot analysis were performed exactly as described in (Vert et al. 2002), except that a 1.1 kb IRT2 probe was used. For RT-PCR experiments, transcripts were treated with DNase before being reverse transcribed. The following primers were designed and used for semi-quantitative PCR :EF1α F 5 -ATGGGTAAAGAGAAGTTTCACATC-3 ; EF1α R 5 -ACCAATCTTGTAGACATCCTGAAG-3 ; IRT2 F 5 -CGTAGCCATTGTTGCCATAC-3 ; IRT2 R 5 -GGT TCCAAGGATGATTCCAG-3. For real-time quantitative

Planta (2009) 229:1171 1179 1173 PCR, the following primers speciwc for the target gene were used : Q-EF-1α F 5 -TGGGTAAAGAGAAGTTTC ACATC-3 ; Q-EF-1α R 5 -ACCAATCTTGTAGACATCC TGAAG-3 ; Q-IRT2 F 5 -CGTAGCCATTGTTGCCATA C-3 ; Q-IRT2 R 5 -GGTTCCAAGGATGATTCCAG-3.; Q-IRT1 F 5 -CGGTTGGACTTCTAAATGC3 ; Q-IRT1 R 5 -CGATAATCGACATTCCACCG-3 ; Q-NR3 F 5 -ACA ATGGGAGTCTCATTCGC -3 ; Q-NR3 R 5 -ATGCAAC CCACAACTCCAAC-3 ; Q-NR4 F 5 -CTTGGATGTTT GGTCAGACG -3 ; Q-NR4 R 5 -TCGACTTCTCTGGAT TGCAC -3. For real-time quantitative PCR, experiments were done in duplicates and results were normalized according to EF-1α ampliwcation. The result of one experiment is shown. Subcellular localization Protoplasts made from an Arabidopsis cell suspension culture (Axelos et al. 1992) were transfected using a PEG protocol derived from (Schirawski et al. 2000). To observe the GFP Xuorescence in transfected protoplasts, a Biorad 1024 CLSM system coupled to a Nikon Optiphot II upright microscope and an Argon Krypton ion laser (15 mw) was used. Excitation wavelength: 488 nm, emission Wlters: 520/ 30 nm for GFP Xuorescence and 680/32 nm for chlorophyll autoxuorescence. Observation was performed using a Nikon objective X20 (O,75NA). Results IRT2 is not involved in iron uptake A previous study indicated that IRT2, like its close homolog IRT1, might be involved in iron uptake from the soil in conditions of iron dewciency (Vert et al. 2001), although no clear evidence from reverse genetics was available. To gain further insight into the biological function of IRT2, we identiwed a T-DNA insertion Arabidopsis mutant in the IRT2 gene, named irt2-2, from the SIGnAL collection (line SALK_106691). A similar strategy previously led us to characterize IRT1 transport properties in planta (Vert et al. 2002). The irt2-2 line contains two copies of T-DNA inserted in opposite orientation in the second exon of the IRT2 gene (Fig. 1a), which prevents synthesis of the corresponding full length mrna (Fig. 1b). The homozygous null mutant plant for IRT2 does not show any macroscopic phenotype when grown in standard soil conditions (Fig. 1c), or in vitro in iron-dewcient conditions. This result is in contrast with the dramatic chlorosis displayed by the irt1-1 knock-out mutant (Fig. 1c, [4]) and shows clearly that the function played by IRT2 in the mechanism of iron homeostasis, if any, is diverent from that of IRT1. We Fig. 1 IRT2 does not share overlapping function with IRT1. a Scheme of the T-DNA integration site in the irt2-2 knock-out mutant. The location of the insertion in the IRT2 coding sequence is shown. Exons are boxed. b Molecular characterization of irt2-2. RT-PCR was performed on total RNA extracted from wild-type (WT) and irt2-2 plants grown in standard conditions. EF-1α ampliwcation was included as a control. c Phenotype of irt2-2 plants. Wild-type (WT), irt1-1, and irt2-2 plants were grown for 3 weeks in soil. d Northern blot analysis. IRT2 coding sequence was hybridized to a blot containing 15 μg of total RNA extracted from wild-type (WT), or T2 lines overexpressing IRT2 cdna in the irt1-1 genetic background under the control of the CaMV35S promoter, grown under standard conditions. e Phenotype of irt1-1/35s::irt2 plants. Wild-type (WT), irt1-1, and irt1-1/35s::irt2 #3 plants were grown for 3 weeks in soil. All the irt1-1/35s::irt2 transgenic lines showed the same phenotype obtained the same result with irt2-3, a second mutated allele of the IRT2 gene identiwed in the SIGnAL collection (line SALK_023336, data not shown). Based on the lack of phenotype of the irt2 knock-out mutants and on IRT2

1174 Planta (2009) 229:1171 1179 Fig. 2 Subcellular localization of the IRT2 protein. Arabidopsis protoplasts transfected with constructs expressing GFP alone (a), IRT2-GFP (b) or IRT1-GFP (c) and observed by confocal microscopy 2 days after transfection. An overlay of chlorophyll and GFP signals is shown relatively low expression level and induction by iron starvation (Vert et al. 2001), the IRT2 transporter might either play a minor role in the overall iron uptake process in the plant or play an entirely diverent role. Because IRT1 is a very eycient metal transporter in low iron conditions, its activity could potentially compensate for the lack of IRT2. To rule out a role of IRT2 in iron uptake, we tested whether its overexpression could rescue the iron uptake defect of the irt1-1 knock-out mutant. We therefore overexpressed IRT2 cdna under the control of the strong and constitutive CaMV 35S promoter in the irt1-1 mutant background to look for any reversion of phenotypes provoked by the lack of iron. Five transgenic lines displaying higher IRT2 expression levels than wild-type plants were selected for further analyses (Fig. 1d). Surprisingly, no complementation of irt1-1 chlorosis and growth defect could be observed with any of these Wve transgenic lines, as presented for line # 3 (Fig. 1e), conwrming previous observations from Varotto et al. (2002). This shows that the two iron transporters IRT2 and IRT1, although co-expressed in root epidermis cells in response to iron dewciency (Vert et al. 2001; Vert et al. 2002), do not share overlapping functions. Furthermore, these data imply that IRT2 does not perform iron uptake from the soil. IRT2 is an intracellular membrane protein We had previously shown that IRT1 is addressed to the plasma membrane when expressed in protoplasts (Vert et al. 2002). The PsortI (Nakai and Kanehisa 1992) and TargetP (Emanuelsson et al. 2000) programs predict plasma membrane localization for both IRT1 and IRT2 proteins. However, the fact that IRT2 overexpression fails to rescue irt1-1 mutant prompted us to test whether IRT1 and IRT2 are localized on diverent cellular membranes. We fused the green Xuorescent protein (GFP) to the C-terminus of IRT2 and transiently expressed the corresponding fusion protein under the control of the CaMV 35S promoter in protoplasts made from an Arabidopsis suspension cell culture. Fluorescence of the GFP was observed by confocal microscopy 24 72 h after transfection, during which period no variation in the pattern of GFP localization was detected. Cells expressing GFP alone labeled homogeneously the cytoplasm and the nucleus (Fig. 2a), whereas IRT2-GFP displayed a punctuate pattern of Xuorescence in the cytoplasm (Fig. 2b). The numerous small Xuorescent vesicles observed for IRT2-GFP were in constant motion in the cytoplasm, and did not co-localize with chloroplasts as shown by the overlay with chlorophyll autoxuorescence (Fig. 2b). The close IRT2 homolog, IRT1, was shown to be targeted to the plasma membrane using the exact same strategy (Fig. 2c; Vert et al. 2002). Feeding protoplasts with iron or adding the iron chelator ferrozine had no evect on IRT2 subcellular localization (data not shown). Therefore, unlike IRT1, IRT2 is not targeted to the plasma membrane. Instead, IRT2 is located to intracellular membrane compartments of unknown nature, thus explaining the lack of signiwcant functional complementation of the irt1-1 knock-out by the overexpression of IRT2. IRT2 overexpression leads to heavy metal overaccumulation in plants Functional complementation of the fet3fet4 and zrt1zrt2 yeast mutants, defective in high and low aynity iron and zinc uptakes respectively, indicated that the Arabidopsis IRT2 protein functions as a transporter of iron and zinc (Vert et al. 2001). In contrast with the well-known broad spectrum metal transporter IRT1, IRT2 appeared unable to transport manganese and cadmium when expressed in yeast. To shed light on the function of IRT2 in planta, we wanted to know whether its overexpression had any evect on the overall plant metal content. Various transgenic lines were analyzed for IRT2 overaccumulation (Fig. 3a). Line 35S::IRT2#3, harboring the strongest IRT2 mrna levels, was selected for further experiments (Fig. 3a). Metal content was measured by ICP-MS on leaves from 35S::IRT2#3 plants cultivated in soil to avoid interference of the large non-speciwc apoplastic metal pools. Plants overexpressing IRT2 accumulated more iron and zinc, respectively, in their

Planta (2009) 229:1171 1179 1175 leaves compared to wild-type plants (Fig. 3b). In addition to Fe and Zn, 35S::IRT2 plants also contained increased levels of Mn and Cd (Fig. 3b), even though IRT2 did not confer Mn and Cd transport activities in yeast (Vert et al. 2001). Metal content analyses of the irt2-2 mutant revealed however no signiwcant diverences with wild-type plants (data not shown). Thus, from these data, we conclude that overexpressing IRT2 leads to overaccumulation of Fe, Zn, Mn and Cd in transgenic plants. IRT1 and IRT2 act together to maintain iron homeostasis The subcellular localization of IRT2 is incompatible with the metal overexpression phenotype observed in 35S::IRT2 transgenic plants. Moreover, the range of metals accumulated in such plants (Fe, Zn, Mn, Cd, including the absence of accumulation of Cu, Fig. 3b), strikingly resembled the substrate speciwcity identiwed for IRT1 (Eide et al. 1996; Korshunova et al. 1999; Vert et al. 2001; Vert et al. 2002). It was therefore tempting to speculate that altering IRT2 expression results in de-regulation of IRT1 expression. To test this hypothesis, we examined IRT1 expression in roots of IRT2 overexpressing plants by real-time quantitative RT- PCR. Figure 4 shows that IRT1 transcript accumulation is enhanced in 35S::IRT2 plants, and correlates with IRT2 overexpression levels. This conwrms that overexpressing IRT2 does lead to an upregulation of IRT1 and explains why such plants show a metal accumulation harboring IRT1 s signature. These experiments also reveal the existence of a tight link between IRT1 and IRT2 transport functions that both take part in the process of iron uptake in response to iron dewciency in root peripheral cell layers. Coordinate regulation of IRT2 and the root iron uptake genes A previous study had shown that IRT2 transcripts accumulate mildly in root epidermal cells upon iron dewciency (Vert et al. 2001). This pattern matches the one of FRO2 and IRT1, encoding the key components of the Arabidopsis high-aynity iron uptake system (Connolly et al. 2003; Vert et al. 2002). To highlight the close relationship between the endomembrane-localized IRT2 transporter and the iron uptake machinery, we further examined the regulation of IRT2. We Wrst investigated if IRT2 was under the control of the master regulator FIT, which has been shown to promote FRO2 and IRT1 mrna accumulation in the root epidermis (Colangelo and Guerinot 2004). Real-time quantitative RT-PCR indicates that IRT2 mrna levels drop in the Wt-1 mutant background compared with wild-type plants (Fig. 5a). This suggests that IRT2 is co-regulated with the Fig. 3 IRT2 overexpression leads to metal overaccumulation. a Northern blot analysis. IRT2 coding sequence was hybridized to a blot containing 10 μg of total RNA extracted from wild-type (WT), or T2 lines overexpressing IRT2 cdna under the control of the CaMV35S promoter, grown under standard conditions. b Elemental analysis. Aerial parts from both wild-type and IRT2 overexpressing lines 35S::IRT2#3 grown for 3 weeks in soil were processed for inductively coupled plasma-mass spectrometry (see Experimental Procedures ). Results are presented as mean of two separate measures. Error bars indicate standard error

1176 Planta (2009) 229:1171 1179 Fig. 4 IRT2-dependent deregulation of IRT1 expression. Real-time quantitative RT-PCR analysis. Total RNA extracted from wild-type (WT), 35S::IRT2 lines #1 and #3 plants grown under standard conditions were subjected to quantitative RT-PCR for IRT1 gene expression. Signals were normalized according to the ampliwcation of EF-1α. Experiments were done in duplicates and results were normalized according to EF-1α ampliwcation. The result of one experiment is shown root iron machinery encoding genes, although not directly involved in iron acquisition from the soil. We have previously used a split-root approach, in which only half of the root system is subjected to iron dewciency, to show that not only the expression of FRO2 and IRT1 is regulated by iron dewciency, but also dually controlled by local and long-distance signals (Vert et al. 2003). Iron serves indeed as a local inducer of the iron dewciency responses which are, in counterpart, submitted to the feedback control by a systemic signal. Using the same approach, we have investigated the regulation of IRT2 to determine whether IRT2 responds to long-distance signals. Real-time quantitative RT-PCR analyses indicated that IRT2 mrna amounts reproducibly increased in the part of roots still supplied with iron similarly to IRT1 and FRO2 (Fig. 5b; Vert et al. 2003). Although IRT2 induction in the +Fe-grown root compartment is milder than what has been reported for IRT1, consistent with the fact that the induction of gene expression by iron dewciency is much lower for IRT2 (Vert et al. 2001), both genes show qualitatively the same response to iron nutrition in split root. We also analyzed the response to split-root conditions of two other ironstarvation induced genes, AtNRAMP3 and AtNRAMP4, which encode iron transporters involved in other processes than iron uptake from the soil (Lanquar et al. 2005; Fig. 5 Co-regulation of IRT2 with the iron uptake machinery genes. a IRT2 mrna levels were monitored by real-time quantitative RT-PCR on total RNA extracted from 10 day-old roots of wild-type and Wt-1 plants. Iron-replete plants (7 days) were transferred for 3 days on iron-dewcient (300 μm ferrozine, Fe) or iron-suycient (50 μm, +Fe) medium. b Real-time quantitative RT-PCR analysis on root total RNA extracted from 10-day-old plants grown in split-root conditions. mrna levels of IRT2, NRAMP3 and NRAMP4 were monitored, and induction fold compared with wild-type plants determined for each gene. Experiments were done in duplicates and results were normalized according to EF-1α ampliwcation. The result of one experiment is shown Thomine et al. 2003). Interestingly, both genes showed a radically diverent regulation to split-root conditions, with their respective mrna accumulating in the -Fe compartment. This indicates that IRT2 is tightly coupled to the iron

Planta (2009) 229:1171 1179 1177 uptake machinery, and suggests that FRO2, IRT1 and IRT2 are part of the same cellular responses under iron dewciency to ensure a proper uptake of iron by Arabidopsis root epidermal cells. Discussion In a previous work, we had shown that the Arabidopsis IRT2 gene encodes a putative transporter of iron and zinc which, like its close homolog the major root Fe 2+ transporter IRT1, is expressed in root epidermal cells in response to iron dewciency (Vert et al. 2001). Preliminary investigations suggested that IRT2 may have a minor function in iron uptake (Varotto et al. 2002; Vert et al. 2002). It is now clear that IRT2 is not directly involved in iron uptake. Indeed, we show that irt2 null mutants show no apparent phenotype, in contrast to the dramatic chlorosis and growth defect displayed by the irt1-1 mutant (Fig. 1c; Vert et al. 2002). Furthermore, overexpression of IRT2 in irt1-1 plants does not reduce their chlorotic symptoms, thus showing that IRT2 is not able to replace IRT1 in the function of iron uptake from the soil (Fig. 1e; Varotto et al. 2002). We also show here that an IRT2-GFP fusion protein is addressed to the membrane of internal vesicles in transfected Arabidopsis cells (Fig. 2b), whereas using the same experimental system we had previously observed a plasma membrane localization of a similar IRT1-GFP fusion (Vert et al. 2002). Because IRT2 is upregulated by iron dewciency, a tempting hypothesis is that IRT2 function in epidermal cells would be to remobilize iron stores from internal storage vesicles when cytosolic iron is low. As yet, we have found no evidence to support this hypothesis. On the contrary, it is contradicted by the fact that 35S::IRT2 plants show a clear upregulation of IRT1 (Fig. 4), arguing in favor of a depletion of cytosolic iron by IRT2. In addition to the deregulation of IRT1 in 35S::IRT2 genetic backgrounds shown in the present study, we had previously observed that IRT2 is upregulated in the irt1-1 knock-out line (Vert et al. 2002). This indicates that the activity of these two transporters is integrated in the root epidermal cell. Expression of IRT1 is controlled by the FIT bhlh transcription factor, as attested by the low level of IRT1 mrna accumulation in the Wt-1 mutant (Colangelo and Guerinot 2004; Jakoby et al. 2004; Seguela et al. 2008; Yuan et al. 2008). Our results show that IRT2 is also controlled by FIT (Fig. 5a). Furthermore, we show in this study, through a split-root experiment, that as previously demonstrated for IRT1 and FRO2 in Arabidopsis and other species (Schikora et al. 2006; Vert et al. 2003), IRT2 expression is controlled both locally and systemically by the iron status (Fig. 5b), further supporting the strict co-regulation with IRT1. The co-regulation of the IRT2 gene with the genes encoding the two major components of the iron uptake machinery argues for the implication of IRT2 in the same cellular process, i.e. iron absorption from the soil, although not involved in iron uptake per se. The investigation of the intracellular localization of the IRT2-GFP fusion protein revealed an endomembrane localization. How to reconcile the fact that IRT2 expression complements the growth defect of yeast mutants avected in metal uptake (Vert et al. 2001) whereas it is localized to an endomembrane compartment in planta? The strong overexpression resulting from the use of a yeast expression vector carrying a 2μ replication origin and the strong phosphoglycerate kinase promoter may lead to mis-sorting of proteins in the cell. This was already reported for the AtNRAMP3/4 vacuolar iron transporters that complement the fet3fet4 iron uptake mutant (Lanquar et al. 2005; Thomine et al. 2003). Therefore, functional heterologous expression in yeast does not necessarily provide information about the intracellular targeting of the protein in its native environment. Other non-plant members of the ZIP family of metal transporters are targeted, as described in the present study for IRT2, to endomembrane compartments and allow sequestration or release of metals in and from such compartments (Huang et al. 2005; Kumanovics et al. 2006). However, we cannot exclude that IRT2 might dynamically localize to diverent membranes in response to a given stimulus, as reported for several ZIP transporters in mammals (Dufner-Beattie et al. 2004). This signal appears to be independent of iron since (1) IRT2-GFP localization in protoplasts is not avected by iron (data not shown) and (2) IRT2 overexpression in the strongly chlorotic irt1-1 mutant did not rescue its phenotype (Fig. 1e). If not involved in external iron uptake through the plasma membrane, what is then the role of IRT2 in the iron dewciency response? It is possible that speciwc processes requiring iron take place in the IRT2-localized vesicles, that are either necessary to the iron dewciency response or to the metabolism of iron-limited cells. A second explanation could reconcile the fact that IRT2 is both involved in iron compartmentalization and synthesized in iron-dewcient roots. Low iron levels, that trigger the iron dewciency response of the roots, lead to fast and strong expression of FRO2 and IRT1. As a result, cytosolic iron content is likely to rapidly rise to ranges that are toxic for the cell. The coinduction of IRT2 with the genes responsible for iron uptake could prevent toxicity through rapid sequestration of iron into vesicles. This phenomenon is known as a proactive adaptation to environmental change. The involvement of such a proactive mechanism was Wrst reported for the control of zinc homeostasis in yeast (MacDiarmid et al. 2003). Under low zinc, plasma membrane zinc uptake transporters are highly expressed, which leads to a rapid accumulation of large amounts of zinc when the metal is

1178 Planta (2009) 229:1171 1179 re-supplied to the cells, a condition referred to as zinc shock. The ZRC1 vacuolar Zn transporter, that is induced prior to the shock in Zn limited cells, is required for tolerance to this Zn shock. A similar scenario was reported in Arabidopsis where the tonoplastic AtMTP3 protein that accumulates in the root epidermis under iron dewciency conditions was proposed to contribute to basic cellular Zn tolerance, particularly under conditions of high Zn inxux into the root symplasm such as Fe (Arrivault et al. 2006). Although we favor this model for a role of IRT2 in planta, we were unable to monitor signiwcant resistance or hypersensitivity to re-supply of iron after starvation in 35S::IRT2 and irt2 plants, respectively. During the past decade, reverse genetics provided a wealth of information on iron homeostasis genes. The next challenge will be to understand how the diverent actors of the iron homeostasis are integrated in the plant to optimize iron uptake, distribution and storage to ensure proper growth and development of the plant. Acknowledgments The authors are grateful to O. Bruguier (IS- TEEM, University of Montpellier, France) for assistance in ICP-MS analyses. Work was supported by a B.D.I. fellowship awarded by the Centre National de la Recherche ScientiWque (G.V) and by a thesis fellowship from the French Ministry of National Education, Research and Technology (M.B., M.S). Research was supported in part by an ACI (2000-51) from MENRT. References Arrivault S, Senger T, Kramer U (2006) The Arabidopsis metal tolerance protein AtMTP3 maintains metal homeostasis by mediating Zn exclusion from the shoot under Fe dewciency and Zn oversupply. Plant J 46:861 879 Axelos M, Curie C, Mazzolini L, Bardet C, Lescure B (1992) A protocol for transient expression in Arabidopsis thaliana protoplasts isolated from cell suspension culture. Plant Physiol Biochem 30: 128 Becker D (1990) Binary vectors which allow the exchange of plant selectable markers and reporter genes. Nucleic Acids Res 18:203 Bienfait HF, Briel W, Mesland-Mul NT (1985) Free space iron pools in roots: generation and mobilization. Plant Physiol 78:596 600 Clough SJ, Bent AF (1998) Floral dip: a simpliwed method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J 16:735 743 Colangelo EP, Guerinot ML (2004) The essential basic helix-loop-helix protein FIT1 is required for the iron dewciency response. Plant Cell 16:3400 3412 Connolly EL, Campbell NH, Grotz N, Prichard CL, Guerinot ML (2003) Overexpression of the FRO2 ferric chelate reductase confers tolerance to growth on low iron and uncovers posttranscriptional control. Plant Physiol 133:1102 1110 Curie C, Briat JF (2003) Iron transport and signaling in plants. Annu Rev Plant Biol 54:183 206 Curie C, Panaviene Z, Loulergue C, Dellaporta SL, Briat JF, Walker EL (2001) Maize yellow stripe1 encodes a membrane protein directly involved in Fe(III) uptake. Nature 409:346 349 Dufner-Beattie J, Kuo YM, Gitschier J, Andrews GK (2004) The adaptive response to dietary zinc in mice involves the diverential cellular localization and zinc regulation of the zinc transporters ZIP4 and ZIP5. J Biol Chem 279:49082 49090 Eide D, Broderius M, Fett J, Guerinot ML (1996) A novel ironregulated metal transporter from plants identiwed by functional expression in yeast. Proc Natl Acad Sci USA 93:5624 5628 Emanuelsson O, Nielsen H, Brunak S, von Heijne G (2000) Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J Mol Biol 300:1005 1016 Henriques R, Jasik J, Klein M, Martinoia E, Feller U, Schell J, Pais MS, Koncz C (2002) Knock-out of Arabidopsis metal transporter gene IRT1 results in iron dewciency accompanied by cell diverentiation defects. Plant Mol Biol 50:587 597 Huang L, Kirschke CP, Zhang Y, Yu YY (2005) The ZIP7 gene (Slc39a7) encodes a zinc transporter involved in zinc homeostasis of the Golgi apparatus. J Biol Chem 280:15456 15463 Jakoby M, Wang HY, Reidt W, Weisshaar B, Bauer P (2004) FRU (BHLH029) is required for induction of iron mobilization genes in Arabidopsis thaliana. FEBS Lett 577:528 534 Korshunova YO, Eide D, Clark WG, Guerinot ML, Pakrasi HB (1999) The IRT1 protein from Arabidopsis thaliana is a metal transporter with a broad substrate range. Plant Mol Biol 40:37 44 Kumanovics A, Poruk KE, Osborn KA, Ward DM, Kaplan J (2006) YKE4 (YIL023C) encodes a bidirectional zinc transporter in the endoplasmic reticulum of Saccharomyces cerevisiae. J Biol Chem 281:22566 22574 Lanquar V, Lelievre F, Bolte S, Hames C, Alcon C, Neumann D, Vansuyt G, Curie C, Schroder A, Kramer U, Barbier-Brygoo H, Thomine S (2005) Mobilization of vacuolar iron by AtNRAMP3 and AtNRAMP4 is essential for seed germination on low iron. EMBO J 24:4041 4051 Lobreaux S, Briat JF (1991) Ferritin accumulation and degradation in diverent organs of pea (Pisum sativum) during development. Biochem J 274:601 606 MacDiarmid CW, Milanick MA, Eide DJ (2003) Induction of the ZRC1 metal tolerance gene in zinc-limited yeast confers resistance to zinc shock. J Biol Chem 278(17):15065 15072 Nakai K, Kanehisa M (1992) A knowledge base for predicting protein localization sites in eukaryotic cells. Genomics 14:897 911 Palmgren MG (2001) Plant plasma membrane H + -ATPases: powerhouses for nutrient uptake. Annu Rev Plant Physiol Plant Mol Biol 52:817 845 Robinson NJ, Procter CM, Connolly EL, Guerinot ML (1999) A ferric-chelate reductase for iron uptake from soils. Nature 397:694 697 Schikora A, Thimm O, Linke B, Buckhout TJ, Müller M, Schmidt W (2006) Expression, localization, and regulation of the iron transporter LeIRT1 in tomato roots. Planta 284:101 108 Schirawski J, Planchais S, Haenni AL (2000) An improved protocol for the preparation of protoplasts from an established Arabidopsis thaliana cell suspension culture and infection with RNA of turnip yellow mosaic tymovirus: a simple and reliable method. J Virol Methods 86:85 94 Seguela M, Briat JF, Vert G, Curie C (2008) Cytokinins negatively regulate the root iron uptake machinery in Arabidopsis through a growth-dependent pathway. Plant J 55:289 300 Thomine S, Lelievre F, Debarbieux E, Schroeder JI, Barbier-Brygoo H (2003) AtNRAMP3, a multispeciwc vacuolar metal transporter involved in plant responses to iron dewciency. Plant J 34:685 695 Varotto C, Maiwald D, Pesaresi P, Jahns P, Salamini F, Leister D (2002) The metal ion transporter IRT1 is necessary for iron homeostasis and eycient photosynthesis in Arabidopsis thaliana. Plant J 31:589 599 Vert G, Briat JF, Curie C (2001) Arabidopsis IRT2 gene encodes a root-periphery iron transporter. Plant J 26:181 189 Vert G, Grotz N, Dedaldechamp F, Gaymard F, Guerinot ML, Briat JF, Curie C (2002) IRT1, an Arabidopsis transporter essential for iron

Planta (2009) 229:1171 1179 1179 uptake from the soil and for plant growth. Plant Cell 14:1223 3 Vert GA, Briat JF, Curie C (2003) Dual regulation of the Arabidopsis high-aynity root iron uptake system by local and long-distance signals. Plant Physiol 132:796 804 Yuan Y, Wu H, Wang N, Li J, Zhao W, Du J, Wang D, Ling HQ (2008) FIT interacts with AtbHLH38 and AtbHLH39 in regulating iron uptake gene expression for iron homeostasis in Arabidopsis. Cell Res 18:385 397