Arabidopsis Yellow Stripe-Like2 (YSL2): a metal-regulated gene encoding a plasma membrane transporter of nicotianamine–metal complexes


  • Raymond J. DiDonato Jr,

    1. Biology Department, University of Massachusetts, Amherst, MA 01003, USA, and
      1Microbiology Department, University of Massachusetts, Amherst, MA 01003, USA
    Search for more papers by this author
  • 1 Louis A. Roberts,

    1. Biology Department, University of Massachusetts, Amherst, MA 01003, USA, and
      1Microbiology Department, University of Massachusetts, Amherst, MA 01003, USA
    Search for more papers by this author
    • Current address: Louis A. Roberts, Biomedical Research Institute, Baystate Medical Center, Springfield, MA 01199, USA.

  • Tamara Sanderson,

    1. Biology Department, University of Massachusetts, Amherst, MA 01003, USA, and
      1Microbiology Department, University of Massachusetts, Amherst, MA 01003, USA
    Search for more papers by this author
  • Robynn Bosler Eisley,

    1. Biology Department, University of Massachusetts, Amherst, MA 01003, USA, and
      1Microbiology Department, University of Massachusetts, Amherst, MA 01003, USA
    Search for more papers by this author
  • Elsbeth L. Walker

    Corresponding author
    1. Biology Department, University of Massachusetts, Amherst, MA 01003, USA, and
      1Microbiology Department, University of Massachusetts, Amherst, MA 01003, USA
    Search for more papers by this author

(fax +1 413 545 3243; e-mail


The Yellow Stripe-Like (YSL) family of proteins has been identified based on sequence similarity to maize Yellow Stripe1 (YS1), the transporter responsible for the primary uptake of iron from the soil. YS1 transports iron that is complexed by specific plant-derived Fe(III) chelators called phytosiderophores (PS). Non-grass species of plants neither make nor use PS, yet YSL family members are found in non-grass species (monocot, dicot, gymnosperm, and moss species) including Arabidopsis thaliana. YSLs in non-grasses have been hypothesized to transport metals complexed by nicotianamine (NA), an iron chelator that is structurally similar to PS and which is found in all higher plants. Here we show that Arabidopsis YSL2 (At5g24380) transports iron and copper when these metals are chelated by NA. YSL2 is expressed in many cell types in both roots and shoots, suggesting that diverse cell types obtain metals as metal–NA complexes. YSL2 transcription is regulated by the levels of iron and copper in the growth medium. Based on its expression pattern, a major function of the YSL2 appears to be in the lateral movement of metals in the vasculature.


Plants, like all organisms, achieve metal ion homeostasis in several ways. The first is by controlling primary uptake so that adequate but non-toxic amounts of metal ions are acquired from the soil. Our understanding of the mechanisms of primary uptake has improved rapidly in recent years, especially for the uptake of iron and zinc (Curie et al., 2001; Eide et al., 1996; Grotz et al., 1998; Henrique et al., 2002; Lasat et al., 2000; Pence et al., 2000; Robinson et al., 1999; Varotto et al., 2002; Vert et al., 2002). Another homeostatic mechanism is sequestration of excess metals to prevent them from causing oxidative damage to cells. A number of such homeostatic mechanisms have been described, for example, the metallothioneins, phytochelatins (recently reviewed by Cobbett and Goldsbrough, 2002) and ferritin (reviewed by Briat et al., 1999) all serve to sequester metals intracellularly. A vacuolar Nramp protein responsible for re-mobilization of vacuolar metal stores has recently been described (Thomine et al., 2003).

Although no less important than uptake and sequestration, the mechanisms controlling the correct distribution of metals into particular organs, tissues and cells of the plant have received less attention. Because many metals are highly reactive, long-distance movement probably involves relatively non-reactive complexed or chelated metals. For example, an iron transport protein that appears to carry iron during phloem movement has recently been described (Kruger et al., 2002). The processes of metal ion partitioning on the whole plant level are not well characterized, and the molecular mechanisms by which metal ions are moved throughout the plant body have not been examined in detail.

The non-proteinogenic amino acid nicotianamine (NA) appears to play a key role in metal partitioning in plants. Nicotianamine is a strong complexor of various transition metals, particularly Fe(II) (Anderegg and Ripperger, 1989) and Fe(III) (von Wiren et al., 1999), as well as Cu(II), Ni(II), Co(II), Mn(II) and Zn(II) (Anderegg and Ripperger, 1989). Fe(II)–NA complexes are relatively inactive Fenton reagents, and so complexation of iron by NA has been proposed as a means by which plants control oxidative damage by iron during transport (Reichman and Parker, 2002; von Wiren et al., 1999). NA is present in shoots and roots at concentrations ranging between 20 and 500 nmol g−1 fresh weight (Stephan et al., 1990), and is present in both xylem (approximately 20 μm; Pich and Scholz, 1996) and phloem (approximately 130 μm; Schmidke and Stephan, 1995), suggesting that NA is a major complexor of metals throughout the plant.

Much of our understanding of the physiologic function of NA comes from the complex phenotype of a tomato mutant called chloronerva. The chloronerva mutation disrupts the function of the single gene in tomato encoding NA synthase (Herbik et al., 1999; Higuchi et al., 1999; Ling et al., 1999). Homozygous chl plants exhibit symptoms of iron deficiency: interveinal chlorosis in young leaves and upregulation of the iron acquisition system (extrusion of protons into the medium and increased ferric reductase activity; Pich and Scholz, 1991; Stephan and Grun, 1989). Mature leaves do become fully green, albeit slowly, probably because iron begins to arrive via xylem once the leaf matures fully and transpiration occurs at sufficient rates. The levels of iron, zinc, manganese and copper are all below normal in shoot apices, which rely on phloem transport for their nutrient supply (Pich et al., 1994). This, together with the deficiency symptoms in young tissues, has led to the conclusion that NA is required for phloem transport of these metals.

In contrast to the metal deficiency of young tissues in chl mutants, measurements of metal concentration in mature tissues indicate that excess iron, zinc and manganese (Pich et al., 1994) are accumulated. Thus, NA does not appear to be required for either primary uptake or xylem translocation of these metals. Most of the excess iron that is transported to mature leaves in chloronerva plants is symplasmic, indicating that iron unloaded from the xylem is taken up by leaf cells in the absence of NA (Becker et al., 1992). However, the high concentration of iron in leaf cells leads to the formation of intracellular precipitates (Becker and Manteuffel, 1995; Liu et al., 1998). The presence of these precipitates has been interpreted as an indication that NA is required for maintaining the solubility of intracellular iron pools, but it is difficult to separate the effect of high overall concentrations of iron from the effect of the absence of NA.

Although NA may not be necessary for root-to-shoot translocation of iron, zinc and manganese, it does appear to be required for translocation of copper to shoots. Homozygous chloronerva mutant plants have excess copper in their roots, but fail to transport normal amounts of copper to mature leaves. Furthermore, xylem exudates from mutant plants contain unusually low levels of copper (Pich and Scholz, 1996). Based on physical measurements, von Wiren et al. (1999) predicted that, in wild-type plants, 100% of the copper in the xylem would occur as NA complexes.

Recently, a family of proteins called Yellow Stripe-Like (YSL) was suggested to transport metals bound to NA (Curie et al., 2001). This suggestion was based on the sequence similarity between the YSL genes deduced from the Arabidopsis thaliana genome sequence and the recently cloned Yellow Stripe1 gene from maize (ZmYS1). YS1 is a proton-coupled symporter (Schaaf et al., 2003) that transports iron complexed by specific plant-derived Fe(III) chelators known as phytosiderophores (PS) that form stable Fe(III) chelates (Curie et al., 2001; Roberts et al., 2004; Schaaf et al., 2004). NA is the direct biochemical precursor to PS, and as such, is structurally similar. Recently, ZmYS1 was shown to transport iron from Fe(II)–NA complexes (Roberts et al., 2004; Schaaf et al., 2004), and was observed to transport metals other than iron (Schaaf et al., 2004). Because PS are neither made nor used by non-grass species like Arabidopsis, the role of the YSL proteins in Arabidopsis is likely in the transport of metals complexed by the PS-related compound NA.

Here, we examine Arabidopsis YSL2 (At5g24380) and find that, as predicted, this protein transports metals that are chelated by NA. Like ZmYS1, Arabidopsis YSL2 transports both iron and copper when these metals are complexed with NA. In contrast to ZmYS1, YSL2 transports only Fe(II), and can also transport copper from Cu–NA, which is not a suitable substrate for copper transport by ZmYS1. The pattern of expression of YSL2 indicates that the transport of metal–NA occurs in diverse cell types in both roots and shoots, suggesting that metals are regularly imported into cells as NA complexes. Consistent with its function as an iron and copper transporter, YSL2 transcription is regulated by the levels of these metals in the growth medium. YSL2 has an unusual pattern of distribution within the plasma membranes of vascular parenchyma cells that is consistent with a function in moving metals laterally within veins. Based on its expression pattern, the major function of YSL2 appears to be in the lateral movement of metals within the vasculature. This movement appears to be most critical when iron is abundantly available, as YSL2 expression under iron-deficient growth conditions is reduced.


Cloning of a YSL2 cDNA and functional complementation in yeast

Eight full-length YSL genes (AtYSLs) can be identified in the Arabidopsis genome. (A ninth partial gene is also evident.) The AtYSL proteins inferred from cloned YSL cDNAs (Roberts and Walker, unpublished data, Genbank accession numbers AY515560, AY515561, AY515562, AY515563, AY515564, AY515565, AY515566), or inferred from the Arabidopsis genome sequence share strong full-length sequence similarity with each other and with the YS1 protein from maize (ZmYS1). Clustal analysis of AtYSL protein sequences (Figure 1a) indicates the existence of three distinct sub-groups within the family. A consensus topology for the nine YSL family members (eight AtYSLs and one ZmYS1) was generated using TMAP (Persson and Argos, 1994, 1996). There are 15 predicted transmembrane domains (Figure 1b). Among the predicted extra-membranous segments of the proteins is one region with very poor sequence conservation (the variable region). Another variable region is found at the amino termini of the proteins. However, in spite of poor sequence similarity, the amino acid composition of the amino termini of all nine YSLs is strikingly similar, being unusually rich in the acidic residues glutamate and aspartate.

Figure 1.

The Arabidopsis YSL family of proteins.
(a) Clustal analysis of AtYSL protein sequences.
(b) Consensus-predicted topology for the YSL family generated using TMAP (Persson and Argos, 1994, 1996).

The AtYSL protein that is most similar to ZmYS1 is YSL2 (At5g24380; 62% identical and 78% similar to ZmYS1). Because of this strong sequence similarity, the YSL2 gene was chosen for further study. A cDNA corresponding to the full coding region of YSL2 was obtained by screening a size-selected hypocotyl cDNA library (Genbank AY515561; see Experimental procedures). This YSL2 cDNA was subcloned into the yeast expression vector pFL61 (Minet et al., 1992) for use in yeast functional complementation assays.

To test whether YSL2 is able to transport iron–NA, we used the iron-uptake-defective yeast strain DEY1453 (fet3fet4 double mutant). This strain cannot grow on iron-limited medium. The YSL2 cDNA was introduced into DEY1453, and growth was tested on a medium containing micromolar amounts of iron with and without NA (Figure 2). As a positive control, the Arabidopsis IRT1 cDNA, which is capable of supporting the growth of the fet3fet4 yeast (Eide et al., 1996), was also introduced into the strain. The viability of each strain is demonstrated by growth under permissive conditions (50 μm iron citrate, pH 4.2; Figure 2a). Expression of IRT1 restored growth under all conditions, as expected. Expression of the YSL2 cDNA restored growth on 4 μm Fe(II)–NA, but not on 4 μm Fe(II) in the absence of NA (Figure 2b,c). This result indicates that YSL2 is an iron transporter that can use NA-bound Fe(II) as a substrate.

Figure 2.

Functional complementation of yeast.
(a–g) The fet3fet4 strain DEY1453 was transformed with constructs expressing YSL2, AtIRT1, or the empty vector, pFL61, and grown on synthetic defined medium containing: (a) 50 μm iron citrate; (b) 4 μm Fe(II)SO4; (c) 4 μm Fe(II)SO4 with 5 μm NA; (d) 10 μm Fe(III)Cl3; (e) 10 μm Fe(III)Cl3 with 10 μm MA; (f) 10 μm Fe(II)SO4 with 10 μm MA; (g) 10 μm Fe(III)Cl3 with 10 μm NA.
(h–k) The ctr1 strain YSC5 was transformed with constructs expressing YSL2, AtIRT1, ScCTR1 or the empty vector, pFL61, and grown on synthetic defined medium containing: (a) glucose and 1 μm CuSO4; (b) glycerol and 1 μm CuSO4; (c) glycerol and 1 μm CuSO4 with 5 μm NA; (d) glycerol and 1 μm CuSO4 with 10 μm MA. Pairs of spots correspond to 10- and 100-fold dilutions of the original cultures.

As a further test of the substrate specificity of YSL2, growth assays were performed using iron-phytosiderophore as a substrate. Non-grass species do not synthesize PS, and so these conditions are not intended to mimic physiologic conditions in Arabidopsis. Instead, the question is whether YSL2 requires NA per se, or can instead use the structurally similar but non-identical compound, mugineic acid (MA). Iron was provided as Fe(III)–MA, which is the substrate for the maize YS1 transporter (Curie et al., 2001), or as Fe(II)–MA, which is not expected to be formed in natural soil conditions where iron is rapidly oxidized. YSL2 did not complement growth on 10 μm Fe(III)Cl3, 10 μm Fe(III)–MA, or 10 μm Fe(III)–NA (Figure 2d,e,g), indicating that Fe(III) is a poor substrate for transport by YSL2, even in the presence of NA. Interestingly, YSL2 did complement growth on a 4 μm Fe(II)–MA-containing medium (Figure 2f). This result indicates that YSL2 requires Fe(II) as a substrate for uptake, but is able to accept iron from two related, yet non-identical chelators.

To test for copper uptake, yeast strain YSC5 (ctr1), was used. The YSC5 strain is unable to take up adequate amounts of copper because of the deletion of the CTR1 gene encoding the yeast high-affinity copper transporter. The growth assay to test copper uptake is complicated by the fact that, in yeast, iron uptake depends on adequate intracellular levels of copper (Askwith et al., 1994; Dancis et al., 1994b). As a result, under ordinary conditions, the growth defect exhibited by YSC5 yeast is caused by secondary failure to take up iron, not the primary copper uptake defect. Clearly, growth defects related to inadequate iron uptake could potentially be reversed through expression of YSL2. To test directly for copper uptake, glycerol was substituted for glucose in the medium, requiring the cells to respire, which in turn leads to an absolute requirement for copper (Dancis et al., 1994a; Zhou and Gitschier, 1997). Thus, inadequate copper uptake restricts the growth of the ctr1 strain on medium containing glycerol as the sole carbon source.

The plasmids expressing YSL2 and control cDNAs were introduced into the ctr1 yeast. The positive control was the yeast high-affinity copper transporter CTR1. AtIRT1, which does not transport copper but does transport iron (Eide et al., 1996), is a negative control in these experiments that would reveal restoration of growth resulting from alleviation of iron deficiency. The viability of each strain is demonstrated by growth under permissive conditions (SD-URA with glucose; Figure 2h). Under the restrictive growth conditions used in this assay (1 μm CuSO4 or 1 μm CuSO4 with 5 μm NA), only the CTR1 positive control and the YSL2 cDNA were able to restore growth. Growth restoration by YSL2 required NA in the medium (Figure 2i,j). The ability of YSL2 to complement on copper complexed with the phytosiderophore, MA was also tested. Under the conditions used in this experiment (1 μm CuSO4, 5 μm MA), MA was not able to substitute for NA in the uptake of copper by YSL2 (Figure 2k). This result indicates that, in addition to being a transporter of iron, YSL2 is also a copper transporter that requires Cu–NA as the substrate for transport.

We also tested whether expression of YSL2 could restore growth of a zinc-uptake-defective strain of yeast [ZHY3 (Zhao and Eide, 1996); zrt1zrt2]. No complementation by YSL2 was observed on a medium with 1.4 μm ZnSO4 either containing or lacking NA (not shown), suggesting that YSL2 does not transport zinc at the low, physiologic concentration used in our experiments.

Pattern of YSL2 expression

To establish the pattern of expression of YSL2, we used a semi-quantitative RT-PCR assay. RNA was isolated from the leaves, roots, flowers, siliques, and inflorescence stems of 3-week-old plants. YSL2 mRNA was detected predominantly in the leaf and root tissues of 3-week-old plants, although expression was also detectable at low levels in the stem, cauline leaves, flowers, and siliques (not shown). In order to refine these observations, we made β-glucuronidase (GUS) reporter constructs containing the upstream sequences of YSL2 fused in-frame to GUS (YSL2p:GUS). This construct contains 564 bp of YSL2 promoter plus 155 bp of 5′ UTR sequence and 13 nucleotides (four amino acids) of coding sequence. In shoots, YSL2p:GUS was most strongly expressed in fully expanded leaves, while young, still-growing leaves expressed only weakly or not at all (Figure 3a). In leaf tissues, YSL2p:GUS was expressed most strongly in the vasculature, but weaker expression was also observed in non-vascular tissues, especially in cells close to the leaf veins (Figure 3b,c). Transverse sections through the midveins showed that YSL2p:GUS was expressed in xylem-associated cells and the mesophyll (Figure 3c).

Figure 3.

YSL2p:GUS expression. Transgenic lines containing YSL2p:GUS were examined for GUS activity by histochemical staining.
(a) 14 d.a.g. plate-grown Arabidopsis. Note darker staining in older leaves.
(b) A mature leaf from a 21 d.a.g. soil-grown plant.
(c) Transverse leaf section from a 21 d.a.g. Arabidopsis rosette leaf. Staining was carried out for a limited time period to highlight vascular staining relative to staining in other cells.
(d, e) Root mature and elongation zones and primary root tip.
(f) Transverse section of a root from 10 d.a.g. plate-grown plant.

In roots, the YSL2 promoter conferred GUS expression in the apical end of the elongation zone (Figure 3e). YSL2p:GUS expression continued throughout the rest of the root up to the root/shoot junction (Figure 3d). No GUS expression was observed in the meristematic zones or the root cap (Figure 3e). YSL2p:GUS was most strongly expressed in the vascular tissue of the primary root, but every cell type within the mature primary root stained positive for GUS activity (Figure 3f). Lateral roots recapitulate this pattern of expression. Thus, the YSL2 promoter used in YSL2p:GUS confers broad expression throughout the root.

As RT-PCR analysis revealed expression in flowers and siliques, YSL2p:GUS expression was also examined in these structures. YSL2p:GUS was expressed in the vasculature of sepals, petals, anthers, and developing siliques (not shown). No YSL2p:GUS expression was observed in developing seeds. Expression was also observed in the stigma. Thus, the YSL2 promoter used in YSL2p:GUS confers expression in many floral organs, implying that movement of NA-bound metals occurs within flowers and developing fruits.

YSL2 expression is altered in response to metal deficiency

To determine whether YSL2 expression changes in response to metal deficiency, plants were grown on a 1X MS medium containing standard micronutrient amounts for a period of 7 days after germination (d.a.g.), then transferred to medium that lacked iron. Control plants were transferred to fresh plates containing standard MS. No special measures were taken to deplete minute amounts of metals contaminating either purified water or other reagents, so the medium undoubtedly contained trace amounts of metal. After 5 days on a metal-limited medium, RNA was prepared from shoots and roots and the relative steady state expression level of YSL2 was determined by semi-quantitative RT-PCR (Figure 4a). In response to limitation for iron, YSL2 steady state mRNA expression was reduced in both shoots and roots. This pattern is in contrast to that observed for the primary Arabidopsis iron uptake transporter IRT1 (Curie et al., 2001; Eide et al., 1996), which showed root-specific expression that was positively regulated by iron deficiency (Figure 4a). Regulation of YSL2 mRNA levels by copper starvation was also tested by growing plants on a complete medium and then shifting to a medium lacking copper for 5 days. No change in YSL2 mRNA expression was observed (not shown). It is important to note, however, that because special measures were not taken to ensure complete copper depletion of the medium, we cannot be certain that these plants were truly experiencing copper limitation.

Figure 4.

Effect of iron status on YSL2 expression. Plants were grown 7 d.a.g. and transferred to iron-deficient medium for 5 days.
(a) Total RNA from: lane 1: shoots of plants grown in full metal conditions; lane 2: shoots of plants grown in iron-deficient conditions; lane 3: roots of plants grown in full metal conditions; lane 4: roots of plants grown in iron-deficient conditions, was reverse transcribed and the resulting cDNA was amplified by PCR with primers specific to YSL2, IRT1 or actin.
(b, c) Shoot (b) and root (c) expression of YSL2p:GUS in plants grown under iron-sufficient (left) or iron-deficient (right) conditions.
(d) GUS activity in protein extracts of roots and shoots of plants grown with (full MS) and without iron (−Fe).

Metal regulation of the YSL2 promoter

To ascertain whether regulation of YSL2 occurs at the level of transcription, we examined YSL2p:GUS transgenic plants grown under iron-limited conditions. YSL2p:GUS plants were grown for 7 d.a.g. on standard MS medium, then transferred to iron-deficient medium for 5 days. GUS expression was monitored by histochemical staining (Figure 4b,c) and as GUS activity in protein extracts (Figure 4d). GUS activity from YSL2p:GUS in both shoots and roots was negatively affected by iron starvation. These results are consistent with the results from the RT-PCR analysis of the native YSL2 gene, and suggest that this negative regulation is, at least in part, a function of the YSL2 promoter. As the YSL2p:GUS transgene includes 155 bp of the YSL2 5′ UTR, it is also possible that post-transcriptional regulation occurring through this region is responsible for the changes in expression that are observed upon iron starvation.

Although YSL2 appeared to transport copper, copper deficiency had no apparent effect on YSL2 mRNA expression. To investigate the effect of copper oversupply on expression of YSL2, we grew wild-type Arabidopsis Col0 plants for 10 d.a.g. on a standard 1X MS medium or on a 1X MS medium supplemented with 50 μm copper sulfate. Under the high copper conditions, plants showed symptoms of toxicity: decreased growth and yellowing of the leaves by day 7 of the copper treatment. RNA was prepared from leaves and roots of these plants, and was used as template for RT-PCR. Under high copper treatment, steady state YSL2 mRNA abundance declined to barely detectable levels in shoots (Figure 5a, lanes 1 and 3) but appeared unchanged in roots (Figure 5a, lanes 2 and 4).

Figure 5.

Effect of elevated copper on YSL2 expression. Plants were grown 10 d.a.g. in the presence or absence of 50 μm CuSO4.
(a) Total RNA from shoot (lanes 1 and 3) and root tissue (lanes 2 and 4) from plants grown under normal (lanes 1 and 2) or high copper (lanes 3 and 4) conditions was reverse transcribed and amplified by PCR. As a control for contaminating genomic DNA, a sample of RNA with no reverse transcription step (−RT) was also amplified (lane 5). The cDNA was amplified using primers specific to YSL2 or to actin.
(b, c) YSL2p:GUS expression in plants grown under normal metal (b) or high copper (c) conditions.

We examined YSL2p:GUS transgenic plants grown in conditions of copper excess identical to those used for the RT-PCR analysis (see above). GUS expression from YSL2p:GUS was markedly reduced in the copper-treated plants (Figure 5b,c). After only 2 h of GUS staining, untreated plants exhibited obvious GUS staining in both the roots and shoots, while the copper-treated plants exhibited little GUS activity in either tissue. The finding that copper treatment affected YSL2p:GUS expression in both the roots and shoots was surprising, because RT-PCR analysis of native gene expression had indicated that YSL2 mRNA levels were decreased in shoots but not in roots. The inconsistency between the native mRNA abundance and GUS expression from YSL2p:GUS in roots may reflect the stability of the native YSL2 mRNA, or alternatively may indicate that the elements of the YSL2 promoter that enhance root expression under these conditions are missing in the YSL2p:GUS construct.

Redundancy of YSL2 function

To examine the in planta role of YSL2 further, we sought mutants in which the YSL2 gene was disrupted. We found a putative T-DNA insertion in YSL2 (SALK_001348) by screening the sequence-indexed Arabidopsis T-DNA insertion lines generated by the Salk Institute Genomic Analysis Laboratory (Alonso et al., 2003). PCR screening and sequencing of the resulting PCR fragment confirmed a T-DNA insertion within YSL2 exon 4 (Figure 6a), so this allele was designated ysl2-1. Homozygous ysl2-1 mutant individuals were identified, and RT-PCR was used to determine whether YSL2 mRNA expression had been abolished by the T-DNA insertion (Figure 6b). Homozygous ysl2-1 mutants had undetectably low levels of YSL2 mRNA, thus ysl2-1 is a true knockout allele.

Figure 6.

Identification of a T-DNA insertion in YSL2.
(a) Map of ysl2-1 indicating the position (above) and orientation (arrow) of the T-DNA within the fourth exon of the YSL2 gene. Introns are indicated as black lines; exons as boxes, with coding portions in gray and non-coding regions in white. Positions of RT-PCR primers are shown by small arrowheads above and below. The forward primer spans an intron/exon boundary.
(b) RT-PCR analysis of ysl2-1. RNA prepared from wild-type Col0 (W) or from an individual homozygous for ysl2-1 (M) was reverse transcribed (+RT) and amplified using primers specific for YSL3 (as a control for cDNA quality) and YSL2 (see Experimental procedures). As a control for contaminating genomic DNA, samples of RNA with no reverse transcription step (−RT) were amplified. Molecular weight markers (3, 2.5, 2, 1.5, 1.2, 1, 0.9, 0.8, 0.7, 0.6, and 0.5 kb) are shown in the middle lane.

Homozygous ysl2-1 plants exhibited no overt mutant phenotypes when grown in standard culture conditions, either on plates or in soil. As YSL2 mRNA levels decrease in response to iron deficiency, we predicted that ysl2-1 mutant plants might be ‘pre-adapted’ to this condition, and display resistance to iron deficiency. However, when ysl2-1 plants were grown under iron deficiency, their growth and appearance were again comparable with wild-type control plants: ysl2-1 plants were neither more susceptible nor more resistant to growth in low iron. The ysl2-1 mutants exhibited iron deficiency chlorosis with the same timing, and to the same extent as wild-type control plants. By the same argument, we postulated that ysl2-1 mutant plants might exhibit resistance to high copper treatment. However, the mutant again exhibited no differences from wild-type plants. Mutant plants were also tested on a medium containing decreased amounts of zinc, manganese and copper. The mutants grew comparably with wild-type controls under all these conditions. The total metal content of ysl2-1 mutant plants was measured by ICP-MS. The levels of all metals were indistinguishable from those present in the wild-type plants. The lack of effect of the ysl2-1 mutation in these conditions may reflect functional redundancy within the Arabidopsis YSL family.

Localization of the YSL2 protein

All of the AtYSL proteins are predicted to be localized to the plasma membrane, as they lack sequences known to target integral membrane proteins to other organelles. We examined the localization of YSL2 using a YSL2–GFP fusion protein. In the fusion construct (YSL2:GFP), the entire YSL2 gene under 564 bp of its native promoter, was fused in-frame at its C-terminus to GFP. In plants stably transformed with this construct, green fluorescence was observed in leaves and in roots (Figure 7a,b). In both the leaves and roots, fluorescence was most apparent in the vascular tissue, which is consistent with the pattern of expression for YSL2 deduced using YSL2p:GUS. Weak fluorescence was observed in non-vascular cells, especially those adjacent to the veins (not shown). In YSL2:GFP plants, fluorescence was strongest in elongated cells within the vasculature, apparently xylem parenchyma. Notably, fluorescence was located exclusively at the edges of these cells, consistent with localization to the plasma membrane (Figure 7). Strikingly, fluorescence was only rarely observed at the apical or basal ends of cells (see Figure 7c for a rare exception), but was instead found almost exclusively on the lateral plasma membranes. This localization implies that YSL2 moves metal–NA complexes laterally within the veins of both leaves and roots.

Figure 7.

Localization of YSL2:GFP fusion protein.
(a) Optical section through a petiole and lower leaf. Images from left to right are false DIC, fluorescence, and a merged image showing the location of fluorescence.
(b) Optical section through a primary root. Images from top to bottom are as described in (a).
(c) Optical section through a leaf midvein showing a rare instance of fluorescence located at the basal end of a cell.


This investigation demonstrates that YSL2 is a transporter of NA-bound metals. YSL2 is capable of transporting Fe–NA and Cu–NA in yeast functional complementation assays. This metal transporter function is further supported by the finding that the patterns of YSL2 expression change in response to metal concentrations in the surrounding medium. YSL2 is downregulated by iron limitation and also by copper excess. We have shown that this regulation occurs at the level of transcription as the YSL2 promoter confers these regulated patterns of expression on a GUS reporter gene.

YSL2 is expressed in many cell types in leaves, roots and reproductive organs. Expression is strongest in the veins, with xylem parenchyma cells having the strongest expression in two reporter constructs. Expression of YSL2 in non-vascular cells implies that the uptake of iron and copper from NA complexes is accomplished by many cells, not just cells within veins. Furthermore, the finding that YSL2 is expressed in diverse cell types throughout leaves and roots shows that many cells can take up metals from NA complexes. Moreover, NA must be present within the apoplast in order to provide substrates for YSL2 transport. Thus, in addition to its previously deduced roles in iron solubility within cells, NA also appears to have a significant role as an extracellular chelator of metals.

YSL2:GFP localization is most consistent with a plasma membrane localization for YSL2. However, discerning between plasma membrane and tonoplast membrane localization in the highly vacuolated vascular parenchyma cells that express YSL2 most strongly is difficult, owing to the two membranes being pressed together at the periphery of these cells. These cells contain plastids, which were readily visible at higher magnifications (not shown), but in no case did we observe a membrane deviating around such a plastid. If YSL2 were located in the tonoplast membrane, we would have expected to see such instances. The localization of YSL2 in membranes at the lateral (but not the apical and basal) ends of the vascular parenchyma cells is also most consistent with the localization of YSL2 to the plasma membrane, as subdomains in the plasma membrane are well documented (Friml et al., 2002a,b; Galweiler et al., 1998; Muller et al., 1998; Swarup et al., 2001), while subdomains of the tonoplast membrane are unknown.

The lateral membrane localization revealed by the YSL2:GFP fusion is consistent with a role in the lateral movement of metals within veins. This lateral movement could be toward or away from the xylem. We hypothesize that the predominant physiologic role for YSL2 is to take up iron that has arrived in tissues via xylem transport, thus moving it away from the xylem vessels. The logic behind this hypothesis rests on YSL2 localization coupled with the pattern of iron-regulated expression observed for YSL2. YSL2 expression is decreased (in both leaves and roots) by iron deficiency. This regulation is in contrast to that of genes that are part of the iron uptake machinery in plants (e.g. AtIRT1, AtFRO2, and ZmYS1). Interestingly, though, the tobacco NA synthase gene is also negatively regulated by iron deficiency (Higuchi et al., 1995).

We hypothesize that the negative regulation of YSL2 allows the plant to restrict exit of metals from the xylem in roots and mature leaves – the two contexts in which YSL2 expression is highest. If the role of YSL2 were to move metals toward the xylem, the opposite pattern of expression would be expected, especially in roots, where sparing amounts of iron would need to be mobilized into shoots via the xylem. The restriction of iron exit from the xylem could be interpreted as having either of two opposite consequences. By restricting uptake in mature tissues, metals would continue to be translocated apically toward younger leaves higher on the plant. Conceivably, mature tissues and cells, which already have had time to attain adequate metal levels, can withstand brief periods of iron limitation, while more apical, younger tissues would more rapidly experience metal starvation. Thus downregulation of YSL2 would represent a mechanism to enhance long-distance iron translocation to shoots by preventing iron sequestration in more basal parts of the plant. Alternatively, the uptake of metals from the xylem stream could represent part of a nutrient cycling mechanism in which metals are passed from the xylem stream into the phloem. In this scenario, preventing the exit of metals from the xylem restricts metal uptake into the phloem, causing metals to remain in mature tissues at the expense of sinks such as the apical meristems and young leaves. It is interesting to note that iron-deficient plants exhibit a pattern of chlorosis in which the veins stay green while the interveinal regions become yellow. It is intriguing to speculate that this pattern reflects iron sequestration within veins that maintains chlorophyll synthesis there.

The finding that ysl2-1 mutants have no discernible phenotypes, even under conditions of metal deficiency, suggests that other YSL family members confer overlapping functions in metal–NA transport. A functional analysis of other YSL family members is needed to establish whether they have identical biochemical and/or physiologic roles. Moreover, the production of double mutants for those YSLs with apparently overlapping functional roles is expected to shed additional light on the physiologic function(s) performed by the YSL family.

Experimental procedures

Plant growth conditions

Plants were grown in sterile culture on plates containing standard 1X MS medium with 1% sucrose. Seeds were surface sterilized in a solution (20% (v/v) Clorox, 0.01% (v/v) Triton X-100), then imbibed in the dark at 4°C for 3–5 days. Plants were grown at 22°C in a Conviron growth chamber under 16 h day, 8 h night conditions. Pot-grown plants were germinated (following imbibition in the dark at 4°C for 3–5 days) at room temperature under continuous light in 4′′ × 4′′ pots filled with Promix (Griffin Greenhouse, Tewksbury, MA, USA).

Growth conditions for analysis of the ysl2-1 mutants

Homozygous mutants were grown in standard MS and soil conditions as described above. In addition, these mutants were grown on MS medium modified by independently altering the iron, zinc, manganese, and copper concentrations to 0.5X, 0.25X, 0.1X, and 0X (no added metal). In one trial, the plants were germinated directly on these metal-deficient media; in another, they were germinated on full MS, then transferred after 14 days to metal-deficient plates. In all tests, wild-type ColO plants were grown on the same plate or in the same pot for comparison.

YSL2 cDNA clones

The Ecker size selected (2–3 kB) cDNA library from 3-day hypocotyls was obtained from the Arabidopsis Biological Resource Center. A probe corresponding to the full-length YSL2 genomic sequence was amplified from Arabidopsis genomic DNA, and used to screen the cDNA library. A single full-length clone was identified and was subcloned into the yeast expression vector pFL61 for use in yeast functional complementation assays.

Yeast growth

The iron-uptake-defective strain DEY1453 was grown on standard YPD supplemented with 25 mm NaCitrate (pH 4.2) and 50 μm FeCl3. Once transformed with pFL61-derived plasmids, DEY1453-derived strains were grown on synthetic defined media lacking uracil supplemented with 25 mm NaCitrate (pH 4.2) and 50 μm FeCl3 (SD-Ura + Fe). For complementation assays, SD-Ura lacking iron was prepared and iron was added, in the forms indicated in the text just prior to pouring plates. Fe(III) and Fe(II) were prepared as 1 m stocks from FeCl3 and FeSO4, respectively, in metal-free RO H2O. The Fe-containing solutions were immediately sterile filtered and appropriate amounts were added to each plate. Desired volumes of previously prepared sterile NA or MA stock solutions were immediately added to the iron solution. Following complex formation, 15 ml of molten medium was added to the plate, contents were mixed, and plates were allowed to solidify. Yeast cells were diluted and immediately plated onto each freshly prepared medium. Nicotianamine was obtained from T-Hasegawa Co. (Tokyo, Japan). The copper-uptake-defective strain, YSC5 and its transformed derivatives were grown on standard YPD, and standard SD-Ura, respectively. For complementation assays, glycerol was used as a carbon source in place of dextrose and medium was prepared without added copper. Copper was added in the forms indicated in the text just prior to pouring plates.

Semi-quantitative RT-PCR

For the detection of relative transcript abundance by semi-quantitative RT-PCR, total RNA was isolated from 50–75 mg of tissue using the Qiagen RNEasy Kit (Qiagen, Inc., Valencia, CA, USA). Total RNA of 500 ng was used in a reverse transcription reaction with poly d(T) oligo using protocols and reagents from the Invitrogen Superscript RT-PCR System (Invitrogen, Carlsbad, CA, USA). Following DNaseI treatment, 1/10th of the total cDNA reaction was used in a 50-μl polymerase chain reaction using Ex-Taq polymerase (Takara-Mirus, Inc., Madison, WI, USA) for 25 cycles: 94°C, 15 sec, 55°C, 30 sec, 72°C, 1 m 30 sec. These conditions were determined following testing for linear amplification (by analysis of products on ethidium bromide-stained agarose gels) of YSL2 at 17, 19, 21, 23, 25, 27, and 29 cycles. Cycle 25 was found to be within the linear amplification range for YSL2 mRNA in both the root and shoot. The following primer sets were used at a concentration of 0.4 μm in a 50 μl reaction volume: YSL2, 5′-caaaaccaacagcatcagcacattgg-3′ (forward), 5′-cgcaaaggcaaagaatccgtagca-3′ (reverse); IRT1, 5′-gagtcattgccatggtcttgga-3′ (forward), 5′-gtatactcagcctggaggatacaaccg-3′ (reverse); actin, 5′-gcaagtcatcacgattggtgc-3′ (forward), 5′-gaaccaccgatccagacactgt-3′ (reverse). Amplified products from 15 μl of PCR reaction were visualized on a 1% TAE agarose gel containing ethidium bromide. Bands were photographed using a Kodak Digital ScienceTM DC40 camera and software (Eastman Kodak, USA).

High sensitivity RT-PCR analysis of ysl2-1

cDNA was prepared from homozygous ysl2-1 individuals and from wild-type plants as described in the previous section. As a control to detect amplified genomic DNA, RNA that was not reverse transcribed was also amplified. To control for cDNA quality, primers specific for a second YSL gene, YSL3, were used in a separate set of RT-PCR reactions. We have found that YSL3 mRNA levels are similar to the levels of YSL2 mRNA in shoots. The YSL3 primer set is 5′-attggccaggaaacaagtgtttgggt-3′ (forward) and 5′-gacaagtcccgcgactacaccattt-3′ (reverse). Inclusion of these reactions ensures that failure to observe an amplified YSL2 cDNA product did not result from poor quality cDNA preparations, but rather from a true absence of mRNA in the mutant plants. PCR was performed in 50 μl reactions using Ex-Taq polymerase (Takara-Mirus, Inc.) for 40 cycles: 94°C, 15 sec, 55°C, 30 sec, 72°C, 1 m 30 sec. Products were visualized as described in the previous section.

GUS histochemical staining and visualization

For analysis of GUS expression under the control of YSL promoters, transgenic plants containing YSL-GUS reporter constructs were stained in GUS assay buffer [50 mm KPO4, pH 7.0, 10 mm EDTA, 0.5 mm K3Fe(CN)6, 0.01% Triton X-100, 0.3 mg ml−1 (w/v) X-gluc (5-bromo-4-chloro-3-indolyl β-d-glucuronide; Rose Scientific, CA, USA)] for 2–24 h. Staining was stopped by the addition of 70% (v/v) ethanol. If sectioning was required, plant tissues were fixed with formalin–acetic acid–alcohol (FAA; 50% ethanol, 5% glacial acetic acid, 3.7% formaldehyde in distilled water) for 10 min at room temperature. Destained plants were visualized using a Nikon SMZ1200 dissecting microscope. For thin sections, root and rosette leaf tissues from adult plants were coarse sliced with a razor blade and infiltrated using the Technovit 7100 kit (Kulzer Histo-Technik, Heraeus, Kulzer, Germany). Eight-micron transverse sections were cut using a glass knife and a Microtome Ultracut. Specimens were photographed using a Spot Insight camera and processed with Spot Advanced software (MicroVideo Instruments, Avon, MA, USA) and Adobe Photoshop Version 7.0.

GUS activity measurements

Protein extracts were prepared by grinding tissue in Extraction Buffer (50 mm sodium phosphate, pH 7.0, 10 mm EDTA, 0.1% Triton X-100, 0.1%N-lauroylsarcosine, 10 mmβ-mercaptoethanol, and 140 μm PMSF), and removing debris by repeated (2×) centrifugation. Extracts were stored at −80°C. Protein concentrations were determined using the Bio-Rad RC DC kit (Bio-Rad, Hercules, CA, USA). GUS activity was measured fluorometrically. Briefly, one-fifth of the total protein extracted was added to GUS assay buffer (50 mm sodium phosphate, pH 7.0, 10 mm EDTA, 0.1% Triton X-100, 0.1%N-lauroylsarcosine, 10 mmβ-mercaptoethanol, and 50 μm 4-methyl-umbelliferyl β-d-glucuronide). Half of the assay was immediately removed and quenched in 0.2 m Na2CO3, to serve as a control for background fluorescence. The other half was incubated at 37°C for 24 h. Fluorescence was measured in a Turner TBS-380 Minifluorometer (Turner Biosystems, Sunnyvale, CA, USA).

GFP visualization

Seeds were stratified and germinated directly on MS medium, and grown for 2 weeks. Whole plants were visualized using a Zeiss Axiovert confocal microscope using a 488 nm argon laser line, and a 520–550 nm bandpass filter.


We thank Satoshi Mori for his very kind gifts of MA and NA, David Salt and Brett Lahner for performing ICP-MS analysis, David Eide and Andy Dancis for generously sharing strains and advice, Joe Ecker for providing his cDNA library, and Tara Burke for diligent technical assistance. We thank Tobias Baskin and Alison DeLong for critical reading of the manuscript.