High expression of zinc (Zn)-regulated, iron-regulated transporter-like protein (ZIP) genes increases root Zn uptake in dicots, leading to high accumulation of Zn in shoots. However, none of the ZIP genes tested previously in monocots could enhance shoot Zn accumulation. In this report, barley (Hordeum vulgare) HvZIP7 was investigated for its functions in Zn transport.
The functions of HvZIP7 in planta were studied using in situ hybridization and transient analysis of subcellular localization with a green fluorescent protein (GFP) reporter. Transgenic barley lines overexpressing HvZIP7 were also generated to further understand the functions of HvZIP7 in metal transport.
HvZIP7 is strongly induced by Zn deficiency, primarily in vascular tissues of roots and leaves, and its protein was localized in the plasma membrane. These properties are similar to its closely related homologs in dicots. Overexpression of HvZIP7 in barley plants increased Zn uptake when moderately high concentrations of Zn were supplied. Significantly, there was a specific enhancement of shoot Zn accumulation, with no measurable increase in iron (Fe), manganese (Mn), copper (Cu) or cadmium (Cd). HvZIP7 displays characteristics of low-affinity Zn transport.
The unique function of HvZIP7 provides new insights into the role of ZIP genes in Zn homeostasis in monocots, and offers opportunities to develop Zn biofortification strategies in cereals.
Zinc (Zn) is an essential micronutrient for all organisms. It is involved in many biochemical processes (Welch, 1995; Maret, 2004). All organisms must maintain adequate Zn in their cells. To achieve sustained Zn uptake from fluctuating environments, plants and other eukaryotes are equipped with a dual-transporter system, which consists of both high-affinity and low-affinity Zn transporters (Eide, 2006). In yeast, the high-affinity transporter gene (Zrt1) is responsible for uptake of Zn in Zn-limiting media. When Zn is abundant in external media, Zrt1 is repressed and the low-affinity transporter (Zrt2) mediates Zn uptake. Zrt1 has a remarkably high affinity for Zn with an estimated apparent Km of 10 nM for free Zn2+ ions, whereas Zrt2 has a lower affinity of apparent Km of c. 100 nM (Eide, 2006). The single-celled alga, Chara corallina, is similar to yeast, and also has a dual transporter system for Zn influx (Reid et al., 1996). The high-affinity system saturates at c. 0.1 μM, and the low-affinity system shows a linear dependence on Zn concentration from 0.5 μM up to at least 50 μM (Reid et al., 1996). A kinetic study of Zn influx in a higher plant, wheat, also shows the presence of a dual Zn uptake system (Hacisalihoglu et al., 2001). Apparent Km values of the high-affinity system are between 0.6 and 2 nM, while the Km values of the low-affinity system are between 2 and 5 μM (Hacisalihoglu et al., 2001).
Soils with low available Zn occur over a significant proportion of the world's agricultural area (Alloway, 2004), which not only affects grain yield but also the Zn concentration of the grain. Low Zn in cereal grains exacerbates Zn deficiency in humans (Cunningham-Rundles et al., 2005; Hotz et al., 2005), and consequently increasing grain Zn content in cereals (biofortification) is considered to be important for the alleviation of human Zn deficiency (Welch & Graham, 2004). Understanding the functional role of the major genes that control Zn uptake and transport is important for improvement of Zn nutrition in plants and enhancement of Zn concentrations in grains. The Zn-regulated, iron-regulated transporter-like protein (ZIP) family is considered to be the primary group of transporters controlling plant Zn influx at the plasma membrane (Grotz et al., 1998; Guerinot, 2000; Eide, 2006). There are 15 members of the ZIP gene family identified in the Arabidopsis thaliana genome (Grotz & Guerinot, 2006), and 16 have been identified in the rice genome (Narayanan et al., 2007; Chen et al., 2008). Only four barley ZIP genes (HvIRT1, HvZIP3, HvZIP5 and HvZIP8; Pedas et al., 2008, 2009) and one from the tetraploid emmer wheat (TdZIP1; Durmaz et al., 2010) have been described so far in temperate cereals, which are major staple foods for humans (Graham & Welch, 2001). Therefore, a better understanding of the contribution of ZIP transporters to Zn homeostasis in cereal plants is important for biofortification (Cakmak, 2008; Palmgren et al., 2008; Zhao & McGrath, 2009; Waters & Sankaran, 2011).
The complementation of yeast mutants defective in both the high-affinity transporter, Zrt1, and low-affinity transporter, Zrt2, has been a key tool in characterization of the Zn transport ability of plant ZIP family transporters (Grotz et al., 1998; Ramesh et al., 2003; Ishimaru et al., 2005; Pedas & Husted, 2009; Pedas et al., 2009; Lee et al., 2010a; Stephens et al., 2011). Kinetic studies on these plant ZIP transporters by expression in yeast indicate that apparent Km values are between 2 and 18.5 μM Zn (Grotz et al., 1998; Ramesh et al., 2003; Stephens et al., 2011); these values are similar to the Km range determined in planta for low-affinity transport systems (Reid et al., 1996; Hacisalihoglu et al., 2001). Whether these plant ZIP transporters are high- or low-affinity Zn transporters remains to be elucidated in planta, although tissue-specific expression using promoter-reporters suggests that some of these ZIP transporters may be high-affinity (Milner et al., 2012).
Studies of metal hyperaccumulating dicotyledonous species, such as Noccaea caerulescens (previously known as Thlaspi caerulescens) and Arabidopsis halleri, show that high expression of members of the ZIP family, NcZNT1, AtZIP4 and AtIRT3, contributes to high Zn accumulation (Hanikenne et al., 2008; Lin et al., 2009; Milner et al., 2012). By contrast, overexpression of OsZIP4, OsZIP5 and OsZIP8 in rice increased Zn concentration in roots, but reduced Zn accumulation in shoots and grains (Ishimaru et al., 2007; Lee et al., 2010a,b). It is not known if any members of the ZIP family in monocotyledonous species can lead to increased Zn accumulation in shoots and/or grains if expression is enhanced.
In this report, we describe the identification of additional ZIP genes from the barley genome, and functional characterization of one HvZIP gene, HvZIP7, in planta. Phylogenetic analysis shows that HvZIP7 is closely related to NcZNT1, AtZIP4 and AtIRT3. HvZIP7 is highly induced in both roots and shoots by Zn deficiency, and its protein is localized in the plasma membrane. Overexpression of HvZIP7 in barley increases root uptake and/or root to shoot translocation of Zn, but only at moderately high concentrations of Zn supply. Our results provide new insights into the functional role of ZIP genes in Zn homeostasis in monocotyledonous species.
Materials and Methods
Sequences of 16 ZIP family members from rice (Oryza sativa L.) and 18 from Arabidopsis (Arabidopsis thaliana L.) were used to search for homologous sequences in barley (Hordeum vulgare L.), Brachypodium (Brachypodium distachyon L.) and wheat (Triticum aestivum L.). Rice and Arabidopsis ZIP family members were retrieved from the Rice Genome Annotation Project Database release 6.1 (http://rice.plantbiology.msu.edu/) at Michigan State University and from A. thaliana genome annotation database release 9 (http://www.arabidopsis.org/) at The Arabidopsis Information Resource (TAIR). Similarity searches were performed primarily by BLASTp (Altschul et al., 1997, 2005) querying the JGI v1.0, 8 × assembly of Brachypodium at Phytozome, version 7 (http://www.phytozome.net/), and the nonredundant database of proteins at NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi) for barley and wheat. ZIP protein sequences were aligned by standalone MAFFT v6.905b using the L-INS-I method with associated default parameters (Katoh et al., 2009), and imported into ClustalW2 v2.1 (Larkin et al., 2007). The unrooted tree was generated in ClustalW2 using the neighbor-joining method with all parameters set to default. One thousand bootstrap datasets were generated to estimate the confidence limits of nodes. The tree was visualized using the Molecular Evolutionary Genetics Analysis (MEGA) package, version 5.05 (Tamura et al., 2011).
Barley (cv Lofty Nijo) seedlings used for examining effects of Zn deficiency on transcript abundance of HvZIP7 (Fig. 1a) were grown in hydroponics with 0.005 (−Zn) or 0.5 μM Zn (+Zn) for 14 d. In brief, barley seeds were surface-sterilized with 70% ethanol for 1 min, 3% hypochlorite for 5 min, rinsed with deionized water and incubated in Petri dishes for 2 d at room temperature. Seeds with emerged radicles were put into a seedling cup which was placed in the lid of a black plastic container. Each container contained four seedling cups and was filled with 1 l of nutrient solution. Each cup contained two plants. The basal nutrients were as follows (in μM); Ca(NO3)2, 1000; KNO3, 1000; NH4H2PO4, 100; MgSO4, 250; KCl, 50; H3BO3, 12.5; Fe-HEDTA, 10; MnSO4, 0.4; CuSO4, 0.1; NiSO4, 0.1; and H2MoO4, 0.1. 2-[N-morpholino] ethane-sulfonic acid-KOH of 2 mM was used to buffer pH to 6.0. Macronutrients and micronutrients were supplied at half and full strength, respectively, until D9, and full strength of all nutrients thereafter. The nutrient solution was aerated continuously and replaced at D9 and D12, respectively. Plants were grown in a growth room at 20 : 15°C day : night temperature and with a photoperiod of 14 : 10 h day : night at 300 μmol m−2 s−1 photon flux intensity at the plant level. At D14, plants were harvested. Roots were briefly rinsed in deionized water, separated from shoots, and excess water was blotted onto fresh laboratory tissues. The roots and shoots were frozen immediately in liquid nitrogen for transcript analysis.
RNA isolation and quantitative real-time reverse transcription polymerase chain reaction (RT-PCR) analysis of transcripts were conducted as described by Huang et al. (2011). The transcript abundances of four control genes (barley a-tubulin, heat shock protein 70, glyceraldehyde-3-phosphate dehydrogenase and cyclophilin) were determined for all cDNA samples, and the most similar three of these four genes were used as normalization controls. The primers 5′-TGGAAGGCATCCTCGACTCTG-3′ and 5′- CAATCAGATGGACACAGGCACAT-3′ were used for endogenous HvZIP7, and the primers 5′- TACTCATATACATGGCGCTGGT-3′ and 5′- TTTATTGCCAAATGTTTGAACG-3′ were used for the HvZIP7 transgene. The normalized copies μg−1 RNA were used to represent transcript abundances. For semiquantitative RT-PCR analysis (Supporting Information, Fig. S2b), RNA isolation and cDNA synthesis were the same as described earlier, transgene-specific transcripts were amplified from cDNA using the primer pair of 5′-GCTGGTTTGTCCTCATTTTACG-3′ and 5′-ATGATAATCATCGCAAGACCG-3′, and the transcripts of the barley glyceraldehydes-3-phosphate dehydrogenase (HvGAPDH) gene were also amplified as a loading control using the primer pair of 5′-GTGAGGCTGGTGCTGATTACG-3′ and 5′-TGGTGCAGCTAGCATTTGACAC-3′.
In situ PCR analysis
Barley plants (cv Golden Promise) were grown in hydroponics with no Zn addition. Plant growth conditions were the same as those described for transcript analysis. Roots and leaves were collected from 16-d-old plants. In situ PCR was performed using the method described by Koltai & Bird (2000) with a few modifications. Roots and leaves were fixed in FAA (63% ethanol, 5% acetic acid and 2% formaldehyde, v/v), embedded in 5% agarose in 1 × PBS (w/v), and then sectioned using a microtome (Leica VT1200S, Germany). The sections were treated with DNase (Ambion), followed by heat inactivation. cDNA synthesis was carried out using Superscript III RT (Invitrogen) with a gene-specific primer (5′-CACGTTATCTGATGTATGTATG-3′). PCR reactions were performed in a volume of 50 μl containing 1 × PCR buffer (Invitrogen), 1.5 mM MgCl2, 200 μM dNTP, 0.2 nM digoxigenin-11-dUTP (Roche), 1.5 U Taq polymerase (Invitrogen) and 150 ng of both forward and reverse primers (5′-ACCAGGTTCTACGAGACCAAG-3′ and 5′-TTAGGAGCGCACGTGTC-3′) for the variable region of the HvZIP7 coding sequence.
Subcellular localization of HvZIP7
HvZIP7 fragments containing the full-length open reading frame were amplified from barley cDNA by PCR with the primer pair of 5′- TCAGGGCATGATGATCGGTGTA-3′ and 5′-CAATCAGATGGACACAGGCACAT-3′. The HvZIP7 fragments were cloned into a pGEM®T-easy vector (Promega) and sequenced. The coding sequence of HvZIP7 without the stop codon was amplified from pGEM®T-easyHvZIP7 plasmid using the primer pair of 5′-ATGATGATCGGTGTAGCAGGCTTC-3′ and 5′-GGCCCAGACTGCAAGCAT-3′. The PCR fragment was ligated into the pCR8-GW-TOPO vector (Invitrogen) and transferred into pMDC83 containing the mGFP gene (Curtis & Grossniklaus, 2003). The resulting plasmid (pMDC83HvZIP7::GFP) places HvZIP7 upstream of mGFP. Transient expression of HvZIP7::GFP in onion epidermal cells and visualization of green fluorescent protein (GFP) in subcellular locations were conducted as described by Preuss et al. (2010). The plasmolysis of onion epidermal cells was performed by immersion in 1 M sucrose for 1 min before confocal image analysis.
Generation of HvZIP7 overexpressing barley plants
The full-length open reading frame of HvZIP7 was amplified from the pGEM®T-easyHvZIP7 plasmid by PCR using the primer pair of 5′-CTGGTACCATGGTGATCGGTGTAGCAG-3′ and 5′-TACTCGAGTTCAGGCCCAGACTGC-3′. The PCR product was ligated into the pCR8-GW-TOPO vector and then transferred into the pMDC32 vector (Invitrogen) by LR reaction according to the manufacturer's instructions. The resulting pMDC32HvZIP7 plasmid was used for barley (cv Golden Promise) transformation using an Agrobacterium-mediated method described by Tingay et al. (1997) and Matthews et al. (2001). The constitutive expression of HvZIP7 was driven by a double CaMV35S promoter. The number of transgene loci in transgenic plants was determined by standard Southern hybridization (Fig. S2a), and transgene segregation in progenies was analyzed by PCR using the transgene-specific primers described earlier for transcript analysis.
Plant growth and accumulation of Zn and Cd in soil
Zinc-deficient siliceous sand (DTPA-extractable Zn, 0.07 mg kg−1 soil) was used in soil experiments. Calcium carbonate of 0.5% (w/w) was added to the sand to simulate calcareous sandy soil for maintaining high pH at 8.0 and reducing bioavailability of Zn (Genc et al., 2007). Basal nutrients were as described by Genc et al. (2007). Micronutrients (mg kg−1 soil) included FeSO4.7H2O (16.8), MnSO4.H2O (3), CuSO4.5H2O (5), H3BO3 (0.3), CoSO4.7H2O (1), MoO3 (0.005), and NiSO4.6H2O (0.15). Four Zn rates (0, 1.0, 5.0, and 12.5 mg Zn kg−1 soil) as ZnSO4 were applied to cover plant Zn nutrition range from deficiency to luxury. Two barley grains with emerged radicles were planted into each pot containing 1 kg soil. Plants were thinned to one per pot 9 d after imbibition (D9). Plants were grown in a growth room at 20 : 15°C day : night temperature and a photoperiod of 14 : 10 h day : night at 300 μmol m−2 s−1 photon flux intensity at the plant level. At D28, plants were harvested. Roots were washed free of soil, briefly rinsed in deionized water and excess water was blotted on fresh laboratory tissues. The roots and shoots were then oven-dried at 80°C for 48 h for nutrient analysis.
For Cd accumulation, the experimental setup was similar to that of the Zn accumulation experiment. Two rates of Cd (3 and 10 mg Cd kg−1 soil as CdCl2) were added into the soil with basal nutrients. At D26, shoots were harvested and oven-dried at 80°C for 48 h for nutrient analysis.
Plant growth and Zn accumulation in nutrient solution
Plant growth conditions and basal nutrients in nutrient solution were the same as described earlier in transcript analysis except for plant number, harvest dates and replacement of nutrient solution. Three of the four seedling cups in each container contained two plants for the first three harvests at 14 d after seed imbibition (D14), D15 and D16, and one plant for the last harvest at D20. All plants were grown with 0.5 μM Zn until D14. Four Zn treatments (0.5, 5, 10, and 20 μM Zn supplied as ZnSO4) were applied at D14. Macronutrients and micronutrients were supplied at half and full strength, respectively, until D10, and at full strength of all nutrients thereafter. The nutrient solution was aerated continuously and replaced at D10, D12, D14, and D17, respectively. The pH of the solution was constant at pH 6.0 during the experiment. Light and temperature settings for plant growth were identical to those for soil-grown plants described earlier. At each harvest, roots were briefly rinsed in deionized water and excess water was blotted onto fresh laboratory tissues. The roots and shoots were separated and then oven-dried at 80°C for 48 h for mineral element analysis.
Plant growth in potting mix
Plants were grown in pots (15 cm diameter ×15 cm height) containing 1 kg of potting mix. The potting mix consisted of coco peat (540 l), sand (60 l), hydrated lime (0.24 kg), agricultural lime (0.6 kg), dolomite lime (calcium magnesium carbonate, 0.18 kg), calcium nitrate (0.45 kg), superphosphate (0.18 kg), iron sulfate (0.45 kg), iron chelate (0.03 kg), Microplus (0.18 kg; Langley Fertilizers, Perth, WA, Australia) and mini-Osmocote (1.8 kg), and the pH of the mix was 6.5. Microplus contains 0.4% (w/w) of soluble Zn (equivalent to 17.8 mg Zn kg−1 potting mix) as well as Fe, Mn, Mo, B and Mg. One barley seed with emerged radicle was planted in the pots. Deionized water was used for watering during the experiment. Plants were grown in a glasshouse under c. 24 : 13°C day : night temperature and 12 : 12 h day : night cycle. Shoots and grains were harvested at maturity and oven-dried at 80°C for 48 h.
Mineral element analysis
Dry plant samples and grains were weighed to c. 0.3 g, digested with an 11 ml nitric acid/perchloric acid mixture (10 : 1 v/v), boiled until the volume was reduced to c. 1.0 ml, added to a final volume of 25 ml using deionized water, and then analyzed for mineral elements by inductively coupled plasma optical emission spectrometry at Waite Analytical Services, Urrbrae, South Australia (Wheal et al., 2011).
All experiments were set up as a completely randomized block design with three to four replicates. The data were analyzed using the Genstat Statistical Program (version 11.1; VSN International Ltd, Hemel Hempstead, UK), and pairwise comparisons of the means were made using the least significance difference (LSD) test at P = 0.05. To overcome the problem of nonhomogeneity of variances, log-transformed data were used for the ANOVA.
Identification of members of the ZIP family from the barley genome
Sequences of 16 ZIP family members of rice (O. sativa) and 15 reported members of A. thaliana, along with three newly identified members (AtIAR1, AtPutZnT and AtZTP29) (Table S1) were used to search for candidate members in barley as well as Brachypodium (B. distachyon) and wheat (T. aestivum). Nine new barley ZIP proteins (HvZIP1, HvZIP2, HvZIP6, HvZIP7, HvZIP10, HvZIP11, HvZIP13, HvZIP14 and HvZIP16) were identified in addition to the four HvZIP proteins (HvIRT1, HvZIP3, HvZIP5 and HvZIP8) reported by Pedas et al. (2008, 2009). NcZNT1, a close homolog of AtZIP4, was also included for phylogenetic analyses of ZIP proteins, as it has been shown to be responsible for the Zn/Cd hyperaccumulation in N. caerulescens (Lasat et al., 2000; Milner et al., 2012). Phylogenetic analyses showed that HvZIP7 was closely related to OsZIP7, forming a unique clade with HvZIP10 and OsZIP10. Three AtZIP proteins (AtZIP4, AtIRT3, and AtZIP9) and NcZNT1 were grouped to the clade of HvZIP7 and OsZIP7 (Fig. S1). HvZIP1 and HvZIP2 (249 amino acid residues), which were 82% identical to the wheat ZIP protein, TaZIP2, were clustered with OsZIP1, OsZIP2, AtZIP2 and AtZIP11 (Fig. S1). HvZIP3 formed a distinct clade together with OsZIP3 and OsZIP4. HvZIP5 and HvZIP8 were 86% similar in protein sequence, and they were grouped with OsZIP9 and OsZIP5 (Fig. S1). HvZIP5 and HvZIP8 were also closely related to OsZIP8 and HvZIP13 (Fig. S1). HvZIP6, HvZIP11, HvZIP14 and HvZIP16 formed a distinct clade with OsZIP6, OsZIP11, OsZIP14 and OsZIP13/OsZIP16, respectively. Each of these four clades also contained one Arabidopsis homolog, AtZIP6, AtPutZnT, AtIAR1 and AtZTP29, respectively (Fig. S1).
Sixteen members of the ZIP gene family were identified in the fully sequenced genome of Brachypodium, a new model for grasses (Brkljacic et al., 2011). Almost all distinct clades of Brachypodium and rice ZIP gene families contain at least one barley homolog (Fig. S1), indicating that barley homologs are represented in the major clades of the ZIP gene family in monocotyledonous species. At least seven wheat ZIP (TaZIP) proteins were found in distinct homologous groups for HvZIP proteins (Fig. S1). The results indicate that the ZIP family proteins are relatively conserved in monocotyledonous lineages. By comparison, some members of the AtZIP family, such as AtZIP1, AtZIP3, AtZIP5 and AtZIP12, branched separately from all other sequences in the tree with low bootstrap values (Fig. S1), suggesting that sequence divergence occurs in some members of the ZIP family between monocotyledonous and dicotyledonous species. HvZIP7 was selected for further studies because it belongs to a phylogenetic clade different from those (OsZIP4, OsZIP5 and OsZIP8) previously studied (Ishimaru et al., 2007; Lee et al., 2010a,b), and its homologs (NcZNT1, AtZIP4 and AtIRT3) have been shown to play an important role in Zn hyperaccumulation (Lasat et al., 2000; Hanikenne et al., 2008; Lin et al., 2009).
Expression of HvZIP7 in both roots and shoots is increased by Zn deficiency
Transcript abundance of HvZIP7 was low in roots and shoots of Zn-adequate barley plants (Fig. 1a). Zn deficiency increased the transcript abundance of HvZIP7 in both roots and shoots by at least ninefold (Fig. 1a). To find out in which cell types HvZIP7 is specifically expressed, in situ RT-PCR was performed in cross-sections of roots and leaves of Zn-deficient plants. There were no HvZIP7 transcripts detected in the no-RT controls of either roots or leaves (Fig. 1b,d,f). HvZIP7 transcripts in the roots were detected mainly in epidermal cells and the cells within the vascular bundle of roots (Fig. 1c,e), while in the cross-section of leaves HvZIP7 transcripts were found primarily in the cells of the vascular bundle (Fig. 1g). These results suggest that HvZIP7 is involved in root uptake and translocation of Zn and Zn translocation in leaves.
HvZIP7 localizes to the plasma membrane
To determine subcellular localization of HvZIP7, an HvZIP7::GFP construct was generated for transient expression in onion epidermal cells. The green fluorescence of HvZIP7::GFP appeared in the location of the plasma membrane (PM; Fig. 2a,b). To distinguish PM from the vacuolar membrane, the onion epidermal cell, which expressed HvZIP7::GFP, was plasmolyzed for visualization of the Hechtian strands connecting PM to the cell wall (Fig. 2c,d). A close-up view shows adhesion of the green fluorescent strands of PM to the cell wall (Fig. 2e,f), confirming that the fluorescent signal derived from HvZIP7::GFP is associated with PM.
HvZIP7 overexpression in plants results in higher Zn accumulation in shoots when Zn supply is moderately high in soil
A number of attempts to complement two yeast strains involved in Zn homeostasis (zrt1zrt2 mutant defective in Zn uptake and zrc1cot1 mutant with a Zn sensitive phenotype) with HvZIP7 were unsuccessful. Therefore, transgenic lines overexpressing HvZIP7 were generated in barley for gain-of-function analyses. Transgenic plants with a single locus insert of the HvZIP7 transgene were identified by Southern blot analysis (Fig. S2a) and selected for propagation. Progenies of three independent, single locus-insert lines (OX-2, OX-7 and OX-10) were genotyped by PCR to identify homozygous lines with or without transgene (null). As the processes of tissue culture, selection, and plant regeneration were employed for Agrobacterium-mediated barley transformation (Tingay et al., 1997; Matthews et al., 2001), these processes could potentially induce genetic alterations in the resulting transformants. If such changes are present in the transformants selected, a null line arising from segregation of the transgene would be a more appropriate control because both of them have been exposed to the same tissue culture conditions, unlike the wildtype plant. Therefore, null lines were included as a control in all further molecular and physiological experiments, and the wildtype control was added whenever necessary. Semiquantitative RT-PCR analysis showed that the HvZIP7 transgene was expressed in all three independent transgenic lines, while no transcripts were detected in either the corresponding null line or the wildtype (Fig. S2b).
The three independent homozygous transgenic lines were initially examined for Zn accumulation in various growth conditions supplemented with low to moderately high Zn supply. There were no differences in plant growth and Zn accumulation observed between transgenic and null lines or the wildtype when Zn supply was low to adequate. However, when the transgenic lines were grown with moderately high Zn supply, a significant difference in plant Zn accumulation was observed between the transgenic and null lines or the wildtype. To substantiate these early findings, one of the three transgenic lines, OX-10, was selected to study plant Zn accumulation in both soil and nutrient solution. When OX-10 was grown in soil supplemented with 0 to 12.5 mg Zn kg−1 soil (covering the plant Zn nutrition range from deficiency to luxury), there were no significant differences in dry weight of shoots or roots between the transgenic and null or wildtype plants after 28 d of growth in any of the four rates of Zn supplementation (Fig. 3a,b). The Zn concentrations of shoots in all three genotypes rose from c. 10 to 500 μg Zn g−1 DW or more with an increase in the Zn supplement rate (Fig. 3c). The shoot Zn concentration of 10 μg Zn g−1 DW at nil Zn is below the critical level required for normal plant growth (Genc et al., 2002), while the shoot Zn concentration of 500–700 μg Zn g−1 DW at 12.5 mg Zn kg−1 soil is close to toxic concentrations reported for barley (Aery & Jagetiya, 1997). Failure to detect significant growth reductions of shoots for either nil Zn or the highest Zn supplement could be as a result of the appearance of Zn deficiency or toxicity only towards the end of the experiment (Fig. 3a,b). At nil Zn, there was no significant difference in shoot Zn concentrations between the transgenic and null or wildtype plants (Fig. 3c). Significant differences in shoot Zn concentrations were observed between the transgenic and null or wildtype plants when the Zn supplement was 1.0 mg Zn kg−1 soil or higher (Fig. 3c). The transgenic plants had at least 60% higher Zn concentrations than those of the null or wildtype (Fig. 3c). The transgenic plants had also higher shoot Zn content than the null or wildtype plants when the Zn supplement was 1.0 mg Zn kg−1 soil or higher (Fig. 3e).
Zinc concentrations in roots of all three genotypes also rose as Zn supplement rates increased (Fig. 3d). There were no significant differences in root Zn concentration and content between the transgenic and null or wildtype plants at either 0 or 12.5 mg Zn kg−1 soil (Fig. 3d,f). Notably, the transgenic plants had a significantly lower Zn concentration and content in roots than the null or wildtype when supplemented with 1.0 and 5.0 mg Zn kg−1 soil (Fig. 3d,f). Despite a small reduction in root Zn content at 1.0 and 5.0 mg Zn kg−1 soil, total Zn accumulation in the HvZIP7 overexpressing plants across all four Zn rates remained significantly higher (P < 0.01) than that of the null or wildtype (Fig. 3e,f). These results indicate that the HvZIP7 overexpressing plants can increase Zn accumulation in plants only when Zn availability in soil is moderately high.
Increasing the rates of Zn supplementation reduced the concentrations of Fe, but increased the concentrations of Mn and Cu in shoots and roots of all three genotypes (Fig. S3). However, the overexpression of HvZIP7 had little effect on the content of Fe, Mn or Cu in either shoots or roots relative to the null or wildtype plants (Fig. 4). These results indicate the specificity of HvZIP7 as a Zn transporter.
HvZIP7 overexpressing plants results in higher accumulation of Zn in shoots when the Zn concentration is 5 μM or above in nutrient solution
To determine what external Zn concentrations are required for increased accumulation of Zn in HvZIP7 overexpressing plants, a short-term accumulation experiment in nutrient solution with Zn concentrations ranging from 0.5 to 20 μM was conducted. The Zn concentration of 0.5 μM in the nutrient solution is adequate for barley seedling growth (Genc et al., 2007). Zn accumulation in null and transgenic HvZIP7 plants of OX-10 increased linearly with Zn concentration in the nutrient solution, and a small but significant increase of Zn accumulation in shoots of HvZIP7 overexpressing plants relative to the null could be detected as early as 24 h after exposure to the 10 μM Zn treatment. After 48 h, there were no significant differences in shoot and root dry weight between the transgenic and null or wildtype plants (Fig. S4a,b), but there was a large and significant increase in shoot Zn concentration and content in the transgenic plants at Zn concentrations of 5 μM and above (Fig. 5a,b). By contrast, there were no differences in Zn concentration and content of roots between null and transgenic plants at any concentrations of Zn supply (Fig. 5c,d). These results indicate that overexpression of HvZIP7 increases Zn accumulation in shoots, which could result from either higher Zn uptake and/or translocation from roots to shoots at external Zn concentrations of 5 μM or higher.
By contrast, and in agreement with the Zn accumulation experiment conducted in soil, there were no effects of overexpression of HvZIP7 on the accumulation of Fe, Mn or Cu in either shoots or roots (Fig. 6), despite changes in Fe, Mn and Cu concentrations of roots and shoots as a result of different Zn supplies (Fig. S5). Again, these results indicate that HvZIP7 is specific for Zn transport.
Overexpression of HvZIP7 has no effect on Cd accumulation in plants
The overexpression of HvZIP7 increased Zn accumulation in shoots (Figs 3e, 5b), but not accumulation of Fe, Mn or Cu in shoots (Figs 4, 6). However, the toxic metal ion Cd is an analogue of Zn, and may be a substrate for HvZIP7. The ability of HvZIP7 overexpressing plants to transport Cd was tested by growing plants in soil, similar to that in Fig. 3, supplemented with two rates of Cd (3 and 10 mg Cd kg−1 soil). There were no significant differences in plant growth between the transgenic and null lines at either rate of Cd (data not shown). Concentrations and content of Cd in the shoots of all plants were increased with Cd supply (Fig. 7a,b), but there were no significant differences in shoot Cd concentration or content between the transgenic and null lines at either rate of Cd supply (Fig. 7a,b). These results indicate that the overexpression of HvZIP7 in barley does not affect Cd accumulation in shoots.
HvZIP7 overexpressing plants have higher grain Zn content
To examine whether the overexpression of HvZIP7 could have impacts on plant growth, grain yield and grain Zn content, all three independent transgenic lines, OX-2, OX-7 and OX-10 lines (Fig. S2), were grown to maturity in a potting mix containing relatively high Zn supply. There were no significant differences in total shoot dry weight or grain yield between the transgenic and null or wildtype lines (Fig. 8a,b). Furthermore, the overexpression of HvZIP7 did not affect total grain number or 1000 grain weight (Fig. S6). These results show that the overexpression of HvZIP7 does not have negative impacts on plant growth and grain yield.
The average grain Zn concentration in the null lines and wildtype was 70 mg kg−1 DW (Fig. 8c), which is comparable to Zn concentrations of field-grown wheat plants with applied soil and foliar Zn supplements (Zou et al., 2012). By contrast, the average grain Zn concentration in the three transgenic lines was at least 60% higher than that of either the null or wildtype (Fig. 8c). The total grain Zn content of all three transgenic lines was also significantly higher (by c. 60%) than that of the null or wildtype (Fig. 8d). These results indicate that overexpression of HvZIP7 can increase grain Zn content when soil is supplemented with Zn fertilizer. Furthermore, higher grain Zn content in the transgenic plants has little impact on the concentrations or content of Fe, Mn or Cu in the grain (Fig. S7). This further indicates a high specificity of HvZIP7 for Zn transport.
HvZIP7 displays characteristics of low-affinity Zn transport in the plant
HvZIP7 belongs to a distinct phylogenetic clade that comprises OsZIP7, NcZNT1, AtZIP4, AtZIP9 and AtIRT3 (Fig. S1), and excludes other monocotyledonous ZIP members previously characterized (OsZIP4, OsZIP5 and OsZIP8). We were unable to restore growth of the zrt1zrt2 yeast mutant defective in Zn uptake, although different yeast expression vectors (pFL61 and pYES2) and yeast strains (ZHY3 and BY4741) were used for the expression of HvZIP7. Notably, yeast mutant complementation for three close homologs of HvZIP7 (i.e. OsZIP7a, AtZIP4 and AtZIP9) could not be established either (Grotz et al., 1998; Yang et al., 2009; Milner et al., 2013). This suggests that some members of this phylogenetic clade may target different cellular membranes in yeast as a result of an absence of plant-specific, interacting proteins/molecules, or that they have a different affinity for Zn transport from the ZIP family members that are able to restore growth of the yeast mutant. However, overexpression of HvZIP7 in barley increased Zn accumulation in shoots at moderately high rates of Zn supply (Figs 3e, 5b), demonstrating that HvZIP7 is able to transport Zn in planta, at least under relatively Zn-abundant conditions. The low micromolar range of external Zn, in which the higher accumulation of Zn was detected in the HvZIP7 overexpressing plants (Fig. 5b), is similar to that described for the low-affinity transporter system in alga and wheat (Reid et al., 1996; Hacisalihoglu et al., 2001). We did not observe higher accumulation of Zn in the shoots of the HvZIP7 overexpressing plants when they were grown under low Zn supply or adequate Zn supply for optimum plant growth (Figs 3e, 5b), suggesting that HvZIP7 functions as a low-affinity Zn transporter in plants. It is interesting to note that higher Zn accumulation in Arabidopsis overexpressing NcZNT1 could be detected similarly only at external Zn concentrations > 5 μM (Milner et al., 2012). This suggests either that NcZNT1 in the plant cell behaves differently from the way it behaves in yeast, or that the complexity of the Zn transport system in whole plants and interactions of different Zn transporters between cell types influence measurements relating to the affinity of a specific transporter protein. Further studies are warranted.
HvZIP7 mediates higher root uptake and/or root-to-shoot translocation of Zn
HvZIP7 was localized in the plasma membrane of onion epidermal cells (Fig. 2a,c). HvZIP7 was highly induced by Zn deficiency in roots (Fig. 1a), and found to be expressed in epidermal cells and vascular tissues of roots (Fig. 1c,e). The membrane localization and tissue expression patterns of HvZIP7 are similar to those of the two closely related homologs from dicotyledonous species, NcZNT1 (Milner et al., 2012) and AtIRT3 (Lin et al., 2009). Overexpression of NcZNT1 has been shown to be responsible for increased accumulation of Zn in root cells at high external Zn concentrations of 30 μM (Milner et al., 2012). A heavy metal ATPase4 (HMA4)-dependent up-regulation of AtIRT3 and AtZIP4 was also observed in the roots of Arabidopsis (Hanikenne et al., 2008), suggesting that high expression of AtIRT3 and AtZIP4 in roots is needed for increased Zn accumulation in HMA4 overexpressing plants. Collectively, these data suggest that HvZIP7 plays a role in Zn uptake by roots in monocotyledonous species similar to its close homologs in dicotyledonous species. Importantly, the increased Zn uptake by the roots of HvZIP7 overexpressing plants can lead to the enhanced accumulation of Zn in shoots (Figs 3e, 5b). This differs from the three ZIP genes (OsZIP4, OsZIP5, and OsZIP8) previously characterized in monocotyledonous species (Ishimaru et al., 2007; Lee et al., 2010a,b). The three OsZIP genes can increase Zn accumulation in roots, but are not able to enhance root-to-shoot translocation. Although the precise mechanism for how HvZIP7 mediates higher Zn accumulation in shoots is not fully understood, our results indicate that higher root uptake and/or higher translocation from roots to shoots drives the increased Zn accumulation. A mechanistic explanation for the involvement of HvZIP7 requires further investigation.
HvZIP7 shows specificity for Zn transport in plants
Several ZIP family proteins have been shown to transport Zn and also other metals such as Fe and Cd in plants (Connolly et al., 2002; Ishimaru et al., 2007). However, the overexpression of HvZIP7 did not increase the content of Fe, Mn, or Cu in either roots or shoots (Figs 4, 6), or in grains (Fig. S7). The overexpression of HvZIP7 did not increase the accumulation of Cd in shoots either (Fig. 7b). This indicates that HvZIP7 is highly selective for Zn ions. NcZNT1 is also fairly selective for Zn in the plant (Milner et al., 2012). Ion selectivity is crucial for cereal biofortification because the accumulation of undesirable, toxic metals such as Cd will adversely affect human health (Zhao & McGrath, 2009). For example, AtIRT1 is capable of transporting Fe and Zn as well as Cd (Connolly et al., 2002). The HMA subfamily of metal transporters, including AtHMA2, AtHMA4, and OsHMA2, can also transport both Zn and Cd in planta (Hussain et al., 2004; Satoh-Nagasawa et al., 2012). The apparent selectivity of HvZIP7 for metal ions would allow manipulation of the specific uptake of Zn from soil for Zn biofortification in cereals.
Higher Zn accumulation in grains of HvZIP7 overexpressing plants
Zinc translocation from roots to shoots has been considered a bottleneck for increasing grain Zn content (Palmgren et al., 2008). This is evident in studies utilizing the overexpression of OsZIP4, OsZIP5, and OsZIP8 (Ishimaru et al., 2007; Lee et al., 2010a,b). The higher accumulation of Zn in shoots of HvZIP7 overexpressing plants demonstrates that the overexpression of HvZIP7 can overcome the restriction on Zn translocation from roots to shoots when the external Zn supply is relatively abundant (Figs 3e, 5b). This should improve Zn fertilizer use efficiency from soil application and lead to more Zn in shoots available for loading into grains (Hegelund et al., 2012). As HvZIP7 is expressed in vascular tissues of leaves (Fig. 1f,g), it is possible that the overexpression of HvZIP7 would also show increased Zn partitioning from vegetative tissues to grains and a response to foliar Zn fertilization. Further investigation is required to assess the relative benefits of soil vs foliar Zn applications in relation to grain Zn loading in transgenic plants. Our results showed an increase of c. 60% in grain Zn content in HvZIP7 overexpressing plants relative to controls when available Zn was relatively high in the growth medium (Fig. 8d). Recent field studies in wheat show that grain Zn concentrations with soil Zn fertilization (Zou et al., 2012) are comparable to those measured in control plants from the current study (Fig. 8c). This suggests that transgenic cereals overexpressing HvZIP7 have the potential to achieve 60% higher grain Zn content in the field when soils are supplemented with Zn fertilizers.
In summary, HvZIP7 is a close homolog of NcZNT1, AtZIP4 and AtIRT3 from dicotyledonous species. It is localized in the plasma membrane, and shares a similar tissue expression pattern with its dicotyledonous counterpart. We demonstrate that high expression of HvZIP7 can mediate specific root uptake and/or root-to-shoot translocation of Zn at moderately high concentrations of Zn supply, suggesting that it functions as a low-affinity Zn transporter. These findings are of significance for the development of Zn biofortification strategies in cereals.
We are grateful to R. Singh, U. Langridge, N. Shirley, A. Ismagul, G. Mayo, and S. Conn for their technical support in generating and growing transgenic plants, real-time RT-PCR, biolistic bombardment, and in situ PCR. This work was supported by the Grains Research and Development Corporation (C.Y.H., J.T., J.E.H., J.T., P.L.), the Australian Research Council (C.Y.H., J.T., J.E.H., J.T., P.L.), the South Australian Government (C.Y.H., J.T., J.E.H., J.T., P.L.), the University of Adelaide (J.W.T., G.K.M., Y.G., C.Y.H., J.E.H., J.T., P.L.), and the Danish Strategic Research Council (PP, NUTRIEFFICIENT; grant no. 10-093498).