Systemic Upregulation of MTP2- and HMA2-Mediated Zn Partitioning to the Shoot Supplements Local Zn Deficiency Responses

The physiological Zn status of the shoot controls root transcript levels of MTP2 and HMA2, which encode membrane transport proteins and act in Zn partitioning to the shoot of Zn-deficient Arabidopsis.

Abstract

Low bioavailable concentrations of the micronutrient zinc (Zn) limit agricultural production on 40% of cultivated land. Here, we demonstrate that plant acclimation to Zn deficiency involves systemic regulation. Physiological Zn deficiency of Arabidopsis thaliana shoots results in increased root transcript levels of the membrane transport protein-encoding genes METAL TRANSPORT PROTEIN2 (MTP2) and HEAVY METAL ATPASE2 (HMA2), which are unresponsive to the local Zn status of roots. MTP2 and HMA2 act additively in the partitioning of Zn from roots to shoots. Chimeric GFP fusion proteins of MTP2 complement an mtp2 mutant and localize in the endoplasmic reticulum (ER) membrane of the outer cell layers from elongation to root hair zone of lateral roots. MTP2 restores Zn tolerance in a hypersensitive yeast mutant. These results are consistent with cell-to-cell movement of Zn toward the root vasculature inside the ER-luminal continuum through the desmotubules of plasmodesmata, under Zn deficiency. The previously described Zn deficiency response comprises transcriptional activation of target genes, including ZINC-REGULATED TRANSPORTER IRON-REGULATED TRANSPORTER PROTEIN genes ZIP4 and ZIP9, by the F-group bZIP transcription factors bZIP19 and bZIP23. We show that ZIP4 and ZIP9 respond to the local Zn status in both roots and shoots, in contrast to the systemic regulation identified here. Our findings are relevant for crop management and improvement toward combating human nutritional Zn deficiency that affects 30 to 50% of the world’s population.

INTRODUCTION

As photoautotrophic organisms, plants depend merely on inorganic nutrients for their growth and reproduction. Insufficient supply of any essential mineral nutrient can cause symptoms of limitation, for example, growth arrest, chlorosis, infertility, and necrosis (Marschner, 1995). Plant acclimation to mineral nutrient deficiencies can include enhanced nutrient mobilization in the soil by plant-mediated alterations of the rhizosphere, alterations in roots for enhanced affinity and efficacy of nutrient uptake into root cells, increased efficacy of nutrient movement across the root toward the xylem for translocation to the shoot, mobilization of stored nutrient pools, and economizing on nutrient use (Marschner, 1995; Burkhead et al., 2009; Kobayashi and Nishizawa, 2012; Sinclair and Krämer, 2012). In response to nutrient deficiency, plants can also alter the permeability of the root endodermal barrier for a facilitated passage of specific groups of nutrient ions into the vasculature (Barberon et al., 2016).

Zinc (Zn) is an essential micronutrient, and ∼10% of the proteins encoded in the genome of Arabidopsis thaliana have been estimated to interact with Zn²⁺ cations (Andreini et al., 2006a, 2006b). Besides the direct participation in enzymatic catalysis, Zn²⁺ ions can have structural roles in the generation of protein tertiary structure and protein-protein interactions. Soils in many of the world’s important agricultural regions are low in Zn, leading to yield losses in the absence of targeted and controlled input of mineral fertilizers (Alloway, 2008). Suboptimal Zn concentration in crops is a prevalent cause of human nutritional Zn deficiency, which can result in immune and neurological dysfunction, as well as growth defects (Prasad, 2013; Hara et al., 2017). As many as two billion people worldwide are estimated to be at risk of Zn deficiency (Caulfield et al., 2004; Wessells and Brown, 2012). Thus, understanding how plants maintain Zn nutrition, growth, and reproductive fitness under Zn deficiency is vital for both agriculture and human health.

The transcriptional upregulation of genes with diverse functions in Zn nutrition is a well-characterized component of plant acclimation to Zn deficiency. Profiling of Arabidopsis Zn deficiency responses identified increased levels of transcripts encoding 10 of the 15 members of the zinc-regulated transporter/iron-regulated transporter-like protein (ZIP) transmembrane metal transporter family of proteins, which generally operate to move divalent metal cations into the cytosol (Grotz et al., 1998; Wintz et al., 2003; Talke et al., 2006; van de Mortel et al., 2006). The biological roles of most plant ZIP family proteins are still poorly understood. Among these, Arabidopsis IRON-REGULATED TRANSPORTER3 (AtIRT3) was shown to localize to the plasma membrane in planta and to complement Zn and iron (Fe) uptake-defective yeast mutants. AtIRT3-overexpressing (Pro35S:IRT3) plants accumulated more Zn and Fe than the wild type (Lin et al., 2009). AtZIP1 complemented a yeast mutant disrupted in Zn uptake and was localized in the vacuolar membrane (Grotz et al., 1998; Milner et al., 2013). While AtZIP2 transcript levels are unresponsive to Zn deficiency, heterologous expression in yeast suggested that AtZIP2 can act as a Zn transporter in vivo. Single Arabidopsis zip1 and zip2 mutants show no phenotype (Milner et al., 2013), and an irt3 mutant has yet to be characterized. It was suggested that functional redundancy among the multiple ZIP family members is responsible for the absence of detectable phenotypes in the characterized single mutants.

Some studies reported Zn deficiency to result in an upregulation of transcript levels of NICOTIANAMINE SYNTHASE (NAS) genes encoding enzymes that catalyze the biosynthesis of the metal chelator nicotianamine and of HEAVY METAL P-TYPE ATPASE2 (HMA2) (Wintz et al., 2003; Talke et al., 2006; van de Mortel et al., 2006). HMA2 localizes to the plasma membrane of root pericycle cells, and it acts to pump Zn²⁺ cations into the xylem for transport to the shoot (Hussain et al., 2004; Wong and Cobbett, 2009). Taking together different published results, it can be hypothesized that increased HMA2 expression in Zn-deficient plants is likely to enhance root-to-shoot Zn flux. A paralog of HMA2, HMA4, is constitutively expressed at low levels and transcriptionally unresponsive to plant Zn status in Arabidopsis (Talke et al., 2006).

The regulation of Zn deficiency responses is not well understood. In Arabidopsis, the basic leucine zipper domain-containing transcription factors bZIP19 and bZIP23 were found to interact with Zn deficiency response elements in the promoters of many genes that are transcriptionally upregulated under Zn deficiency, including eight ZIP and two NAS family genes (Assunção et al., 2010). A bzip19 bzip23 double mutant lacks these central transcriptional Zn deficiency responses and shows severe symptoms of limitation under Zn-deficient growth conditions. To date, we do not know how the bZIP19/bZIP23 transcription factors are activated (Assunção et al., 2013).

It has not been examined whether known Zn deficiency responses, for example, the upregulation of ZIP4 or ZIP9 transcript levels, are regulated locally or systemically. Here, we employ the Arabidopsis hma2 hma4 double mutant as a tool for identifying candidate systemic Zn deficiency responses. On media containing normal levels of Zn, hma2 hma4 double mutants exhibit a phenotype of Zn limitation in shoots, whereas root Zn concentrations are moderately increased compared with the wild type. Accordingly, HMA2 and HMA4 proteins were concluded to redundantly contribute to the efflux of Zn from root pericycle and xylem parenchyma cells into the apoplastic xylem for mass flow-driven translocation of Zn from the root to the shoot (Hussain et al., 2004; Sinclair et al., 2007). It can be expected that any locally regulated Zn deficiency responses of roots are repressed in hma2 hma4 double mutants. By contrast, we expect that any shoot-governed systemically regulated Zn deficiency responses of roots are enhanced in hma2 hma4 plants compared with the wild type.

Here, we identified, among other genes, root transcript levels of HMA2 and of METAL TRANSPORT/TOLERANCE PROTEIN2 (MTP2) to be regulated systemically in response to shoot physiological Zn status. By contrast, the known target genes of the Zn deficiency response-regulating transcription factors bZIP19/23, ZIP4 and ZIP9, responded strictly to local Zn status in both roots and shoots. Our results indicate that HMA2 and MTP2 function in the roots of Zn-deficient plants to enhance the partitioning of Zn to the shoot. Furthermore, we propose that MTP2 contributes to an unsuspected pathway of symplastic radial cell-to-cell movement of Zn toward the root stele within the lumen of the endoplasmic reticulum (ER). In summary, we have identified a bZIP19/23-independent component of Zn deficiency responses that involves the systemic regulation of transcript levels of genes contributing to root-to-shoot Zn translocation.

RESULTS

Experimental Design and Validation

We aimed to identify the subset of root transcriptional Zn deficiency responses that is regulated systemically dependent on shoot Zn status. For this, we took advantage of the hma2 hma4 double mutant and its tissue type-specific alterations in Zn content (Supplemental Figure 1A). To validate the chosen experimental strategy under our growth conditions, we confirmed that shoot Zn concentrations were similarly lowered in wild-type seedlings grown on Zn-deficient medium as in hma2 hma4 seedlings grown on control medium (16–40 µg Zn g⁻¹ dry biomass), by comparison to wild-type seedlings grown on control medium (190 µg Zn g⁻¹ dry biomass; Figure 1A). Despite lowered Zn concentrations in hma2 hma4 shoot tissues, roots moderately overaccumulated Zn as expected (Supplemental Figure 1B) (Hussain et al., 2004). To assess physiological Zn status, we conducted RT-qPCR analyses of physiological Zn-deficiency marker gene expression. Relative transcript abundance of the bZIP19/23 targets ZIP4 and ZIP9 reflected local tissue Zn concentrations (Figure 1B), indicating Zn deficiency in shoot tissues of hma2 hma4 seedlings cultivated under control conditions. An analogous response to local tissue Zn concentrations was observed at the level of ZIP4 promoter activity in ZIP4 promoter-GUS (ProZIP4:GUS) reporter lines in the wild-type and hma2 hma4 genetic backgrounds upon visualizing GUS activity through histochemical staining (Figure 1C). Taking these results together, upon cultivation in control medium, shoots of the hma2 hma4 double mutant were physiologically Zn deficient comparable to the wild type grown in zinc-deficient medium. Conversely, roots of the double mutant were physiologically more Zn-sufficient than roots of wild-type plants grown in control medium.

Figure 1.

Experimental Design and Validation, and Results of Transcriptomics Targeting Systemic Zn Deficiency Response of Roots.

(A) Shoot Zn concentrations in 21-d-old wild-type seedlings grown on Zn-deficient medium (−Zn) and hma2 hma4 plants grown on control agar-solidified medium for 14 d, subsequent to an initial cultivation on agar-solidified 0.5× MS medium for 7 d.

(B) ZIP4 and ZIP9 relative transcript levels in shoot (S) and root (R) of WT and hma2 hma4 seedlings, based on RT-qPCR. Seedlings were grown as in (A).

(C) GUS activity in ProZIP4:GUS wild-type and hma2 hma4 seedlings. Pictures are representative of 12 independent transgenic lines in Col-0 background, of which one representative line was used to introduce ProZIP4:GUS into the hma2 hma4 background by crossing. Shown are F3 seedlings that are homozygous for the ProZIP4:GUS transgene and for the hma2 and hma4 mutations. Seedlings were grown as in (A). Bars = 5 mm (WT) and 10 mm (hma2 hma4).

(D) Summary of microarray data. Shown are the number of genes showing ≥2-fold differences in transcript levels (in parentheses for same direction of changes in both comparisons).

Bar graphs show arithmetic mean ± sd (n = 3 pools of 16 plants, with each pool from one plate [A]; n = 4 technical replicates using tissue pooled from three plates, each with 16–18 seedlings [B]). Distinct letters indicate significant differences (P < 0.05) according to a Bonferroni-corrected Student’s t test.

The Systemic Zn Deficiency Response

To define the systemically Zn deficiency-responsive transcriptome of roots, we employed hybridization of ATH1 microarrays to identify differences between the hma2 hma4 double mutant and the wild type under control conditions, and we compared these with differences between Zn-deficient and control wild-type plants (Supplemental Data Sets 1 and 2). In more detail, we first identified transcripts that differed in abundance between roots of hma2 hma4 plants and the wild type cultivated under control conditions (Supplemental Figure 1A). Among these, we expected root transcripts that respond systemically to the deficient Zn status of shoots, as well as transcripts responding to the locally elevated Zn status in roots characteristic of the double mutant (Supplemental Figure 1B, left and right, and 1D, left circle). To identify the subset of root transcripts that respond systemically to the Zn status of the shoot, we intersected the former list of transcripts with the transcripts that differed in abundance between roots of Zn-deficient and Zn-sufficient wild-type plants (Supplemental Figure 1B, left and center, and Figure 1D, right circle). Among the probe sets showing differential hybridization signals in these two comparisons (corresponding to 646 and 528 genes, respectively), 54 genes were identified in both comparisons. Out of these, 25 genes showed codirectional changes indicative of systemic regulation (Figure 1D, Table 1). Root transcript abundances of these 25 genes were thus responsive to Zn deficiency in shoots and not to the local Zn status in the root. Consequently, these transcripts represent candidate targets in the root of a shoot-derived long distance signal. Of these, 19 genes exhibited higher hybridization signal intensities in response to Zn deficiency (upregulation) and six exhibited lower signal intensities (downregulation). Genes of unknown functions (11) predominated among the putative targets of systemic regulation, which also included several genes with functions in environmental responses, such as protein turnover (ARABIDOPSIS S-PHASE KINASE ASSOCIATED PROTEIN1 LIKE19 [ASK19], AT4G39753), signaling (STRUBBELIG RECEPTOR FAMILY2 [SRF2], AT5G06820, and MILDEW RESISTANCE LOCUS O 9 [MLO9], AT1G42560), nutrient storage (SEED STORAGE ALBUMIN1 [SESA1], AT1G03890) and metal transport (METAL TRANSPORT/TOLERANCE PROTEIN2 [MTP2], AT3G61940). An analogous comparison was performed in shoots (Supplemental Data Sets 3 and 4 and Supplemental Table 1). The resulting group of transcripts comprises the local Zn deficiency response of shoots, for example, ZIP4, ZIP5, ZIP9, IRT3, and NAS3, including candidates that may participate in the generation of a long-distance signal, for example, PHLOEM PROTEIN 2-B15 (AT1G09155), SERINE CARBOXYPEPTIDASE-LIKE genes (SCPL15, AT3G12240, and SCPL22, AT2G24000), and genes encoding kinases, phosphatases, and transcription factors.

Table 1. Fold Changes of 25 Transcripts That Are Candidates for Systemic Regulation in Roots, as Identified in Microarray Analyses in This Study.

Probe Set ID	hma2 hma4 vs. WT Fold Changea	WT −Zn vs. WT Control Fold Change	Locus	Gene Name	Annotation	Putative Function
259556_at	23	3.2	AT1G21300	–	Transposable element gene	Not assigned
250904_at	9.4	2.4	AT5G03620	–	Subtilisin-like serine endopeptidase	Proteolytic activity
255239_at	7.1	4.8	AT4G05580	–	Transposable element gene	Not assigned
265094_at	6.7	3.5	AT1G03890	–	RmlC-like cupin	Nutrient reservoir
245779_at	6.3	8.3	AT1G73510	–	Unknown protein	Not assigned
251294_at	6.2	1283	AT3G61940	MTP2	Efflux transmembrane transporter/ zinc ion transmembrane transporter	Zinc ion transport
258301_at	5.8	13.0	AT3G30510	–	Transposable element gene	Not assigned
256586_at	5.8	7.6	AT3G28770	–	Unknown protein	Not assigned
257458_at	5.7	4.9	AT2G05400	–	Meprin and TRAF homology domain-containing protein	Ubiquitin-specific proteolysis
267334_at	5.7	5.3	AT2G19420	–	Unknown protein	Not assigned
256544_at	5.6	4.0	AT1G42560	MLO9	Mildew resistance locus O 9	Calmodulin binding in pathogen response
252865_at	5.5	6.1	AT4G39753	–	Galactose oxidase/kelch repeat superfamily protein	Protein regulation via E3 ubiquitin ligase complex
265199_s_at	5.4	4.6	AT2G36780	–	UDP-glucoronosyl/UDP-glucosyl transferase family protein	Transfer of glucose to flavonol, required for anthocyanin biosynthesis
266726_at	4.7	3.9	AT2G03160	ASK19	Ubiquitin-protein ligase	Transfer of ubiquitin to proteins targeted for degradation
263645_at	4.4	4.7	AT2G04720	–	Pseudogene, GTP-binding protein-related, similar to GTP-binding protein	Not assigned
245890_at	3.9	3.3	AT5G09490	RPS15B	Ribosomal protein	Ribosome component
260188_at	3.1	5.0	AT1G35995	–	Transposable element gene	Not assigned
253894_at	2.4	4.4	AT4G27150	NWMU2-2S	2S seed storage protein 2, 2S albumin storage protein 2	Nutrient reservoir
259332_at	2.2	7.1	AT3G03830	–	SAUR-like auxin-responsive protein family	Developmental responses
250699_at	0.1	0.2	AT5G06820	SRF2	Strubbelig receptor family 2	transmembrane receptor, protein tyrosine kinase signaling pathway
263674_at	0.1	0.3	AT2G04790	–	Unknown protein	Not assigned
263905_at	0.2	0.3	AT2G36190	ATCWINV4	Hydrolase, hydrolyzing O-glycosyl compounds	Involved in sugar catabolism
255609_s_at	0.2	0.2	AT4G01180	–	XH/XS-domain-containing protein	Not assigned
248812_at	0.2	0.2	AT5G47330	–	Palmitoyl protein thioesterase family protein	Involved in lipid catabolism
255970_s_at	0.2	0.1	AT3G31540	–	Transposable element gene

Shown are means of two independent replicate experiments. Entities were considered regulated if log₂(fold change) > 1 or < −1 in both experiments and significant using a Mann-Whitney unpaired, corrected P value < 0.2. –, Not assigned.

^aGrown in control conditions.

Root MTP2 Transcript Levels Depend on Shoot Zn Status

Our transcriptomics suggested that MTP2 (AT3G61940) is among the quantitatively most highly regulated candidate targets of systemic regulation in roots. MTP2 transcript levels were ∼6-fold upregulated in hma2 hma4 roots relative to wild-type roots on agar-solidified control medium (Table 1). MTP2 transcript levels were ∼1300-fold upregulated in roots of 8-week-old wild-type plants grown in Zn-deficient relative to plants cultivated in Zn-sufficient control hydroponic media (Table 1). The MTP2 protein is a member of the so-called Cation Diffusion Facilitator (CDF) family of membrane transporters believed to operate as metal-cation antiporters, in the subgroup associated with Zn²⁺ or Co²⁺ transport (Montanini et al., 2007; Ricachenevsky et al., 2013).

We confirmed the regulation of MTP2 observed in the microarray experiments by RT-qPCR in an independent experiment (Figure 2A). MTP2 transcript abundance was highest in roots of hma2 hma4 grown on Zn-deficient medium and ∼10-fold lower in wild-type seedlings grown under the same conditions. The MTP2 transcript was undetectable in roots of wild-type seedlings grown on control medium. Importantly, by comparison, MTP2 transcript levels were clearly elevated in hma2 hma4 seedlings grown under control conditions (Figure 2A). This result supported our hypothesis that MTP2 transcript levels in the root respond to the physiological Zn status of the shoot and not the Zn status of the root (Figure 1B). We determined the global expression pattern of MTP2 by RT-qPCR in different tissues of mature 10-week-old plants grown hydroponically in control and Zn-deficient medium. The transcript was detectable exclusively in roots and siliques of Zn-deficient plants (Supplemental Figure 1C). Root MTP2 transcript levels were more than 2-fold upregulated upon 1 week exposure to elevated concentrations of Co and Cd, but less in response to excess Mn, Fe, Zn, or Mn deficiency (Supplemental Figure 1D).

Figure 2.

Root MTP2 Transcript Abundance Responds to Shoot Zn Status and Is Regulated at the Level of Promoter Activity.

(A) MTP2 transcript levels in roots of 19-d-old wild-type and hma2 hma4 seedlings cultivated on control and Zn-deficient (−Zn) agar-solidified media (n.d., not detected).

(B) GUS activity in wild-type and hma2 hma4 ProMTP2:GUS seedlings grown as in (A). Arrows indicate areas of GUS activity. Twenty independent ProMTP2:GUS transgenic lines were examined in the T2 generation. In the T3 generation, two representative homozygous lines were crossed into the hma2 hma4 background, and representative seedlings homozygous for the ProMTP2:GUS transgene and hma2 and hma4 mutations are shown.

(C) ZIP4 and MTP2 transcript levels in roots of 8-week-old hydroponically cultivated wild-type plants. Plants were transferred to Zn-sufficient (5 µM; Suf.) or Zn-deficient (0 µM added Zn; −Zn) hydroponic solutions for 3 weeks, or Zn (5 µM) was resupplied to −Zn-grown plants for the indicated periods of time, before harvest.

(D) ZIP4 and MTP2 transcript levels in roots of wild-type plants. Plants were cultivated in Zn-deficient hydroponic solutions for 4 weeks, followed by Zn resupply to rosettes as a foliar spray every 2 d for 14 d (Control: 0.01% [v/v] Triton X-100; Spray: 1 mM ZnSO₄ in the same solution).

Bar graphs show arithmetic means ± sd (n = 4 technical replicates using tissue pooled from 3 replicate plates consisting of 16–18 seedlings each [A]; n = 4 technical replicates using cDNA pooled from 6 replicate hydroponic growth vessels, each containing 3 plants [C]; n = 5 technical replicates using tissue pooled from 13 replicate hydroponic vessels, each containing 2 plants each [D]). Asterisk indicates statistically significant difference between control and spray (P = 0.0013; Student’s t test). NS, not significant.

To examine MTP2 expression patterns in planta, we generated MTP2 promoter-GUS (ProMTP2:GUS) lines in a wild-type genetic background. Histochemical GUS staining was detectable upon cultivation on Zn-deficient medium in both a small segment above the meristematic zone of root tips and a more extensive segment in the maturation/root hair zone of emerging lateral roots, but no staining was visible in 19-d-old wild-type seedlings cultivated on control medium (Figures 2B; Supplemental Figures 1E to 1H). We additionally introduced ProMTP2:GUS into the hma2 hma4 genetic background by crossing. GUS staining was observed in the elongation zone of lateral roots of hma2 hma4 seedlings grown on both control and Zn-deficient medium (Figure 2B). This confirmed the systemic regulation of MTP2 transcript levels in roots, and it suggested that this regulation occurs at the level of MTP2 promoter activity.

To investigate the dynamics of regulation of MTP2, we performed a Zn resupply experiment in hydroponics. Total RNA was extracted from roots after 3 weeks of cultivation in Zn-deficient hydroponic medium followed by transfer into Zn-sufficient control medium for 0, 2, 8, and 24 h, as well as from plants grown in Zn-sufficient control medium throughout their entire cultivation period. RT-qPCR demonstrated that root transcript levels of ZIP4 of the locally regulated Zn deficiency response (Figure 1B) decreased to 25% of the maximum within 2 h of resupply and reached steady state levels of Zn-sufficient plants ∼8 h after resupply (Figure 2C). MTP2 transcript abundance, however, remained unchanged on average after 2 h, was at 28% after 8 h, and returned to steady state levels present in Zn-sufficient plants only 24 h after resupply (Figure 2C). Upon Zn resupply, the delayed decrease in transcript abundance of MTP2 by comparison to ZIP4 further supports the premise that transcript abundance of MTP2 is regulated in a different manner than previously identified locally regulated Zn deficiency-responsive transcripts.

To further test the hypothesis that MTP2 regulation is controlled by a shoot-derived signal, we administered Zn re-supply exclusively to rosette tissues as a foliar spray. Wild-type plants grown in Zn-deficient hydroponic medium for 4 weeks were sprayed with a solution containing 1 mM ZnSO₄ or a control solution without added Zn every 2 d for 2 weeks. There was no significant difference in root ZIP4 transcript levels between Zn-sprayed and mock solution-sprayed plants, whereas root MTP2 transcript levels of Zn-sprayed plants were 41% lower than those of mock solution-sprayed plants (Figure 2D). This finding supported the systemic regulation of root transcript levels of MTP2 bidirectionally in dependence on the physiological Zn status of the shoot. Expression levels of Zn deficiency-responsive genes were overall lower in this experiment (Figure 2D), in which we cultivated plants in 0.25× Hoagland solutions, than in the Zn resupply experiment, during which we cultivated plants in 1× Hoagland solution (Figure 2C).

Enhanced Sensitivity to Zn Deficiency in mtp2 Mutants

To assess the contribution of MTP2 to the Zn deficiency response of Arabidopsis, a T-DNA knockout in the MTP2 locus was obtained in which MTP2 transcript levels were undetectable (Supplemental Figures 2A and 2B). Nineteen-day-old seedlings cultivated on Zn-deficient medium produced less fresh biomass than on control medium (Figures 3A to 3C). The reduction in biomass production on Zn-deficient medium was more severe in mtp2 seedlings than in the wild type. Zn deficiency induced an increase in primary root length in both genotypes, but this increase was significantly attenuated in mtp2 when compared with the wild type (Figure 3D). We also observed that the difference between the wild type and mtp2 grown on Zn-deficient medium increased over time (Supplemental Figure 2F). Finally, on Zn-deficient medium, mtp2 accumulated significantly less Zn than the wild type in shoots, whereas no difference was observed between the cultivated genotypes on control medium (Supplemental Figures 2C and 2D). In order to account for the lower uptake capacity of the smaller root surface area in mtp2 mutants, we normalized total shoot Zn content to root biomass. Even then, the resulting shoot Zn partitioning was significantly lower in mtp2 than in the wild type upon growth on Zn-deficient medium (Figure 3E), in accordance with shoot:root Zn concentration ratios (Supplemental Figure 2E).

Figure 3.

Growth of mtp2 Seedlings Is Impaired and Zn Partitioning Altered on Zn-Deficient Medium.

(A) and (B) Images of representative 19-d-old wild-type (A) and mtp2 (B) seedlings on control (left) or Zn-deficient (−Zn) (right) agar-solidified media.

(C) and (D) Shoot fresh biomass (C) and root length (D) of wild-type and mtp2 seedlings as shown in (A) and (B).

(E) Shoot Zn partitioning in wild-type and mtp2 seedlings as shown in (A) and (B).

Bar graphs show arithmetic means ± sd (n = 17 seedlings [C] and [D]; n = 4 independent plates, from each of which tissues were pooled from 16 to 18 seedlings [E]). Distinct letters indicate significant differences (P < 0.05) according to a Bonferroni-corrected Student’s t test. Bars = 20 mm in (A) and (B).

To determine the effect of an extended period of Zn deficiency on the mtp2 mutant, plants were grown in control and Zn-deficient hydroponic medium in short days and then harvested at 7 weeks of age (Figures 4A and 4B). Upon cultivation in a Zn-deficient hydroponic solution, the mtp2 mutant showed a significant decrease in shoot chlorophyll content compared with the wild type (Figure 4C). Shoot and root dry biomass were decreased in both the wild-type and mtp2 cultivated under Zn deficiency compared with control conditions. However, the biomass decrease in mtp2 was not statistically significant and proportionally lower than in the wild type (Figures 4D and 4E). Concentrations of Ca, Cu, Fe, and Mn were independent of MTP2 function in both roots and shoots (Supplemental Figures 3A to 3D). Zn concentrations were ∼50% lower in shoots of mtp2 than in shoots of the wild type cultivated in Zn-deficient hydroponic medium (Supplemental Figure 3E). Root Zn concentrations were not significantly different between genotypes upon cultivation in Zn-deficient or -sufficient hydroponic solutions (Supplemental Figure 3F). Shoot-to-root ratios of Zn concentrations and shoot Zn partitioning were reduced by ∼50% in mtp2 relative to wild-type plants cultivated in Zn-deficient hydroponic medium (Figures 4F; Supplemental Figure 3G). Taken together, these data strongly suggest that mtp2 is impaired in root-to-shoot translocation specifically of Zn under Zn-deficient growth conditions, resulting in impaired shoot growth or leaf chlorophyll content dependent on plant size at the onset of Zn deficiency.

Figure 4.

Reduced Shoot Zn in mtp2 upon Long-Term Exposure to Zn Deficiency in Hydroponics.

(A) and (B) Photographs of representative rosettes of 7-week-old hydroponically cultivated wild-type (A) and mtp2 (B) plants cultivated in control or Zn-deficient (−Zn) conditions.

(C) to (F) Leaf chlorophyll concentrations (C) measured in a between 50 and 100% fully expanded leaf from each plant, shoot dry biomass (D), root dry biomass (E), and shoot Zn partitioning (F) in plants grown as in (A) and (B).

Bar graphs show arithmetic means ± sd (n = 16 individuals [C], [D], and [F]; n = 7 pools of roots of 2 to 3 plants per pool [E]). Distinct letters indicate significant differences (P < 0.05) according to a Bonferroni-corrected Student’s t test. Bars = 10 mm in (A) and (B).

Physiological Role of MTP2 in Zn Deficiency

Previously, other members of the same subgroup of paralogous CDF family transporters, MTP1 and MTP3, were proposed to act in the transport of Zn into the vacuole, thus contributing to basal Zn tolerance of Arabidopsis (Clemens, 2001; Desbrosses-Fonrouge et al., 2005; Arrivault et al., 2006). MTP1 and MTP3 cDNAs were able to complement Zn-hypersensitive yeast mutants (Desbrosses-Fonrouge et al., 2005; Arrivault et al., 2006; Kawachi et al., 2008). The MTP1 protein was demonstrated to act as a Zn²⁺/H⁺ antiporter in the vacuolar membrane (Kawachi et al., 2008). To assess whether MTP2 may act as a Zn transporter in vivo, the MTP2 cDNA was cloned into the yeast expression plasmid pFL61, and this construct was used to transform the Zn/Co-sensitive yeast mutant zrc1 cot1, the Cd-sensitive mutant ycf1, and the Fe-sensitive mutant ccc1. MTP2 partially complemented Zn hypersensitivity of the zrc1 cot1 mutant and also conferred slight Co tolerance (Figure 5). Expression of MTP2:GFP in the zrc1 mutant suggested the localization of MTP2 to an intracellular membrane (Supplemental Figure 4A). In yeast, MTP2 is thus likely to act as a metal transporter that complements genetic defects in the sequestration of excess cytoplasmic metal into a subcellular compartment. Considering the substantially lower concentrations of Co relative to Zn in planta, MTP2 is likely to act as a Zn transporter in Arabidopsis.

Figure 5.

Heterologous Expression of MTP2 in Yeast Mutants.

Heterologous expression of the MTP2 cDNA can rescue the Zn-sensitive yeast mutant zrc1 cot1, but not the Cd-sensitive ycf1 or the Fe-sensitive ccc1 mutants. Serial 1:10 dilutions are shown from left to right for S. cerevisiae wild type (BY4741), mutant transformed with empty vector pFL61 (ev), and mutant transformed with MTP2.

To determine the subcellular localization of MTP2 in planta, the MTP2 cDNA was translationally fused to GFP at either its N or C terminus and expressed under the control of the native MTP2 promoter (ProMTP2:GFP:MTP2 and ProMTP2:MTP2:GFP). Confocal laser scanning microscopy was used to investigate MTP2 localization in roots of homozygous T3 lines grown on Zn-deficient medium for 14 d. GFP fluorescence was restricted to intracellular membranes for both N- and C-terminal GFP fusions of MTP2. To identify the intracellular compartment giving rise to GFP fluorescence, we crossed one representative transgenic line of each the N- and C-terminal GFP fusion to MTP2 with each of a series of 21 transgenic lines containing proteins of known localization fused to mCherry (Geldner et al., 2009). We found colocalization exclusively with Pro35S:NIP1;1:mCherry (Figures 6A and 6B). NIP1;1 is an aquaporin that localizes to the ER membrane and to the plasma membrane upon prolonged immersion in water (Boursiac et al., 2005; Geldner et al., 2009). These data suggested that MTP2 is an ER membrane-localized protein in planta.

Figure 6.

Subcellular Localization of MTP2.

Roots of 14-d-old wild-type seedlings homozygous for both ProMTP2:MTP2cDNA:GFP and ER membrane marker Pro35S:NIP1;1cDNA:mCherry transgenes (A) or ProMTP2:GFP:MTP2cDNA and Pro35S:NIP1;1cDNA:mCherry transgenes (B), showing from left to right, GFP fluorescence, mCherry, and merged images. All seedlings were grown on Zn-deficient agar-solidified medium. Bars = 25 µm.

Importantly, the same MTP2-GFP fusion lines as used for subcellular localization were able to complement the fresh biomass production and root elongation phenotypes of the mtp2 mutant (Supplemental Figures 4B to 4D). This indicated that the phenotypic alterations in this mutant were indeed caused by a lack of MTP2 protein. Furthermore, these results supported the subcellular localization of the MTP2 protein in the ER membrane of Arabidopsis. Taken together, our data suggested that MTP2 acts in the transport of Zn between the cytoplasm into the ER lumen and that it is required for the full extent of acclimation to Zn deficiency in Arabidopsis.

Mutant Combinations Show That HMA2 and MTP2 Act Together in Zn Deficiency

Transcript levels of several ZIP and NAS genes of the Arabidopsis Zn-deficiency regulon were shown to be controlled by the transcription factors bZIP19 and bZIP23 (Assunção et al., 2010). We show here that ZIP4 and ZIP9 of this regulon respond to local Zn status (Figure 1B) (Talke et al., 2006). Besides known targets of bZIP19/bZIP23, transcript abundance of additional genes respond to Zn deficiency (Wintz et al., 2003), for example, HMA2 and MTP2 (van de Mortel et al., 2006) (Supplemental Data Set 1). Root MTP2 transcript levels responded systemically to shoot Zn status (Figure 2), and the promoters of MTP2 and HMA2 lack a Zn deficiency response element required for regulation by bZIP19 and bZIP23 (Assunção et al., 2010).

Our transcriptomic approach for the identification of targets of systemic regulation did not allow the detection of the HMA2 transcript as such a target because the hma2 hma4 double mutant does not produce any HMA2 transcript (Hussain et al., 2004). We thus tested for the systemic regulation of HMA2 in the same experiment in which we demonstrated the systemic regulation of MTP2 transcript levels in roots (Figure 2D). We observed that foliar Zn spraying significantly reduced HMA2 transcript levels in the roots of wild-type plants grown in a Zn-deficient hydroponic solution by ∼70%, similar to MTP2. This was in contrast to root HMA4 transcript levels, which were unaffected by foliar Zn application (Supplemental Figure 5).

To address the relationship between the roles of MTP2 and HMA2 in root-to-shoot Zn partitioning, we generated an mtp2 hma2 double mutant. On control medium, there were no statistically significant differences in shoot Zn partitioning between genotypes (Figure 7A). On Zn-deficient agar-solidified medium, however, shoot Zn partitioning of both mtp2 and hma2 single mutants was reduced, and the reduction was even more severe in the mtp2 hma2 double mutant (Figure 7B). This suggests that HMA2 and MTP2 have additive effects on shoot Zn partitioning under Zn-deficient growth conditions. Moreover, under Zn-sufficient growth conditions, we found no evidence for a contribution of either of these genes to shoot Zn partitioning.

Figure 7.

Genetic Interaction between mtp2 and hma2.

Shoot Zn concentrations in 19-d-old wild-type, mtp2, hma2, and mtp2 hma2 seedlings grown on control (A) and Zn-deficient medium (B). Bar graphs show arithmetic means ± sd (n = 4 plates, each with 16 to 18 seedlings). Distinct letters indicate significant differences (P < 0.05) according to a Bonferroni-corrected Student’s t test.

To assess the long-term effects of Zn deficiency on older plants, shoot Zn accumulation was measured in 7-week-old plants grown in control and Zn-deficient hydroponic solutions. Upon cultivation in control solutions, neither of the two single mutants showed a difference from the wild type in shoot Zn partitioning (Supplemental Figure 5B). However, there was a decrease in shoot Zn partitioning in the mtp2 hma2 double mutant in Zn-sufficient control conditions. We attribute this to a slight Zn depletion in hydroponic culture when plants are larger. When cultivated in Zn-deficient hydroponic solutions, shoot Zn partitioning was reduced in both single mutants and even more so in the double mutant (Supplemental Figures 5B and 5C). These data indicate that both MTP2 and HMA2 are required for the acclimation of Arabidopsis seedlings as well as adult plants to Zn deficiency and that the functions of these genes in root-to-shoot Zn translocation are additive.

DISCUSSION

Local versus Systemic Regulation of the Zn Deficiency Response

The defect in root-to-shoot Zn translocation in the hma2 hma4 double mutant provides a unique tool for dissecting Zn homeostasis in plants, enabling us to separate locally regulated from systemically regulated Zn deficiency responses (Figures 1; Supplemental Figure 1). Employing this system, in combination with Zn resupply to deficient plants in hydroponic solution and as a foliar spray, we identify a systemically regulated transcriptional Zn deficiency response in roots, which includes both MTP2 and HMA2 (Table 1, Figures 1 and 2; Supplemental Figure 5). This systemic Zn deficiency response is controlled by the nutritional Zn status of the shoot. Published HMA2 promoter-reporter studies were conducted in the hma2 hma4 double mutant background (Hussain et al., 2004; Sinclair et al., 2007), in which shoots were physiologically Zn deficient so that systemic Zn deficiency responses were induced in roots under standard growth conditions (Figure 1). This explains why the authors were able to detect HMA2 promoter activity. We show that the known bZIP19/bZIP23 transcription factor target genes ZIP4 and ZIP9 (Assunção et al., 2010) exemplify local transcriptional Zn deficiency responses of both roots and shoots (Figures 1 and 2). The sensor proteins and molecular mechanisms through which plants perceive Zn deficiency remain unknown. It was hypothesized that the bZIP19 and bZIP23 transcription factor proteins, which activate locally regulated transcriptional Zn deficiency responses, may directly function as Zn²⁺ binding sensors (Assunção et al., 2013).

The IRT3, ZIP1, ZIP4, and NAS2 genes of Arabidopsis are targets of the bZIP19 and bZIP23 transcription factors and are transcriptionally activated as components of the local Zn-deficiency response. ZIP4 and IRT3 were proposed to act in cellular Zn uptake across the plasma membrane and ZIP1 in the remobilization of stored Zn from the vacuole into the cytoplasm (Grotz et al., 1998; Lin et al., 2009; Milner et al., 2013; Zhang et al., 2014). Transient heterologous expression of chimeric GFP fusion proteins in tobacco (Nicotiana tabacum) BY-2 cells suggested a cytoplasmic localization of Arabidopsis NAS proteins (Nozoye et al., 2014). NAS proteins function in the biosynthesis of nicotianamine that was proposed to facilitate the intra- and intercellular mobility of Zn in the root symplasm (Deinlein et al., 2012) and vacuolar Zn storage (Haydon et al., 2012). Root tips carry the predominant sites of Zn uptake in Arabidopsis. Tips of growing roots explore the soil solution and encounter soil microenvironments of low or high Zn availability. Local regulation of gene expression in roots is presumably important to allow rapid responses in this dynamic environment and to prevent the local accumulation of excessive, toxic levels of Zn in roots. In shoots, local regulation allows each individual cell, or multicellular symplastic unit, to meet its nutritional needs.

The transcriptional upregulation of root Fe deficiency responses, comprising in particular root Fe uptake systems, is systemically controlled by the physiological Fe status of the shoot (Vert et al., 2003; Enomoto et al., 2007). It was suggested that repartitioning to roots of shoot-derived Fe, mediated by OLIGOPEPTIDE TRANSPORTER3, could act as a shoot-derived signal that suppresses root Fe deficiency responses in Fe-sufficient plants (Mendoza-Cózatl et al., 2014; Zhai et al., 2014). More recently, a shoot-derived signal dependent on the YELLOW-STRIPE-LIKE Fe-nicotianamine transporters YSL1 and YSL3 was proposed to mediate the transcriptional upregulation of root Fe-deficiency responses in Fe-deficient plants (Kumar et al., 2017). The molecular nature of the involved signal is unknown.

In addition to the known systemic regulation of plant iron deficiency responses, some macronutrient deficiencies are also regulated in a systemic manner. These include the communication of phosphate status via the phloem-mobile miR399 that is generated in shoots in response to P_i deficiency and targets PHOSPHATE2 (PHO2) transcripts for degradation in roots (Lin et al., 2008; Pant et al., 2008; Zhang et al., 2014). This results in a dramatic reduction in the abundance of PHO2 protein, an E2 ubiquitin conjugase that marks several transmembrane phosphate transporter proteins for ubiquitination and turnover (Fujii et al., 2005; Bari et al., 2006; Chiou et al., 2006; Zhang et al., 2014). Thus, both processes of phosphate uptake into root cells via PHOSPHATE TRANSPORTER1 family proteins and the export of phosphate from the root symplasm into the xylem via the phosphate transporter PHO1 protein are enhanced under P_i deficiency (Liu et al., 2014).

The Function of MTP2 in Zn-Deficient Plants

The phenotypic characterization of an mtp2 mutant, and its complementation by an intact MTP2 transgene, suggests a role for MTP2 in root-to-shoot Zn partitioning (Figures 3 and 4; Supplemental Figures 2 to 4). In the context of the role of the systemically regulated MTP2 gene in root-to-shoot Zn translocation, the localization of MTP2 in the ER membrane is unexpected (Figure 6; Supplemental Figure 4). The general direction of metal transport by members of this protein family is out of the cytosol (MacDiarmid et al., 2002; Blaudez et al., 2003; Desbrosses-Fonrouge et al., 2005; Montanini et al., 2007; Kawachi et al., 2008; Podar et al., 2012). Consequently, MTP2 is predicted to transport Zn²⁺ into the ER lumen under Zn deficiency, in accordance with the yeast complementation and localization data shown here (Figure 5; Supplemental Figure 4A). In the same protein family, Schizosaccharomyces pombe Zhf1 and the Saccharomyces cerevisiae heteromeric complex of ScMSc2 and ScZrg17 were shown to transport of Zn²⁺ into the ER (Clemens et al., 2002; Ellis et al., 2004). Similar to ScMSc2 and ScZrg17, MTP2 may be required to ensure the delivery of Zn to Zn-dependent metalloproteins in the ER lumen under Zn limitation (Ellis et al., 2005). However, the highly localized expression pattern of MTP2 in roots (Figure 2; Supplemental Figure 1) is in disagreement with this hypothesis, given that Zn metalloproteins are required in all cells and not just those of emerging lateral roots. Our working model for the role of MTP2 is thus the transport of Zn into the ER for facilitating Zn movement across root cell layers through desmotubules. Desmotubules traverse through the plasmodesmata of plant cells and comprise the portion of the ER that is contiguous between two adjacent cells (Overall et al., 1982; Roberts and Oparka, 2003). Cell-to-cell movement of biomolecules through plasmodesmata has been demonstrated for transcription factors, metabolites, and small RNAs (Maule, 2008). Cell-to-cell movement of a 10-kD fluorescent molecule was reported to be more efficient via the desmotubule than via the cytoplasm of plasmodesmata (Barton et al., 2011). Thus, the expression of MTP2 in roots of Zn-deficient plants may allow a more efficient symplastic movement of Zn inside the lumen of the ER from the epidermis through the internal cell layers of the root toward the xylem (Figure 8). Cell-to-cell movement inside the ER can circumvent a possible scavenging or utilization of Zn by cytoplasmic Zn-dependent metalloproteins of root cells along the symplastic radial pathway of Zn toward the stele. Such prioritization would help to maintain essential Zn-dependent functions in the shoot, for example, in protein biosynthesis, photosystem II repair, and chloroplast protein import, during episodes of extended Zn deficiency (Krämer and Clemens, 2005). Although we consider it to be less likely, an opposite direction of Zn²⁺ transport by MTP2 is possible in principle, or even alternate transport directions depending on the electrochemical gradients across the ER membrane. This possibility awaits future direct experimental testing with more sensitive Zn fluorophores than are available to date.

Figure 8.

Working Model.

Zn deficiency in shoots results in the systemic upregulation of both MTP2 and HMA2 transcript abundance in roots. This leads to enhanced root-to-shoot Zn translocation. HMA2 was previously shown to mediate Zn export from the root symplasm into the apoplastic xylem. In the early differentiation zone of lateral roots, MTP2-mediated transport of Zn²⁺ into the ER allows cell-to-cell movement of Zn through desmotubules into the contiguous ER lumen of the neighboring cell layer via plasmodesmata. This enhances the passage of Zn from the outer epidermal cell layer toward the xylem and consequently to the shoot. Ep, epidermis; C, cortex; En, endodermis; P, pericycle; Xy, xylem.

Interaction of MTP2 with HMA2 Encoding a Known Transporter in Root-to-Shoot Zn Translocation

AtHMA2 transcript levels are shown here to be upregulated as part of the systemically regulated Zn deficiency response (Supplemental Figure 5). AtHMA2 was reported to localize to the plasma membrane of root pericycle cells and to function in the export of Zn from the pericycle into the xylem for mass flow-driven Zn translocation to the shoot (Sinclair et al., 2007). The physiological phenotype of the hma2 single mutant, namely, a reduction in root-to-shoot Zn partitioning in Zn-deficient medium, is similar to that of the mtp2 mutant (Figures 7; Supplemental Figure 5). This was not observed in standard medium, in line with previously published results, and with strongly increased expression of both the MTP2 and HMA2 genes under Zn deficiency (van de Mortel et al., 2006) (Supplemental Data Set 1 and Supplemental Figure 1).

Regulation and roles of MTP2 and HMA2 suggest that the systemic regulation of root Zn deficiency responses dependent on shoot physiological Zn status functions to mobilize an adequate proportion of Zn from the root to the shoot. This is in line with our observation that Zn-deficient wild-type plants allocate a larger proportion of total plant Zn to their shoots than Zn-sufficient plants (Figures 3 and 4) (Talke et al., 2006). We propose that MTP2 mediates an enhanced intra-ER “symplastic” movement of Zn from outer root cell layers to inner cell layers, whereas HMA2 is known to function in enhancing xylem loading of Zn. MTP2 is expressed in lateral roots only and in a more distal zone of the root tip compared with HMA2, whereby the zone of expression of MTP2 overlaps partially with that of HMA2 (Supplemental Figure 1; Figures 2 and 8) (Hussain et al., 2004). The additive effects observed of both mutations in the mtp2 hma2 double mutant are consistent with contributions of the two genes to distinct pathways of root-to-shoot Zn movement or with both genes having additive effects on Zn flux through a common pathway. As demonstrated previously, enhanced root-to-shoot translocation of Zn can activate local Zn-deficiency responses in roots (Hanikenne et al., 2008). Consequently, the upregulation of MTP2 and HMA2 expression in roots triggered by the shoot-derived Zn deficiency signal can further enhance local transcriptional Zn deficiency responses of roots controlled by bZIP19/bZIP23.

Outlook

In the acclimation of plants to iron deficiency, a shoot-derived signal and a proposed, poorly characterized local signal act on an overlapping set of genes (Vert et al., 2003; Kumar et al., 2017). By contrast, our results suggest that Zn-deficient plants activate two different regulatory pathways mediating distinct local and systemic Zn deficiency responses, respectively (Figure 1). The systemic signal is either generated (Zn deficiency signal) or eliminated (Zn sufficiency signal) in physiologically Zn-deficient shoots. Candidate genes for an involvement in signal generation or propagation are transcripts coregulated in abundance in shoots of the wild-type cultivated in Zn-deficient medium and in shoots of hma2 hma4 cultivated in Zn-sufficient medium, when compared with shoots of the wild type grown in Zn-sufficient medium (Supplemental Table 1). Future work will address, for example, whether specific candidate genes, such as PHLOEM PROTEIN 2-B15, NAS3, and CPK19, are involved. This group of genes evidently also includes well-known locally regulated Zn deficiency-responsive genes, for example, ZIP4, IRT3, and ZIP5 (Supplemental Table 1).

Both HMA4 and HMA2 have additional roles in reproductive development by mediating the export of Zn from the maternal seed coat to supply Zn to the developing filial tissues of seeds, with an apparently predominant role for HMA4 (Olsen et al., 2016). In analogy with the results of this study, the function of HMA2 may be more prevalent under physiological Zn deficiency. In addition to roots, MTP2 transcript was detectable only in siliques of Zn-deficient plants (Supplemental Figure 1). The biological implications of this, also in relation to HMA2, deserve further study.

In summary, we report here a number of candidate target genes of Zn deficiency-responsive transcriptional regulation that operates in roots but is controlled systemically by the physiological Zn status of the shoot. Moreover, we show here that the previously characterized, bZIP19/bZIP23-dependent Zn deficiency responses are governed by local physiological Zn status in roots and shoots, respectively. We identify MTP2 and HMA2 as two key targets of shoot-to-root communication of Zn status. We demonstrate an additive function of these genes in enhancing root-to-shoot Zn partitioning. Based on the cell-type-specific expression of MTP2 and HMA2, the subcellular localization and transport function of MTP2, and existing knowledge on HMA2, we propose a working model for how these proteins enhance root-to-shoot Zn flux under conditions of Zn deficiency. Further research in crop plants and under field conditions will be required in order to address whether the common practice of foliar Zn application adversely affects crop yield and quality on Zn-deficient soils.

METHODS

Plant Materials and Growth Conditions

Col-0 seeds were obtained from the NASC. The hma2, hma4, and hma2 hma4 lines were described previously (Hussain et al., 2004), and the mtp2 mutant corresponds to SALK_003649. Pro35S:NIP1;1:mCherry seeds were obtained from NASC in the Wave Line set (Geldner et al., 2009).

For sterile growth, seeds were surface sterilized with 2% (w/v) NaOCl and 0.01% (v/v) Triton X-100 for 5 min before washing three times with ultrapure water (Milli-Q; Merck). Seedlings were grown on vertically positioned 120-mm square polystyrene Petri dishes (Greiner Bio-One), each containing 50 mL agar-solidified modified Hoagland solution, in 16-h-light (22°C, 120 µmol m⁻² s⁻¹ white light) and 8-h-dark (18°C) cycles in a growth cabinet (Percival CU-41L4; CLF Climatics). In all experiments that included a Zn deficiency (−Zn) treatment, contaminant Zn was removed from agar (Agar Type M; Sigma-Aldrich) by washing three times with 10 mM EDTA for 12 to 16 h, then six times with ultrapure water, all at 1% (w/v) agar. Final agar-solidified media contained 1% (w/v) sucrose and 0.8% (w/v) agar in modified Hoagland solution composed as described (Bernal et al., 2012), either with no added Zn (Zn-deficient medium) or with 5 µM ZnSO₄ (control medium). Sterilized seeds (∼20 per plate) were plated directly on agar-solidified treatment media and stratified at 4°C for 48 h, unless stated otherwise.

For the experiments shown in Figures 1A to 1C and Supplemental Figure 1B, seeds were germinated on agar-solidified 0.5× MS medium (containing 15 µM Zn) supplemented with 1% (w/v) sucrose for 7 d and subsequently transferred to agar-solidified treatment media as described above. For the microarray comparison between the wild type and the hma2 hma4 double mutant, seedlings were germinated for 7 d on, and subsequently transferred onto, plates of agar-solidified modified Hoagland medium containing 1 µM ZnSO₄ (referred to as control in this experiment) and grown for another 14 d before harvest.

Hydroponic plant cultivation for microarray-based transcriptome comparisons between Zn deficiency and control conditions in wild-type Arabidopsis was as described (Talke et al., 2006). Briefly, tissues were sampled of 6-week-old vegetative plants cultivated in 11 h light (20°C, 145 µmol m⁻² s⁻¹)/13 h dark (18°C), with −Zn (no Zn added to solution) or +Zn (control, 5 µM Zn) conditions for 3 weeks before harvest. Two independent replicate experiments were conducted for all microarray hybridizations. The rationale for the use of different cultivation systems in microarray experiments was as follows: At the outset of our study, we were not technically capable of generating Zn deficiency reliably and reproducibly in seedlings cultivated on agar-solidified media without adding an excess of a chelator, which we judged as nonphysiological based on pilot studies. The plate cultivation system, however, had to be chosen for the comparison of the hma2 hma4 mutant with the wild type on Zn-sufficient media. This enabled us to capture transcriptional responses representative of physiological Zn deficiency and not transcriptional responses associated with severe damage including chlorosis, growth inhibition, developmental alterations, and necrosis, which occur when hma2 hma4 plants are cultivated without substantial additional Zn supplementation for an extended period of time.

Non-microarray hydroponic experiments shown in Figures 2C and Supplemental Figures 1C and 1D were conducted under growth conditions as described by Talke et al. (2006). For all other hydroponic experiments, seedlings were transferred at the age of 10 d from agar-solidified control or Zn-deficient 1× Hoagland solutions into 400 mL of hydroponic solution per four plants, with the same mineral nutrient composition, but 0.25× concentrated, in short days (8 h of 120 µmol m⁻² s⁻¹ white light at 22°C, 16 h dark at 18°C) in a growth chamber (Model TC; Conviron Adaptis), and solutions were exchanged weekly.

Multi-Element Analysis of Plant Tissues

Freshly harvested root and shoot tissues were desorbed by incubating twice in 10 mM EDTA, then twice in ultrapure water, each 25 mL, for 10 min on ice. Tissue was blotted dry, then dried at 60°C for 3 d and allowed to equilibrate at RT for one further day. Tissue was weighed into acid-washed Duran glass tubes and mineralized upon addition of 2 mL 65% (w/w) HNO₃ and standing at RT overnight, in a heating block at 80°C for 1 h and at 120°C for 1.5 h. Tubes were left to cool to below 50°C and cleared by addition of 1 mL 30% (v/v) H₂O₂, kept at RT for 0.5 h, 60°C for 0.5 h, 100°C for 0.5 h, left to cool to RT, and subsequently filled up to a final volume of 10 mL with ultrapure water. Multi-element analysis of digests was conducted using inductively coupled plasma optical emission spectrometry of these acid digests using an iCAPDuo 6500 instrument (Thermo Fisher Scientific), calibrated with a blank and a series of five multi-element standards manually pipetted from single-element standard solutions for 17 elements commonly detected in Arabidopsis (AAS Standards; Bernd Kraft). The precision of measurements was validated by measuring a sample blank and an intermediate calibration standard solution, as well as digests of a certified reference material (Virginia tobacco leaves, INCT-PVTL 6; Institute of Nuclear Chemistry and Technology), before and after each set of ∼50 samples. Recoveries were 92% ± 10% (mean ± relative sd) for Zn.

Generation of Transgenic Plants

The region directly upstream of the ATG of ZIP4 to the nucleotide preceding the end of the coding region of the neighboring gene (597 bp) was PCR-amplified and cloned into pENTR (Gateway; Invitrogen; see Supplemental Table 2 for primer sequences and cycling conditions). After sequencing, LR Clonase II (Gateway; Invitrogen) was used to transfer this fragment into pMDC162 (Curtis and Grossniklaus, 2003) upstream of the GUS gene to form the ProZIP4:GUS construct. The resulting binary plasmid was used to transform Agrobacterium tumefaciens (strain GV3130 pMP90RK), and wild-type Arabidopsis thaliana Col-0 was transformed using the floral dip method (Clough and Bent, 1998). The transgene was subsequently crossed into the hma2 hma4 background.

To generate the ProMTP2:GUS construct, a 2542-bp-long sequence fragment was amplified from genomic DNA, introducing NheI and NgoMIV sites using the primers MTP2prom_for and MTP2prom_rev. The PCR product was cloned into the XbaI and XmaI sites of pGPTV-kan (DNA Cloning Service; Herman Schmidt). This fragment comprised part (722 bp) of the transcribed region of the upstream gene including its 3′-untranslated region, the 1263-bp intergenic region, the 183-bp-long 5′-untranslated region of MTP2, an intron of 93 bp in length just before the start of the coding sequence, and 281 bp of the coding sequence of MTP2, which encode the N terminus including the first two transmembrane helices. To generate a fusion of eGFP to the N terminus of the MTP2 protein (ProMTP2:GFP:MTP2cDNA), MTP2 including stop codon was subcloned by in vitro site-directed recombination (Gateway; Invitrogen) into the binary vector pK7WGF2 (Karimi et al., 2002). Using Arabidopsis gDNA as template, a fragment including 1277 bp upstream of the MTP2 coding sequence and the first four codons of MTP2 was amplified using the primers pMTP2f_Sac1 and pMTP2r_Spe1, introducing recognitions sites for SacI and SpeI, respectively. The amplification product (ProMTP2) was cloned into pCR2.1 and verified by sequencing. The CaMV 35S promoter sequence (Pro35S) in a pK7WGF2 construct harboring MTP2 was replaced by the native ProMTP2 using SacI and SpeI restriction sites. Similarly, a construct encoding a translational fusion of eGFP to the C terminus of MTP2 (ProMTP2:MTP2cDNA:GFP) was obtained: Using Arabidopsis gDNA as template, a fragment including 1277 upstream of MTP2 and the entire MTP2 ORF including stop codon was amplified using the primers pCDF2f_Sac1 and pCDF2r_Asc1, introducing recognitions sites for SacI and AscI, respectively. The amplification product (ProMTP2:MTP2) was cloned into pCR2.1, verified by sequencing, and cloned into pK7FWG2 (Karimi et al., 2002) using SacI and AscI restriction sites. Arabidopsis Col-0 was transformed as described above.

Root Length and Chlorophyll Measurements

Root lengths were quantified in images generated on a flatbed scanner using ImageJ (http://imagej.nih.gov/ij/). Leaf chlorophyll content of seedlings cultivated was determined after extracting pigments from rosette tissues homogenized in methanol at room temperature in the dark for 10 min while shaking at 200 rpm on an orbital shaker (Model 3005; GFL). Extracts were centrifuged for 5 min at 14,000g, and the absorbance of the supernatant used to calculate chlorophyll concentration (chlorophyll [µg mL⁻¹] = 22.5 × A₆₅₀ + 4.0 × A₆₆₅) (Porra et al., 1989).

Yeast Constructs, Strains, and Growth

The MTP2 cDNA including translational stop codon was subcloned into the yeast expression vector pFL61 adapted for Gateway cloning (Desbrosses-Fonrouge et al., 2005) and into a Gateway-adapted derivative of pUG35 (Güldener and Hegemann, 1998) in which a translational fusion is generated of the encoded protein to an C-terminal yEGFP-tag (Desbrosses-Fonrouge et al., 2005). Fluorescence was visualized using a confocal laser scanning microscope (Leica TCS SP2; Leica Microsystems), exciting with a 488-nm argon laser and capturing fluorescence using Leica EGFP settings.

For complementation assays, the following Saccharomyces cerevisiae strains were transformed with constructs as indicated: ccc1 (Chen and Kaplan, 2000), ycf1 (Petrovic et al., 2000), zrc1 cot1 (Becher et al., 2004), BY4741 (Mat a, his3D∆1, leu2D∆0, met15D∆0, ura3D∆0) or BY4742 (Mat a, his3D∆1, leu2D∆0, ura3D∆0) (Brachmann et al., 1998). Transformants were selected on solid synthetic complete medium (SC) lacking uracil and histidine as appropriate (Sherman et al., 1986) with 2% (w/v) D-glucose as carbon source.

For each construct, several independent transformant colonies were grown overnight at 28°C in 2 mL SC-URA to early stationary phase (OD₆₀₀ ≈ 1.0). To induce expression of MTP2-yEGFP using pUG35, methionine was removed from media by washing yeast cells twice in water and resuspending them in SC-Ura-Met at OD₆₀₀ = 0.5. For drop-test complementation assays, yeast cells were washed once in water, and a serial dilution of OD₆₀₀ = 0.5, 0.05, 0.005, and 0.0005 was made using water + 0.2 ‰ (v/v) Tween 20. Drops of 10 μL were spotted on solid low-sulfate/phosphate media (Conklin et al., 1992) lacking uracil with 2% (w/v) d-glucose, supplemented with the indicated concentrations of various metal ions using sterile stock solutions of 0.1 M ZnSO₄, 0.1 M CoCl₂, 1 M MnSO₄, 1 M CdSO₄, 1 M NiSO₄, 0.1 M FeSO₄, or 1 M MnCl₂. Plates were incubated at 28°C for 3 to 5 d before photographs were taken.

Microarray Analysis

For the comparison of transcriptomes of Arabidopsis grown under Zn-deficient and -sufficient conditions, RNA extraction, labeling, and ATH1 hybridization were performed as described previously (Talke et al., 2006). For the comparison of 21-d-old wild-type and hma2 hma4 seedlings grown on plates containing control (+Zn) agar-solidified modified Hoagland solution, root and shoot tissues were harvested separately and immediately frozen in liquid nitrogen. RNA was extracted using the RNeasy Plant Mini Kit (Qiagen). Labeling and hybridization ATH1 microarrays (Affymetrix) was performed at Gene Core, EMBL, Heidelberg, according to the instructions of the manufacturer. All data were imported into Genespring (Version 8; Agilent Technologies), and a MAS5 summarization/normalization was performed. To account for random fluctuations that influence low signal intensities, entities in the 25th percentile of signal intensities in all microarray hybridizations were filtered out. Probe sets were identified that showed a 2-fold or larger average difference between conditions or between genotypes. Probe sets of interest were further narrowed down by selecting only those that showed a 2-fold or larger difference in each of both replicate experiments. These data were then subjected to an unpaired Mann-Whitney test with a cutoff value of P < 0.2 with Benjamini and Hochberg adjustments for multiple comparisons. Thus, we obtained a list of candidate genes from the comparison between roots (or shoots) of Zn-deficient and Zn-sufficient plants and a second list of candidate genes from the comparison between roots (or shoots) of hma2 hma4 and wild-type seedlings grown in control conditions. Gene lists were then cross-referenced with one another to identify coregulated transcripts as described in the Results. By choosing those genes commonly regulated in both cultivation systems (see Figure 1D and Supplemental Figure 1A), we identified the subset of corresponding genes irrespective of plant age and environmental conditions.

RT-qPCR Analysis

Tissues were harvested, and total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific), pooled in equal quantities from the two independent replicate experiments, and 800 ng RNA used in first-strand cDNA synthesis using the SuperScript III kit and oligo(dT) primer (Thermo Fisher Scientific) according to the manufacturer’s instructions. Four microliters of a 1/30 dilution of the cDNA was used as template for qPCRs, which were performed using a LightCycler 480 real-time PCR system (Roche Diagnostics) in 384-well plates. SYBR Green PCR Mastermix (Applied Biosystems) was used to monitor amplification of cDNA using an annealing temperature of 60°C. Reactions with primer amplification efficiencies below 1.8 were discarded, and mean primer amplification efficiency (PE) for each primer pair calculated from all reactions. These were used together with the individual threshold cycles (C_T) to calculate individual expression values (Ev) for each reaction Ev = PE^-CT. The mean Ev (mEv) of the constitutively expressed reference genes UBQ10 (At4g05320) or HELICASE (HEL; At1g58050) was used to calculate relative transcript levels (RTLs) for each individual gene tested RTL = Ev/mEv(UBQ10), and the arithmetic mean and sd of these RTLs was calculated from technical replicates (see Supplemental Table 2 for primer sequences).

Histochemical Staining for GUS Activity

For ProMTP2:GUS plants, 19-d-old seedlings were immersed in GUS staining solution [50 mM Na₂HPO₄, 50 mM NaH₂PO₄, 3 mM K₃Fe(CN)₆, 3 mM K₄Fe(CN)₆, and 1% (v/v) Triton-X-100] for 18 h at 37°C in the dark (Jefferson et al., 1987). ProZIP4:GUS seedlings were stained in the same manner, but for 3 h instead. Chlorophyll was then removed by incubation in 80% (v/v) ethanol before imaging seedlings on a Leica MZ12 dissecting microscope coupled to a Zeiss AxioCan MRc digital camera (Carl Zeiss).

Confocal Microscopy

ProMTP2:GFP:MTP2 and ProMTP2:MTP2:GFP lines were grown on Zn-deficient agar-solidified medium for 14 d and then mounted in ultrapure water. GFP fluorescence was imaged in emerging lateral roots using the 488-nm laser for excitation on an SP5 confocal laser scanning microscope (Leica) using a PLAN APO 40× water-immersion lens and Leica emission settings for eGFP and propidium iodide, respectively. For mCherry imaging, the 561-nm laser with Leica TRITC emission settings was used. GFP and mCherry fluorescence was collected using sequential scanning.

Accession Numbers

Sequence data for the genes mentioned in this article can be found on The Arabidopsis Genome Initiative (TAIR) or GenBank website. The microarray data set accession number is GSE111443 (https://www.ncbi.nlm.nih.gov/geo/). Others are as follows: ZIP4, AT1G10970.1; ZIP9, AT4G33020.1; MTP2, AT3G61940.1; HMA2, AT4G30110.1; HMA4, AT2G19110.1; UBQ10, AT4G05320.2; ACT2, AT3G18780.2; and Helicase, AT1G58050.1.

Supplemental Data

• Supplemental Figure 1. Experimental Design and MTP2 Gene Expression Patterns.

• Supplemental Figure 2. Zinc Concentrations in Roots and Shoots of mtp2 Mutant and Wild-Type Seedlings.

• Supplemental Figure 3. Zinc Concentrations in Roots and Shoots of Hydroponically Cultivated mtp2 Mutant and Wild-Type Plants.

• Supplemental Figure 4. Localization of MTP2 in MTP2:yEGFP Expressing Yeast, and Complementation of the mtp2 Mutant Phenotype with Fusion Proteins of MTP2 and GFP.

• Supplemental Figure 5. Systemic Regulation of Root HMA2 Transcript Levels, Zn Partitioning in Hydroponically Cultivated Zn-Deficient Wild-Type Plants, and Genetic Interaction between mtp2 and hma2.

• Supplemental Table 1. Transcripts That Are Candidates for Involvement in Generating the Systemic Signal in Shoots, as Identified in Microarray Analyses in This Study.

• Supplemental Table 2. Primer Sequences.

• Supplemental Data Set 1. Transcriptional Response to Zn Deficiency in Roots of Wild-Type Arabidopsis, as Identified in Microarray Analyses in This Study.

• Supplemental Data Set 2. Genes Differentially Expressed in Roots of the hma2 hma4 Mutant Compared with the Wild Type, as Identified in Microarray Analyses in This Study.

• Supplemental Data Set 3. Transcriptional Response to Zn Deficiency in Shoots of Wild-Type Arabidopsis, as Identified in Microarray Analyses in This Study.

• Supplemental Data Set 4. Genes Differentially Expressed in Shoots of the hma2 hma4 Mutant Compared with the Wild Type, as Identified in Microarray Analyses in This Study.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• ACT2 Gramene: AT3G18780

• ACT2 Araport: AT3G18780

• ALBUMIN1 Gramene: AT1G03890

• ALBUMIN1 Araport: AT1G03890

• ARABIDOPSIS Gramene: AT4G39753

• ARABIDOPSIS Araport: AT4G39753

• AT4G39753 Gramene: AT4G39753

• AT4G39753 Araport: AT4G39753

• AtHMA2 Gramene: AT4G30110

• AtHMA2 Araport: AT4G30110

• AtIRT3-overexpressing Gramene: At1g60960

• AtIRT3-overexpressing Araport: At1g60960

• AtZIP2 Gramene: At5g59520

• AtZIP2 Araport: At5g59520

• CPK19 Gramene: AT1G61950

• CPK19 Araport: AT1G61950

• FAMILY2 Gramene: AT5G06820

• FAMILY2 Araport: AT5G06820

• HEL Gramene: AT1G58050

• HEL Araport: AT1G58050

• HMA4 Gramene: AT2G19110

• HMA4 Araport: AT2G19110

• IRT3 Gramene: AT1G60960

• IRT3 Araport: AT1G60960

• MLO9] Gramene: AT1G42560

• MLO9] Araport: AT1G42560

• MTP1 Gramene: At2g46800

• MTP1 Araport: At2g46800

• MTP2-yEGFP Gramene: AT3G61940

• MTP2-yEGFP Araport: AT3G61940

• MTP3 Gramene: At3g58810

• MTP3 Araport: At3g58810

• NAS2 Gramene: AT5G56080

• NAS2 Araport: AT5G56080

• NAS3 Gramene: AT1G09240

• NAS3 Araport: AT1G09240

• NICOTIANAMINE CHEBI: CHEBI:25520

• OLIGOPEPTIDE CHEBI: CHEBI:25676

• PHO1 Gramene: AT3G23430

• PHO1 Araport: AT3G23430

• Pro35S:NIP1;1:mCherry Gramene: At4g19030

• Pro35S:NIP1;1:mCherry Araport: At4g19030

• Pro35S:NIP1;1cDNA:mCherry Gramene: At4g19030

• Pro35S:NIP1;1cDNA:mCherry Araport: At4g19030

• ProMTP2 Gramene: AT3G61940

• ProMTP2 Araport: AT3G61940

• ProMTP2:GFP:MTP2 Gramene: AT3G61940

• ProMTP2:GFP:MTP2 Araport: AT3G61940

• ProMTP2:GFP:MTP2cDNA Gramene: AT3G61940

• ProMTP2:GFP:MTP2cDNA Araport: AT3G61940

• ProMTP2:GUS Gramene: AT3G61940

• ProMTP2:GUS Araport: AT3G61940

• ProMTP2:MTP2 Gramene: AT3G61940

• ProMTP2:MTP2 Araport: AT3G61940

• ProMTP2:MTP2:GFP Gramene: AT3G61940

• ProMTP2:MTP2:GFP Araport: AT3G61940

• ProZIP4:GUS Gramene: AT1G10970

• ProZIP4:GUS Araport: AT1G10970

• SCPL15 Gramene: AT3G12240

• SCPL15 Araport: AT3G12240

• SCPL22 Gramene: AT2G24000

• SCPL22 Araport: AT2G24000

• SRF2] Gramene: AT5G06820

• SRF2] Araport: AT5G06820

• YSL1 Gramene: AT4G24120

• YSL1 Araport: AT4G24120

• YSL3 Gramene: AT5G53550

• YSL3 Araport: AT5G53550

• ZIP1 Gramene: AT3G12750

• ZIP1 Araport: AT3G12750

• ZIP4 Gramene: At1g10970

• ZIP4 Araport: At1g10970

• ZIP5 Gramene: AT1G05300

• ZIP5 Araport: AT1G05300

• ZIP9 Gramene: AT4G33020

• ZIP9 Araport: AT4G33020

• chlorophyll CHEBI: CHEBI:28966

• hma2 Gramene: AT4G30110

• hma2 Araport: AT4G30110

• mtp2 Gramene: AT3G61940

• mtp2 Araport: AT3G61940

• physiological role CHEBI: CHEBI:52211

• root AmiGo: PO:0009005

• root length Planteome: TO:0000227

• root tip AmiGo: PO:0000025

• zinc CHEBI: CHEBI:27363

• zinc concentration Planteome: TO:0006053

• zip2 Gramene: AT5G59520

• zip2 Araport: AT5G59520

• bZIP19 Gramene: At4g35040

• bZIP19 Araport: At4g35040

• bZIP23 Gramene: At2g16770

• bZIP23 Araport: At2g16770