Physiological effect of zinc deficiency in Arabidopsis thaliana

To investigate the effect of Zn deficiency on plant physiology, we cultivated WT plants in Murashige and Skoog-based zinc-sufficient (15 µM Zn; + Zn) and deficient (0 µM Zn; -Zn) media. After three weeks of growth on horizontal plates under -Zn conditions, the plants showed stunted growth and chlorosis, resulting in a significant reduction in both fresh weight (FW) and chlorophyll content compared to those grown in +Zn conditions (S1A–C Fig). As anticipated, the diminished chlorophyll content correlated with a decrease in the maximum quantum yield of photosystem II (PSII, F_v/F_m), which reflects the potential efficiency of PSII, further confirming the impact of Zn-deficiency on photosynthetic performance (S1D Fig). While mineral deficiencies often induce significant compensatory changes at the whole plant level, the specific effect of Zn deficiency on root and shoot ionome profiles have not been extensively documented. To address this gap, we conducted an ICP-OES (Inductively Coupled Plasma – Optical Emission Spectrometry) analysis of mineral content in the roots and shoots of 3-week-old plants grown in +Zn or -Zn conditions. As expected, Zn was nearly undetectable in both roots and shoots of WT plants under -Zn conditions (S1E and S1F Fig and S1 Table). Moreover, we observed significant reductions in the content of potassium (K), magnesium (Mg), Mn, sodium (Na), strontium (Sr), calcium (Ca), and Cu in the shoots of plants grown in -Zn compared to those grown in +Zn. Conversely, the levels of Fe and molybdenum (Mo) remained largely unaffected by Zn deficiency in shoots. Notably, Zn-deficient roots accumulated over three times more Mn than control roots, suggesting that the transport systems induced under Zn deficiency may also facilitate Mn uptake. Additionally, we found that Mg, Sr, Fe and K levels were significantly reduced in the roots under -Zn conditions while the contents of Ca, Cu, Mo and Na remained unchanged. We also assessed the translocation of minerals from roots to shoots under both +Zn and -Zn conditions. The root-to-shoot translocation of Ca, Cu, K, Mn, Mo, Na and Sr decreased under -Zn condition. In contrast, the translocation of Fe and Zn increased under -Zn conditions. The translocation of Mg from roots to shoots remains consistent between +Zn and -Zn conditions (S1G Fig). Overall, these findings establish conditions under which Zn deficiency significantly affects plant development, physiology and overall mineral accumulation. Furthermore, they suggest that compensatory mechanisms may be at play for other minerals in -Zn conditions.

Spatio-temporal expression analysis of the AtZIP family reveals potential candidates for metal acquisition from the soil

To gain insight into the transcriptional response of plants under Zn-deficient conditions and identify candidates for Zn uptake from the soil, we first assessed the expression of the entire ZIP family in WT plants grown for 1 week in either +Zn or -Zn conditions (Fig 1A). The expressions of IRT1, ZIP6, ZIP7, ZIP10 and ZIP11 were not significantly altered by Zn deficiency (Figs 1B and S2A). In contrast, we observed increased mRNA accumulation of IRT3, ZIP1, ZIP3, ZIP4, ZIP5, ZIP9 and ZIP12 in WT roots grown under Zn deficiency compared to control conditions, while IRT2, ZIP2 and ZIP8 showed decreased expression, similarly to what previous studies reported [30–33]. These results suggest that the upregulated transporters may serve as potential candidates for Zn uptake in conditions of limited Zn availability. Notably, ZIP2 was the only member of the ZIP family that exhibited high expression under control conditions (S2A Fig). This observation aligns with previous reports of high ZIP2 expression in control conditions [18].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Spatio-temporal expression of the AtZIP family in roots.

(A) Phylogenetic tree of the 15 members of AtZIP family in Arabidopsis thaliana. The tree was generated using protein sequences with Phylogeny.fr. (B) Relative transcript levels of the fifteen ZIP members in WT plant roots grown 1 week under control conditions (+Zn) or Zn-deficient conditions (-Zn). Results are presented as a heatmap, highlighting fold changes compared to control conditions (+Zn). Numeric values are provided in S2A Fig. Data represent 4 biological replicates, with each replicate being a pool of at least 40 roots. Statistical differences were determined using Student’s t-test or Mann Whitney test with significance levels indicated (*P < 0.05; **P < 0.01; ***P < 0.0001). (C) Representative images of ZIP::NLS-3xmVenus expression pattern from five-day-old plants grown in control conditions. Images taken in different root zones: zone I (root tip, lower pictures), zone II (early differentiated zone, middle pictures), and zone III (mature zone, upper pictures). Nuclear-localized mVenus signal is shown in green while PI used to stain the cell wall is shown gray. Images were captured with identical settings for each line along the root, but settings varied between lines due to differences in promoter activity among ZIPs. For zone I, a singe focal plant for mVenus and PI is displayed (XY), while for zone II and III, mVenus signal is presented as a maximum projection of Z-stacks (XYZ), overlaid with a single orthogonal view of PI extracted from the Z-stacks. Images correspond to T1 plants; at least 5 independent T1 plants were imaged. The signal was confirmed in T2 plants. Scale bars: 25 µm for all images.

https://doi.org/10.1371/journal.pgen.1011796.g001

Anticipating that transporters involved in Zn uptake from the soil would be expressed at the root periphery, we next examined the spatial expression pattern of the 15 ZIP transporters using promoter-reporter lines. We investigated promoter activity in the root tip (zone I), the fully elongated zone (zone II) and the differentiated root (zone III) in control conditions (i.e.,; containing 15 µM Zn). The promoters were used to express a fluorescent reporter, NLS-3xmVenus, in 5-day-old seedlings (Fig 1C). Promoters of IRT1, IRT2, ZIP2, ZIP3, ZIP5 and ZIP8 displayed activity primarily at the root periphery in zone II and III (i.e., mainly in epidermal and cortical tissues), while the promoter of ZIP7 was predominantly active in central root tissues such as the endodermis and the stele. ZIP11 also showed activity in central root tissues but was mostly active in cortical cells. IRT3, ZIP1 and ZIP10 exhibited almost exclusively endodermal promoter activity (with ZIP10 being active only in zone III), while ZIP6 promoter activity was primarily detected in the pericycle in both zones II and III. Promoter activity for ZIP4, ZIP9 and ZIP12 was minimal under control conditions (as indicated by the absence of nuclear signal; the faint green signal likely corresponds to background fluorescence), with ZIP12 displaying extremely low signals at the root periphery and in the inner tissue of the zone I. These observations suggest that ZIP2, ZIP3, ZIP5 and ZIP8 may be involved in metal uptake from the soil at the root periphery, similarly to the well-characterized IRT1 in this condition. Given IRT3, ZIP1 and ZIP10 endodermal promoter activity in control conditions, these three transporters might play a role in metal uptake and transport within the endodermis. Interestingly, six ZIP promoters (IRT2, IRT3, ZIP1, ZIP2, ZIP3 and ZIP7) showed activity in cells at the root periphery in zone I, indicating their potential role in mediating the transport of divalent transition metal ions in the root tip. Having observed that zinc deficiency led to a higher expression for seven ZIP transporters (Fig 1B), we then investigated their respective promoter activity in -Zn conditions in zone III of the root (S2B Fig). The localization of promoter activities for ZIP3 and ZIP5 were similar to the ones observed in control conditions (Fig 1C). However, zinc deficiency led to an expansion to additional cell layers for ZIP1 promoter (in the stele), ZIP4 promoter (in the stele, cortex and epidermis), ZIP9 promoter (in the cortex and epidermis) and IRT3 promoter (in the stele, cortex and epidermis), similarly to previous observations [30]. We could not detect ZIP12 promoter activity with our construct in our zinc deficiency conditions. Overall, these data suggest a putative additional role of ZIP4, ZIP9 and IRT3 in Zn uptake from the soil at the root periphery under low zinc availability.

Based on our expression analyses and promoter reporter lines characterization, we selected ZIP2, ZIP3, ZIP5 and ZIP8, expressed at the root periphery already in control conditions (i.e., non-deficient), for further investigation as potential candidates for metal uptake from the soil. ZIP2 was of particular interest due to its outer promoter activity, its strong expression in control conditions, and its ability to complement ctr1 and smf1 yeast strains. Given these features, we hypothesize that ZIP2 may play a role not in Zn uptake, but in the acquisition of Cu and Mn from the soil (Figs 1B, 1C, S2A and S2C) [18,22]. ZIP8, on the other hand, was investigated as a candidate for Fe acquisition due to its specific epidermal promoter activity, increased expression under Fe deficiency, and its close homology to IRT1 (Figs 1A, 1C, and S2D). ZIP3 and ZIP5 were selected as prime candidates for Zn acquisition from the soil based on multiple lines of evidence. Both genes were expressed at higher levels under Zn deficiency (Figs 1B and S2A) and are known to be regulated by the transcription factors bZIP19 and bZIP23 [31]. Additionally, their promoter activity was primarily detected at the root periphery (Figs 1C and S2B), and both have been shown to complement the yeast zrt1zrt2 strain, further supporting their role in Zn acquisition [18,28]. In addition, both were very recently shown to redundantly act in Zn uptake in an independent study [30].

ZIP2, ZIP3, ZIP5 and ZIP8 localize to the epidermis-soil interface

To confirm the potential role of ZIP2, ZIP3, ZIP5 and ZIP8 in metal acquisition from the soil, we first examined their localization in roots. Transgenic lines were generated expressing a fusion of the ZIP genomic DNA with the coding sequence for mCitrine (mCit), under the control of their endogenous promoters, within their respective loss-of-function mutant backgrounds to ensure functional complementation (see below and S2 Table). Based on previous studies that demonstrated the successful integration of mCitrine in the second extracellular loop of IRT1 while maintaining its functionality [34], we applied the same strategy for ZIP2/3/5/8-mCitrine constructs. Structural modeling of the fusion proteins predicted that inserting mCitrine into these ZIPs transporters did not alter their 3D conformation, similar to IRT1-mCitrine (S3 Fig). Since we expected these transporters to localize at the PM under certain conditions, we co-localized the ZIP transporters with FM4–64, a marker for the PM and endocytic system.

In control, Cu- or Mn-deficient conditions, ZIP2-mCitrine predominantly accumulated in the epidermis of root zone II and in both epidermal and cortical cells of the differentiated root zone III (Fig 2A). Some signal was also detected in the vasculature. ZIP2-mCitrine showed partial PM localization under control conditions but was also present in intracellular structures, which co-localized with FM4–64 after 10 minutes, suggesting its localization in endosomal compartments (S4A and S4B Fig). This dual PM-endosomal localization is common for PM proteins fused to highly sensitive fluorophores, possibly reflecting intermediate trafficking stages from the endoplasmic reticulum (ER) to the PM, as observed for transporters like IRT1, NRAMP1 (NATURAL RESISTANCE-ASSOCIATED MACROPHAGE PROTEIN 1) and PHO1 (PHOSPHATE 1) [35–37]. For IRT1, such localization also reflects substrate-controlled cycling between the PM and early endosomes/Trans-Golgi Network (EE/TGN) [34,37,38]. In ZIP2, Cu and Mn deficiency did not notably alter cell-specific localization, except for the near absence of ZIP2-mCitrine in the stele under Cu-deficient conditions (Fig 2A). Considering that transporter polarity can affect their function, we evaluated the potential polarity of ZIP2 in the elongation zone using UBQ10::ZIP2-mCitrine overexpression line (Fig 2B). This line was used because ZIP2 expression under its endogenous promoter was undetectable in the elongation zone, where polarity can be more accurately assessed. By calculating the ratio of outer to inner PM fluorescence, we found ZIP2 to be apolar in the epidermis, with a fluorescence ratio close to 1 (Fig 2C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Localization of ZIP2, ZIP8, ZIP3 and ZIP5 in the outer root domain.

(A, D, E and G) Representative confocal images showing mCitrine fluorescence in 5-day-old seedlings expressing ZIP2, ZIP8, ZIP3 or ZIP5 fused to mCitrine under the control of their endogenous promoters. Seedlings were grown in either control or metal-deficient conditions (-Cu, -Mn, -Fe + 100 µM Ferrozine, -Zn). The upper panels display root transversal views (extracted from Z-stacks), while the lower panels show longitudinal views taken at the median focal plane of the root. The corresponding root zone (II or III, as defined in Fig 2C) is indicated below the images. All images were acquired from T2 plants. (A, B, D-H) mCitrine fluorescence is displayed using a Fire look-up-table (LUT). (B, F, H) Confocal images of mCitrine fluorescent signal in 5-day-old seedlings expressing UBQ10::ZIP2/3/5-mCitrine grown under control (ZIP2) or -Zn conditions (ZIP3 and ZIP5). Roots were imaged in the median focal plane of epidermal (Ep) cells within the elongation zone to assess polar localization. (C) Fluorescence ratio analysis for ZIP2, ZIP3, and ZIP5-mCitrine shown in panels B, F and H (pink). The ZIP8-mCitrine ratio (orange) was calculated in the epidermis of root zone III. The fluorescence ratio was determined by dividing the average fluorescence intensity at the outer (Out) PM by the average intensity at the inner (In) PM for each epidermal cell. The same region of interest (ROI) was used for both outer and inner fluorescence quantification. An example of the ROI used to quantify Out/In ratio is shown with the dashed square for each genotype (Panels B, D, F, H). Ratios were calculated from at least 5 cells per root and at least 3 independent roots. The dashed lines in the violin plots represent the median, while the dotted lines represent the first and third quartiles (n ≥ 15). Scale bars: 25 µm for all images.

https://doi.org/10.1371/journal.pgen.1011796.g002

Next, we investigated ZIP8 localization, given its potential role in Fe acquisition. Under, control conditions, ZIP8-mCitrine was barely detectable in root zone II but was observed at low levels in the epidermis of root zone III (Fig 2D). Under Fe deficiency (-Fe + 100 µM Ferrozine), ZIP8-mCitrine strongly accumulated at the PM of the epidermal cells in both zone II and III (Figs 2D and S4C). Interestingly, ZIP8-mCitrine displayed pronounced outer polar localization in the epidermis under both control and Fe-deficient conditions, reinforcing its possible involvement in metal acquisition from the soil (Figs 2C and S4C).

We also examined the localization of ZIP3 and ZIP5. In control conditions, ZIP3-mCitrine was detected in a few cells in the vasculature (white arrows, Fig 2E). Under Zn-deficient conditions, ZIP3-mCitrine accumulation significantly increased, with prominent expression in the epidermis and cortex, in addition to the vasculature (white arrows). ZIP3-mCitrine predominantly localized to the PM in the elongation zone, though subcellular localization was also observed, co-localizing with FM4–64 indicating these structures were part of the endocytic system (S4D and S4E Fig). Due to the low accumulation of ZIP3-mCitrine in the root tip, we assessed ZIP polarity in the elongation zone using the overexpression line, revealing outer polar localization (facing the soil) in the epidermis under Zn deficiency (Fig 2C and 2F). Lastly, ZIP5-mCitrine localization was explored. In control conditions, ZIP5-mCitrine was undetectable in roots, but it accumulated in the epidermis of root zone II under Zn deficiency (Fig 2G). In zone III, ZIP5-mCitrine accumulated in the epidermis, cortex and some cells in the vasculature under Zn-deficient conditions (white arrows, Fig 2G). ZIP5-mCitrine predominantly localized to the PM but unlike ZIP3, did not show polar localization in the epidermis (Figs S4F, S4G, 2C, and 2H).

ZIP2 mediates Cu acquisition

Given ZIP2’s expression at the root periphery and its PM localization in epidermal cells under Cu and Mn deficiency, as well as its ability to complement smf1 and ctr1 yeast strains, we hypothesized that ZIP2 could facilitate Cu and Mn acquisition from the soil [18,22]. To investigate this, we first isolated the previously described T-DNA insertion line zip2–1, which exhibits no detectable ZIP2 expression (S5A and S5B Fig) [18] and generated a new CRISPR/Cas9-induced allele, zip2–2_cr (S5A Fig). The zip2–2_cr mutant harbors a 68 bp deletion resulting in a premature stop codon (S5C Fig). When grown in soil, zip2–2_cr mutants exhibited a slight increase in shoot weight compared to both WT and zip2–1 (S5D and S5E Fig). However, when plants were grown on Murashige and Skoog-based agar plates containing 0.05 µM Cu (+Cu), shoot FW, root length and chlorophyll content were similar across all genotypes (WT, zip2–1, and zip2–2_cr) (Fig 3A–D). In contrast, when Cu was omitted from the growth medium (-Cu), both zip2–1 and zip2–2_cr exhibited reduced shoot FW and root length compared to WT plants (Fig 3A–D). This increased sensitivity to Cu deficiency in the zip2 mutants suggests a contribution of ZIP2 to Cu acquisition. To explore whether ZIP2 might also mediate Mn uptake, we grew zip2 mutants and WT plants on Mn-deficient plates. However, unlike under Cu deficiency, neither zip2–1 nor zip2–2_cr displayed notable differences compared to WT under Mn-deficient condition (-Mn) (S5F–H Fig). The lack of a pronounced phenotype suggests that ZIP2 plays no or a limited role in Mn uptake. To further confirm ZIP2’s role in Cu and Mn acquisition, we measured Fe, Zn, Cu and Mn content in both shoots and roots of zip2 mutants grown for 3 weeks on control agar plates (+Cu). These metals were selected as they represent the primary substrates for ZIP transporters. Notably, both zip2–1 and zip2–2_cr mutants accumulated less Cu than WT in both roots and shoots, reinforcing the idea that ZIP2 plays a role in Cu acquisition (Fig 3E and 3F). Given the reduction in Cu content in zip2–1, we used this phenotype to assess the functionality of the ZIP2-mCitrine fusion protein used in localization studies (Fig 2A). Complementation of Cu content in roots of two independent zip2–1/ZIP2::ZIP2-mCitrine lines confirmed the functionality of the fusion protein and the causal role of ZIP2 in this phenotype (S5I Fig). The increased root Mn content observed in the complemented lines suggests that ZIP2 may also facilitate Mn transport into the root (S5I Fig). Overall, our data strongly suggest that ZIP2 primarily mediates Cu acquisition from the soil, with a potential, though more limited, role in Mn acquisition.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Impaired Cu accumulation in the zip2 mutant.

(A) Representative images of WT and two zip2 mutant alleles grown vertically for 10 days under control conditions (+Cu) and copper deficient conditions (-Cu). Three representative plants for each genotype were transferred to agar plates for imaging. Scale bars: 1 cm. (B-D) Violin plots showing the distribution of (B) fresh weight (FW), (C) root length, and (D) chlorophyll content of plants grown in +Cu or -Cu for 10 days (n ≥ 15 for B and C, n ≥ 5 for D). Dashed lines of the violin plots represent the median, while dotted lines mark the first and third quartiles. Different letters indicate statistically significant differences between genotypes, as determined by one-way ANOVA followed by Tukey’s post hoc test or Kruskal–Wallis test followed by Dunn’s test (P < 0.05). (E-F) Metal content (Fe, Cu, Mn, Zn) in shoots (left) and roots (right) of WT, zip2–1 and zip2–2_cr mutant grown vertically for 3 weeks on control agar plates (+Cu). For each biological replicate, 3–4 plants were pooled (n = 4). Data are represented as mean ± SD, with significant differences between genotypes determined one-way ANOVA followed by Tukey post hoc test (P < 0.05).

https://doi.org/10.1371/journal.pgen.1011796.g003

ZIP8 mediates Fe acquisition

Given ZIP8’s protein localization and expression patterns (Figs 2D and S2D), we hypothesized that ZIP8 may contribute to Fe uptake from the soil, similarly to its closest homologue, IRT1 (Fig 1A). As little functional data are currently available for ZIP8, we tested this hypothesis using a yeast complementation assay in the fet3fet4 double mutant, which lacks both high- and low-affinity iron-uptake systems. We expressed IRT1 cDNA as a positive control, along with two predicted versions of ZIP8 cDNA (Fig 4A). As expected, IRT1 expression restored growth of the fet3fet4 mutant even in the presence of 10 µM BPDS, a strong Fe chelator. Notably, ZIP8 expression also partially rescued the growth defect, indicating that ZIP8 can mediate Fe transport in this heterologous system.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Impaired Fe accumulation in zip8 mutant.

(A) Growth of the fet3fet4 mutant transformed with empty vector pDR196 (negative control) or with pDR196 containing IRT1 (positive control), or two ZIP8 cDNA version (ZIP8–1 and ZIP8–2). Ten-fold serial dilutions of overnight grown liquid cultures were spotted on -Ura medium, containing or not 10 µM BPDS. (B) Representative images of WT and two zip8 mutant alleles grown vertically for 10 days under control conditions (+Fe) and Fe-deficient conditions (-Fe). Three plants per genotype were transferred onto an agar plate for imaging. Scale bars: 1 cm. (C-E) Violin plots showing the distribution of (C) FW, (D) root length, and (E) chlorophyll content of plants grown as described in panel A (n ≥ 15 for B and C, n ≥ 5 for D). Dashed lines of the violin plots represent the median, while dotted lines mark the first and third quartiles. Different letters indicate statistically significant differences between genotypes, as determined by one-way ANOVA followed by Tukey’s post hoc test or Kruskal–Wallis test followed by Dunn’s test (P < 0.05). (F-G) Metal content (Fe, Cu, Mn, Zn) in shoots (left) and roots (right) of WT, zip8–1_cr and zip8–2_cr mutants grown vertically for 3 weeks on control conditions. For each biological replicate, 3–4 plants were pooled (n = 4). Data are represented as mean ± SD, with significant differences between genotypes determined by Student’s t-test or Mann-Whitney test for shoot data (no significant differences observed), while root data were analyzed using one-way ANOVA followed by Tukey post hoc test and different letters indicate statistically significant differences between genotypes for a given element (P < 0.05). For shoots analysis (F), data for zip8-1_cr and zip8-2_cr were obtained from 2 independent experiments.

https://doi.org/10.1371/journal.pgen.1011796.g004

To investigate whether ZIP8 can also contribute to Fe transport in planta, we generated two CRISPR/Cas9-induced alleles, as no T-DNA insertion line were available for the 5’ coding region of ZIP8. The zip8–1_cr allele carries a deletion in the promoter and the 5’ coding sequence, predicted to produce a truncated 2 aa peptide (S6A and S6B Fig), while zip8–2_cr harbors a single A insertion in the 5’ coding sequence, leading to a premature stop codon and a 43 aa protein (S6A and S6C Fig). We began by investigating the phenotype of WT, zip8–1_cr, zip8–2_cr, and irt1–2 plants grown on soil for 3 weeks. Since IRT1 is the only known ZIP transporter involved in Fe uptake, we used irt1–2 as a reference mutant. Unlike irt1–2, which exhibited strong chlorosis, neither of the zip8 mutants showed obvious chlorotic symptoms when grown on soil (S6D Fig). However, both zip8 mutant lines displayed a slight reduction in shoot FW compared to WT (S6E Fig). Next, we grew WT and the two zip8 mutants for 10 days on control (+Fe, 50 µM Fe-EDTA) and Fe-deficient (-Fe, 0 µM Fe) Murashige and Skoog-based agar media to assess their sensitivity to Fe deficiency. We analyzed shoot FW, root length and chlorophyll content, as indicators of Fe deficiency response. However, no significant difference in these traits were observed between WT and zip8 mutants under either Fe condition (Fig 4B–E), suggesting that ZIP8 does not play a major role in Fe deficiency. To further explore ZIP8’s role in metal acquisition, we measured metal content (Fe, Cu, Mn, and Zn) in shoots and roots of WT and zip8 mutants grown on +Fe agar plates for 3 weeks. While Mn, Zn and Cu levels were comparable between WT and zip8 mutants in roots, Fe content in the roots of zip8 alleles was reduced compared to WT (Fig 4F and 4G). Interestingly, no significant changes in metal content were observed in the shoots of zip8 mutants, suggesting that ZIP8 primarily mediates metal acquisition in the outer root tissues without substantially affecting root-to-shoot translocation. To confirm the functionality of the ZIP8-mCitrine fusion protein, we used the UBQ10::ZIP8-mCitrine reporter line (S2 Table). This line showed increased Zn and Mn content in roots, indicating that the fusion protein used for protein localization was functional (Figs 2D, S4C and S6F). Similarly, overexpression of IRT1 increased Zn and Mn levels in roots, without significantly altering Fe content (S6G Fig).

In conclusion, our results demonstrate that ZIP8 is involved in Fe acquisition in the outer root tissues and, similar to IRT1, is likely involved in the transport of other divalent ions such as Cu²⁺, Mn²⁺, and Zn²⁺. However, ZIP8’s role in Fe transport appears to be more restricted than IRT1.

ZIP3 and ZIP5 mediate Zn acquisition

Given the localization of ZIP3 and ZIP5 at the root periphery under Zn-deficient conditions (Fig 2E and 2G), along with prior evidence from yeast zrt1zrt2 complementation assays, we hypothesized that these transporters could be involved in Zn uptake from the soil [18,28]. Due to the close phylogenic relationship between ZIP3 and ZIP5 (Fig 1A), we generated a double mutant (zip3–1zip5–1) by crossing the single KO T-DNA insertion mutants zip3–1 and zip5–1 to account for potential functional redundancy (S7A–C Fig). Additionally, we created CRISPR/Cas9 mutants for ZIP3 and ZIP5 (zip3–2_cr and zip5–2_cr, S7B and S7C Fig). The zip3–2_cr mutant carries a substitution of two CC nucleotides with a single A, resulting in a premature stop codon, while the zip5–2_cr mutant has a single T insertion that also introduces a premature stop codon (S7B and S7C Fig). We first assessed the growth of WT, zip3–1, zip5–1, zip3–1zip5–1 and the CRISPR-induced mutants zip3–2_cr and zip5–2_cr on soil. No significant differences in FW were observed between WT and the mutants (S7D and S7E Fig). When grown on Murashige and Skoog-based agar plates with sufficient Zn (15 µM Zn, Control), the zip3–1zip5–1 double mutant exhibited larger overall growth compared to WT and the single mutants, but no significant differences in root length or chlorophyll content were noted (Fig 5A–E). Under Zn-deficient conditions (-Zn), WT plants and the single mutants displayed increased root length, consistent with previous findings [32,39]. Strikingly, the double mutant showed around 40% reduction of shoot FW and chlorophyll content compared to WT under Zn deficiency, indicating heightened sensitivity to low Zn (Fig 5C–5E). Because the reduction in FW and chlorophyll content of the zip3zip5 double mutant remained modest after 10 days of growth compared to WT plants, and no phenotype was observed in the single mutants, we extended the treatment for an additional 10 days under both +Zn and -Zn conditions. As expected, after prolonged Zn deficiency, the zip3zip5 double mutant displayed reduced FW and chlorophyll content compared to WT, while the single mutants still showed no detectable phenotype (S7F–H Fig), suggesting that ZIP3 and ZIP5 might function redundantly in Zn acquisition.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Phenotype of the zip3, zip5 and zip3zip5 double mutant(A) Representative images of WT plants, zip3-1, zip5-1 and the double mutant zip3-1zip5-1 grown vertically for 10 days under control conditions (+Zn) and Zn-deficient conditions (-Zn).

Three representative plants from each genotype were transferred to an agar plate for imaging. Scale bars: 1 cm. (B) Zoom of WT and zip3-1zip5-1 shown in (A). scale bar: 0.5 cm. (C-E) Violin plots showing the distribution of (C) FW, (D) root length, and (E) chlorophyll content of plants grown as described in panel A (n ≥ 15 for C and D, n ≥ 5 for E). Dashed lines of the violin plots represent the median, while dotted lines mark the first and third quartiles. Different letters indicate statistically significant differences between genotypes, as determined by one-way ANOVA followed by Tukey’s post hoc test or Kruskal–Wallis test followed by Dunn’s test (P < 0.05).

https://doi.org/10.1371/journal.pgen.1011796.g005

To further support the role of ZIP3 and ZIP5 in Zn acquisition, we measured metals content in roots and shoots of 3-week-old plants grown on control Murashige and Skoog-based agar plates containing 15µM Zn. While Fe, Zn, Cu and Mn levels in shoots were similar between WT and zip3–1zip5–1, Zn levels in roots were reduced in the double mutant, confirming that ZIP3 and ZIP5 are involved in Zn acquisition (Fig 6A and 6B). Because Murashige and Skoog based medium contains high levels of Zn that could potentially distort metal homeostasis, we also measured shoot metal content in soil-grown plants. Interestingly, we observed lower Zn (and, to a lesser extent, Cu) levels in zip3zip5 double mutants’ shoots compared to WT, whereas Fe and Mn levels remained unaffected, confirming the role of ZIP3 and ZIP5 in Zn acquisition, especially in conditions where Zn concentration is low, and possibly under conditions involving evapotranspiration (Fig 6C). Transpiration and root pressure can indeed impact mineral translocation as shown particularly well for boron [40].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Ionomic analysis on the zip3zip5 double mutant.

(A-B) Metal content (Fe, Cu, Mn, Zn) in shoots (left) and roots (right) of WT and zip3-1zip5-1 mutants grown for 3 weeks on control agar plates. For each biological replicate, 3-4 plants were pooled (n = 4). Data are represented as mean ± SD, with significant differences between genotypes determined by Student’s t-test or Mann-Whitney test (*P < 0.05; **P < 0.01; ***P < 0.0001). (C) Metal content (Fe, Cu, Mn, Zn) in shoots of WT, zip3-1zip5-1 and zip3-2_crzip5-2_cr mutants grown for 3 weeks on soil. For each biological replicate, 3-4 plants were pooled (n = 4). Data are presented as mean ± SD, with different letters indicating statistically significant differences between genotypes, determined by one-way ANOVA followed by Tukey’s post hoc test (P < 0.05). (D) Root Zn uptake in WT and zip3-1zip5-1 double mutant before (-Zn) and after being resupplied with 10µM Zn for 3h (3h Zn) and 6h (6h Zn). For each biological replicate, 3-4 plant roots were pooled (n = 4). Data are presented as mean ± SD, with different letters indicating statistically significant differences between genotypes, determined by one-way ANOVA followed by Tukey’s post hoc test (P < 0.05). (E) Zn content in roots and shoots from WT, zip3-1, zip5-1 and zip3-1zip5-1 grown for 3 weeks on soil. Shoots and roots were harvested from the same plants. For each biological replicate, 3-4 plants were pooled (n = 4). Data are presented as mean ± SD, with different letters indicating statistically significant differences between genotypes, determined by one-way ANOVA followed by Tukey’s post hoc test (P < 0.05). (F). Representative confocal images of 5-day-old WT and zip3-1zip5-1 roots labeled with 20 µM Zinpyr-1 for 3 h. Plants were grown under Zn-deficient conditions for 5 days prior to staining. Upper panels show cross-sectional views and the lower panels show maximum projections of the differentiated root. Scale bars: 25µm. (G) Violin plots showing the distribution of endodermal Zinpyr-1 signal quantified in region of interest (ROI) in WT and zip3-1zip5-1 roots. At least five independent roots were analyzed (n ≥ 15). In the violin plots dashed lines represent the median and dotted lines represent the first and third quartiles. Statistical difference between genotypes was determined using Mann-Whitney test (*P < 0.05; **P < 0.01; ***P < 0.0001).

https://doi.org/10.1371/journal.pgen.1011796.g006

Given the accumulation of ZIP3 and ZIP5 in the vasculature, we wondered if they could contribute to Zn distribution to sink tissues. However, the similar root-to-shoot Zn translocation ratios between WT and zip3–1zip5–1 suggest that their contribution to Zn partitioning is limited and that other transporters are also likely involved (S8A Fig) [24]. In this respect, the higher mRNA abundance of IRT1, IRT3, ZIP9 and ZIP12 in the roots of zip3–1zip5–1 compared to WT could reflect a role of these transporters in root-to-shoot Zn translocation in the double mutant where they might take over ZIP3 and ZIP5 function in Zn distribution to sink tissues (S8B Fig). This is also supported by the activity of IRT1 and IRT3 promoters in the vasculature, which would be compatible with a role in root to shoot translocation (Fig 1C) [27,41].

To confirm the involvement of ZIP3 and ZIP5 in Zn uptake from the soil, we performed short-term zinc uptake assays using roots of WT and zip3–1zip5–1 double mutants grown under hydroponic conditions. Plants were cultivated for 20 days in Zn-replete hydroponic medium and then transferred to Zn-deficient medium for 1 week to deplete internal Zn stores and induce Zn acquisition systems. Upon Zn resupply, we measured Zn accumulation in roots after 3 and 6 hours. Six hours after resupply, Zn accumulation in zip3–1zip5–1 mutant was a threefold lower than in WT roots (Fig 6D). This result confirmed that ZIP3 and ZIP5 are required for efficient Zn uptake in roots.

To dissect the individual contributions of ZIP3 and ZIP5, we examined metal content in the single mutants. Since Zn accumulation defects were more visible on soil-grown plants compared to plants grown on Murashige and Skoog-based agar plates (Fig 6A–C), we analyzed Zn content of zip3 and zip5 mutants shoots and roots grown on soil. This analysis revealed that ZIP3 has a primary role over ZIP5 for the Zn deficiency observed in zip3–1zip5–1, as zip5–1 mutants accumulated Zn at levels comparable to WT (Figs 6E and S8C–E). In contrast to zip3 mutant, no significant differences in root and shoot Zn content were detected between WT and the zip5 mutants, both in Murashige and Skoog-based agar plates and soil conditions, suggesting that ZIP5 plays a more limited role in Zn acquisition (Figs 6E and S8D–F). Importantly, elevated Zn levels in the zip3–1/ZIP3::ZIP3-mCitrine grown on soil and to a lower extend in zip5–1/ZIP5::ZIP5-mCitrine roots grown on plates confirmed the functionality of the fusion proteins (Figs 2E-H, S8G, and S8H).

To gain further insight into the distribution of Zn in the zip3–1zip5–1 mutant, we used Zinpyr-1 fluorescent probe [11]. In WT differentiated roots, Zinpyr-1 fluorescence was primarily localized in the stele and endodermis, indicating Zn accumulation in these tissues. In contrast, fluorescence intensity was significantly reduced in the zip3–1zip5–1 double mutant, though the overall pattern of distribution remained unchanged (Fig 6F and 6G).

In addition, metal measurements and Zinpyr-1 staining in ZIP3 and ZIP5 overexpression lines provide further evidence to the role of ZIP3 and ZIP5 in metal accumulation (Fig 7A and 7B). Indeed, overexpression of ZIP3 using the 35S promoter leads to increased Zn content in roots from plants grown in Murashige and Skoog-based agar plates (+Zn). However, no overaccumulation of other metals typically transported by ZIP family members (e.g., Cu, Mn, Fe) was observed. Moreover, overexpression of both ZIP3 and ZIP5 lead to an increased Zn accumulation across root cell layers as visualized after Zinpyr-1 staining (Fig 7B). Overall, these data indicate that ZIP3 and ZIP5 are partially redundantly involved in Zn acquisition from the soil in Arabidopsis, with ZIP3 playing a more dominant role.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. ZIP3, ZIP5 and IRT1 mediate Zn acquisition.

(A) Metal content (Fe, Cu, Mn, Zn) in roots of WT, ZIP3 and ZIP5 overexpression lines grown for 3 weeks on control agar plates. For each biological replicate, 3-4 T1 plants were pooled (n = 4). Data are presented as mean ± SD, with different letters indicating statistically significant differences between genotypes, determined by one-way ANOVA followed by Tukey’s post hoc test (P < 0.05). (B) Representative confocal images of 5-day-old WT, 35S::ZIP3 (T1) and 35S::ZIP5 (T1) plants stained with 20 µM Zinpyr-1 for 3 h. Plants were grown for 5 days on 1 µM Zn prior to staining. Upper panels show a cross section of the differentiated root (scale bars: 25 µm), and lower panels show maximum projections from the root tip (scale bars: 200 µm). (C) Phenotype of WT, zip3-1zip5-1 and zip3-1zip5-1/ZIP_2/3/5::IRT1 lines. For each transgenic line, 9 independent T1 were grown per pot. (D) Metal content of the plants grown as described in panel C. For each biological replicate 3-4 independent T1 plants were pooled (n = 4). Different letters indicate statistically significant differences between genotypes for a given element using one-way ANOVA followed by Tukey’s post hoc test (P < 0.05).

https://doi.org/10.1371/journal.pgen.1011796.g007

The fact that irt1 mutants accumulate WT-like Zn levels in several studies, raises questions about the extent of IRT1’s contribution to Zn uptake in soil-grown plants [27,37]. To address this, we employed a promoter-swap approach, expressing IRT1 under ZIP2, ZIP3 and ZIP5 promoters (Figs 7C, 7D, S9A, and S9B). Plants expressing ZIP2::IRT1 and ZIP3::IRT1 in the zip3–1zip5–1 background restored Zn accumulation to WT levels in shoots, confirming IRT1’s capacity for Zn transport in planta (Fig 7D). Overexpression of IRT1 (35S::IRT1) leading to an increase in root Zn levels further corroborated this finding (S6G Fig). However, ZIP5::IRT1 failed to rescue Zn levels in zip3–1zip5–1 shoots, likely due to the lack of ZIP5 promoter activity in the vasculature under our conditions (Fig 2C). Aligning with these data, the root length of zip3–1zip5–1 mutants expressing ZIP2::IRT1 and ZIP3::IRT1 were similar to WT plants in -Zn conditions, contrary to ZIP5::IRT1 expression in zip3–1zip5–1 where no complementation was observed.

Overall, these results suggest that Zn acquisition predominantly depends on ZIP3 activity at the root periphery and that ZIP5 and IRT1 can also contribute to Zn acquisition in certain conditions.

Zn excess triggers ZIP3 endocytosis

Nutrient transporters are frequently regulated at multiple levels, including transcriptional, post transcriptional and post translational mechanisms, in response to nutrient availability [19,34,37,42–44]. To investigate if Zn availability influences ZIP3 stability in the root, we exposed UBQ10::ZIP3-mCitrine plants to 150 µM Zn, a 10-fold excess relative to our standard conditions. We first used Propidium Iodide (PI) to control that expositing the plants to 150 µM Zn for 16 h do not cause cell death at the root tip, which could interfere with the localization and expression of ZIP3-mCitrine (S10A Fig). Since we wanted to focus on Zn post transcriptional effect on ZIP3, we used the constitutive UBQ10 promoter, to bypass potential transcriptional regulation of ZIP3 and focus on post-transcriptional effects. ZIP3 was chosen because of its predominant role in Zn acquisition from the soil. Confocal microscopy revealed that, despite the Zn excess, the overall ZIP3-mCitrine fluorescence did not significantly decrease compared to Zn-deficient condition (Fig 8A and 8B). This suggests that ZIP3 protein abundance is not markedly affected by Zn levels, or that higher Zn concentrations might be required to influence its stability. Although ZIP3-mCitrine accumulation was unchanged, we hypothesized that Zn availability might impact ZIP3 at the post-translational level affecting its subcellular localization, as it is well-established for other transporters that undergo endocytosis in response to excess substrates [34,36,43,45]. Previous colocalization with FM4–64 indicated that ZIP3 is present both at the PM and within the endocytic system (S4E Fig). We examined ZIP3 localization in the root tip, an optimal region for subcellular compartment imaging due to its small cell size. For endosomes detection, we imaged the surface focal plane and, for assessing ZIP3 polarity, we focused on the median plane of epidermal cells. Plants were grown in the presence of 15 µM Zn for 5 days, then exposed to either Zn deficiency (0 µM Zn, -Zn) or a 10-fold Zn excess (150 µM Zn, 10xZn) for 16 hours. Under -Zn conditions, ZIP3-mCitrine predominantly localized to the PM, displaying outer polarization, as described earlier (Figs 2C, 2F, and 8C). However, upon exposure to excess Zn, the intracellular ZIP3 signal increased while PM localization diminished (Fig 8C and 8D). Quantification of the PM-to-intracellular (PM/Intra) ratio of ZIP3-mCitrine revealed a significant reduction in this ratio, indicating delocalization of ZIP3 in response to Zn excess (Fig 8E). Notably, we also observed that ZIP3 internalization is triggered at lower Zn concentration, as indicated by a Zn concentration-dependent reduction in the ZIP3 PM/intracellular ratio, with stronger effects observed at higher Zn levels (100 and 150 µM, S10B and S10C Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. ZIP3 localization depends on Zn availability.

(A-G) UBQ10::ZIP3-mCitrine seedlings were grown on agar plates containing Zn (15 µM) for 5 days, then transferred to liquid medium for 16 h under either Zn-deficient conditions (-Zn) or with a 10-fold excess of Zn (10x Zn; 150 µM) prior to imaging. (A) Representative maximum projections of 5-day-old roots expressing UBQ10::ZIP3-mCitrine. The mCitrine signal is shown in yellow. Scale bars: 125 µm. (B) Quantification of ZIP3-mCitrine signal intensity from maximum projection of several independent roots, as shown in panel A (n ≥ 5). Data are represented as mean ±SD. No significant difference between conditions was observed using Mann-Whitney test (P < 0.05). (C, D) Close-up views of the elongation zone in roots shown in panel A, with images captured at the median focal plane of the epidermis (C) and at the surface of the epidermis (D). The mCitrine signal is displayed using the Fire (LUT) color scale. Scale bars: 12.5 µm. (E) Quantification of the ratio of plasma membrane (PM) to intracellular (intra) of ZIP3-mCitrine signal (n ≥ 20). Data are represented as mean ± SD, and statistical differences between conditions were determined using Mann-Whitney test (*P < 0.05; **P < 0.01; ***P < 0.0001). (F) Surface views of epidermal cells from plants grown as described above and subsequently exposed to 100 µM CHX and 50 µM BFA prior to imaging. The mCitrine signal is displayed using the Fire (LUT) color scale. Scale bars: 25 µm. (G) Quantification of the ratio of PM to intracellular (intra) of ZIP3-mCitrine signal (n ≥ 20) in plants treated as described in panel F. Data are represented as mean ± SD, and statistical differences between conditions were determined using Mann-Whitney test (*P < 0.05; **P < 0.01; ***P < 0.0001).

https://doi.org/10.1371/journal.pgen.1011796.g008

To determine if this intracellular signal was due to ZIP3 endocytosis upon Zn excess, we treated the plants with Brefeldin A (BFA), a fungal toxin that inhibits vesicle recycling between endosomes and PM. Since BFA also blocks the secretion of newly synthesized PM proteins, we pre-treated plants with cycloheximide (CHX) to inhibit protein synthesis, enabling us to track only the pool of already translated ZIP3-mCitrine and thus focus specifically on post-translational effects. Under -Zn conditions, ZIP3-mCitrine was found at the PM as well as in BFA-induced intracellular bodies, suggesting that ZIP3 cycles between the PM and endocytic compartments under Zn deficiency (8F and S10D Figs). Following Zn excess, the majority of the ZIP3-mCitrine signal accumulates in BFA bodies, as indicated by the reduced PM/Intra ratio (Fig 8F and 8G). This confirms that ZIP3 undergoes rapid endocytosis in response to Zn excess, likely as a protective mechanism to prevent excessive Zn acquisition as previously shown for IRT1 [34,37,38]. Whether ZIP3 is subsequently targeted for vacuolar degradation remains an open question. Interestingly, unlike IRT1 whose localization is regulated by its secondary substrates, ZIP3 localization appears to be modulated at the post translation level by the availability of its primary substrate, Zn.

Plant material

All experiments were conducted using Arabidopsis thaliana wild type (WT) Columbia-0 background. Mutants and transgenic plants used in this study are as follow: zip2–1 (SALK_094937, [18]), zip3–1 (salk_35_B08, [63]), zip5–1 (SALK_009007C, [29]), all of which were obtained from the Nottingham Arabidopsis Stock Center (NASC). The double mutant zip3–1zip5–1 was generated by crossing the single mutants zip3–1 and zip5–1 and the mutant irt1–2 was obtained from [64]. Additionally, the following mutants were generated for this study using CRISPR-Cas9 technology: zip2–2_cr, zip3–2_cr, zip5–2_cr, zip8–1_cr and zip8–2_cr. The double mutant zip3–2_crzip5–2_cr was generated by crossing the two single zip3–2_cr and zip5–2_cr. Transgenic lines generated for this research include: pIRT1::NLS-3xmVENUS, pIRT2::NLS-3xmVENUS, pIRT3::NLS-3xmVENUS, pZIP1::NLS-3xmVENUS, pZIP2::NLS-3xmVENUS, pZIP3::NLS-3xmVENUS, pZIP4::NLS-3xmVENUS, pZIP5::NLS-3xmVENUS, pZIP6::NLS-3xmVENUS, pZIP7::NLS-3xmVENUS, pZIP8::NLS-3xmVENUS, pZIP9::NLS-3xmVENUS, pZIP10::NLS-3xmVENUS, pZIP11::NLS-3xmVENUS, pZIP12::NLS-3xmVENUS, pZIP2::ZIP2-mCitrine, pZIP3::ZIP3-mCitrine, pZIP5::ZIP5-mCitrine, pZIP8::ZIP8-mCitrine, UBQ10::ZIP2-mCitrine, UBQ10::ZIP3-mCitrine, UBQ10::ZIP5-mCitrine, UBQ10::ZIP8-mCitrine, pZIP2::IRT1, pZIP3::IRT1, pZIP5::IRT1, p35S::ZIP3, p35S::ZIP5 and p35S::IRT1. The construct pIRT1 was ordered from NASC (N2106308).

Constructs

Plasmids were generated using Gibson and Multisite Gateway cloning techniques (Thermo Fisher Scientific). A list of primers used for cloning is provided in S3 Table. All DNA were amplified using Phusion Plus DNA Polymerase (ThermoFischer) following manufacturer instructions. ZIP promoter sequences upstream of the ATG start codon were amplified from Arabidopsis Col-0 genomic DNA and inserted into a modified pDONRP4-P1R (Thermo Fisher Scientific) via Gibson assembly. Briefly, the ccdB and the CmR cassettes between the AttP1R and attP4 recombination sites were removed in pDONRP4-P1R (Thermo Fisher Scientific) and replaced by EcoRV, BglII, XbaI, BamhI cloning sites (S11A Fig). For the generation of promoter-reporter constructs (promoter::NLS-3xmVenus), entry plasmids containing the promoter regions were recombined with pDONRL1-NLS-3xmVenus-L2 and pEN-R2-tHSP18.2-L3 into the destination vector pFR7m34GW [65]. For yeast complementation assays, two ZIP8 cDNA versions (ZIP8–1: 888 bp and ZIP8–2: 897 bp) predicted from Aramemon (ARAMEMNON, plant membrane protein database release 8.1) were synthetized by Genewiz. IRT1 cDNA was cloned using specific primers. IRT1 and ZIP8 cDNA were then inserted by Gibson in the pDR196 yeast expression vector [66]. To study endogenous protein localization, ZIP genomic sequences were amplified from Col-0 genomic DNA and inserted by Gibson in a modified pDONR221, containing cloning sites instead of ccdB and CmR cassettes (S11B Fig). The mCitrine fluorescent protein was inserted into the second extracellular loop of each ZIP transporter using Gibson assembly [34]. For ZIP2-mCitrine, mCitrine was inserted 153 bp downstream of the ATG start codon; for ZIP3-mCitrine, it was inserted 95 bp from ATG; for ZIP5-mCitrine, 89 bp from ATG; and for ZIP8-mCitrine, 87 bp from ATG. The final destination vectors for expression in plants were obtained using multisite Gateway recombination system (Life Technologies), employing the pFR7m34GW destination vector along with the various pDONR221-ZIP-mCitrine entry clones to generate the ZIP::ZIP-mCitrine constructs. For promoter swap analysis, entry vector containing the promoters of ZIP2, ZIP3 and ZIP5 were recombined with pDONRL1-gIRT1-L2 and pEN-R2-tHSP18.2-L3 into the destination vector pFR7m34GW. Vectors carrying ZIP2/3/5::IRT1 were transformed in zip3–1zip5–1 background to evaluate IRT1 function in Zn acquisition. For CRISPR/Cas9 mutant generation, single guide RNA (sgRNA) for Cas9 were designed using webtool Benchling (https://www.benchling.com). Pairs of annealed oligonucleotides of the sgRNA were cloned into the BbsI linearized entry vector recombined into the destination vector containing Cas9 expression cassette controlled by the PcUBi4-2 promoter and a FastRed selection marker [67]. To generate the zip2–2_cr mutant, a single sgRNA targeting the 5’ genomic region was combined with the ZCas9i controlled by the pEC1.2 promoter [68]. For the generation of the zip3–2_cr, zip5–2_cr, zip8–1_cr and zip8–2_cr, triple sgRNA targeting the 5’ genomic region were employed with Cas9. All constructs were sequenced (Sanger DNA sequencing) in pDONR and in the destination plasmid before transformation into Agrobacterium strain GV3101 by electroporation and were subsequently used for Arabidopsis plant transformation via the floral-dip method. T1 plants were first selected based on the red fluorescence of transformed seeds and fluorescent mono-insertional T2 were subsequently selected. All experiments were performed on mono-insertional T2 after red fluorescence sorting unless otherwise specified in figure legends. CRISPR-induced mutation experiments were performed on T3 or T4 homozygous lines.

Growth conditions

For in vitro culture, seeds were surface sterilized in 70% ethanol for 5 minutes with agitation, followed by 3 rinses with absolute ethanol before drying. The sterilized seeds were then sown in lines (density of 12–20 seeds depending of the experiment) at 1.5 cm from the upper size of 12x12 cm square plates containing 50 ml half-strength Murashige and Skoog (MS) based media with 0.8% agar (Duchefa) and no sucrose (pH 5.7, adjusted with KOH). For metal deficiency or sufficiency, plants were grown in the presence (+Zn, + Cu, + Fe + Mn) or absence (-Zn, -Cu,-Fe, -Mn) of 15 µM Zn, 0.05 µM Cu, 50 µM Fe and 50 µM Mn provided as ZnSO₄, CuSO₄, Fe-EDTA and MnSO₄ respectively. For –metal conditions, the respective metal was not added to the medium, but trace amounts are likely present in the agar. Additional nutrients were supplied at following concentration: 9.4 mM KNO₃, 10.3 mM NH₄NO₃, 625 mM KH₂PO₄, 1.49 mM CaCl₂, 0.75 mM MgSO₄, 50 µM H₃BO₃, 0.5 µM Na₂MoO₄, 2.5 µM KI, 0.05 µM CoCl₂. After stratification in the dark for 3 days at 4°C, the plates were transferred to growth chamber under continuous day conditions (light intensity ~100 μE) at 22 °C with ~50% humidity and the plants grown vertically. For physiological analysis, root length, FW and chlorophyll contents were assessed in 10-day-old plants grown on control or metal-deficient agar plates. For elemental analysis, 12 plants per plate or 6 plants per pot were grown either on agar control media for three weeks or in soil for the same duration before harvesting. For soil grown plants, plants were harvested at Zeitgeber time (ZT 8) (8 h after lights on). For mRNA relative quantification, twenty 7-day-old seedlings per plate were directly sown and vertically grown on control or metal deficient plates before harvesting. Live-microscopy analyses were conducted on 5-day-old seedlings. For promoter activity, expression analysis and protein localization, plants were directly sown and grown for 5 days on control or metal-deficient agar plate. For short-term metal deficiency or excess treatments, 5-day-old plants were transferred to liquid medium without Zn (-Zn) or with an excess of Zn (10xZn; 150 µM of Zn provided as ZnSO₄) for 16 hours before imaging. CHX was applied to plants at a final concentration of 100 μM for 1 hour before treatment with 50 μM BFA, which was maintained for the duration of the BFA treatments (3 hours). For amplification and experiments in soil, plants were grown in long-day conditions (16 hours of light, 8 hours of darkness) with a light intensity of 150–180 μE at 60–70% humidity, and at 20 ± 2 °C.

Short time Zinc uptake assay

For short time Zn uptake assay, plants were hydroponically grown in a controlled growth chamber (22°C, 70% relative humidity, ~ 100 μE light intensity) under long day conditions (16 hours of light, 8 hours of darkness). A density of one plant per 80 ml of MS/2 solution was used (same composition and pH as described for in vitro conditions). Following 3 weeks of growth, in presence of Zn, plants were transferred to a MS/2 solution without Zn and grown for 1 week in this Zn-deficient condition. This Zn deprived media was renewed once after 4 days. After 1 week of Zn-deficiency, plants were incubated in a MS/2 solution containing 10 µM ZnSO₄ for 3 and 6 hours, before roots were harvested. Roots were washed and analyzed as described for the ionomic analysis.

Confocal Microscopy and image analysis

Cell walls were stained with a 10 μg.ml^-1 solution of propidium iodide (PI; Sigma-Aldrich, Cat# P4170) for 10 minutes, while the PM and endocytic compartments were stained with a 4 µM solution of the styryl dye FM4–64 (ThermoFischer scientific, Cat# T13320) for 10 minutes. Zn visualization was conducted with 10 µM Zinpyr-1 (Abcam, Cat# 145349) staining as described by [11]. To prevent excessive Zn accumulation in the roots and minimize fluorescent signal saturation, plants were grown under control conditions with 1 µM Zinc (+Zn) instead of 15 µM. Confocal laser scanning experiments were conducted using a Zeiss LSM 780 system. The excitation (ex) and detection (em) settings were configured as follows: mVenus/mCitrine ex: 514 nm, em: 515–570 nm; propidium iodide (PI) ex: 514 nm, em: 586–679 nm; Zinpyr-1: ex: 488nm, em: 495–535 nm; and FM4–64 ex: 514 nm, em: 650–742 nm. Images were captured using a 40X objective for Z-stack imaging and a 20X dry objective for longitudinal views. All images were processed using the Fiji software (LUT, orthogonal views, maximum projections, merge, fluorescence quantification, etc) [69]. For Z-stack images of promoter-reporter lines, protein localization and Zinpyr-1 staining, the Z-stack images were resliced vertically to generate transversal views. For promoter reporter lines only, a maximum projection of the mVenus marker channel was then merged (3D; X, Y, Z) with a representative single stack of the PI-stained cell wall channel (2D; X, Y). For T1 lines, at least five independent lines were observed (selected based on red seed fluorescence) and for T2 lines, at least two independent mono-insertional lines were observed. In all case a representative line is shown in the manuscript. For Fig 8, PM/intra signal were calculated as described in [70]. The fluorescence intensity of Zinpyr-1 in the endodermis was determined using identical regions of interest (ROI). The Zinpyr-1 signal in three to four endodermal cells from at least five independent roots was measured to obtain the average Zinpyr-1 fluorescence intensity.

qRT-PCR

For gene expression analysis, seedlings were sown (20 per plate) and vertically grown for 7 days. Roots from approximately forty seedlings were harvested and pooled to form one biological replicate. RNAs were extracted using a TRIzol-adapted RNeasy MinElute CleanupKit (Qiagen). Subsequently, RNAs were reverse transcribed with a Thermo Scientific Maxima First Strand cDNA Synthesis Kit following the manufacturer’s protocol. Quantitative Real-time PCR was conducted on an Applied Biosystems QuantStudio5 thermo-cycler using Applied Biosystems SYBR Green master mix. The relative expression of ZIP genes was quantified with the 2^−ΔΔCt method with Clathrin as a reference gene (AT5G46630). A list of primers used for qRT-PCR is provided in S3 Table, the primer efficiency is provided in S4 Table and Clathrin expression in S5 Table. Each experiment included 3 technical and at least three biological replicates.

Ionomic analysis

Three weeks-old seedlings (roots and shoots) were used to quantify metal content. For shoots and roots from Murashige and Skoog-based agar plates and shoot from soil conditions, samples were harvested separately and desorbed by washing in 10 mM Na₂EDTA (pH 5.7) for 10 minutes, followed by 3 rinses with ultrapure Milli-Q water for 1 minute each. For soil grown roots analysis, 3 weeks old plants were roughly washed with distilled water to remove most of the soil attached to the roots. Roots were then carefully washed with MilliQ water. Roots were finally immersed in 4°C MilliQ water and soil particles were removed from roots using tweezers until fully cleaned. Once cleaned, roots were desorbed as described above. After desorbing, the tissues were blotted dry with paper towels and placed at 65°C overnight. Once dried, the tissues were ground, and 5–10 mg of dry tissue per biological replicate was mixed with 750 µl of nitric acid (65% [v/v]) and 250 µl of hydrogen peroxide (30% [v/v]). The samples were left at room temperature overnight and then mineralized at 85°C for 24 hours. Following mineralization, 4 ml of Milli-Q water was added to each sample. Four to five biological replicates were analyzed for each genotype. Elemental analysis was performed using inductively coupled plasma optical emission spectroscopy (ICP-OES 5800; Agilent Technologies) using calibration standard solutions provided by the manufacturer. The ICP-OES is calibrated in wavelength each time the machine is switched on, using a 5 ppm calibration solution supplied by the manufacturer (ICP-OES&MP-AES wave cal 6610030100, Agilent Technologies) and performance tests are proceeded (optics, sensitivity and resolution). Quality control was performed with Certified Reference Material strawberry leaf powder (LGC7162), corresponding perfectly to the type of matrices usually handled. Atomic emission wavelengths are selected using the precision obtained from CRM measurements within +/- 10% of the theoretical value. The following elements quantified with wavelength: calcium (Ca 393.366nm), copper (Cu 324.754nm), iron (Fe 238.204nm), potassium (K 766.491nm), magnesium (Mg 280.270nm), manganese (Mn 257.610nm), molybdenum (Mo 202.032nm), sodium (Na 589.592nm), strontium (Sr 407.771 nm) and zinc (Zn 202.548nm).

Yeast complementation assay

For yeast complementation, the fet3fet4 double mutant was used [71].The fet3fet4 strain was transformed using one-step transformation protocol [72], either with the empty pDR196 vector, or pDR196 vector containing IRT1 or ZIP8 cDNAs. Positive clones were recovered on -URA (Uracil) selection medium containing 0.078% CSM-URA (Complete Supplement Mixture), 2% glucose, 0.67% YNB (yeast nitrogen base without amino acids) and 2% agar, and verified by PCR. For drop tests, yeast clones were incubated overnight at 28°C in an agitating incubator in 2ml of -URA liquid media containing 50 µM FeCl₃. Cells were then collected, washed twice with sterile MilliQ water, the OD measured and adjusted to the same for all constructs before realizing serial dilutions. For each dilution, 5 μl drops were deposited on the same -URA selection media with or without 10µM BPDS (bathophenanthroline disulfonic acid disodium salt). Yeasts were incubated 5 days at 28°C and pictures taken with an Epson perfection V800 scanner.

Chlorophyll quantification and F_v/F_m measurements

Chlorophylls were extracted from 3 weeks (only for Fig 1C) or 10-day-old plants tissues by incubating sample overnight at 4°C in 10 volumes of dimethylformamide (DMF) (v/w). Following extraction, samples were centrifugated, and chlorophyll content was quantified using a NanoDrop spectrophotometer at wavelength 647 nm and 664 nm. Chlorophyll concentrations were calculated according to the methods previously described by [73]. For Fv/Fm measurement, plants were dark-adapted for 15 minutes before measurements were taken with a FluorCam 800 MF, as described in [74]. The maximum quantum efficiency of photosystem II (F_v/F_m) was measured by applying saturating pulses of 2000 µmol m⁻² s⁻¹ (white light, 400–720 nm) for 960 ms, followed by detection using orange-red light pulses (620 nm,10 µs). All measurements were performed at room temperature.

Phylogenetic tree

Protein sequences for Arabidopsis thaliana ZIP family members were retrieved from TAIR (https://www.arabidopsis.org) or Aramemnon (https://aramemnon.botanik.uni-koeln.de/). Phylogenetic relationships were defined using Phylogeny.fr (http://www.phylogeny.fr/).

AlphaFold and AlphaFold2 protein interaction prediction

Wild type ZIP and ZIP-mCitrine models were predicted using AlphaFold (a) (https://alphafold.ebi.ac.uk/) or the ColabFold v1.5.5: AlphaFold2 using MMseqs2 (b) [75–77]. The number of recycles were set to 3 for ColabFold predictions.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism software or within the R environment (R Core Team 2023). Data were first assessed for normality and homoscedasticity using the Shapiro-Wilk test and Brown-Forsythe test, respectively. For data meeting these assumptions, pairwise comparisons were performed using a two-tailed Student’s t-test, while multiple comparisons were conducted using a one-way ANOVA followed by a Tukey’s post hoc test. When the assumptions of normality or homoscedasticity were violated, non-parametric tests were employed. In such cases, pairwise comparisons were carried out using the two-tailed Mann-Whitney U test, and multiple comparisons were analyzed using the Kruskal–Wallis test followed by Dunn’s post hoc test or a non-parametric version of Tukey’s test (R). A significance threshold of P < 0.05 was applied in all tests. All experiments were repeated at least 3 independent times except for experiments in S5B and S7A Figs and representative experiments are presented in the manuscript. In all the manuscript, we used the term biological replicates to refer to independently collected pools of plants (ranging from a minimum of 3–4 plants to up to 40 seedlings, depending on the experiment, as specified in the figure legends). These pools were harvested from the same experiment, in which plants were grown and samples analyzed in parallel.

Accession numbers

The sequence data used in this article (for generating constructs and for qRT-PCR) can be found in the GenBank/EMBL databases under the following numbers: IRT1 (AT4G19690), IRT2 (AT4G19680), IRT3 (AT1G60960), ZIP1 (AT3G12750), ZIP2 (AT5G59520), ZIP3 (AT2G32270), ZIP4 (AT1G10970), ZIP5 (AT1G05300), ZIP6 (AT2G30080), ZIP7 (AT2G04032), ZIP8 (AT5G45105), ZIP9 (AT4G33020), ZIP10 (AT1G31260), ZIP11 (AT1G55910), ZIP12 (AT5G62160).