AtIAR1 is a Zn transporter that regulates auxin metabolism in Arabidopsis thaliana

Abstract

Root growth in Arabidopsis is inhibited by exogenous auxin–amino acid conjugates, and mutants resistant to one such conjugate [indole-3-acetic acid (IAA)–Ala] map to a gene (AtIAR1) that is a member of a metal transporter family. Here, we test the hypothesis that AtIAR1 controls the hydrolysis of stored conjugated auxin to free auxin through zinc transport. AtIAR1 complements a yeast mutant sensitive to zinc, but not manganese- or iron-sensitive mutants, and the transporter is predicted to be localized to the endoplasmic reticulum/Golgi in plants. A previously identified Atiar1 mutant and a non-expressed T-DNA mutant both exhibit altered auxin metabolism, including decreased IAA–glucose conjugate levels in zinc-deficient conditions and insensitivity to the growth effect of exogenous IAA–Ala conjugates. At a high concentration of zinc, wild-type plants show a novel enhanced response to root growth inhibition by exogenous IAA–Ala which is disrupted in both Atiar1 mutants. Furthermore, both Atiar1 mutants show changes in auxin-related phenotypes, including lateral root density and hypocotyl length. The findings therefore suggest a role for AtIAR1 in controlling zinc release from the secretory system, where zinc homeostasis plays a key role in regulation of auxin metabolism and plant growth regulation.

Regulation of intracellular zinc compartmentation by IAR1 plays a key role in the control of auxin activity through zinc-sensitive release of IAA from IAA conjugate stores.

Introduction

Auxin (indole-3-acetic acid, IAA) is a major plant hormone involved in many developmental processes in plants, including embryonic (Schiavone and Cooke, 1987; Liu et al., 1993) and post-embryonic development through maintenance of the root apical meristem (Ioio et al., 2008) and the shoot apical meristem (Reinhardt et al., 2000; Vernoux et al., 2011). Both polar transport and metabolism operate as major controls on auxin action.

Polar transport between and within cells and tissues is used to create concentration gradients which coordinate specific developmental patterns. In order to create auxin concentration gradients, various auxin transporters are required including the polarly localized long PINFORMED (PIN) auxin exporters (Okada et al., 1991; Friml et al., 2002). In addition, subcellular compartmentalization of IAA is also controlled by transporters. Transporters such as short PINs and PIN-LIKES (PILS) family members localize to the endoplasmic reticulum (ER) and coordinate IAA transport between the ER and cytosol (Mravec et al., 2009; Barbez et al., 2012; Simon et al., 2016).

The major routes for synthesis of IAA originate with Trp and then proceed either via indole-3-acetaldoxime (IAOx) and indole-3-acetonitrile (IAN) (Normanly et al., 1997; Hull et al., 2000; Mikkelsen et al., 2000) or via indole-3-pyruvic acid (IPA). The route via IPA involves TAA/TAR and YUCCA enzymes (Mashiguchi et al., 2011; Stepanova et al., 2011; Won et al., 2011), and severe developmental phenotypes in mutants suggest that the pathway involving YUCCA is the major pathway to synthesize IAA (Cheng et al., 2007; Stepanova et al., 2008).

The concentration of active IAA is then also controlled through its conjugation into inactive storage forms and degradation via oxidation (Mellor et al., 2016). The role and importance of conjugates in IAA concentration modulation and auxin action is currently poorly understood. These inactive conjugates can be irreversibly or reversibly formed. Irreversibly formed conjugates include acidic amino acid-linked IAAs such as IAA–Asp and IAA–Glu (Östin et al., 1998) formed from IAA through the action of enzymes in the Gretchen Hagen 3 (GH3) family (Staswick et al., 2005; Di Mambo et al., 2019). Reversibly formed conjugates are hypothesized to act as a temporary inactive reservoir of IAA, and include ester-linked IAAs such as IAA–glucose (IAA–Glc) (Jackson et al., 2001; Jin et al., 2013), methylated IAA (Qin et al., 2005; Yang et al., 2008), and some IAA–amino acid conjugates (IAA–Ala, IAA–Leu, and IAA–Phe among others). In the case of these non-acidic IAA–amino acids, the conjugation is reversible through the activity of IAA–amino acid amidohydrolases (Bartel and Fink, 1995; Davies et al., 1999; LeClere et al., 2002; Sanchez Carranza et al., 2016).

In Arabidopsis, IAA–amino acid hydrolysis to IAA is governed by four of a seven-member Ilr1-like amidohydrolase family (LeClere et al., 2002). The four enzymes shown to have IAA–amino acid hydrolysis activity are IAA-LEUCINE RESISTANT 1 (ILR1) (Bartel and Fink, 1995), ILR-LIKE 1 (ILL1), ILL2, and IAA-ALANINE RESISTANT 3 (IAR3) (Davies et al., 1999), all of which show enhanced in vitro activity with increased Mn and Co and are inhibited by Zn ions (LeClere et al., 2002). The IAA–amino acid-hydrolysing enzymes were all determined to be localized to the ER through bioinformatic, proteomic, and green fluorescent protein (GFP) tagging analysis (Bartel and Fink, 1995; Davies et al., 1999; LeClere et al., 2002; Sanchez Carranza et al., 2016).

High concentrations of IAA conjugates elicit a characteristic shortening of root lengths in wild-type plants (Bartel and Fink, 1995; Lasswell et al., 2000). To identify genes that control auxin conjugate hydrolysis, screening conditions were therefore established with high concentrations of IAA–amino acid conjugates to find mutants insensitive to the effects of IAA conjugates (Bartel and Fink, 1995; Lasswell et al., 2000). Among many mutants identified, these screens revealed genes involved in metal homeostasis, such as IAA-Leu resistance 2 (AtILR2) which when mutated increased Mn sequestration (Magidin et al., 2003), and AtILR3 (Rampey et al., 2006) which is involved in Fe deficiency (Li et al., 2016) and Fe excess responses (Tissot et al., 2019). Mutants insensitive to IAA–Ala resulted in identification of the IAA-Ala resistance1 (AtIAR1) gene (Lasswell et al., 2000). Seven Atiar1 mutants with a range of phenotypic severity were characterized. The Atiar1 mutants were all shown to be less sensitive to all IAA–amino acid conjugates tested including not only IAA–Ala, but also IAA–Leu and IAA–Phe, among others. High auxin Atsuperroot1 plants (Boerjan et al., 1995; Celenza et al., 1995) had their high auxin phenotype of reduced root growth partially suppressed after 10 d in Atsuperroot1 Atiar1 double mutants, suggesting that Atiar1 mutants were low in auxin (Lasswell et al., 2000).

Evidence available at the time of the discovery of AtIAR1 and subsequently has indicated that ion homeostasis might play a role in IAA–Ala insensitivity of Atiar1 mutants. Increasing the Mn content of the media gradually suppressed this IAA–amino acid insensitivity in Atiar1 plants (Lasswell et al., 2000). AtIAR1 was identified as a member of the ZRT IRT1-related protein (ZIP) transporter family, and expression of the AtIAR1 mouse homologue MmKE4 (MmZIP7) complemented the Atiar1 phenotype (Lasswell et al., 2000). Further work has shown that MmZIP7 functions as a Golgi-localized Zn transporter in mice (Huang et al., 2005), suggesting that the Atiar1 phenotype is metal transport related.

In addition, unlike the Atiar3 and Atilr1 hydrolase mutants, the Atiar1 phenotype was partially compensated by mutations in the METAL TOLERANCE PROTEIN 5 gene (AtMTP5) (Rampey et al., 2013). AtMTP5 was predicted and later established (Fujiwara et al., 2015) to act as a cytoplasmic Zn exporter, which is hypothesized to act antagonistically to AtIAR1 in the secretory pathway to control the activity of IAA–amino acid amidohydrolases (Rampey et al., 2013).

AtIAR1 is therefore thought to play a key role in controlling hydrolysis of inactive auxin conjugates, and hence buffering of free auxin. However, the physiological significance of this control, the metal specificity of AtIAR1, and the relationship of AtIAR1 to auxin metabolism and auxin-related phenotypes remains to be established. Here, we investigate these relationships and show a major role for AtIAR1 in linking Zn homeostasis to regulation of auxin activity.

Materials and methods

Phylogenetic analysis of the ZIP transporter family

To determine the relationship of AtIAR1 to other plant ZIPs as well as to the wider ZIP transporter family, a phylogenetic tree was constructed. Amino acid sequences of ZIP proteins were gathered using the tBLASTn function on the BLAST server (Boratyn et al., 2013) using the amino acid sequences of AtIAR1 (Uniprot accession no. Q9M647) and AtIRT1 (Uniprot accession no. Q38856) as the queries and collecting sequences from organisms within animal (Drosophila melanogaster and Homo sapiens), plant (Arabidopsis thaliana, Oryza sativa, and Medicago truncatula) and fungal (Saccharomyces cerevisiae) kingdoms. These sequences were aligned and phylogeny calculated within the MEGA-X software (Kumar et al., 2018). For alignment, MUSCLE (Edgar, 2004) was used, utilizing default settings and phylogeny calculated with the Maximum Likelihood method and the Jones–Taylor–Thornton (JTT) matrix-based model (Jones et al., 1992) with a gamma distribution of five categories to model different evolutionary rates at different sites. The phylogenetic tree was then constructed from 100 bootstrap replicates and initial trees using BioNJ and Neighbor–Joining algorithms on the JTT-generated matrix.

Saccharomyces cerevisiae strains and constructs for expression

AtIAR1 (Arabidopsis thaliana IAA-Ala resistant 1) functionality including transport capabilities was assessed by transforming AtIAR1-based constructs into S. cerevisiae transporter mutants using BY4741 and DY1457 strains as controls as appropriate. Details of these mutants are shown in Table 1.

Table 1.

List of Saccharomyces cerevisiae strains used in this study

Strain	Description	Genotype	Source
BY4741	Background for smf1Δ	MATa; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0	Podar et al. (2012)
smf1Δ	Mn import mutant	BY4741; MATa; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0; YOL122c::kanMX4	Euroscarf
DY1457	Background for zrt1Δ zrt2Δ and fet3Δ fet4Δ	DY1457; MATa; ade1/+; can1; his3; leu2; trp1; ura3	Evens et al. (2017)
zrt1Δ zrt2Δ	Zn import mutant	DY1457; MATa; ade1/+; can1; his3; leu2; trp1; ura3; zrt1::LEU2, zrt2::HIS3	Evens et al. (2017)
fet3Δ fet4Δ	Fe import mutant	DY1457; MATa; ade1/+; can1; his3; leu2; trp1; ura3; fet3-2::HIS3, fet3-1::LEU2	Evens et al. (2017)

To assess the function of AtIAR1, constructs for heterologous expression of AtIAR1 in S. cerevisiae were assembled into a pYES2 (ThermoFisher, V82520) vector, allowing for selection in bacterial and yeast hosts with ampicillin resistance and uracil autotrophy, respectively. Constructs were synthezised (Integrated DNA Technologies) to represent codon-optimized AtIAR1 cDNA from the A. thaliana Col-0 ecotype as well as for the Atiar1-3 mutant identified by Lasswell et al. (2000).

Saccharomyces cerevisiae growth and transformation

Growth of non-transformed S. cerevisiae strains was conducted in YPAD medium with or without agar. YPAD medium consisted of 1% (w/v) yeast extract (Millipore, 70161), 2% (w/v) peptone (Formedium, PEP02), 2% (w/v) glucose (Fisher Scientific, G/0500/53), 0.002% (w/v) adenine (Sigma-Aldrich, A2786), and 2% (w/v) agar (Formedium, AGA03) if required.

Once transformed with recombinant plasmids, S. cerevisiae gained uracil autotrophy and were grown on synthetic complete medium lacking uracil (SC-U). In addition, expression of the AtIAR1-based construct was controlled by glucose or galactose addition. Unless otherwise stated, SC-U was made using glucose to repress expression of the AtIAR1-based construct. SC-U medium consisted of 0.69% (w/v) yeast nitrogen base (Formedium, CYN0405), 2% (w/v) glucose or galactose (Formedium, GAL03), 0.19% (w/v) Kaiser mixture synthetic complete uracil drop-out (Formedium, DSCK102), and 2% (w/v) agar if required. For fet3Δ fet4Δ transformants, the pH of SC-U liquid and solid media was set to 4.5 because of slow growth, whilst all other strains were grown at pH 5.3.

Non-transformed S. cerevisiae strains were streaked onto solid YPAD plates and incubated at 28 °C for 2–4 d until individual colonies were visible. These strains were then transformed using the lithium acetate method (Gietz and Wood, 2002). Transformed cells were then grown on SC-U selective plates for 2–4 d at 30 °C, with colonies replated and confirmed by PCR.

Saccharomyces cerevisiae drop assays

To assay complementation of mutant strains on restrictive media, a drop assay was performed with the transformed S. cerevisiae strains. Individual colonies of the transformed S. cerevisiae were grown in 10 ml of liquid SC-U overnight at 30 °C, shaking at 210 rpm. These cultures were pelleted by centrifugation at 1500 g for 10 min at room temperature, with the supernatant discarded and the pellet resuspended in dH₂O. After repeating the centrifugation step and discarding the supernatant, the pellet was then resuspended in dH₂O to an OD₆₀₀ of 1.0. A 10 µl aliquot of this diluted culture and serial dilutions up to 10⁵ were then pipetted onto SC-U agar plates containing 2% (w/v) galactose to induce expression of the AtIAR1-based construct.

Restrictive media for transformants were based on those from previous publications (Podar et al., 2012; Milner et al., 2013), with some minor alterations as detailed below. The smf1Δ strain was sensitive to Mn deficiency, and complementation therefore was assayed in medium with 20 mM EGTA (Millipore, 324626) and 50 mM MES (Sigma-Aldrich, M8250) at pH 5.5. The zrt1Δ zrt2Δ strain was sensitive to Zn deficiency, and complementation was therefore assayed in medium with 500 µM ZnSO₄ (Sigma-Aldrich, Z1001) and 1 mM EDTA (Sigma-Aldrich, E9884). The fet3Δ fet4Δ strain was sensitive to Fe deficiency, and complementation was therefore assayed in medium with 25 µM bathophenanthroline disulfonate (BPDS; Sigma-Aldrich, 146617) and 20 mM MES.

Plant materials

To assess the impact of loss of AtIAR1 activity in A. thaliana, two different mutant lines were sourced. The Atiar1-3 mutant used by Lasswell et al. (2000) was kindly provided by the Bartel lab. The second mutant was a T-DNA insertion mutant (Alonso et al., 2003) (SALK_ 047876C, obtained through the Nottingham Arabidopsis Stock Centre) in a Col-0 background, hereafter referred to as Atiar1-t. The A. thaliana Col-0 ecotype was therefore used as a wild-type reference for both these mutants. Genotyping of the Atiar1-3 plants was conducted as described by Lasswell et al. (2000) whilst genotyping of the Atiar1-t plants was conducted using left border primer GTCTCTTGCGTGAATGAGAGG, right border primer CATTTCTGCAAGAACTCCAGC, and T-DNA border primer ATTTTGCCGATTTCGGAAC.

Arabidopsis seed sterilization and vernalization

Arabidopsis thaliana seeds were surface-sterilized using 70% (v/v) ethanol followed by treatment with 5% (v/v) bleach and 0.05% (v/v) Tween-20 (Sigma-Aldrich, P2287) solution for 10 min. After three 10 min washes with dH₂O, the seeds were placed into sterile 0.1% (w/v) agarose (Melford Biolaboratories Ltd, A20080) solutions. Sterilized seeds were then vernalized for 48 h at 4 °C in darkness, before being placed into the relevant medium.

EDTA-washed Hoagland’s medium-based agar

In order to create Zn-deficient conditions on solid media, agar was washed with EDTA as described in Sinclair et al. (2018). For this process, 30 g of agar was washed three times with 1 litre of 50 mM EDTA pH 8.0, twice for 5 h then once for 14 h. The agar was then washed again with 1 litre of dH₂O four times for 4 h then once for 14 h. To maintain suspension of agar during washes, the agar solution was shaken at 130 rpm. To exchange media, the agar suspension was filtered through Miracloth (475855, Merck Millipore Ltd, Watford, UK). After the water washes, the agar was air-dried on filter paper (185 mm diameter, FT3205185, Sartorius Ltd, UK) for 1 d at room temperature, before adding the modified Hoagland growth medium (Hoagland and Arnon, 1938) (Supplementary Table S1) with 0, 1, or 150 µM Zn and sucrose (Formedium, S/8600/60) to 1% (w/v) to yield medium with 1.5% (w/v) agar, with pH set to 5.7 using KOH.

Root and shoot growth phenotyping

To assess phenotypes, sterilized and vernalized seeds were sown on Hoagland medium-based EDTA-washed agar media with 0, 1, or 150 µM ZnSO₄ and with either no auxin, 100 nM IAA, or 20 µM IAA–Ala added, and grown vertically. After 10 d, images of plants were taken for measurement of primary root length, shoot hypocotyl length, and lateral root (LR) density. For calculating the percentage change in auxin-containing medium, the phenotypic measurements were compared with the mean measurement in the same biological replicate grown on medium without any auxin added.

Skotomorphogenesis was induced as previously described by Mateo‐Bonmatí et al. (2021) by transferring vernalized seeds to light conditions at 21 °C for 8 h to allow germination, before transferring to darkness for an additional 4 d at 21 °C.

All images were analysed with Fiji (ImageJ) software.

Plant metabolite quantification

To investigate the influence of Zn status and AtIAR1 activity on auxin metabolism, the abundance of several auxin-related metabolites was measured (Supplementary Table S2; Supplementary Fig. S1). These were chosen to represent precursors of IAA (Trp and IAN), IAA itself, and IAA conjugation and degradation products (IAA–Ala, IAA–Asp, IAA–Glc, and oxIAA).

Sterilized and vernalized seeds were sown on Hoagland solution-based EDTA-washed agar medium with 0, 1, or 150 µM ZnSO₄, and grown vertically for 16 d. In a protocol adapted from Sugahara et al. (2020), 100 mg pools of 16-day-old whole seedlings were snap-frozen in liquid nitrogen and homogenized with 3 mm tungsten carbide beads with a TissueLyser LT homogenizer. A 333 µl aliquot of 80% (v/v) methanol containing the internal standard deuterated IAA (D₂-IAA) at 500 nM was added to the homogenate, then centrifuged at 10 800 g for 1 min, and the supernatant was extracted and retained in a separate 2 ml Eppendorf vial on ice. Two further rounds of adding 333 µl of 80% (v/v) methanol, homogenization, and collecting supernatant were performed, resulting in 1 ml of 80% (v/v) methanol solution containing extracted metabolites and 166 nM D₂-IAA.

The 80% (v/v) methanol solution was then subjected to liquid–liquid extraction using hexane. After adding 700 µl of hexane and inverting several times, the mixture was separated by centrifugation at 16 000 g for 3 min. The upper hexane layer was removed, and the hexane addition, inversion, and centrifugation steps repeated a further twice.

The resulting fraction was dried in a Genevac EZ-2 Elite Evaporator (SP Scientific), with a temperature limit of 45 °C, evaporation time of 90 min, and a drying time of 30 min. Dried samples were then resuspended in 100 µl of 80% (v/v) methanol containing no D₂-IAA and passed through a 0.22 µm pore spin filter (8161, Corning Ltd).

Liquid chromatography–mass spectrometry

Samples were analysed on an Acquity UPLC equipped with a Xevo TQ-S mass spectrometer (Waters). Separation was on a 100 × 2.1 mm 2.6μ Kinetex EVO C18 column (Phenomenex) using the following gradient of methanol (B) versus 0.5% formic acid in water (A), run at 0.4 ml min^–1 and 40 °C: 0 min, 5% solvent B; 4 min, 90% B; 5 min, 90% B; 5.2 min, 5% B; 8 min, 5% B. Detection was by positive mode electrospray with spray chamber conditions of 1.5 kV capillary voltage, 500 °C desolvation gas temperature, 1000 l h^–1 desolvation gas, 150 l h^–1 cone gas, and a nebulizer pressure of 70 bar. Mass transitions and detection parameters for each analyte are shown in Supplementary Tables S3–S7. Analytes were quantified using internal standard calibration (D₂-IAA as internal standard) using TargetLynx software (Waters) and are expressed in mol mg^–1 FW.

Expression analysis of AtIAR1

Whole 16-day-old seedlings were pooled into 100 mg samples which were snap-frozen in liquid nitrogen. The samples were then homogenized using 3 mm tungsten carbide beads and a TissueLyser LT homogenizer. RNA was extracted using an RNeasy Plant Mini Kit (74904, QIAGEN, Hilden, Germany) according to the manufacturer’s specifications, with RNA concentration measured with a NanoDrop™ 2000 spectrophotometer. To synthesize cDNA, a SuperScript™ IV First-Strand Synthesis System (Invitrogen, 18091050) was used according to the manufacturer’s specifications with each 20 µl reaction containing 4 µl of RNA (400 ng µl^–1). Reactions were run at 37 °C for 50 min, then 70 °C for 15 min in a Mastercycler® pro thermocycler (Eppendorf UK Ltd, Stevenage, UK). cDNA was diluted 10-fold prior to use as a template in PCR. PCR was conducted using primers for cDNA of the full-length AtIAR1 gene (ATGTCGTTCTCGCTGAGAAAG and TCATTCTATAAGAGAGATGCAAAGAG) and for cDNA of the AtUBC fragment (CTGCGACTCAGGGAATCTTCTAA and TTGTGCCATTGAATTGAACCC).

Chlorophyll and shoot fresh weight analysis

Sterilized and vernalized seeds were sown onto Hoagland solution-based EDTA-washed agar medium with 0, 1, or 150 µM ZnSO₄ and grown for 16 d before having the shoot fresh weight measured, and then were snap-frozen in 50 mg pools. These pools were homogenized using 3 mm tungsten carbide beads and the TissueLyser LT homogenizer. Samples were dissolved in 100% methanol (Fisher Scientific, 10674922) for 30 min at 40 °C, before being clarified by centrifugation at 10 000 g for 1 min. The resulting supernatant was then diluted as appropriate, and absorbances at 650 nm (A₆₅₀) and 665 nm (A₆₆₅) were measured using a spectrophotometer to calculate chlorophyll concentration based on the following equation (Porra et al., 1989): Chl (µg ml^–1)=22.5(A₆₅₀)+4(A₆₆₅)

Chlorophyl content was then reported as µg mg^–1 FW.

Results

AtIAR1 is a LIV-1 subfamily member, divergent from other Arabidopsis ZIP family genes

The ZIP family can be split into four subfamilies (Jeong and Eide, 2013), with phylogenetic analysis showing that AtIAR1 is the only ZIP gene in Arabidopsis found in the LIV-1 subfamily (Fig. 1A). The closest characterized homologues within this subfamily are human and mouse ZIP7, DmCatsup, and ScYke4, rather than the other ZIP genes found in Arabidopsis. Constitutive expression of the mouse homologue MmZIP7 in Atiar1 mutants restored IAA–Ala sensitivity (Lasswell et al., 2000), and in ScYKE4 mutants restored calcofluor white resistance (Kumánovics et al., 2006), suggesting that AtIAR1 and its homologues possess a high degree of conservation of function across kingdoms.

Fig. 1.

AtIAR1 phylogeny, structure, and Atiar1 mutants used in this study. (A) Phylogenetic tree of ZIP family members. Sequences taken from within animal (Drosophila melanogaster and Homo sapiens), plant (Arabidopsis thaliana highlighted in green, Oryza sativa, and Medicago truncatula), and fungal (Saccharomyces cerevisiae) kingdoms. Numbers by branches show bootstrap support values. AtIAR1 is indicated by the red star. (B) AlphaFold (Jumper et al., 2021) structural model of AtIAR1 oriented in the cytosol-down configuration and colour-coded by feature, with transmembrane helices in green, signal peptide cleavage site as predicted by signalP-5.0 (Almagro Armenteros et al., 2019b) in dark blue, H-rich loops shaded light blue, acidic helix shaded orange, and ‘variable loop’ shaded grey. (C) AlphaFold structural model of AtIAR1, rotated from (B) and annotated similarly, with the addition of the red region representing the amino acids substituted in the Atiar1-3 variant of AtIAR1. (D) Amino acid sequences at the C-terminus in AtIAR1, AtIAR1 plant homologues, and Atiar1-3. Asterisks indicate STOP codons while the residues in red are those altered through a frameshift mutation in the Atiar1-3 gene. Residues highlighted in grey are those showing 100% conservation in the sample of plant AtIAR1 homologues shown (Oryza sativa, Nicotiana tabacum, Zea mays, Glycine max, and Marchantia polymorpha). (E) Electrostatic surface of the AtIAR1 protein model rotated to the same degree of that in (C) generated by Adaptive Poisson–Boltzmann Solver (APBS) (Dolinsky et al., 2004), with red and blue corresponding to negatively charged and positively charged surfaces, respectively. The black triangle indicates the negatively charged pocket that might be disrupted in Atiar1-3.

The AlphaFold (Jumper et al., 2021) structural model of AtIAR1 predicts eight transmembrane helices in addition to a signal peptide, two histidine-rich loops on the luminal side of the membrane, and a large ‘variable loop’ on the cytosolic side found in many ZIP transporters (Guerinot, 2000) (Fig. 1B).

The illustrated topology of AtIAR1 is supported by immunofluorescence experiments placing the N-terminus (without signal peptide) of hZIP2 on the extracellular side of the membrane (Franz et al, 2014), by evidence of cytosolic metal binding and post-translational modifications on the cytosolic variable loop in AtIRT1 (Dubeaux et al, 2018), and by that of bioinformatics using the TOPCONS methodology (Tsirigos et al, 2015).

To investigate the role of AtIAR1, two different Arabidopsis mutants were used in this study, Atiar1-t and Atiar1-3. Atiar1-t is a T-DNA insertion line (SALK_ 047876C) not yet characterized, with the T-DNA insertion site within exon 10 encoding part of central helix 5. Therefore, Atiar1-t is predicted to be a full knockout which is confirmed by the lack of expression of full-length AtIAR1 (Supplementary Fig. S2). By contrast, Atiar1-3 is a partially characterized mutant found during a screen for IAA–Ala insensitivity by Lasswell et al. (2000) and contains a frameshift mutation, close to the 3' end of the coding region, which leads to changing only the final 11 amino acids (highlighted in red in Fig. 1C) whilst maintaining expression of a full-length Atiar1-3 transcript (Supplementary Fig. S2). This C-terminus is highly conserved in plants (Fig. 1D) and may also play a role in AtIAR1 functionality by forming dimerization contacts as predicted through analysis of evolutionary coupling of residues in the ZIP family (Wiuf et al., 2022). The wild-type C-terminus is modelled to contribute to a negatively charged pocket on the luminal side of the membrane which might also contribute to activity (Fig. 1E).

AtIAR1 transports Zn into the cytosol, possibly from the secretory system

In the model of AtIAR1 and AtMTP5 function proposed by Rampey et al. (2013), the metal transported by AtIAR1 was not directly assayed, although Zn was suspected, as evidenced by the metal transport capability of MTP proteins (Debrosses-Fonrouge et al., 2005), and later MTP5 itself in complex with MTP12 (Fujiwara et al., 2015). ZIP proteins in plants have previously been characterized as being capable of transporting Fe in the case of AtIRT1 (Korshunova et al., 1999) and Mn in the case of AtZIP2 (Milner et al., 2013) in addition to Zn ions. Complementation of yeast strains defective in metal import is a commonly used method for assessing the metal transport capabilities of plant ZIP proteins including from Arabidopsis (Milner et al., 2013). Therefore, a collection of strains defective in the import of Zn²⁺, Mn²⁺, and Fe²⁺ was transformed with AtIAR1-based constructs. The transformed strains were then assessed for their growth on Zn, Mn, or Fe restrictive media.

As Fig. 2 shows, the AtIAR1 construct lacks the ability to complement the Mn and Fe uptake mutants, while it is able to complement the Zn uptake mutant partially, implying that AtIAR1 shows specific transport activity for Zn as predicted. Transformation with the Atiar1-3 construct did not lead to complementation, which could indicate either reduced localization to the plasma membrane in yeast or reduced transport activity of Atiar1-3 protein.

Fig. 2.

Complementation of metal import yeast mutants by AtIAR1 constructs. Serial 10-fold dilutions (left to right) of transformed yeast cultures were pipetted onto restrictive media. Wild-type (WT) or mutant strains defective in the import of Zn (zrt1Δ zrt2Δ), Mn (smf1Δ), or Fe (fet3Δ fet4Δ) were transformed with empty vector (e.v.), AtIAR1, or Atiar1-3 constructs. Transformed strains were then grown on (A) Zn-restrictive medium, (B) Mn-restrictive medium, or (C) Fe-restrictive medium.

Several points of evidence indicate that AtIAR1 is localized to the secretory pathway in Arabidopsis. Hydrolases specific for IAA–amino acid conjugates such as IAA–Ala have been shown to localize to the ER, through bioinformatic, proteomic, and GFP tagging experiments (Bartel and Fink, 1995; Davies et al., 1999; LeClere et al., 2002; Sanchez Carranza et al., 2016). Since Atiar1 mutants are insensitive to IAA–Ala (Lasswell et al., 2000), it is therefore likely that AtIAR1 is present at least on the ER membrane. In addition, the Atiar1 phenotype was partially complemented by additional mutations in AtMTP5 (Rampey et al., 2013). AtMTP5 was found to encode a cis-Golgi-localized protein that interacts with AtMTP12 to facilitate transport of Zn into the Golgi (Fujiwara et al., 2015). Furthermore, expression of the AtIAR1 mouse homologue MmKE4 (MmZIP7), which localizes to the ER/Golgi in mice (Taylor et al., 2004; Huang et al., 2005), complemented the Atiar1 phenotype (Lasswell et al., 2000).

A range of software has been developed for predicting localization of proteins in plants. PProwler v1.2 (Hawkins and Bodén, 2006), TargetP v2.0 (Almagro Armenteros et al., 2019a), and DeepLoc v2 (Thumuluri et al., 2022) were used in this study. All localization software tested showed that the most likely localization of AtIAR1 was in the ER or the secretory pathway (Table 2).

Table 2.

AtIAR1 is predicted to localize to the secretory pathway

Prediction software	Subcellular prediction	Probability
PProwler v1.2	Secretory pathway	1.0
TargetP v2.0	Secretory pathway	0.923
DeepLoc v2.0	ER	0.648

Subcellular prediction with highest probability is listed along with the associated probability of this location.

Therefore, the current evidence suggests that AtIAR1 imports Zn to the cytosol from the secretory pathway, thereby playing a role in regulating Zn concentration within the lumen of the secretory pathway. The effect of the disruption of this transport activity in Atiar1 plants on wider auxin metabolism, including that of conjugation and its relationship with media Zn levels, was therefore investigated.

Atiar1 mutants show disrupted wider auxin metabolism

Precise quantification of metabolite levels is essential to understanding changes in IAA metabolism. LC-MS was used to quantify IAA and IAA-related metabolites using D₂-IAA as an internal standard based on previous approaches (Sugahara et al., 2020). A methodology adapted for the current study (Sugahara et al., 2020), as shown in the Materials and methods, produced an appropriate linear response range for metabolite analysis in Arabidopsis whole seedlings for a variety of IAA-related metabolites including IAA precursors (Trp and IAN), IAA, IAA conjugates (IAA–Ala, IAA–Asp, and IAA–Glc), and degradation pathway intermediates (oxIAA) (Supplementary Table S7). Levels of IAA–Ala were too low to detect in this analysis, a result that aligns with previous work suggesting that the content of IAA–Ala is below that of other conjugates including IAA–Glc and IAA–Asp (Sugahara et al., 2020). This differing abundance suggests some degree of selection between conjugates as an IAA store, although the mechanisms and significance of this selection remain unclear.

In wild-type (Col-0) plants, Zn-deficient conditions induced dramatic increases in Trp and IAA–Glc concentrations (Fig. 3). These results align with findings that Trp levels increase in Zn deficiency as a result of reduced protein synthesis (Cakmak et al., 1989). By contrast, under Zn excess, the largest changes were the increases in IAA–Glc and IAA–Asp concentrations. Interestingly, IAA-Glc shows an increase in Zn-deficient and Zn-excess conditions in both Col-0 and Atiar1-3 plants. However, no significant increase in IAA–Glc was seen in Atiar1-t mutants, and the levels in Zn-deficient conditions were less in Atiar1-3 and Atiar1-t genotypes than in Col-0 plants. This suggests that the predicted disruption of IAA–amino acid hydrolysis in Atiar1 plants leads to compensatory reductions in IAA–Glc pools under high and low Zn conditions. Genotype-specific changes in the levels of oxIAA and IAN were also observed between Zn-excess and Zn-deficient conditions, which further suggests the sensitivity of wider auxin metabolism to Zn and Atiar1 functionality (Supplementary Fig. S3).

Fig. 3.

Auxin-related metabolite content. Contents of IAA precursor (Trp), IAA, and IAA conjugates (IAA–Glc, IAA–Asp) were measured from whole Col-0 (grey), Atiar1-3 (blue), and Atiar1-t (green) plants grown for 16 d in Zn-deficient (0 Zn added), Zn control (1 µM), and Zn excess (150 µM) conditions on modified Hoagland’s medium containing EDTA-washed agar. A 100 mg aliquot of fresh tissue was used in each of three biological replicates. Lower case lettering indicates statistically significant differences between groups (labelled sequentially from ‘a’ in order of estimated mean) as calculated using ANOVA utilizing Tukey’s method for P-value adjustment (Tukey, 1949) for nine groups with a P-value cut-off of 0.05. For analysis of IAN and oxIAA levels, see Supplementary Fig. S3.

Despite these changes in IAA precursors and conjugates, the concentration of IAA in whole-plant samples remained stable across Zn conditions and genotypes (Fig. 3). As this analysis measured whole-plant IAA levels, it remains possible that IAA distribution or IAA levels at a different developmental stage may be perturbed in Zn- and genotype-specific manners. Since auxin metabolism disturbances in Atiar1 plants and the activity of IAA–amino acid amidohydrolases are Zn dependent, the Zn dependence of the IAA–Ala sensitivity of Atiar1 plants was investigated.

Atiar1 mutants show an impaired IAA–Ala response and disruption of a novel excess Zn:exogenous auxin interaction

The lack of root growth inhibition by IAA–Ala in Atiar1 plants can be complemented by excess Mn, possibly as a result of activation of IAA–amino acid hydrolases by increased Mn binding (Lasswell et al., 2000). However, the IAA–Ala sensitivity of the mutants has not been assessed as a function of different concentrations of Zn, which is hypothesized to inhibit the IAA–amino acid hydrolases (LeClere et al., 2002). To test the IAA–Ala sensitivity of the Atiar1 mutants, primary root lengths 10 d after germination were measured in control medium or in the presence of 20 µM IAA–Ala. Any effect of IAA–Ala could be a function of IAA–Ala uptake, hydrolysis, and subsequent direct or indirect IAA activity. To determine where within this sequence any Zn- and genotype-related phenotypes are acting, the primary root length in medium with 100 nM IAA was also assayed (Fig. 4).

Fig. 4.

Primary root length of Atiar1 mutants grown at different Zn concentrations. Primary root length was measured for Col-0 (grey fill), Atiar1-3 (blue fill), and Atiar1-t (green fill) plants 10 d post-germination with medium conditions as in Fig. 3 in control conditions (black outline, with no auxin added), or in the presence of 20 µM IAA–Ala (red outline) or 100 nM IAA (purple outline). For clarity, statistically significant differences between groups (labelled sequentially from ‘a’ in order of estimated mean) are shown in a separate table and were calculated as for Fig. 3 for 27 groups using a P-value cut-off of 0.05. For further details on the interaction of genotype, Zn, and auxin condition, see Supplementary Table S8.

Primary root length has been reported to be inhibited in mutants sensitive to Zn deficiency [Atbzip19 Atbzip23 double mutants (Assunção et al., 2010)] and Zn excess [Atmtp1 mutants (Kobae et al., 2004)]. In Fig. 4 it is shown that in Zn-deficient conditions, with no auxin present, there is a significant reduction in primary root length in Atiar1-3 mutants relative to the control, although this difference persists across all Zn conditions and so is not Zn specific. On the other hand, the Atiar1-t mutant shows increased primary root length only in Zn-deficient conditions. This suggests the two mutants are responding to Zn differently and neither Atiar1 mutant shows a strong root growth phenotype in Zn-deficient or Zn-excess conditions. Overall, this implies that AtIAR1 is not a major player in Zn homeostasis, further evidenced by chlorophyll and shoot fresh weight analysis which showed a lack of a strong Zn-dependent phenotype (Supplementary Fig. S4).

It is also shown in Fig. 4 that the primary root growth across all Zn conditions in the presence of IAA–Ala is greater in Atiar1 mutants than in Col-0. The IAA–Ala insensitivity phenotype is more marked in Atiar1-t than in Atiar1-3 mutants, which might suggest that the Atiar1-3 variant protein maintains some partial functionality.

The interaction between genotype and media auxin level was differentially influenced by Zn in the media (F statistic 3.8774 and P-value 0.0001435: see Supplementary Table S8 for further details). Through conversion of the data in Fig. 4 into percentage primary root growth in Supplementary Fig. S5, it is apparent that for Col-0 the percentage primary root growth is further inhibited by excess Zn compared with control Zn levels in the presence of exogenous IAA–Ala and IAA, an effect which is lost in Atiar1-3 plants and lessened in Atiar1-t plants. As this interaction of Zn, genotype, and auxin condition occurs in IAA-containing media, this three-way interaction occurs post-IAA–Ala hydrolysis.

To determine whether further auxin-related phenotypes show similar Zn dependence of the AtIAR1 and auxin level interaction, shoot hypocotyl length (Supplementary Fig. S6) in media with IAA–Ala and IAA was also measured. The insensitivity to IAA–Ala of Atiar1 mutants was conserved, as shoot hypocotyls were elongated more in Col-0 than in Atiar1 mutants in IAA–Ala-containing medium. Additionally, a Zn dependence in the effect of different auxin media on different genotypes was observed (F statistic 5.6037, P-value <2.2e-16 for the interaction between Zn, genotype, and auxin variables). In excess Zn conditions, shoots of Col-0 were longer in IAA-containing medium than in medium with no auxin—an increase which did not occur in either of the Atiar1 mutants.

Overall, it seems that two factors are involved in AtIAR1-mediated auxin regulation: conjugate hydrolysis and also a novel Zn excess-mediated effect which appears to act post-hydrolysis in a tissue-specific way. Despite maintaining constant IAA levels across Zn conditions after 16 d, it is possible that changes in auxin levels were seen at different time points or in specific tissues. Therefore, to investigate how auxin-related growth phenotypes might be altered by the disruption of auxin metabolism and altered auxin activity in Atiar1 mutants, LR density and shoot hypocotyl length during skotomorphogenesis were measured across different Zn conditions.

Atiar1 mutants show auxin-related phenotypes

It is shown in Fig. 5A that LR density is selectively increased in Atiar1-t mutants and not in Atiar1-3 mutants, in all Zn conditions. Additionally, LR density is not Zn responsive in wild-type and Atiar1-3 genotypes whereas in high Zn conditions the LR density of Atiar1-t decreases. In Zn deficiency and control Zn conditions, Atiar1-t plants showed a reduction in hypocotyl length compared with both Col-0 and Atiar1-3 plants, further suggesting that Atiar1-3 retains at least partial functionality. For Col-0 and Atiar1-3 plants at 10 d, Zn excess conditions caused a reduction in shoot hypocotyl length (Fig. 5B). The lack of difference across genotypes in excess Zn conditions could indicate complementation of the Atiar1-t phenotype or that hypocotyl length had reached a lower threshold in this condition.

Fig. 5.

Auxin-related phenotypes are altered in the Atiar1-t mutant. (A) Lateral root density or (B) hypocotyl length 10 d after germination was measured for each seedling grown in medium conditions as in Fig. 3. At least 50 plants were measured in each of three biological replicates, with statistically significant differences between groups calculated and displayed as in Fig. 3.

During skotomorphogenesis, auxin plays a key role in hypocotyl elongation through biphasic cell expansion (Du et al., 2022). To determine the potential role of Zn and AtIAR1 in hypocotyl elongation in the dark, Col-0 and Atiar1 mutants were grown in light for 8 h then either kept in the light or transferred to darkness for 4 d, after which hypocotyl lengths were measured (Fig. 6).

Fig. 6.

Shoot hypocotyl length in light and dark conditions. Hypocotyl length after growth in light (A) or darkness (B) was measured 5 d after germination for each seedling grown in medium conditions as for Fig. 3. At least 30 plants were measured in each of three biological replicates, with statistically significant differences between groups calculated and displayed as in Fig. 3.

After growth for 5 d in light under Zn deficiency and control Zn conditions, Atiar1-t plants showed a larger shoot hypocotyl length than the wild type (Fig. 6A). This contrasts with data after 10 d where Atiar1-t plants showed smaller hypocotyls (Fig. 5B). Interestingly, the hypocotyl length in skotomorphogenesis was significantly greater compared with Col-0 in both Atiar1 mutants in all Zn conditions (Fig. 6B). This increased skotomorphogenic hypocotyl length in both Atiar1 mutants persists across all Zn conditions, including excess Zn conditions where all genotypes show reductions in hypocotyl length.

Discussion

AtIAR1 releases Zn ions into the cytosol probably from the secretory system

The complementation of the Zn import mutant strain zrt1Δ zrt2Δ by AtIAR1 indicates that AtIAR1 transports Zn. Complementation of this yeast plasma membrane transporter mutant does not necessarily indicate that AtIAR1 is plasma membrane localized in Arabidopsis. For example, the Arabidopsis TPC1 vacuolar membrane protein complements yeast plasma membrane transport mutants (Furuichi et al. 2002; Peiter et al., 2005).

Our attempts to visualize the intracellular location of AtIAR1 directly with a GFP-expressing construct were not successful because of low expression levels (data not shown). We therefore utilized software prediction packages to predict IAR1 location. The three packages chosen uniformly predicted a secretory system location, although resolution between predicted ER and Golgi localization is not clear (DeepLoc, in addition to the ER, predicted Golgi localization, with a lower probability of 0.62). Therefore, we cannot clearly determine where within the secretory system AtIAR1 may be present.

Atiar1 mutants show disruption in auxin metabolism

Disruption of AtIAR1 activity is hypothesized to reduce the hydrolysis and so increase the concentration of reversibly formed IAA–amino acid conjugates by increasing the ER concentration of Zn ions (Lasswell et al., 2000). In this study, IAA–Ala was not detectable, but a reduction of reversibly formed IAA–Glc was seen in the Atiar1-t mutants. Despite changes in IAA precursors and conjugates across the different Zn conditions and genotypes, the IAA levels did not significantly change.

High levels of auxin promote conjugation to IAA–amino acids and IAA–Glc, as well as reduced synthesis (Suzuki et al., 2015). Conversely, ‘low auxin’ metabolic state mutants with the inability to metabolize indole butyric acid, methylated IAA, and IAA–amino acids show compensatory increases in IAA synthesis through TAA/YUCCA induction (Spiess et al., 2014). As Atiar1 mutants are hypothesized to accumulate hydrolysable IAA–amino acids including IAA–Ala but show reduced IAA–Glc levels, it can be hypothesized that the metabolic state of Atiar1 mutants is of ‘low auxin’. In the ‘low auxin’ model of the Atiar1-t mutant, IAA levels are maintained not by the IAA–amino acid hydrolysis which is reduced, but instead by feedback to reduce levels of IAA–Glc. This potential crosstalk between conjugates should be taken into account when analysing mutants with genes responsible for controlling these different conjugation steps. Furthermore, these considerations emphasize the role of IAA conjugates as buffers that smooth IAA concentration changes to establish a required developmental response—albeit in a genotype- and Zn-sensitive manner.

In addition to alteration of auxin metabolism, Atiar1 mutants also show a disruption of auxin action post-IAA–amino acid hydrolysis in Zn excess conditions. Although the precise mechanisms of Zn excess-induced growth changes are mostly unknown, there is some evidence of altered PIN distribution causing altered auxin accumulation in the root meristem (Zhang et al., 2018; Wan et al., 2019; Wang et al., 2021). Therefore, it is possible that Atiar1 mutants show altered auxin distribution as well as metabolism.

Disruption of auxin metabolism in Atair1 mutants correlates with auxin-related phenotypes

Auxin overproducers such as AtYUCCA1-overexpressing plants have increased hypocotyl length in light conditions and decreased hypocotyl length in dark conditions (Zhao et al., 2001). Attmk loss-of-function mutants show reduced response to exogenous IAA along with reduced cell elongation in shoot and root tissue (Dai et al., 2013) and reduced LR number (Huang et al., 2019). Similarly, Atilr1 Atiar3 Atill2 IAA–amino acid hydrolase triple mutants show reduced root length, LR number, and hypocotyl length compared with the wild type (Rampey et al., 2004). Previous work has established that the Atsur1 auxin-overproduction short root phenotype is complemented when crossed with Atiar1 mutants (Lasswell et al., 2000), indicating that Atiar1 mutants have an effect of reducing auxin action after 10 d. This finding aligns with the suspected low auxin metabolic state of Atiar1-t mutants described in this study.

However, the complete loss of function in Atiar1-t plants is associated with an unusual combination of auxin-related phenotypes such as increased LR density, a time-dependent shoot hypocotyl length in photomorphogenesis, and increased hypocotyl length in skotomorphogenesis. This resembles neither a typical ‘high auxin’ or a ‘low auxin’ phenotype. These phenotypes are particularly interesting as they occur without any changes in whole-plant IAA levels, suggesting auxin distribution may be perturbed in Atiar1 mutants due to misregulated auxin conjugates.

Throughout this study, it is evident that Atiar1-3 and Atiar1-t mutations result in different phenotypic severity. As Atiar1-3 shows intermediate sensitivity between Atiar1-t and Col-0 to exogenous IAA–Ala, it is hypothesized that ER Zn levels in Atiar1-3 are intermediate between those of Col-0 and Atiar1-t, and that this leads to milder phenotypes of Atiar1-3 plants than those of Atiar1-t plants, including metabolite levels and developmental changes across different Zn and auxin conditions. Some phenotypes of Atiar1-3 mutants have, however, been distinct from both Atiar1-t and Col-0 plants, which probably relates to the disrupted functionality of the C-terminal region in Atiar1-3.

AtIAR1 mediates an important intersection of Zn homeostasis and auxin homeostasis in the secretory pathway

Zn transport between the secretory system and the cytosol has previously been shown to be important for both Zn deficiency and the salt stress response (Wang et al., 2010; Sinclair et al., 2018). Here, we expand the roles of Zn transport between the secretory system and cytosol to auxin metabolic balance and auxin-related growth phenotypes through the activity of AtIAR1. AtIAR1 activity is proposed to release Zn into the cytosol, increasing the response of plants to IAA–Ala through promoting hydrolysis via a mechanism of decreased Zn-mediated inhibition of conjugate hydrolysis within the secretory system. This work on AtIAR1 has also shown the importance of auxin conjugate balance for a variety of physiological auxin responses including LR density and skotomorphogenesis in Arabidopsis.

Supplementary data

The following supplementary data are available at JXB online.

Table S1. Components of modified Hoagland solution.

Table S2. Metabolites measured in this study and sources.

Table S3. Properties of the liquid chromatography column used in this study.

Table S4. Liquid chromatography conditions used in this study.

Table S5. Inlet method for liquid chromatography used in this study.

Table S6. Mass spectrometry conditions used in this study.

Table S7. LC-MS detection parameters for IAA-related metabolites.

Table S8. ANOVA table for primary root length.

Fig. S1. Structures of metabolites measured in this study.

Fig. S2. Atiar1 mutant expression of AtIAR1.

Fig. S3. oxIAA and IAN metabolite content.

Fig. S4. Shoot phenotypes of Atiar1 mutants grown on different Zn levels.

Fig. S5. Root length percentage change in IAA and IAA–Ala media.

Fig. S6. Shoot hypocotyl length in control, IAA-, and IAA–Ala-containing media.

Glossary

Abbreviations

Col-0

Columbia ecotype

D₂-IAA

deuterated IAA

dH₂O

distilled water

ER

endoplasmic reticulum

GFP

green fluorescent protein

IAA

indole-3-acetic acid

IAN

3-indoleacetonitrile

LR

lateral root

Author contributions

TG: conceptualization, data curation, formal analysis, investigation, validation, visualization, and writing—original draft; TG, LH, AJM, and DS: methodology; AJM and DS: project administration, resources, and supervision: TG, AJM, and DS: writing—review & editing.

Conflict of interest

The authors declare no conflict of interest.

Funding

This work was generously funded by the John Innes Foundation.

Data availability

All raw data underlying the results presented in this study are available upon request.