2.1 Biological material, growth conditions, and Fe treatments

Arabidopsis thaliana WT (ecotype Columbia) and ntrc mutant (Serrato et al. 2004) were routinely grown in soil in growth chambers under short-day (8/16 h light/ dark) photoperiod at 22 and 20°C during light and dark periods, respectively, and light intensity of 125 μmol m⁻² s⁻¹. WT and ntrc seedlings were grown on Murashige and Skoog (MS) medium containing 0.35% (w/v) Gelrite (Duchefa) under continuous light at 22°C and 125 μmol m⁻² s⁻¹.

Fe treatments were performed in synthetic media (pH 5.5, 0.7% agar) containing 5 mM KNO₃, 2.5 mM KH₂PO₄, 2 mM MgSO₄, 2 mM Ca(NO₃)₂, 70 μM H₃BO₃, 14 μM MnCl₂, 10 μM NaCl, 1 μM ZnSO₄, 0.5 μM CuSO₄, 0.2 μM Na₂MoO₄, 4.7 mM MES (Rodríguez-Celma et al. 2013) and a variable concentration of Fe-EDTA as source of Fe. For Fe treatments at the seedling stage, seeds were germinated in synthetic media containing either 50 μM Fe-EDTA (Fe-sufficient, control condition), 500 μM Fe-EDTA (Fe-excess), or medium without Fe-EDTA and supplemented with 100 μM ferrozine (sodium 4-[3-(pyridin-2-yl)-6-(4-sulfophenyl)-1,2,4-triazin-5-yl]benzene-1-sulfonate) to chelate residual Fe (Fe-deficient). Fe treatments at the adult stage were performed on plants grown in Fe-sufficient medium for 21 days and transferred to Fe-sufficient (control conditions), Fe-excess or Fe-deficient media for 9 days.

2.2 RNA extraction, RNA-Seq and RT-qPCR analyses

Two different transcriptomics experiments, at the adult and seedling stages of plant development, were carried out. For adult samples, eight-week-old short-day grown WT and ntrc plants were randomly selected, and young rosette leaves were collected during the day period (after 2 h of illumination at 125 μmol m⁻² s⁻¹). The experimental design consisted of three biological replicates for each genotype, each of them containing leaves from three individual plants. WT raw RNA-Seq data from the adult stage can be found in Gene Expression Omnibus, identified with accession number GSE147793. RNA extraction was performed from pooled leaf samples using the Sure Prep kit (Fisher), following the manufacturer's instructions. RNA concentration and purity were tested by an Agilent 2100 Bioanalyzer, a microfluidics-based platform that performs quality control of DNA and RNA samples before sequencing. Library construction of cDNA molecules was carried out following the manufacturer's instructions using the TruSeq Stranded Total RNA with Ribo-Zero kit (Illumina). The generated DNA fragments were sequenced with the Illumina HiSeq 4000 platform at STAB-VIDA (Caparica, Portugal), yielding approximately 60–80 million 150 base pairs long paired-end reads for each sample. For seedling samples, six-day-old WT and ntrc mutant seedlings, grown under continuous light in MS medium and lacking true leaves, were collected. The experimental design consisted of three biological replicates for each genotype. RNA extraction and quality control were carried out following the above indicated procedure for adult samples. The cDNA libraries were constructed according to the manufacturer's instructions using the Illumina Stranded mRNA prep ligation kit (Illumina). Sequencing of the DNA fragments was performed with the Illumina NextSeq 500 platform at Centro Andaluz de Biología Molecular y Medicina Regenerativa (CABIMER, Seville, Spain), yielding approximately 18–26 million 75 base pairs long single-end reads for each sample.

For all samples, the FastQC software package was used to control the quality (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Once the good quality of all samples had been determined, read mapping to The Arabidopsis Col-0 TAIR 10 reference genome (https://plants.ensembl.org/) and transcript assembly were performed with the software tools HISAT2 and StringTie (Pertea et al. 2016) using default parameters. Differential expression analysis was carried out with the Ballgown (Pertea et al. 2016) and LIMMA (Ritchie et al. 2015) R packages. Differentially expressed genes (DEGs) were selected using a fold change of ± log₂(2) and P-value <0.05 computed according to a moderated t-test. Volcano plots and Venn diagrams were generated using the R packages ggplot2 and eulerr. Gene Ontology (GO) term enrichment analysis over the different gene sets was performed with the enrichGO function of the Bioconductor R package ClusterProfiler (Yu et al. 2012). From the same R package, the compareCluster function was used to compare the different gene ontology terms of adult and seedling samples. Principal Component analysis was carried out using the prcomp function from the R package stats of the R Core Team.

For Real-time quantitative PCR (RT-qPCR) analysis, total RNA was extracted using the Trizol reagent (Invitrogen) and cDNA synthesis was performed with 1 μg of total RNA using the Maxima first-strand cDNA synthesis kit (Thermo Scientific) according to manufacturer's instructions. RT-qPCR was performed using an IQ5 real-time PCR detection system (Bio-Rad) with oligonucleotides listed in Supplemental Table S1. Expression levels were normalized using ACTIN2, UBIQUITIN10 and UBIQUITIN-CONJUGATING ENZYME9 as reference genes.

2.3 Measurements of chlorophyll levels and determination of photosynthetic parameters

Chlorophyll levels were measured as previously described (Pérez-Ruiz et al. 2006). Room temperature chlorophyll a fluorescence was measured using a pulse-amplitude modulation fluorometer IMAGING-PAM M-Series (Walz). The maximum quantum yield of PSII was assayed after incubation of plants in the dark for 30 min by calculating the ratio of the variable fluorescence, F_v, to maximal fluorescence,  F_m (F_v/F_m). Induction-recovery curves were performed using red (635 nm) actinic light at 81 μmol m⁻² s⁻¹ for 10 min. Saturating pulses of red light at 10,000 μmol m⁻² s⁻¹ intensity and 0.6 s duration were applied every 60 s, and recovery in darkness was recorded for up to 10 min. The parameter Y(II), corresponding to the respective quantum yield of PSII photochemistry, was calculated according to reported equations (Kramer et al. 2004).

2.4 Determination of anion superoxide content

The detection of superoxide anion content was performed with nitroblue tetrazolium (NBT) staining, as previously described (Cordoba et al. 2016), with some modifications. Briefly, at least six seedlings of WT and ntrc mutant grown in Fe-deficient, Fe-sufficient (Fe-control) and Fe-excess media were randomly selected and incubated in staining buffer (10 mM phosphate pH 7.6, 10 mM NaN₃ and 0.1% NBT (Sigma)) for 2.5 h at room temperature in agitation. Then, the NBT solution was removed, and a bleaching solution (ethanol: acetic acid: glycerol; 3:1:1) was added. Once the chlorophyll had been removed, a new bleaching solution was added to the seedlings, which were stored at 4°C until images were taken with an Olympus Microscope. For the quantification of the area of cotyledons stained with NBT, colour thresholds were used in the ImageJ software.

2.5 Ionomics determinations

For WT and the ntrc mutant, whole rosette samples of 8-week-old plants grown in soil under short-day conditions and 4-day-old seedlings grown under continuous light in MS medium were lyophilized for determination of ion content by inductively coupled plasma mass spectrometry (ICP-MS) at the Servicio de Análisis, Instituto de Recursos Naturales y Agrobiología de Sevilla (IRNASE, CSIC), Seville, Spain. In brief, 4 mL of HNO₃ (65%) was added to the weighed lyophilized material. The digestion was carried out under pressure in a microwave oven (MARS ONE CEM). The treated samples were diluted with ultrapure water and passed through nylon filters. Ion determination was performed using an AGILENT 7800 ICP-MS. Principal Component analysis was carried out using the prcomp function from the R package stats of the R Core Team.

2.6 Statistical analysis

Statistical analyses were performed using GraphPad Prism version 6.01 (GraphPad Software, Inc.), and the experimental results were analyzed using two-tailed Student's t-tests.

3.1 Comparative transcriptomic analyses reveal a poor impact of NTRC at the seedling stage of plant development

With the aim of determining the impact of NTRC at different stages of plant development, we have performed a comparative transcriptomic analysis of Arabidopsis WT and the ntrc mutant at seedling and adult stages. To that end, genome-wide transcriptomes were determined by RNA-Seq analysis of three independent biological samples harvested from young leaves of 8-week-old Arabidopsis WT and ntrc mutant plants grown under short-day photoperiod (adult stage) and from seedlings that had been grown for 6 days on MS medium under continuous light (seedling stage) (Figure 1A, C). Though the ntrc mutant contained lower chlorophyll levels at both developmental stages, the difference was more significant at the adult stage (Figure 1B, D). The sequencing quality of the samples was assessed by the high percentage of mapped reads and correlation between biological replicates (around 99%) (Figure S1A-D). A principal component analysis showed that the ntrc and WT seedling replicates clustered together, whereas replicates from adult ntrc clustered separately from the WT (Figure S2), thus indicating a more relevant impact of NTRC on the transcriptomes at adult than at seedling stages.

The comparison of the transcriptomes from leaves of the adult plants of the WT and the ntrc mutant, using a fold change of gene expression of ±log₂(2) and p < 0.05, identified a large set of 535 DEGs in the ntrc mutant, 358 upregulated and 177 downregulated (Figure 2A). Strikingly, a relatively low number of DEGs (94), of which 35 were upregulated and 59 downregulated (Figure 2B), were identified in the comparison of the WT and the ntrc mutant transcriptomes at the seedling stage. The lists of upregulated and downregulated genes in the ntrc mutant at both developmental stages are available as supplemental information (Supplemental Appendix S1). A Venn diagram representation of DEGs at adult and seedling stages identified an overlapping set of 24 DEGs common to both stages (Figure 2C). Thus, the higher number of DEGs between the ntrc mutant and the WT at the adult stage confirms the deep impact of NTRC on plant growth, in line with the severe growth inhibition phenotype of the ntrc mutant (Figure 1A, B), whereas the modest transcriptomic changes shown by the ntrc seedlings reveals the poor impact of NTRC at this early developmental stage, which is in line with the almost WT-like phenotype of the ntrc seedlings (Figure 1C, D). Moreover, the low number of DEGs commonly detected at seedling and adult stages indicates that the participation of NTRC throughout plant development is restricted to a few biological processes.

3.2 The ntrc mutant shows transcriptomic changes resembling the response to Fe excess

To identify the processes affected by the lack of NTRC at adult and seedling stages, a gene ontology (GO) term enrichment analysis was performed (Supplemental Appendix S2). In line with the severe growth inhibition phenotype of the ntrc mutant, the comparison of the WT and ntrc mutant transcriptomes at the adult stage showed the pleiotropic effects caused by the absence of NTRC, affecting a wide range of biological processes, as previously reported (Lepistö et al. 2009). Upregulated genes in the ntrc mutant at this developmental stage are enriched in GO terms involved in metabolism (carbohydrates, amino acids and tetrapyrroles), abiotic stress response (hypoxia, nutrient level, nitrate, light, heat, metal ion), biotic stress response (insect, salicylic acid), and detoxification (Figure 3A). In contrast, GO terms related to glucosinolate and glycerolipid metabolisms, cell wall organization, cellular homeostasis, response to starvation and ion transport were the biological processes identified among the downregulated genes in the ntrc mutant (Figure 3B). At the seedling stage, most enriched GO terms among the upregulated genes in the ntrc mutant were related to metal ion homeostasis (Figure 3C), whereas GO terms associated with cellular homeostasis, response to starvation and ion transport, as well as response to metal ion and jasmonic acid biosynthesis, were identified among the downregulated genes (Figure 3D).

To further analyze the impact of NTRC at different stages of plant development, Venn diagrams were constructed identifying common GO terms among upregulated and downregulated genes in the ntrc mutant at adult and seedling stages (Figure 4A, B and Supplemental Appendix S3). This analysis identified GO terms related to detoxification and homeostasis of Fe enriched at both stages of development (Figure 4C). The fact that many of the DEGs between the ntrc mutant and the WT from adult and seedling stages identified terms such as “Intracellular iron homeostasis”, “Iron ion homeostasis”, “Multicellular organismal-level iron homeostasis”, “Response to iron ion” and “Response to iron starvation” (Figure 4C) revealed the effect of NTRC on the transcription of genes involved in Fe homeostasis at both developmental stages here analyzed, a previously unrecognized function of NTRC, which seems more relevant in seedlings than in leaves from adult plants.

In plants, Fe homeostasis involves different organs and tissues including roots and the vascular system and, in leaf photosynthetic cells, different cell compartments (Liang, 2022). Thus, to figure out the effect of NTRC on Fe homeostasis at different developmental stages, DEGs involved in Fe homeostasis or response to Fe in the transcriptomic analyses of seedling and adult samples were selected, and their fold-changes in the level of expression were represented in different organs and cell compartments (Figure 5A). Genes encoding peptides (FEP1, FEP2, IMA4, IMA6) that positively regulate the transcription factors bHLH38, bHLH39, bHLH100 and bHLH101, all of them involved in activation of the expression of genes responsive to Fe deficiency, were found among the most downregulated genes in seedlings and, at lower extent, in adult leaves of the ntrc mutant (Figure 5A), hence indicating that the lack of NTRC triggers transcriptomic changes resembling the response to Fe excess. This notion was further supported by the identification of genes involved in the transport of Fe through the phloem and xylem, such as OLIGO PEPTIDE TRANSPORTER 3 (OPT3) among the downregulated genes at both stages. Furthermore, genes encoding the Fe transporter in root cells IRON REGULATED TRANSPORTER 1 (IRT1) and enzymes of the sideretin biosynthetic pathway were also downregulated in seedlings of the ntrc mutant (Figure 5A). These genes showed differences only at the seedling stage due to the nature of the adult samples, which excluded root tissues. In leaves cells, the comparative transcriptomic analyses revealed upregulation of genes encoding transport of Fe to the vacuoles, VACUOLAR IRON TRANSPORTER-LIKE 1, 2 and 5 (VTL1, VTL2, VTL5) and the chloroplast PERMEASE IN CHLOROPLASTS 1 (PIC1), both at adult and seedling stages (Figure 5A). Furthermore, genes involved in Fe-S assembly (NEET) and tetrapyrrole biosynthesis (HEMA1) into the chloroplast and, remarkably, genes encoding ferritins (FER1, FER3 and FER4) were upregulated in adult and seedlings samples of the ntrc mutant, further supporting the notion that the lack of NTRC triggers a Fe excess transcriptomic response. On the contrary, CONSERVED IN THE GREEN LINEAGE AND DIATOMS 27 (CGLD27), which is required for growth under Fe starvation (Urzica et al. 2012), was downregulated in the ntrc mutant at both developmental stages. Among genes encoding enzymes involved in the scavenging of reactive oxygen species (ROS), formed in the presence of Fe via the Fenton reaction, stromal APX (sAPX) was upregulated in the ntrc mutant, whereas FeSOD1-3 genes were downregulated in mutant seedlings and FeSOD1 upregulated in adult leaves. Finally, several of the DEGs between the ntrc mutant and the WT were chosen to validate the results of the RNA-Seq data by RT-qPCR analysis. The lower level of transcripts of the bHLH38, bHLH39, FEP1 and OPT3 genes and the higher levels of the FER1 and NEET genes in adult leaves (Figure 5B) and seedlings (Figure 5C) of the ntrc mutant validated the results of the transcriptomic analysis.

3.3 Effect of Fe excess and deficit treatments on photosynthetic performance

To gain a deeper insight into the function of NTRC in metal ion homeostasis, we sought to analyze the effect of Fe deficiency and excess treatments on the ntrc mutant performance at the two developmental stages analyzed. For seedling treatments, WT and ntrc seeds were germinated and grown for 4 days under continuous light in a synthetic medium supplemented with 50 mM Fe-EDTA (Fe-control, sufficiency conditions), 100 mM ferrozine and no added Fe-EDTA (Fe-deficiency conditions) and 500 mM Fe-EDTA (Fe-excess conditions) (Figure 6A). To analyze the effect at adult stage, WT and ntrc plants grown for 21 days under short-day photoperiod and Fe-sufficient conditions (Fe-control) were transferred to Fe-control, Fe-deficiency or Fe-excess synthetic media and grown for 9 days. The photosynthetic parameters Fv/Fm, which reflects PSII stability, and quantum yield of PSII photochemistry (Y(II)) were determined as readouts of the treatments (Figure 6A-C).

The Fe deficiency treatment caused no significant differences in the levels of Fv/Fm either at adult or seedling stages of the WT and ntrc mutant (Figure 6A, B). However, the quantum yield of PSII (Y(II)) revealed a slight effect of this treatment at the adult stage in the ntrc mutant, which was more severe at the seedling stage (Figure 6A, C), hence indicating the sensitivity of the ntrc mutant to Fe deficiency, which is higher at early stages of development. The Fe excess treatment, which did not significantly affect the WT at the adult stage, caused slightly higher levels of Fv/Fm (Figure 6A, B), but no significant variation of Y(II) (Figure 6A, C) in adult ntrc plants. In contrast, this treatment had a severe effect on WT seedlings, as revealed by the lower levels of Fv/Fm (Figure 6A, B) and Y(II) (Figure 6A, C). Remarkably, the lack of NTRC triggered a significant tolerance to Fe excess at the seedling stage, as shown by the higher levels of Fv/Fm (Figure 6A, B) and Y(II) (Figure 6A, C) in seedlings of the ntrc mutant.

It is known that Fe excess may be potentially toxic for plants by triggering ROS formation by the Fenton reaction (Halliwell and Gutteridge 1992). Thus, the better performance of the ntrc mutant seedlings under Fe excess might be due to altered ROS homeostasis caused by the lack of NTRC. To test this possibility, seedlings of the WT and the ntrc mutant subjected to Fe excess and deficiency treatments were stained with NBT to detect the contents of superoxide anion (Figure 7A). No significant differences in NBT staining between the WT and ntrc mutant seedlings were observed under control conditions or in response to Fe deficiency; in contrast, the treatment with Fe excess caused higher levels of NBT staining, hence of superoxide anion, in WT, but not in the ntrc mutant seedlings (Figure 7B). Thus, Fe excess provoked oxidative stress in seedlings of the WT, but not of the ntrc mutant, suggesting that the alteration of ROS homeostasis caused by the absence of NTRC protects from the negative effects of Fe excess.

Finally, to further explore the impact of NTRC on metal ion homeostasis, a global ionomic analysis was performed in the WT and the ntrc mutant at the seedling and adult stages. As in the transcriptomic study (Figure S2), a principal component analysis clustered the WT and ntrc seedling samples, while the adult samples were found separately (Figure 8A), further confirming the different impact of the lack of NTRC in these developmental stages. Remarkably, despite the transcriptomic changes affecting genes involved in Fe homeostasis caused by the lack of NTRC, no significant differences in the contents of Fe in the WT and the ntrc mutant were detected either at adult (Figure 8B) or at seedling stages (Figure 8C). However, the lack of NTRC caused higher contents of Na, Cu and S, both at adult (Figure 8B) and seedling stages (Figure 8C), whereas the levels of K, which were higher in adult leaves of the ntrc mutant (Figure 8B), showed no significant differences in WT and ntrc mutant seedlings (Figure 8C). In contrast, opposite changes in the levels of Mn and P were detected in adult (Figure 8B) and seedlings (Figure 8C) of the ntrc mutant, as compared with the WT.