Towards the discovery of novel genetic component involved in stress resistance in Arabidopsis thaliana

Authors

  • Michal Juraniec,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Hélène Lequeux,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    2. Groupe de Recherche en Physiologie Végétale, Earth and Life Institute, Université catholique de Louvain, Louvain-La-Neuve, Belgium
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Christian Hermans,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    Search for more papers by this author
  • Glenda Willems,

    1. Gregor Mendel Institute of Molecular Plant Biology, Austrian Academy of Sciences, Vienna, Austria
    2. Max Planck Institute for Plant Breeding Research, Cologne, Germany
    Search for more papers by this author
  • Magnus Nordborg,

    1. Gregor Mendel Institute of Molecular Plant Biology, Austrian Academy of Sciences, Vienna, Austria
    Search for more papers by this author
  • Korbinian Schneeberger,

    1. Max Planck Institute for Plant Breeding Research, Cologne, Germany
    Search for more papers by this author
  • Pietrino Salis,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    Search for more papers by this author
  • Maud Vromant,

    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    Search for more papers by this author
  • Stanley Lutts,

    1. Groupe de Recherche en Physiologie Végétale, Earth and Life Institute, Université catholique de Louvain, Louvain-La-Neuve, Belgium
    Search for more papers by this author
  • Nathalie Verbruggen

    Corresponding author
    1. Laboratory of Plant Physiology and Molecular Genetics, Université Libre de Bruxelles, Brussels, Belgium
    • Author for correspondence:

      Nathalie Verbruggen

      Tel: +32 (0)2 650 21 28

      Email: nverbru@ulb.ac.be

    Search for more papers by this author

Summary

  • The exposure of plants to high concentrations of trace metallic elements such as copper involves a remodeling of the root system, characterized by a primary root growth inhibition and an increase in the lateral root density. These characteristics constitute easy and suitable markers for screening mutants altered in their response to copper excess.
  • A forward genetic approach was undertaken in order to discover novel genetic factors involved in the response to copper excess. A Cu2+-sensitive mutant named copper modified resistance1 (cmr1) was isolated and a causative mutation in the CMR1 gene was identified by using positional cloning and next-generation sequencing.
  • CMR1 encodes a plant-specific protein of unknown function. The analysis of the cmr1 mutant indicates that the CMR1 protein is required for optimal growth under normal conditions and has an essential role in the stress response. Impairment of the CMR1 activity alters root growth through aberrant activity of the root meristem, and modifies potassium concentration and hormonal balance (ethylene production and auxin accumulation).
  • Our data support a putative role for CMR1 in cell division regulation and meristem maintenance. Research on the role of CMR1 will contribute to the understanding of the plasticity of plants in response to changing environments.

Introduction

Copper (Cu) is an essential metal for normal plant development but becomes rapidly toxic in excess. It is the cofactor of enzymes involved in many biochemical processes, including photosynthesis, respiration, detoxification of peroxide anions, ethylene perception and cell wall (CW) metabolism (Burkhead et al., 2009; Cohu & Pilon, 2010). The average content of Cu in plant tissue ranges from 2 to 50 μg g−1 DW (Epstein & Bloom, 2005; Cohu & Pilon, 2007). Cu is highly toxic as the redox cycling between Cu(I) and Cu(II) catalyses the production of hydroxyl radicals via Fenton's reaction (Drążkiewicz et al., 2004). Symptoms of toxicity usually appear when the Cu concentration exceeds 20 μg g−1 DW in vegetative tissues (Marschner, 1995). Most of the genes up-regulated by Cu excess are not specific to Cu, probably because they respond to the production of reactive oxygen species (ROS; Zhao et al., 2009). Different strategies have evolved in plants to regulate Cu homeostasis in response to available environmental Cu, and several key players have already been identified, such as the family of Cu transporters (COPT), a Cu-transporting P-type ATPase (HMA5) or two Cu chaperones, antioxidant protein1 (ATX1) and ATX1-like Cu chaperone (CCH; Puig & Thiele, 2002; Sancenón et al., 2004; Hanikenne et al., 2005; Andrés-Colás et al., 2006, 2013; Shin et al., 2012).

Plants can adapt to an excess of trace metals in many ways; among others, they are capable of reorganizing their root system architecture (RSA) by inducing primary root (PR) growth inhibition and an increase in the lateral root (LR) density (Potters et al., 2007; Lequeux et al., 2010; Gruber et al., 2013; Verbruggen & Hermans, 2013). Although the morphological changes are generic, they may not be induced through the same signaling pathway (Potters et al., 2007, 2009). Plant hormones, mainly auxin, cytokinin and ethylene, control RSA and remodel characteristics of the root, including PR and LR growth as well as root hair (RH) formation (Aloni et al., 2006; Nibau et al., 2008; Moubayidin et al., 2009; Potters et al., 2009).

The aim of this work is to better understand the mechanisms of Cu excess tolerance in plants by identifying novel genetic factors via forward genetics. A phenotypic screening, based on PR growth on high-Cu medium, was performed in Arabidopsis thaliana in order to find mutants with altered resistance to Cu excess. One Cu2+-sensitive mutant called copper modified resistance1 (cmr1) was isolated and the CMR1 gene was identified. CMR1 is involved in meristem maintenance, which is necessary for optimal growth under control conditions and is required for survival under various environmental stresses.

Materials and Methods

Plant material

Three collections of Arabidopsis thaliana (L.) Col-0 seeds were screened: M2 ethyl methane sulfonate (EMS), fast neutron (FN)-mutagenized (Lehle Seeds, Round Rock, TX, USA) and T4 T-DNA Weigel's lines (NASC, University of Nottingham, Loughborough, UK). The DR5::GUS reporter construct was provided by C. Périlleux (University of Liège, Belgium); pCYCB1;1::GUS and p35S::GUS by T. Beeckman (VIB, University of Ghent, Belgium); pme3 (Gabi-Kat line 00210, T-DNA in the first exon) by V. Lionetti (Sapienza University of Rome, Italy); SALK_070337 transformed with At3g14190 by R. Mercier (INRA Versailles-Grignon, France); and pACS4::GUS by Hong Kong University of Science and Technology (China). T-DNA lines for the At3g14190 locus (SALK_035661, SALK_070337, CS811152) were obtained from NASC and genotyped using appropriate primers (Supporting Information Table S11). The cmr1-1 was backcrossed three times to wild-type (WT) Col-0, while cmr1-2 (SALK_070337) was backcrossed once.

In vitro growth conditions

Seeds were surface-sterilized and plated on half-strength medium MS/2 (Murashige & Skoog, 1962) agar media as described in Lequeux et al. (2010). The seedlings were grown on vertical plates under an 8 : 16 h, dark : light regime (50 μmol m−2 s−1) at 20°C. For most experiments, seedlings were transferred 5 d after germination (DAG) onto MS/2 supplemented with or without selected stress agents. Root length measurements were performed using image analysis software RootSnap CI-690 (CID Bio-Science Inc., Camas, WA, USA).

Mineral analysis

Three-week-old in vitro-grown plants were harvested. Roots and shoots were washed with deionized water, dried at 60°C and crushed into fine powder. Elemental concentrations were determined as described in Gruber et al. (2013).

Ethylene measurements

Seedlings were grown as described in Lequeux et al. (2010). Ethylene production was quantified with the ETD300 photoacoustic ethylene detector (Sensor-Sense, Nijmegen, the Netherlands) as described in Cristescu et al. (2013).

β-glucuronidase expression analysis

β-Glucuronidase (GUS) expression analysis was performed as described in Lequeux et al. (2010).

Confocal and scanning electron microscopy

Root meristems of 1-wk-old seedlings were analysed using a Leica SP2 AOBS 405 microscope (Leica Microsystems, Mannheim, Germany). The material was incubated for 2 min in 10 μM propidium iodide (PI) to stain cell walls (CWs) and was observed under an epifluorescence light with the appropriate filter set (excitation filter, BP 540–552 nm; dichroic mirror, 565 nm; barrier filter, 580–620 nm). To quantify the size of meristem ≥ 30 seedlings were collected from three growth experiments in two biological repeats. GFP fluorescence was analysed using a Zeiss LSM 710 microscope and detected with a 500–550 nm bandpass emission filter.

DNA isolation, amplification and sequencing

Frozen plant material was homogenized in extraction buffer (200 mM Tris-HCl, pH 7.5; 250 mM NaCl; 25 mM EDTA, pH 8; 0.5% SDS). DNA was isolated using phenol extraction and precipitated with isopropanol. DNA used for genotyping or sequencing was amplified by PCR using Taq polymerase according to the manufacturer's protocol (Fermentas, Brussels, Belgium). DNA samples for sequencing were purified with DNA Clean & Concentrator™-5 Kit (Zymo Research, Orange, CA, USA) according to the manufacturer's instructions and sequenced by MWG Biotech (Eurofins MWG Operon, Ebersberg, Germany).

RNA extraction and cDNA synthesis

Total RNA was extracted using the Aurum Total RNA Mini Kit (Bio-Rad, Nazareth Eke, Belgium) according to the manufacturer's instructions. cDNA was prepared using the RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas) according to the manufacturer's protocol.

Genetic mapping

Genetic mapping was performed as described in Hermans et al. (2010).

Transcriptomic analysis

Seeds of cmr1-1 and WT were germinated as described earlier. At 7 DAG, seedlings were transferred onto MS/2 supplemented with or without 25 μM CuSO4. Shoots and roots were harvested 24 h after the transfer and frozen in liquid nitrogen. Total RNA was extracted from 100 mg of tissue powder using TRIzol according to the manufacturer's instructions (Rneasy, Qiagen, Venlo, the Netherlands). The RNA labelling and the microarray hybridization and scanning were performed at the Institute of Life Sciences at the Université catholique de Louvain-la-Neuve (Agilent microarray platform, http://www.uclouvain.be/en-276229.html). Genes whose expression was three times more induced or repressed (cutoff = 3) in the mutant compared with the WT were used for further analysis. The MIPS functional catalogue database (http://mips.helmholtz-muenchen.de/proj/funcatDB/) was used to determine functional categories significantly regulated in cmr1 compared with WT.

Illumina sequencing

DNA was isolated from c. 300 mutant F2 plants using Genomic DNA Purification Kit (Fermentas) according to the manufacturer's instructions. The Illumina library was prepared using the Genomic Sample Prep Kit (Illumina, San Diego, CA, USA), according to the manufacturer's instructions. Insert size of the libraries correspond to c. 200 bp. Sequencing was performed by the sequencing core facilities at the Gregor Mendel Institute (Vienna, Austria). Paired-end reads of 75 bp length were generated on Illumina GAII. Reads were aligned to the Col-0 reference genome sequence using Burrows-Wheeler Aligner software (Li & Durbin, 2009). Read alignments were visualized using Integrated Genomics Viewer (www.broadinstitute.org/software/igv/home).

CMR1 coding sequence cloning

At3g14190 coding sequence (CDS) was PCR-amplified using a Pfu polymerase according to the manufacturer's protocol (Fermentas). The blunt-end PCR product was cloned using the pENTR/D-TOPO cloning system. The LR reaction using a pK7WGF2 destination Gateway vector (Karimi et al., 2002) was performed according to the manufacturer's instructions (Invitrogen).

Cellular localization of CMR1 in tobacco protoplast

Seeds of tobacco (Nicotiana plumbaginifolia) were sown on MS medium containing 3% sucrose. The plants were grown under an 8 : 16 h, dark : light regime (100 μmol m−2 s−1) at 20°C. Protoplasts were isolated from 2-month-old tobacco leaves and transformed as described in Hichri et al. (2010).

Statistical analysis

Results obtained with small sample size were analyzed using the nonparametric Wilcoxon–Mann–Whitney exact test for two independent groups (StatXact-9; Cytel Studio, Cambridge, MA, USA). For three or more independent groups, the nonparametric Kruskal–Wallis exact test was used, followed by a post hoc test. For relative means, statistical analyses were investigated using Student's t-test. For other results (obtained with large sample size), a parametric ANOVA with Tukey's multiple comparison test was performed using SAS 9.1 (SAS Institute Inc., Cary, NC, USA).

Results

Isolation of a copper-sensitive mutant and identification of its causative mutation

As, upon Cu2+ excess, root growth is significantly more affected than shoot growth (Lequeux et al., 2010) and PR length is easily quantifiable in vitro, a screening on vertical plates was performed in order to isolate mutants exhibiting root growth impairment in response to Cu2+ excess. Two different concentrations of copper sulphate were used: 25 and 50 μM. At 10 d after transfer (DAT), PR growth of WT plants was only slightly affected by 25 μM CuSO4, which was chosen to screen for sensitive mutants. By contrast, PR growth of WT plants was severely inhibited in the presence of 50 μM CuSO4, which was used to identify tolerant mutants. About 42 500 M2 seedlings were screened on 25 μM CuSO4 and 86 putative sensitive mutants were selected. However, only two mutants were confirmed in M3 (Table S1): one mutant was partially fertile and difficult to regenerate; the other one, derived from the FN-mutagenized population and named cmr1, which showed a relative PR growth reduction of 55% at 10 DAT (Fig. 1b,f), was studied further. About 16 500 M2 seedlings were screened on 50 μM CuSO4 and 12 putative tolerant mutants were selected, but none of them was confirmed in M3 (Table S1).

Figure 1.

Effect of Cu2+ doses on the in vitro growth of cmr1-1 and cmr1-2 Arabidopsis mutants. (a, b) Primary root (PR) elongation in seedlings transferred 5 d after germination onto MS/2 medium (0.1 μM CuSO4) (a) or on the same medium supplemented with 25 μM CuSO4 (b). (c, d) Root length (c) and shoot FW of 10 seedlings (d) at day 7 after transfer (DAT) onto MS/2 medium in the presence of increasing CuSO4 supply. Average values ( 30 seedlings in (a–c) and = 10 in (d) collected from three growth experiments in two biological repeats) ± SD; asterisks denote significant differences from the wild-type (WT) at  0.05; WT, black closed symbols; cmr1-1, open circles; cmr1-2, open triangles. (e, f) Cu2+-sensitive phenotype of cmr1-1 and cmr1-2. Seedlings were transferred (black line) 5 d after germination onto MS/2 medium (e) or medium supplemented with 25 μM CuSO4 (f). Pictures were taken 10 DAT (from left to right: WT, cmr1-1, cmr1-2 and F1 progeny derived from the cross between cmr1-1 and cmr1-2). Bar, 10 mm.

A map-based cloning was undertaken in order to identify the mutated locus responsible for the cmr1 phenotype. An F2 population was generated from a cross between cmr1 in the Col-0 background and WT Landsberg erecta (Ler-1). F1 seedlings grown in the presence of CuSO4 exhibited the WT phenotype and 25% of F2 were Cu-sensitive, indicating that the cmr1 mutation was recessive (data not shown). The DNA samples of F2 Cu-sensitive mutant individuals were analysed. The lowest recombination frequencies were calculated within the upper arm of the third chromosome. Subsequently, fine-mapping (> 1500 F2 individuals tested) allowed the delimitation of a 184-kb region of interest comprised between IN464 and IN482 markers (Fig. S1; Table S7) containing 53 genes (Table S2). Genes with a putative function in stress tolerance were sequenced in the cmr1 background, and corresponding T-DNA knockout mutants were phenotyped on high Cu, but no candidate could be confirmed. Therefore, other approaches were followed to define the causative mutation. To verify the expression of candidate genes within the 184 kb mapped region, a microarray analysis of cmr1 was undertaken. At3g14310, which encodes PECTIN METHYLESTERASE 3, was strongly repressed in cmr1 as compared with WT in control and high-Cu conditions (Table S3). However, the pme3 knockout mutant was not Cu2+-sensitive (Fig. S2a). Among the genes differentially expressed between cmr1 and WT, no differences in sequences were found, except for At3g14190, which could not be PCR-amplified in cmr1 (using primers at positions −487 and +314; Fig. S3), indicating the presence of a possible mutation at this locus. The At3g14190 mRNA levels were higher in cmr1 than in WT in all tested conditions (Table S3), which was also confirmed by reverse transcription polymerase chain reaction (RT-PCR) (data not shown).

In parallel, a SHOREmap approach was followed. By deep sequencing of a pool of F2 mutant individuals, the location of the mutation can be inferred based on the relative allele frequency observed at the single nucleotide polymorphism (SNP) markers between Col-0 and Ler-1 accessions (Schneeberger et al., 2009). However, since the mutant cmr1 line was obtained through FN mutagenesis, which mostly induces big insertions or deletions rather than point mutations (Li & Zhang, 2002; Alonso & Ecker, 2006; Belfield et al., 2012), SHOREmap was unsuccessful in identifying candidate SNPs at which the mutation could have taken place in the 184 kb interval. Consequently, after aligning paired-end reads using BWA (Li & Durbin, 2009), coverage and read alignments were analysed in the 184 kb interval. Visual inspection of cmr1 read coverage to the WT Col-0 reference genome revealed the presence of two low-coverage regions: one in the At3g14190 gene and the other in the At3g14310 gene (Fig. S4). Moreover, some read pairs, normally spaced c. 200 bp apart, were broken. One end of these read pairs was mapped in the At3g14190 gene, while the other end of the read pairs was aligned to the At3g14310 gene. Those genes are spaced in the reference genome 65 kb apart (Fig. S5). Sequencing of DNA segments amplified using primers belonging to both genes (Fig. S6a–d) showed that a 65 kb inversion occurred between At3g14190 and At3g14310 loci (Fig. S6e). As a consequence, the promoter region and the first 160 nucleotides of the At3g14310 CDS were placed upstream the At3g14190 CDS but not in frame (Fig. S6f). Moreover, because of the reading frame shift, a stop codon appeared 165 bp downstream of the At3g14310 start codon, explaining why the expression of At3g14310 was strongly down-regulated in cmr1 (Table S3). As a result, both genes are disrupted in cmr1.

To confirm the role of the At3g14190 gene in the cmr1 stress-sensitive phenotype, the response to CuSO4 excess was investigated using available T-DNA lines (Table S4). Three lines carrying the T-DNA in the At3g14190 CDS (SALK_035661, SALK_070337 and CS811152) exhibited a Cu2+-sensitive phenotype similar to that of cmr1 (Fig. 2a), confirming the involvement of the At3g14190 gene product in CuSO4 tolerance. A test for allelism was performed by crossing cmr1 to the T-DNA lines. The F1 exhibited a similar Cu2+-sensitive phenotype to cmr1 (Fig. 2a,b). The SALK_070337 was chosen for further work and was designated the cmr1-2 allele, while the original FN-induced mutant allele was named cmr1-1. Additionally, a functional complementation of cmr1-2 with the genomic sequence of CMR1 was performed to confirm that the loss of function of the At3g14190 gene was responsible for Cu-modified resistance. Four independent homozygous transformants were selected and tested on 25 μM CuSO4. All lines exhibited restored root length, further supporting the idea that mutation at the At3g14190 locus was responsible for the Cu2+-sensitive phenotype in cmr1 (Fig. 2c).

Figure 2.

Allelism test between Arabidopsis cmr1-1 and T-DNA mutants at the At3g14190 locus. (a) Pictures were taken 10 d after germination on MS/2 or the same medium supplemented with 25 μM CuSO4. (b) Quantification of root lengths in cmr1-1, cmr1-2 (SALK_070337) and their F1 progeny grown in control and high-Cu conditions for 7 d after transfer. Average values ( 30 seedlings from three different growth experiments) ± SD; asterisks denote significant differences from the wild-type (WT) within each treatment at  0.05. (c) Root length in the cmr1-2 knockout mutant and four independent lines transformed with the WT copy of the CMR1 gene in response to Cu excess in vitro. Seedlings were transferred 5 d after germination onto MS/2 medium supplemented with 25 μM CuSO4 for 1 wk. Average values ( 30 seedlings from three growth experiments) ± SD; asterisks denote a significant difference in the mutant as compared with the WT at  0.05. WT, black bar; cmr1-2 KO, dashed bar; cmr1-2 transformed lines with CMR1 genomic sequence, grey bars.

In silico analysis of the At3g14190 gene

The At3g14190 gene, which we called CMR1 once the mutant was isolated, encodes a plant-specific protein of unknown function, predicted to be involved in cell proliferation and localized in the nucleus (www.arabidopsis.org). The corresponding deduced CMR1 protein has a predicted molecular weight of 21.7 kDa and a size of 193 amino acids. The sequence comprises a destruction box motif (D-box) RKALNDITN in the N-terminal region of the protein (http://bioinfo.weizmann.ac.il/~danag/d-box/main.html), which is a target for ubiquitin-dependent proteasome proteolysis (Vandepoele et al., 2002). Cyclebase.org indicates that the expression of the At3g14190 gene is modulated during the cell cycle with a peak at the early G1 phase (Gauthier et al., 2008). CMR1 shares 43% identity with the At5g12360 gene. Interestingly, a homolog of At3g14190 was identified in rice, sharing 33% identity within the N-terminal sequence comprising the D-box, and the corresponding mutant rice salt sensitive1 (rss1) was described (Ogawa et al., 2011). In addition, CMR1 shares significant identity with proteins from other plant species, such as Arabidopsis lyrata, Populus trichocarpa, Vitis vinifera and Glycine max, all with unknown functions. No CMR1-homologue was found in any other organisms than plants. According to Genevestigator.org, the organs with the highest CMR1 expression level are the ovule, the inflorescence shoot apex, the rosette shoot apex and axillary shoot, the silique replum, the root tip (RT) meristem and the shoot apical meristem.

Subcellular localization of CMR1

A nuclear localization of CMR1 was predicted using PSORT software (http://wolfpsort.org/). The presence of a nuclear localization signal in its sequence (PIHRKKS) was detected. We fused the CDS of the At3g14190 gene to the GFP sequence under the control of the 35S promoter in the pK7WGF2 binary vector and checked its expression in tobacco protoplasts. The same but non-recombined vector was used as a control. Fluorescence was localized in both the nucleus and the cytoplasm (Fig. 3).

Figure 3.

Nuclear localization of the GFP–CMR1 fusion protein. Tobacco protoplasts were transformed with an empty vector pK7WGF2 as a control or with the same vector containing the GFP fused to the N-terminal of the CMR1 gene and placed under control of the 35S promoter. Triangles indicate the nucleus, while arrows indicate inclusion bodies. Bar, 10 μm.

Growth of cmr1 in control and stress conditions

The phenotypes of cmr1-1 and cmr1-2 were further characterized in control and on Cu excess conditions. In all tested conditions, the root growth of cmr1-1 and cmr1-2 was not significantly different. On control medium, both mutants displayed identical seed germination rates but exhibited significant ( 0.05) reduction of root elongation rate by −20% at 10 DAT compared with WT (Figs 1a,e, 2a,b). Growth on a Cu gradient was measured at 7 DAT. Both mutants showed ± 40% root length reduction after transfer onto 10 μM CuSO4 (Fig. 1c). Above 25 μM CuSO4, the WT root growth sharply decreased. Cu sensitivity was also observable in shoots as their biomass was reduced by ± 20% in both mutants as compared with WT at 7 DAT onto MS/2 supplemented with 25 μM CuSO4 (Fig. 1d). No difference was measured in shoot biomass between WT and mutants at 7 DAT in control conditions. The stress-sensitive phenotype of cmr1-1 was also confirmed in older plants grown in hydroponic culture; however, the growth inhibition was weaker than the one observed in in vitro-grown seedlings (Fig. S7).

To verify the specificity of the cmr1 phenotype, root growth in response to various abiotic stresses, especially trace metallic elements and osmoticum, was analysed (Fig. 4). Applied stresses were selected on the basis of their inhibitory effect on root growth in WT, to be comparable to 25 μM CuSO4. cmr1-2 was significantly ( 0.05) sensitive to the excess of MnCl2, CoCl2, CdSO4 and ZnCl2, as shown by a root length reduction of 8, 53, 56 and 75%, respectively, as compared with WT. cmr1-2 root growth was particularly sensitive to salt stress, with an inhibition of 90% on 50 mM NaCl (Fig. 4b). The cmr1-2 PR growth was also inhibited upon both 50 mM KCl and 10 mM LiCl treatment as compared with WT, as well as on Na2SO4 and K2SO4 excess and sorbitol, reaching 85, 92, 72, 82 and 54%, respectively, of length reduction. A similar range of sensitivity was measured in cmr1-1 (data not shown). All these data indicate that both cmr1 alleles are sensitive to various abiotic stresses and without any specificity to mono- or divalent ions. The root growth on 50 mM NaCl was also checked in the F1 derived from the cross between the two cmr1 mutant alleles (Fig. S8a). The PR length measured 30% of that in WT (Fig. S8b). Ca2+ ions are known to ameliorate Na+ toxicity symptoms by decreasing Na+ influx (Shabala et al., 2006). We checked whether the addition of Ca2+ had a beneficial impact on the Na+-sensitive phenotype of cmr1-1. Although the addition of 10 mM Ca(NO3)2 slightly reduced the PR growth in cmr1-1 upon control conditions, it only partially restored the PR length during NaCl stress (data not shown).

Figure 4.

Relative primary root (PR) growth of cmr1-2 Arabidopsis plantlets in response to various external nutritional stimuli (a) and osmoticum (b) measured 10 d after transfer onto MS/2 medium (0.1 μM CuSO4) or on the same medium supplemented with an excess of different chemicals. Average values ( 30 seedlings from three independent experiments) ± SD; asterisks denote significant differences between the cmr1-2 mutant and the wild-type (WT) within each treatment at  0.05 (WT, black bars; cmr1-2, dashed bars).

Mineral profile of cmr1

To investigate the possible impact of the cmr1 mutations on Cu and other essential element homeostasis, mineral analyses of 3-wk-old seedlings grown on control, 25 μM CuSO4- or 25 mM NaCl-supplemented media were assayed (Fig. 5; Tables S5,S6). No significant ( 0.05) differences in Cu concentration of root and shoot organs were observed between WT and mutant genotypes, regardless of the Cu or salt treatments. Interestingly, potassium (K) content was significantly lower in both control and stress conditions (Fig. 5, Table S6). In control conditions, the K concentrations of root and shoot tissues were one-tenth less in cmr1-1 and one-fifth less in cmr1-2, respectively, compared with WT. Upon Cu exposure, cmr1-1 contained 44 and 26% K less in shoots and roots, respectively, than the WT (Table S6). Upon NaCl exposure, K concentration in tissues was lower in cmr1-2 than in WT, with a difference of one-fourth in both shoots and roots. Besides, cmr1-2 exhibited significantly higher Zn concentration in roots upon NaCl excess and higher Mo in both shoots and roots in control conditions and during salt stress (Table S5).

Figure 5.

Macronutrient and sodium content in in vitro-grown cmr1-2 Arabidopsis plants. Shoots and roots were harvested separately 3 wk after transfer onto MS/2 medium (a, c) or on the same medium supplemented with 25 mM NaCl (b, d). Mineral concentrations are expressed in μg g−1 DW. Values are means of five pools of c. 80 organs each ± SE. Asterisks indicate significant differences from the wild-type (WT) at  0.05 (WT, black bars; cmr1-2, dashed bars).

The root tip phenotype of cmr1

As a particular swelling of RT was systematically noticed in the cmr1 mutants after transfer to Cu- or NaCl-enriched media, a detailed observation of RT was conducted using light and confocal microscopy. In control conditions, except for the presence of longer RHs, cmr1 RT showed no specific morphological changes (Fig. 6b). However, at 6 DAT onto 25 μM CuSO4, cmr1 RTs were swollen and formation of numerous RHs was induced. In cmr1, the RH initiation region was closer to the root meristem and RHs were longer than those of WT. In contrast to WT, in which RH formation was inhibited upon salt stress (Fig. 6a), cmr1 had abundant RHs close to the root apex (Fig. 6b). Furthermore, the root phenotype of cmr1 upon salt stress was much more pronounced than that upon Cu stress: RHs were longer and the swelling of RT was greater. To get a better insight into RT anatomy and its cellular organization, PI staining of CWs was performed in in vitro-grown seedlings of cmr1-2 in control, Cu and saline conditions (Fig. 6c–g). In control conditions, the shape of cells, especially those in the cortical and epidermal layers, was less regular in cmr1-2 and cells were often slightly bigger than those in the WT (Fig. 6d). Growth upon 25 μM CuSO4 induced a reduction of meristem size in mutant roots; additionally, the cell shape and size were strongly altered. Cell layers within each tissue were irregular. This phenotype was even stronger upon NaCl stress conditions, where cells were highly irregular and the differentiation zone appeared just after the reduced meristem. Detailed examination of the meristem revealed perturbations during cell proliferation in cmr1-2, and asymmetric divisions were detected in control conditions (Fig. 6f). When the mutant was exposed to salt stress, the functioning of the meristem was damaged and new CWs were inserted arbitrarily, causing a highly irregular cellular pattern (Fig. 6g). The size of root meristem can be examined by measuring the number of cortical cells in a file extending from the quiescent centre to the first elongated cell (Perilli & Sabatini, 2010). A small but significant ( 0.05) decrease in the number of meristematic cortex cells of 7-d-old seedlings in control conditions was observed in cmr1-2 compared with WT (25 ± 4 vs 29 ± 3 cells). Upon exposure to high Cu or salt, the RT was visibly reduced in the mutant (Fig. 6d). However, owing to the irregular cellular pattern, it was difficult to quantify the number of cells.

Figure 6.

Impact of Cu2+ and Na+ excess on root tip morphology and root meristem in cmr1-1 and cmr1-2 Arabidopsis mutants. (a, b) Pictures of representative primary root (PR) tips of wild-type (WT) (a) and cmr1-1 (b), 6 d after transfer onto MS/2 (0.1 μM CuSO4) medium supplemented with 25 μM CuSO4 or 50 mM NaCl. Bar, 500 μm. (c–g) Pictures of representative propidium iodide-stained root tips and magnifications of their quiescent centres in 7-d-old WT (c, e) and in the cmr1-2 mutant (d, f, g) seedlings 24 h after transfer onto MS/2 medium, or the same medium supplemented with 25 μM CuSO4 or 50 mM NaCl. Triangles indicate arbitrarily inserted cell walls. Bars: (c) 50 μm; (e) 20 μm.

In order to check the effect of CMR1 loss of function on cell divisions, an introgression of pCYCB1;1::GUS construct into the cmr1-1 background was generated. In WT cells, CYCB1;1 is expressed in late S to mid-M phase cells (Colón-Carmona et al., 1999; Dohmann et al., 2008; Wu et al., 2010). In WT roots, independently of growth conditions, GUS expression under the activity of CYCB1;1 promoter was weak and punctuate, reflecting the population of cells undergoing mitosis at the time of staining (Fig. 7). By contrast, in the cmr1-1 root meristem, the pattern of staining was still patchy, but the number of blue-stained cells and staining itself increased considerably after 48 h of Cu2+ treatment.

Figure 7.

Effect of Cu2+ excess on CYCB1;1 expression in primary root (PR) tips of the Arabidopsis cmr1 mutant. The pCYCB1;1::GUS reporter line was used as a control to pCYCB1;1::GUS introgression into the cmr1-1 background. Seedlings were transferred 5 d after germination onto plates containing MS/2 medium supplemented with or without 25 μM CuSO4 for 48 h. Bar, 500 μm.

Hormonal imbalance in cmr1-1

Root morphological characteristics of cmr1 could be linked to ethylene, namely inhibition of root growth (Le et al., 2001; Swarup et al., 2007), RT swelling (Smalle & Van Der Straeten, 1997) and induction of ectopic RHs (Tanimoto et al., 1995). Therefore, we monitored the gaseous hormone production of whole seedlings grown vertically on agar media (Fig. 8a–c). Ethylene production of cmr1-1 was similar to that of WT under control conditions and twice as much after treatment with 25 μM CuSO4. Under salt stress, no difference between cmr1-1 and WT could be detected. To examine the production of the ethylene precursor, 1-aminocyclopropane-1-carboxylate (ACC), the ACC SYNTHETASE 4 pACS4::GUS construct was introgressed into the cmr1 background. In contrast to normal conditions, where weak GUS staining was visible in the root meristem, strong and patchy pattern staining was observed in cmr1 at 2 DAT onto 25 μM CuSO4 (Fig. 8d). That observation further supports higher ethylene production. In both control and high-Cu conditions, weak staining was detected in the vascular regions of the differentiated zone in the mutant root, in contrast to WT. The impact of silver (an inhibitor of ethylene biosynthesis) on the root mutant phenotype was tested. The addition of 1 μM Ag2SO4 was not able to ameliorate the cmr1 root growth considerably (Fig. 8e) or to restore normal pACS4::GUS staining (data not shown) upon high-Cu exposure. A similar result was observed for another inhibitor, aminoethoxyvinylglycine (data not shown). As ethylene up-regulates root auxin biosynthesis in order to maximize its ability to inhibit root cell expansion (Swarup et al., 2007), the activity of GUS in cmr1 crossed to the DR5::GUS reporter line was analysed to monitor auxin signals (Ulmasov et al., 1997). Compared with the WT background, RT in cmr1-1 exposed to 25 and 50 μM CuSO4 showed stronger GUS activity, reflecting an increased accumulation of IAA (Fig. 8f,g). In the control medium, the staining in cmr1-1 was similar to that in the WT.

Figure 8.

Impact of Cu2+ and Na+ excess on hormonal imbalance in in vitro-grown Arabidopsis cmr1-1 plantlets. (a–c) Effect of Cu2+ or salt excess on ethylene production by 9-d-old cmr1-1 seedlings grown on MS/2 medium (control), or the same medium supplemented with 25 μM CuSO4 (a), 50 μM CuSO4 (b) or 50 mM NaCl (c). Wild-type (WT), black bars; cmr1-1, white bars. Average values (= 30 seedlings from three independent growth experiments) ± SE. For each growth condition, measurements were concurrently monitored and repeated at least three times. Asterisks denote significant differences between genotypes at  0.05. (d) Impact of Cu2+ excess on ACS4 expression in primary root (PR) tips in cmr1-1. Seedlings were transferred 5 d after germination onto MS/2 or the same medium supplemented with 25 μM CuSO4 for 48 h. Bar, 500 μm. (e) Impact of silver ions on the cmr1-1 Cu2+-sensitive root phenotype. PR length was measured 13 d after germination (DAG); 5-d-old seedlings were transferred onto MS/2 or the same medium supplemented with 25 μM CuSO4 and/or 1 μM Ag2SO4. WT, black bars; cmr1-1, white bars. Average values ( 30 from three different growth experiments) ± SE. Letters denote significant differences between treatments at  0.05. (f, g) Effect of Cu2+ excess on auxin distribution in PR tips in the cmr1-1 mutant. β-Glucuronidase (GUS) staining of the root apex of the DR5::GUS reporter (f) and DR5::GUS × cmr1-1 lines (g). Seedlings were transferred 5 d after germination onto plates containing MS/2 medium or media supplemented with different CuSO4 concentrations for 24 h. Bars, 500 μm.

Discussion

Plants have evolved different strategies to cope with high concentrations of Cu. The goal of this work was to identify novel genetic components involved in Cu tolerance. Up to now, forward genetics have only allowed the identification of one Cu2+-sensitive mutant, cup1-1, on high-cadmium medium, of which the responsible gene was not identified (van Vliet et al., 1995). Here we present the characterization of cmr1, a Cu2+-sensitive mutant, and the cloning of the corresponding gene. CMR1 is not only involved in Cu tolerance but also more generally in growth under normal, and especially under abiotic, stress conditions.

Identification of the cmr1 mutation and cloning of the CMR1 gene

The frequency of identified Cu2+-sensitive mutants appeared relatively low (4.7 × 10−5) as compared with those reported in the literature (Howden & Cobbett, 1992; Wu et al., 1996; Zhu et al., 1998). Several explanations can be found: mutants displaying weaker phenotype may have gone undetected; we can assume that the mechanisms of Cu tolerance are of such importance for plant survival that mutants affected in Cu detoxification mechanisms are lethal (Howden & Cobbett, 1992); and functional redundancy between the components of Cu detoxification mechanisms could have hampered the identification of sensitive mutants, a hypothesis also suggested by Howden & Cobbett (1992) to explain their low frequency observed in the screen for Cd-sensitive mutants. Nonetheless, our work describes the first successful identification of a mutant on high-Cu medium.

A map-based approach was undertaken to identify the position of the cmr1-1 mutation and the gene was eventually cloned using transcriptomics and next-generation sequencing on a mutant pool. Based on the read mappings, a large rearrangement in cmr1-1 involving a 65 kb inversion between At3g14190 and At3g14310 loci was identified. To our knowledge, such a genomic rearrangement after FN mutagenesis is quite exceptional in a plant genome. Deletions and inversions ranging from several bp to 30 kb have been described previously (Shirley et al., 1992; Li & Zhang, 2002; Belfield et al., 2012). However, another example, a large 460 kb deletion, was detected in a soybean supernodulation FN37 mutant (Men et al., 2002). As a consequence of the inversion, the At3g14190 CDS was placed under the activity of the At3g14310 promoter in cmr1-1. According to Genevestigator, the At3g14310 promoter activity is higher than the one of At3g14190, explaining why expression of the At3g14190 gene was higher in cmr1-1. Furthermore, the translation initiation codon of the At3g14310 gene was not in frame, implying that the At3g14310-encoded pectin methylesterase 3 was not functional in cmr1-1. However, the disruption of the At3g14310 gene had no impact on the cmr1-1 Cu2+-sensitive phenotype. This observation is consistent with the recent results of Weber et al. (2013) on pme3 loss-of-function mutants. An allelism test using cmr1-1 and cmr1-2, as well as complementation of the SALK_070337 line with the At3g14190 genomic fragment both validated the identity of At3g14190 as the CMR1 gene. Ogawa et al. (2011) have described an RSS1 protein that maintains meristematic activity in both shoots and roots under stress conditions in rice. RSS1 is homologous to the protein encoded by the At3g14190 gene. Like CMR1, RSS1 possesses a D-box in its N terminus and is localized both in the nucleus and in the cytosol.

Role of CMR1 in the regulation of growth in control and stress conditions

The fact that the root growth was slightly but significantly affected in control conditions supports a role for CMR1 in normal plant growth and development. Detailed observation of the root meristem showed an impairment in the establishment of the cell division plane, but only with a weak impact on root length and growth. There was no significant growth reduction of the cmr1 shoot in control conditions, suggesting that shoot apical meristem was not affected by the loss of function of CMR1. The analysis of shoot phenotype requires further studies. By contrast, the phenotype was more pronounced, but to a similar level, in both cmr1 mutant alleles under Cu excess, indicating that the presence of CMR1 is crucial for development during stress conditions. Similarly to cup1-1, stress sensitivity of cmr1 was not restricted to Cu. A strong growth reduction of PRs and LRs was observed upon all tested abiotic stresses (CdSO4, CoCl2, KCl, K2SO4, LiCl, MnCl2, NaCl, Na2SO4, ZnCl2 and sorbitol). Furthermore, upon exposure to CdSO4, CuSO4, NaCl and ZnCl2, the root phenotype of cmr1 mutants also included a swelling of the RT, an outgrowth of RHs close to the root apex and a lengthening of RHs. Several observations, such as the higher ethylene production, might account for the root phenotype. It was demonstrated that the ethylene-induced root growth inhibition was associated with a reduction in meristem size caused by a premature differentiation of cells (Thomann et al., 2009). Here we showed that upon Cu2+ excess, cmr1 overproduced ethylene in vitro. However, the inhibitors of ethylene biosynthesis were not able to restore the cmr1 phenotype significantly, suggesting that ethylene did not directly induce growth inhibition. GUS staining of pACS4::GUS lines indicated an induction of ACS4 expression in the cmr1 root apex 48 h after transfer onto Cu-enriched medium (Arteca & Arteca, 2007). This result showed a localized induction of ethylene production in the cmr1 root apex, which could be related to the RT swelling and the formation of RHs close to the root apex. In addition, ethylene may also be responsible for the RH elongation. Indeed, eto mutants which overproduce ethylene were shown to produce longer RHs than the WT (Pitts et al., 1998). Synergistic effects of auxin and ethylene on root growth have been extensively studied using Arabidopsis mutants defective in ethylene and auxin signalling (Růžička et al., 2007; Stepanova et al., 2007; Swarup et al., 2007). Ethylene was previously shown to promote auxin biosynthesis in roots by the activation of several auxin biosynthesis genes (Stepanova & Alonso, 2005). Accordingly, an increase in DR5::GUS staining in RTs of Cu2+-treated cmr1 may reflect the ethylene-induced IAA accumulation. In addition, mutants with altered responses to auxin also show defects in RH length, suggesting that, apart from ethylene, auxin also plays a role in controlling RH growth (Pitts et al., 1998; Rahman et al., 2002).

Since maintenance of cellular potassium concentrations is critical for salt tolerance (Zhu et al., 1998), the Na+-sensitive phenotype of the cmr1 mutant could be associated with impaired K+ homeostasis. Roots deprived of K+ were shown to induce the expression of genes involved in ethylene biosynthesis and signalling (Shin & Schachtman, 2004). More recently, Jung et al. (2009) demonstrated a role of ethylene signalling in low K+-induced plant responses in Arabidopsis. It was shown that the low K+-induced ethylene synthesis in turn stimulated the production of ROS in the RH-forming zone, resulting in RH elongation and induction of the high-affinity K+ uptake transporter HAK5, which contributes to the plant survival (Jung et al., 2009). Ethylene overproduction in cmr1 upon Cu stress is probably not directly induced by altered K concentration, as K concentration was also lower in control conditions without concomitantly increased ethylene concentration. Besides, the addition of K+ upon Cu excess did not restore WT pACS4::GUS expression in the RT (data not shown), strongly suggesting that elevated ethylene production in cmr1-1 was not a direct consequence of lower K+ content. The fact that the addition of ethylene inhibitors or Ca(NO3)2 only partially restored the root growth in cmr1 in stress conditions suggests that another affected process impairs root growth.

There are several features of the cmr1 phenotype pointing to a deregulation in the cell cycle: inhibition of root growth, cell division defects and the outgrowth of RHs close to RT, which is a sign of a root meristem shrinkage (Culligan et al., 2004; De Schutter et al., 2007). We have observed similarities between the phenotype of cmr1 and wee1. WEE1 is a cell cycle regulatory kinase that is activated upon cessation of DNA replication or DNA damage (De Schutter et al., 2007). WEE1 knockout plants displayed root growth arrest in the presence of DNA-damaging chemicals, similar to that of Cu- or salt-grown cmr1 (De Schutter et al., 2007; Cools et al., 2011). The phenotype in wee1 was attributable to a failure to block its cell cycle in response to DNA stress. The cells progressed into mitosis prematurely, resulting in a loss of genome integrity. The overexpression of WEE1 resulted in a strong reduction of the meristematic zone, cell cycle arrest, outgrowth of RHs close to RT and premature cell differentiation. It is thus not excluded that, analogously to WEE1, CMR1 is involved in mitosis regulation. An increased number of cells expressing the pCYCB1;1::GUS construct, especially upon Cu2+ excess, was observed in the PR meristem of cmr1. Considering growth inhibition, this result may reflect a G2/M delay or arrest, resulting in a higher number of cells expressing the mitotic marker CYCB1;1 (Zhu et al., 2006; De Schutter et al., 2007; Dohmann et al., 2008). Similar results were observed in csn affected in the COP9 signalosome, which plays a role in the G2 phase progression (Dohmann et al., 2008) or in plants overexpressing the WEE1 kinase (De Schutter et al., 2007). However, Wu et al. (2010) have recently suggested that an accumulation of CYCB1;1 could also be associated with a defect in microtubule organization rather than an arrest in the cell cycle. They noticed that many mutants which accumulated CYCB1;1 were affected in organ polarity and presented a root swelling. PI staining revealed severe alterations in cell division plane establishment in root meristems, which could indeed suggest a microtubule defect during formation of the mitotic spindle. Moreover, Wu et al. (2010) showed an accumulation of CYCB1;1 in a mutant affected in microtubule organization without any arrest in the cell cycle.

Interestingly, the At3g14190 gene was differentially expressed across several microarrays related to the cell cycle (Beemster et al., 2005; Menges et al., 2005; Dewitte et al., 2007; Cools et al., 2011; Heyndrickx & Vandepoele, 2012). The analysis of cis-acting regulatory elements, by means of PLACE (www.dna.affrc.go.jp/PLACE/), in the upstream region of CMR1 revealed several motifs found in the promoter of genes involved in the cell cycle or expressed in meristematic regions (Table S8; Planchais et al., 2002; Ramirez-Parra et al., 2003; Trémousaygue et al., 2003). According to Genevestigator, CMR1 was shown to be highly expressed in shoot and root meristems, which are the main sites of dividing cells. Moreover, the presence of a D-box supports the involvement of the protein in cell cycle-dependent protein turnover. We also observed that the cell cycle category was overrepresented in genes differentially regulated in cmr1 relative to WT (Table S9). Finally, the transient expression of CMR1 fused to GFP in tobacco protoplasts confirmed the predicted nuclear localisation.

The resemblance between the cmr1 phenotype and that of rss1 is also significant. The latter was shown to be NaCl-, LiCl- and sorbitol-sensitive. Root growth and meristem size were strongly reduced under salinity conditions, so that the differentiation zone appeared close to RT. However, in contrast to cmr1, rss1 has no particular phenotype in normal growth conditions. RSS1 is predominantly expressed in root and shoot meristems, most abundantly during the G1 and S phases. RSS1 was proposed to antagonize the G1/S checkpoint in response to stress and induce a slower cell cycle progression to equilibrate cell division with cell differentiation. It is therefore not excluded that CMR1 plays a similar role to RSS1 in maintaining meristematic activity under saline conditions, but CMR1 also plays a role, though to a lesser extent, in root growth under control conditions.

In contrast to rss1, a defect in K homeostasis in cmr1 was shown. By using K+ channel blockers on synchronized BY-2 cells, Sano et al. (2007) demonstrated that a cellular K+ threshold was required for cells to re-enter the cell cycle from the G1 to the S phase. This K+ threshold is thought to be necessary for the proper turgor regulation of cycling cells. Interestingly, they showed that the expression of some K+ transporters was modulated during the cell cycle progression. For example, the inward-rectifying channel gene NKT1 was shown to be predominantly expressed in the G1 phase and responsible for K+ uptake during the G1-to-S phase transition. An Arabidopsis NKT1 ortholog, AKT1, is involved in K+ uptake by roots (Hirsch et al., 1998). It is therefore possible that the putative cell cycle defect in cmr1 has an impact on the K+ uptake detected in normal and stress conditions. In agreement with this, our microarrays revealed several differences in the expression of K+ transporters between the WT and cmr1-1. For instance, KCO1 and AKT5, outward- and inward-rectifying K channels, respectively, were more strongly expressed in Cu2+-treated roots of cmr1-1 than in the WT (Table S10), which was confirmed by RT-PCR (data not shown). In addition, the Na+ transporter AtHKT1, which controls Na+ homeostasis and in turn affects K+ acquisition (Rus et al., 2004), was repressed in Cu2+-treated roots of cmr1-1.

In summary, the present work describes a novel genetic factor involved in plant growth and stress response. Impairment of cmr1 activity alters root growth, meristem activity, K content, and ethylene and IAA accumulation. Research into the role of CMR1 in maintaining meristematic activity is ongoing and will contribute to understanding the plasticity of plants in response to changing environments.

Acknowledgements

This work was supported by a grant from the Belgian National Fund for Scientific Research FNRS-FRFC 2.4.527.10.F ‘Mechanisms of resistance to copper in bacteria and plants; evaluation of plant–bacteria interactions in plant resistance’. The authors thank V. Lionetti (Sapienza University of Rome, Italy) for pme3 seeds; R. Mercier (INRA Versailles-Grignon, France) for SALK_070337 transformed with At3g14190; N. von Wiren and S. Reiner (IPK Gatersleben, Germany) for ICP-OES analysis; F.J. Harren and S. Cristescu (Radboud University, Nijmegen, the Netherlands) for assistance with ethylene measurements carried out within the framework of the EU-FP6-Infrastructures-5 programme (project FP6-026183 Life Science Trace Gas Facility); O. Grandjean (INRA Versailles-Grignon, France) for confocal microscopy assistance; and P. Saumitou-Laprade (University of Lille, France) for access to the Light-Cycler 480. C.H. is an F.R.S-FNRS research associate.

Ancillary