The Arabidopsis LSD1 gene plays an important role in the regulation of low temperature-dependent cell death

Authors

  • Xiaozhen Huang,

    1. State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing 100193, China
    Search for more papers by this author
  • Yansha Li,

    1. State Key Laboratory of Plant Genomics and National Plant Gene Research Center, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Xiaoyan Zhang,

    1. State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing 100193, China
    Search for more papers by this author
  • Jianru Zuo,

    1. State Key Laboratory of Plant Genomics and National Plant Gene Research Center, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Shuhua Yang

    1. State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing 100193, China
    Search for more papers by this author

Author for correspondence:
Shuhua Yang
Tel: +8610 6273 4838
Email: yangshuhua@cau.edu.cn

Summary

  • In higher plants, the crosstalk between cold stress responses and reactive oxygen species (ROS) signaling is not well understood.
  • Two chilling-sensitive mutants, chs4-1 and chs4-3, were characterized genetically and molecularly.
  • The CHS4 gene, identified by map-based cloning, was found to be identical to LESION SIMULATING DISEASE RESISTANCE 1 (LSD1). We therefore renamed these two alleles lsd1-3 and lsd1-4, respectively. These two mutants exhibited an extensive cell death phenotype under cold stress conditions. Consistently, lsd1-3 plants exposed to cold showed up-regulation of the PR1 and PR2 genes, and increased accumulation of salicylic acid. These results indicate that low temperature is another trigger of cell death in lsd1 mutants. Furthermore, lsd1-3 plants accumulated higher concentrations of H2O2 and total glutathione under cold conditions than wild-type plants. Genetic analysis revealed that PAD4 and EDS1, two key signaling regulators mediating resistance responses, are required for the chilling-sensitive phenotype of lsd1-3.
  • These findings reveal a role of LSD1 in regulating cell death trigged by cold stress and a link between cold stress responses and ROS-associated signaling.

Introduction

Temperature is one of the major environmental factors that influence plant growth and development as well as distribution. In order to survive, plants respond and adapt to stress through various biochemical and physiological processes. When exposed to low temperatures, plant cytosolic Ca2+ concentrations increase transiently, followed by altered expression of diverse cold-regulated (COR) genes (Gilmour et al., 1992; Nordin et al., 1993; Yamaguchi-Shinozaki & Shinozaki, 1994). Some of these genes are regulated by C-repeat/dehydration responsive element-binding factor (CBF) transcription factors, which are considered to be the central components responsible for cold tolerance (Shinwari et al., 1998).

Several lines of evidence suggest that plant responses to cold stress are directly linked to reactive oxygen species (ROS) signaling (Lee et al., 2002; Davletova et al., 2005a; Vogel et al., 2005; Einset et al., 2007a,b, 2008). ROS are key signal transduction molecules during responses to environmental stresses and developmental stimuli (Mittler et al., 2004). During environmental stresses, ROS activate stress-response pathways and induce defense mechanisms. On the other hand, excess ROS produced under stress and the failure to maintain ROS balance cause oxidative damage to cells, leading to growth defects or initiation of cell death (Torres & Dangl, 2005). Cold stress increases transcript abundance and protein concentrations of ROS-scavenging enzymes, as well as ROS accumulation (O’Kane et al., 1996). Several genes have been implicated in both cold stress and ROS signaling. For example, overexpression of ZAT12, a C2H2 zinc finger-type transcription factor gene, induces cold-inducible genes and confers increased cold tolerance in plants when overexpressed. Furthermore, ZAT12 plays a central role and regulates a number of genes involved in plant responses to oxidative stress (Davletova et al., 2005b). Glycine betaine (GB) is a chemical that improves tolerance to stress caused by chilling (Park et al., 2004). A membrane trafficking protein RabA4c and a ferric reductase FRO2 play roles in GB’s effect on ROS accumulation during chilling conditions (Einset et al., 2007a, 2008). An ethylene response factor, JERF3, enhances superoxide dismutase activity, and reduces ROS content when overexpressed, thus enhancing tolerance to freezing and other abiotic stresses (Wu et al., 2008).

The Arabidopsis lesion simulating disease resistance 1 (lsd1) mutant shows abnormal regulation of ROS-dependent cell death, characteristics of a runaway cell death (RCD) phenotype when exposed to long photoperiod or infected with an avirulent pathogen (Dietrich et al., 1994). LSD1 encodes a C2H2 zinc finger protein with homology to GATA-type transcription factors (Dietrich et al., 1997). Cell death conditioned by lsd1 requires EDS1 and PAD4, two key regulators that are needed for specific pathogen resistance (Rusterucci et al., 2001). LSD1 is thought to limit the spread of cell death via up-regulation of CuZn-superoxide dismutase that scavenges superoxide (Kliebenstein et al., 1999). Moreover, a basic leucine zipper transcription factor, AtbZIP10, is retained outside the nucleus by LSD1, and these two proteins act antagonistically in both cell death and basal defense responses (Kaminaka et al., 2006). LSD1 is also required for acclimation to conditions that promote excess excitation energy (Mateo et al., 2004), and is involved in regulation of lysigenous aerenchyma formation (Muhlenbock et al., 2007). A recent study suggests that LSD1 serves redox signals generated from redox changes in the plastoquinone pool in chloroplast and leads to cell death under high-light conditions (Muhlenbock et al., 2008).

In this study, we report that two previously isolated chilling-sensitive 4 (chs4) mutants are allelic to lsd1. Cold induction of CBF genes and their targets was reduced in the lsd1-3 mutant to varying extents. Under cold stress, the mutant showed a cell death phenotype, with elevated PR gene expression and ROS imbalance. The mutant phenotypes under cold conditions are fully dependent on PAD4 and EDS1. Our study suggests an important role for LSD1 in the regulation of low temperature-dependent cell death and a signaling crosstalk between cold stress responses and ROS-associated cell death.

Materials and Methods

Plant material and growth conditions

Arabidopsis thaliana plants in accession Columbia (Col), Wassilewskija (Ws) and Landsberg erecta (Ler) were used in this study. lsd1-1 and eds1-1 are in the Ws background, and pad4-1 and PR1::GUS transgenic plant are in the Col background. lsd1-1c was generated by introgressing lsd1-1 into the Col background (Rusterucci et al., 2001). The chs4-1 and chs4-3 (Schneider et al., 1995) in the Col background were obtained from ABRC (stock numbers CS8001 and CS8003). The mutation in chs4-1 was identified by derived cleaved amplified polymorphic sequence (dCAPS) primers LSD1-2 and LSD1-3. For generation of double mutants, single mutants were crossed and the double mutants were identified by PCR genotyping in the F2 population.

Plants were grown at 22 or 4°C, respectively, under long-day (LD, 16 h light : 8 h dark) or short-day photoperiods (SD, 10 h light : 14 h dark) at 100 μmol m−2 s−1, with 50–70% relative humidity in soil or on MS medium (Sigma) containing 2% sucrose and 0.8% agar.

Genetic mapping and cloning of the CHS4 Gene

For genetic mapping of the chs4-1 mutation, the mutant was crossed to the Ler plant. F1 plants from the cross were self-fertilized, and the resulting F2 seeds were collected. A total of 850 chs4-1 mutant plants were selected from the segregating F2 population based on the chilling-sensitive phenotype of the mutant. Genomic DNA from these F2 plants was extracted and used for PCR-based mapping with simple sequence length polymorphism (SSLP) and cleaved amplified polymorphic sequence (CAPS) markers. Initial mapping placed the mutation to the markers F28A21 and F1N20 on the chromosome IV. New mapping markers were developed based on insertion/deletions identified from Cereon Arabidopsis polymorphism and Ler sequence collection (http://www.arabidopsis.org). Genomic DNA corresponding to candidate genes was PCR amplified from the mutants and sequenced to identify the mutation.

Plasmid construction and plant transformation

For the complementation assay, a genomic LSD1 fragment encompassing the LSD1 native promoter (3616 bp) was amplified with primer LSD1-2 and LSD1-p1F by PCR using Probest High fidelity DNA polymerase (Takara, Japan), and cloned into the binary vector pCAMBIA1300 (CAMBIA, Canberra, Australia) with the GFP tag to generate the LSD1::LSD1-GFP.

To construct CHS4::GUS fusion, a 2138 bp genomic fragment upstream of the LSD1 ATG start codon was amplified with primer LSD1-p1F and LSD1-p1R by PCR, and fused with the β-glucuronidase (GUS) reporter gene in the vector pPZPGUS2 (Diener et al., 2000).

The Agrobacterium tumefaciens strain GV3101 carrying different constructs was used to transform wild-type or chs4-1 plants via floral dip transformation (Clough & Bent, 1998).

Ion leakage assay

The electrolyte leakage test was performed as described previously (Lee et al., 2002). Three-week-old plants grown in soil under normal conditions were treated at 4°C for different lengths of time. The percentage of electrolyte leakage was calculated as the ratio of the percentage of the conductivity before autoclaving to the percentage after autoclaving.

Measurement of total glutathione

To compare total glutathione concentration and the redox state of glutathione in wild-type and lsd1-3 plants, reduced glutathione [GSH] and oxidized glutathione [GSSG] in wild-type and lsd1-3 plants were determined as described previously (Rao & Ormrod, 1995). Three-week-old plants grown in soil were treated at 4°C for 10 d.

Salicylic acid measurement

Free salicylic acid (SA) and total SA were extracted and measured from 3-wk-old plants grown at 22°C or treated at 4°C for 10 d as described previously (Li et al., 1999) with some modifications. The last extract residue was dissolved in acetonitrile, and analyzed by high-performance liquid chromatography (HPLC) using 5% acetate (pH 3.2) as the mobile phase.

Histochemical staining assay

Trypan blue staining and 3,3′-diaminobenzidine (DAB) staining were performed as described previously (Bowling et al., 1997; Thordal-Christensen et al., 1997). Histo-chemical detection of GUS activity was performed as described previously (Yang et al., 2006).

Quantitative real-time PCR

Total RNA was isolated from 10-d-old seedlings using TRIzol (Invitrogen), followed by treatment with RNase-free DNase I (Takara) at 37°C for 20 min to degrade genomic DNA. Two micrograms of RNA were subjected to first-strand cDNA synthesis using M-MLV reverse transcriptase (Promega), and an oligo (dT21) primer. The primers used for real-time PCR are listed in the Supporting Information, Table S1. Real-time PCR was performed in 20 μl reactions containing 2 μl fivefold-diluted cDNA, 0.4 μM of each primer, and 10 μl SYBR Green PCR Master Mix (Takara). Analysis was performed using the ABI PRISM 7500 real-time PCR system (Applied Biosystems, Foster City, CA). Primer efficiencies were measured and relative expression level was calculated using the comparative CT method (User Bulletin 2 for ABI PRISM 7700 sequence detection system). All experiments were repeated at least three times with similar results.

Preparation of anti-LSD1 antibodies and western blot analysis

An LSD1 cDNA fragment was cloned into pGEX-4T1 vector (Pharmacia, USA), and purified GST-LSD1 recombinant protein was used to immunize rabbits to generate the antiserum. The anti-LSD1 antibody was then affinity-purified using purified MBP-LSD1 recombinant protein as a ligand. The purified antibody was extensively characterized using protein extracts prepared from wild-type and lsd1 mutant plants.

Total plant protein was isolated from tissues by pulverizing the tissues in ice-cold protein isolation buffer (60 mM Tris-HCl, pH 8.0, 50 mM NaCl, 10 mM EDTA, 30 mM β-mercaptoethanol, and 0.5 mM phenylmethanesulfonyl fluoride). Pulverized tissues were thoroughly mixed and cell debris was cleared by centrifugation at 12 000 g for 10 min at 4°C. Soluble and nuclear proteins were isolated as previously described. Protein concentrations were determined according to the method described by Bradford (1976) with bovine serum albumin as standard.

Extracts were fractionated by SDS-PAGE on 10% (w/v) gel using a minigel system (Bio-Rad). After electrophoresis, the separated proteins were transferred on to a polyvinylidene fluoride (PVDF) nitrocellulose membrane (Bio-Rad), and then blotted with anti-LSD1 as primary antibody (1 : 1000 dilution). Immunodetection was carried out using the secondary antibody (horseradish peroxidase IgG [H + L]) and Western Blotting Luminol Reagent (Santa Cruz Biotechnology, Inc, Santa Cruz, CA, USA).

Results

The chilling sensitive mutant chs4 is allelic to lsd1

Arabidopsis chilling-sensitive 4 (chs4) mutants, chs4-1 and chs4-3, were previously isolated from plants grown at low temperatures (Schneider et al., 1995). In general, chilling sensitivity has been referred to as a variety of types of physiological damage, including wilting, necrosis, chlorosis, and/or leakage of ions from cell membranes (Lyons, 1973). These two chs4 mutants had no obvious growth defects when grown at 20–22°C (Fig. 1a). However, when chs4 mutants were treated at 4°C, the old leaves became yellowish and wilted, and exhibited necrosis (Fig. 1a). Nevertheless, the chs4 plants could flower and set seeds normally if they were transferred to normal temperatures after cold treatment for no longer than 3 wk. Prolonged treatment of chs4-1 plants with low temperatures caused growth arrest and an eventual plant death (Fig. 1a).

Figure 1.

 Map-based cloning of CHS4. (a) Chilling sensitivity of chs4 mutants. Wild-type Arabidopsis Col and chs4 mutants were grown at 22°C for 5 wk (upper panel, top row) or at 22°C for 3 wk and cold-treated at 4°C for 2 wk (upper panel, bottom row). Four-week-old 22°C-grown plants were cold-treated at 4°C for 8 wk (lower panel). Pictures were taken 3 d after plants were transferred back to 22°C. (b) Genetic map of the CHS4 locus region in Col. The corresponding nucleotide positions (in kb) are indicated below the line. CHS4 is positioned between markers F18F4 and T7J7. The number of recombinants is indicated below the markers. Predicted genes are shown by arrows indicating the direction of transcription. (c) Expression of LSD1 transcripts in chs4 mutants analyzed by reverse transcription PCR. Total RNA was isolated from 3-wk-old plants grown at 22°C with or without 10 d cold treatment at 4°C. The β-tubulin gene TUB8 was used as an internal control. (d) Western blot analysis of LSD1 in chs4 and lsd1 mutants. Total protein was extracted from 3-wk-old seedlings and analyzed by immunoblotting with antibody against LSD1. Loading consistency was determined by a nonspecific band (c. 32 kD) recognized by the anti-LSD1 antibody. (e) Complementation test of the chs4-1 mutant at 4°C. A genomic fragment of LSD1 was transformed into chs4-1 plants. Three-week-old T1 plants were cold-treated at 4°C for 21 d.

The chs4-1 mutation was shown to be a recessive mutation in a single nuclear gene (Schneider et al., 1995). A map-based cloning approach was used to identify the CHS4 gene. The chs4-1 mutation was mapped initially between markers F28A21 and F1N20 on chromosome IV. Further mapping narrowed CHS4 to an interval of c. 300kb spanning BACs F18F4 and T7J7 (Fig. 1b). Sequencing analysis detected a single nucleotide substitution of G-to-A in the last exon–intron junction site of LSD1. This mutation presumably leads to abnormal splicing, resulting in a predicted protein with 29 additional amino acid residues. This prediction was verified by reverse transcription PCR and sequencing analyses (Figs 1c, S1). In chs4-3, a single C-to-T mutation was found within the seventh exon of LSD1, resulting in a conversion of Pro to Leu at residue 167 (P167L). No differences in LSD1 mRNA levels were detected between wild-type and chs4-3 plants (Fig. 1c). Immunoblot analyses revealed reduced levels of the mutated proteins in chs4 mutant plants (Fig. 1d), suggesting that these two mutants might be weak alleles of lsd1. The defective phenotypes of chs4-1 and chs4-3 are likely the result of a mutation-induced deactivation of LSD1 protein.

To verify if the chs4 mutant phenotype is caused by mutations in LSD1, a wild-type genomic fragment of LSD1 under the control of its own promoter was transformed into chs4-1 plants. All 15 T1 transgenic plants analyzed showed wild type-like morphology (Fig. 1e). In addition, F1 plants from a cross of chs4-1 and lsd1 (Dietrich et al., 1997) had the lsd1 phenotype (Fig. S2), demonstrating that these two mutants are indeed alleles. For conciseness, we refer to lsd1, lsd1c (Rusterucci et al., 2001), chs4-1, and chs4-3 as lsd1-1, lsd1-1c, lsd1-3, and lsd1-4, respectively, hereafter.

Expression pattern of LSD1

To further elucidate the physiological function of LSD1, we monitored the LSD1 expression pattern by GUS staining of transgenic plants expressing GUS driven by a 2.1 kb fragment of the LSD1 promoter. GUS expression was detected in cotyledons, roots, rosette leaves, stems, inflorescences, and flowers, but was not detected in siliques (Fig. 2a–f). Western blot analyses showed similar patterns of LSD1 protein accumulation (Fig. 2g), indicating that LSD1 may function in most tissues and during most developmental stages.

Figure 2.

 Expression pattern of the Arabidopsis LSD1 gene. (a–f) GUS expression in the LSD1::GUS transgenic lines. β-glucuronidase (GUS) activity was analyzed in seedlings at the two-leaf (a) and eight-leaf (b) stages, cauline leaf (c), stem (d), inflorescence and flowers (e), and silique (f) of LSD1::GUS transgenic plants. (g) Immunoblot analysis of LSD1. Total protein was extracted from various tissues and analyzed by immunoblotting with antibodies against LSD1. 1, 5-d-old seedlings; 2, 14-d-old seedlings; 3, roots; 4, stems; 5, mature rosette leaves; 6, cauline leaves; 7, flowers; 8, siliques. (h) Immunoblot analysis of LSD1 under various treatments. Total protein was extracted from plants treated with cold (4°C), methyl violgen (MV; 5 μM), or benzothiadiazole (BTH; 0.5 mM) at 0, 3, 6, and 12 h, respectively. Protein was analyzed by immunoblotting with antibody against LSD1. Loading consistency in (g) and (h) was determined by a nonspecific band (c. 32 kD) recognized by the anti-LSD1 antibody.

We then used immunoblot analysis to determine whether LSD1 expression was responsive to different stimuli. During a 12 h treatment period, LSD1 was highly induced by methyl violgen (MV), an oxidative inducer. Treatment with benzothiadiazole (BTH), a biologically active analog of SA, slightly increased the LSD1 protein level, whereas LSD1 expression was not altered in response to cold treatment (Fig. 2h). However, LSD1 protein accumulated to higher levels after cold treatment at 4°C for >3 d (data not shown). Consistently, more LSD1 transcripts were detected in plants treated at 4°C for 10 d (Fig. 1c). Thus, it appears that LSD1 expression is rapidly responsive to SA and MV, whereas the cold-induced alteration of LSD1 protein is likely a secondary effect.

Cell death occurs in lsd1-3 plants under cold stress conditions

Visible necrosis was observed on leaves of lsd1-3 and lsd1-4 plants under cold stress conditions (Figs 1a, 3a). Trypan blue staining verified that extensive cell death occurred in cold-treated lsd1-3 and lsd1-4 plants, but not in wild-type plants (Fig. 3a). Cold-treated lsd1-3 plants stained with DAB showed the accumulation of higher levels of hydrogen peroxide (H2O2) than wild-type plants (Fig. 7c). Furthermore, PR1 and PR2 were highly expressed in lsd1-3 plants under cold conditions as revealed by northern blot analysis (Fig. 3b). lsd1-3 plants carrying a PR1::GUS construct showed a strong increase in GUS activity after cold treatment as compared with wild type PR1::GUS transgenic plants (Fig. 3c). Therefore, lsd1-3 plants may constitutively activate defense responses at cold stress.

Figure 3.

 Cell death phenotype of Arabidopsislsd1-3 plants under cold conditions. Plants were grown at 22°C for 4 wk and cold-treated at 4°C for 10 d. (a) Cold-induced cell death in lsd1-3. Detached leaves from wild-type Col and lsd1-3 plants were stained with trypan blue. Dead cells were stained blue. (b) Expression of PR1 and PR2 in wild-type and lsd1-3 plants by northern blot analysis. (c) β-glucuronidase (GUS) analysis of PR1 in wild-type Col and lsd1-3 plants. PR1::GUS transgenic plants were crossed with lsd1-3 plants. F2 homozygous lines were used for GUS staining analysis. (d) Free and total salicylic acid (SA) concentrations in wild-type Col (black bars) and lsd1-3 plants (gray bars). Values represent averages of four replicates ± SD. Experiments were repeated three times with similar results.

Salicylic acid is an essential signaling molecule in plant defense responses (Kunkel & Brooks, 2002). Therefore, endogenous SA levels were measured to examine the role of SA in the lsd1-3 phenotype. The endogenous levels of both free and total SA in lsd1-3 plants were five- and threefold higher than in wild-type plants at 22°C. Cold treatment enhanced accumulation of free and total SA in both wild-type and lsd1-3 plants. However, compared with wild-type plants, the lsd1-3 mutant showed c. nine- and 10-fold increases in concentrations of free and total SA, respectively (Fig. 3d). This result indicates that lsd1-3 plants accumulated more SA under cold conditions, leading to hypersensitive response-like cell death phenotype.

The cell death phenotype of lsd1 is dependent on temperature

Runaway cell death of the null mutant lsd1-1 can be induced by LD conditions (16 h light : 8 h dark) (Dietrich et al., 1994). However, lsd1-3 plants were morphologically indistinguishable from wild-type plants grown either in LD or SD (10 h light : 14 h dark) conditions at 22°C (Figs 1a, 4a). RCD occurred only in lsd1-1 plants but not in lsd1-3 plants 3 d after being shifted from SD to LD at 22°C (Fig. 4a). DAB staining revealed that no obvious H2O2 was accumulated in the shifted lsd1-3 plants, whereas H2O2 accumulation was detected in lsd1-1 leaves (Fig. 4b). We then asked if lsd1-1c and lsd1-3 were sensitive to cold stress during SD conditions. Fig. 4(c) shows that cold triggered RCD in both lsd1-1c and lsd1-3 mutants during the SD condition. PR1 and PR2 genes were also highly induced in lsd1-1c plants by cold stress (Fig. 4e). Hence, lsd1-3 represents a weak allele of lsd1 and both alleles have a chilling-sensitive phenotype.

Figure 4.

 Phenotypes of Arabidopsis lsd1 mutants in response to different light periods and temperatures. (a) Wild-type, lsd1-1, and lsd1-3 plants were grown at 22°C and under short-day (SD) conditions for 50 d and then transferred to 22°C and long-day (LD) conditions for 3 d. (b) 3,3′-diaminobenzidine (DAB) staining of leaves from plants in (a). (c) Wild-type and lsd1-1c plants were grown at 22°C and under short-day (SD) conditions for 21 d or cold-treated at 4°C for 21 d. (d) Wild-type and lsd1-1c plants were grown at 22°C and in SD conditions for 28 d and then placed at 28°C and in LD conditions for 10 d. (e) Analysis of PR1 and PR2 expression in wild-type Col (black bars) and lsd1-1c (gray bars) plants by real-time PCR. Four-week-old plants grown at 22°C and under SD conditions were transferred to 22°C and LD conditions for 10 d, 28°C and LD conditions for 10 d, or 4°C and SD conditions for 10 d.

Previous studies have demonstrated that temperature-dependent cell death in several mutants is compromised by high temperatures (Yang & Hua, 2004; Noutoshi et al., 2005; Yang et al., 2006, 2007; Zhou et al., 2008; Gao et al., 2009; Gou et al., 2009). Next we asked whether high temperatures had any effect on the lsd1-conferred phenotype. The RCD phenotype of lsd1-1c triggered by LD was partially suppressed by high temperatures (28°C) (Fig. 4d). Consistently, induction of the PR1 and PR2 genes in lsd1-1c grown at 22°C was abolished when grown at 28°C (Fig. 4e). These results suggest that temperature is an important factor to modulate cell death conferred by the LSD1 mutation.

lsd1 mutants have defective membrane integrity and cold-regulated gene expression

To determine the extent of chilling injury on membrane integrity in lsd1 plants, lsd1-3 leaves excised from plants treated at 4°C for varying lengths of time were subjected to electrolyte leakage assays. Electrolyte leakage did not change significantly in wild-type plants during a 21 d cold treatment. However, lsd1-3 plants showed a dramatic increase in electrolyte leakage from 6 d after cold treatment, and the electrolyte leakage displayed a time-dependent increase during 6–21 d of cold treatment (Fig. 5a). This indicates that the membrane integrity of lsd1-3 plants was damaged by cold stress.

Figure 5.

 Expression of cold-regulated genes in Arabidopsis lsd1-3 plants. (a) Ion leakage was assayed using wild-type Col (diamonds) and lsd1-3 mutant (squares) leaves from 21-d-old plants grown in soil at 22°C and cold-treated at 4°C in the light for the indicated times. Data represent means of five replicates ± SD. (b) Relative mRNA concentrations in wild-type Col (black bars) and lsd1-3 (gray bars) seedlings were determined by real-time PCR. Ten-day-old seedlings grown at 22°C were treated at 4°C for the indicated times. Data represent means of three replicates ± SD. Similar results were observed in at least three independent experiments. *, < 0.05 (t-test), significant difference from Col.

Next, we examined whether the expression of cold-regulated genes contributes to the lsd1-conferred phenotype. Quantitative real-time PCR analysis showed that cold induced a higher expression of CBF1, CBF2, and CBF3 and their targets RD29A and COR47, which was consistent with previous studies (Jaglo-Ottosen et al., 1998; Miura et al., 2007). CBF1, CBF2, and CBF3 were rapidly induced in lsd1-3 and wild-type plants following cold treatment. Compared with wild-type plants, cold induction of CBF1, CBF2, and CBF3 genes was decreased in lsd1-3 plants after 3 and 6 h of cold treatment (Fig. 5b). RD29A and COR47 were induced at low levels in lsd1-3 plants to varying extents after 12 and 24 h of cold treatment (Fig. 5b). Therefore, LSD1 might modulate plant responses to cold stress, at least partially by altering the expression of cold-regulated genes.

The total glutathione concentration is increased in lsd1-3 plants under cold stress conditions

Glutathione plays important roles in many processes, including the G1-to-S transition in the cell cycle (Vernoux et al., 2000), transcriptional and translation regulation (Kan et al., 1988; Baena-Gonzalez et al., 2001), cell death (Henmi et al., 2001), and disease resistance (May et al., 1996). Glutathione is also implicated in chilling tolerance (Kocsy et al., 2001). To determine the role of glutathione in lsd1-3 plants, cellular redox changes were analyzed by comparing total glutathione ([GSH] + [GSSG]) and the ratio of reduced glutathione : total glutathione ([GSH] : ([GSH] + [GSSG])) of lsd1-3 and wild-type plants under cold stress conditions. Comparable amounts of total glutathione were detected in wild type and lsd1-3 plants at 22°C (Fig. 6a). Cold stress induced the accumulation of total glutathione in both wild type and lsd1-3 plants. lsd1-3 plants accumulated more total glutathione than wild type plants under cold stress (Fig. 6a). By contrast, no significant difference of the [GSH] : ([GSH] + [GSSG]) ratio was observed between lsd1-3 and wild-type plants following cold treatment (Fig. 6a). Therefore, the lsd1-conditioned phenotype is correlated with endogenous glutathione levels, rather than the glutathione redox state. These findings are consistent with a previous study showing that the induction of PR1 accompanied by RCD in lsd1 plants is dependent on glutathione levels, but independent of the redox state of glutathione (Senda & Ogawa, 2004).

Figure 6.

 Concentrations of glutathione and expression of reactive oxygen species (ROS)-associated genes in lsd1-3 plants under cold stress conditions. Three-week-old wild-type Col (black bars) and lsd1-3 plants (gray bars) grown at 22°C were treated at 4°C for 10 d. (a) Changes of cellular glutathione concentrations in lsd1-3 plants under cold stress conditions. Data represent the means of three replicates ± SD. *, < 0.05 (t-test), significant difference from Col. (b) Expression of ROS-associated genes by real-time PCR. Data represent means of three replicates ± SD. All experiments were repeated three times with similar results. *, < 0.05 (t-test), significant difference from Col.

ROS-associated gene expression of lsd1-3 plants during cold stress

As noted previously, greater H2O2 accumulation in lsd1-3 than in wild-type plants was observed under cold stress (Fig. 7c). Excess H2O2 was shown to induce expression of genes involved in oxidative stress (Iba, 2002; Mittler et al., 2004; Rizhsky et al., 2004). Therefore, the transcript abundance of genes encoding ROS-producing enzyme NADPH oxidase (RbohD), and ROS-scavenging enzymes, such as ascorbate peroxidase (APX1), catalase (CAT1), and glutathione reductase (GR1), in lsd1-3 plants was examined. Under normal growth conditions, RbohD expression was at a lower level in lsd1-3 plants than in wild-type plants (Fig. 6b). No notable differences in expression of APX1, CAT1, and GR1 were detected between wild-type and lsd1-3 plants at 22°C (Fig. 6b). The transcript abundances of RbohD, APX1, CAT1, and GR1 were substantially higher in lsd1-3 plants than in wild-type plants upon exposure to cold for 10 d. ZAT12 transcripts were more abundant in lsd1-3 plants than in wild-type plants after 10 d of 4°C treatment (Fig. 6b). Ferritins are considered to be essential for cell protection against oxidative damage (Ravet et al., 2009). FER1 was also up-regulated in lsd1-3 plants in cold conditions (Fig. 6b). Moreover, the extent of cold induction of these genes was similar in lsd1-3 and lsd1-1c plants (Fig. S3). Taken together, these results suggest that the lsd1-3 phenotype in cold conditions is associated with impairment of the ROS detoxification pathway and the oxidative signaling pathway caused by disruption of LSD1.

Figure 7.

 Phenotype of lsd1-3 and double mutant plants at low temperatures. Three-week-old wild-type, lsd1-3, lsd1-3 pad4-1, and lsd1-3 eds1-1 mutants grown at 22°C were treated at 4°C for 10 d. (a) Morphological phenotype of leaves. (b) Trypan blue staining. (c) 3,3′-diaminobenzidine (DAB) staining. (d) Real-time PCR analysis of PR expression.

pad4 and eds1 mutations suppress the chilling sensitivity of lsd1-3 plants

LSD1-regulation of basal defense and cell death is dependent on EDS1 and PAD4, two key regulators that are required for many R gene-mediated defenses (Rusterucci et al., 2001; Aviv et al., 2002). Previous studies have also demonstrated that temperature-dependent growth defects and cell death in some mutants are compromised by eds1 and pad4 (Yang & Hua, 2004; Yang et al., 2006, 2007; Gao et al., 2009; Gou et al., 2009). To test whether the chilling sensitive phenotype of lsd1-3 requires EDS1 and PAD4, lsd1-3 eds1-1 and lsd1-3 pad4-1 double mutants were generated and analyzed. Neither lsd1-3 eds1-1 nor lsd1-3 pad4-1 showed visible cell death phenotypes in cold conditions (Fig. 7a,b). In addition, H2O2 accumulation in lsd1-3 plants was completely suppressed by the eds1 and pad4 mutations (Fig. 7c). In accordance with the cell death phenotype, the increased expression of PR genes at 4°C was also completely abolished by eds1 and pad4 (Fig. 7d). These data indicate that the chilling sensitivity of lsd1-3 is dependent on EDS1 and PAD4.

Discussion

In this study, we characterized two weak alleles of lsd1, lsd1-3 and lsd1-4, which were previously named chs4-1 and chs4-3, respectively, based on their phenotype at chilling temperatures (Schneider et al., 1995). lsd1 mutant plants display a cell death phenotype within their leaves when exposed to 4°C. lsd1 mutants also show constitutive defense responses, including elevated PR gene expression, and accumulation of endogenous SA and excess ROS at 4°C. Furthermore, the lsd1 mutant phenotype is completely suppressed by eds1 and pad4. These characteristics indicate that low temperature is another trigger of cell death in lsd1 mutants dependent on PAD4 and EDS1.

LSD1 is a negative regulator of SA-dependent programmed cell death (PCD) and plant defense responses to pathogens (Dietrich et al., 1997; Rusterucci et al., 2001; Aviv et al., 2002). LSD1 is also necessary for acclimation to conditions that promote photooxidative stress, including long-day photoperiods, high-light conditions, and photorespiratory conditions (Dietrich et al., 1997; Mateo et al., 2004). Here, lsd1 plants displayed growth inhibition and visible PCD within their leaves under cold stress. The LSD1 mutation caused reduced cold induction of CBFs and their target genes, suggesting that LSD1 has specific effects on expression of cold-regulated genes to modulate plant responses to cold stress. How LSD1 regulates the expression of cold-regulated genes remains elusive. LSD1 protein expression was not regulated by cold during a 12 h time period. However, prolonged cold exposure (at least 3 d) can increase LSD1 protein levels (data not shown), implying that LSD1 may not respond to low temperatures directly. Compared with the wild-type, the lsd1 mutation-activated defense response may have a higher threshold to additional stress signals, which may explain a reduced response of lsd1 mutants to cold treatment, as we have observed.

Consistent with a previous study (Kaminaka et al., 2006), we found that LSD1 was largely localized in the cytoplasm at 22°C, and low temperatures had no effect on the subcellular localization of LSD1 (Fig. S4). LSD1 interacts with a bZIP transcription factor, AtbZIP10, thereby retaining AtbZIP10 in the cytoplasm, which results in modulation of cell death and the basal defense (Kaminaka et al., 2006). In the cold signaling pathway, HOS1 has been shown to translocate from the cytoplasm to the nucleus following cold stress. Once in the nucleus, HOS1 interacts with and polyubiquitinates ICE1, a positive regulator of CBF3, thus negatively regulating the expression of cold-regulated genes that contribute to cold acclimation (Lee et al., 2001; Dong et al., 2006). Therefore, LSD1 might regulate the nuclear localization of ICE1 and/or HOS1 proteins by binding them either directly or indirectly through other proteins functioning in this pathway. Further localization studies of the HOS1 and ICE1 proteins in lsd1 mutants under cold stress may help elucidate the LSD1 function in regulation of temperature-dependent plant growth and cell death.

lsd1-3 and lsd1-4 are weak alleles of lsd1-1c, as they do not show obvious defects in response to LD conditions, under which lsd1-1c exhibits an RCD phenotype. Sequencing analysis revealed that the point mutations in lsd1-3 and lsd1-4 did not produce stop codons, and that both mutated proteins could presumably be translated. Indeed, mRNA concentrations of LSD1 were not altered in these two mutants. However, the mutated protein levels were dramatically reduced in lsd1-3 and lsd1-4 plants, which is probably the result of the decreased stability of the mutated proteins. Nevertheless, it appears that the residual mutated proteins in lsd1-3 and lsd1-4 plants are still functional, and that they repress RCD under normal conditions when no excess ROS are accumulated. However, under unfavorable environmental conditions, such as cold stress, the residual activity of the mutated LSD1 proteins was insufficient to inhibit RCD caused by overaccumulated ROS in lsd1-3 and lsd1-4 mutants, leading to their cold-sensitive phenotypes.

Our data show that LSD1 plays an essential role in avoidance of temperature-induced oxidative stress. In lsd1 mutants, excess ROS lead to up-regulation of genes ezncoding ROS-scavenging enzymes and enhance the concentration of the nonenzymatic antioxidant glutathione, which can contribute to plant protection against adverse situations. However, even the increased amount of these protectants was unable to prevent the damage caused by excess ROS accumulation during cold conditions in lsd1 mutants.

Several environmental factors can initiate cell death. For example, lsd1 plants initiate RCD during long photoperiods and high-light conditions, whereas lsd2, lsd5, and len1 plants develop lesions under short-day conditions but not under long-day conditions (Dietrich et al., 1994; Ishikawa et al., 2003). Furthermore, lesion formation and cell death are compromised by high temperature and/or humidity in transgenic plants overexpressing the tomato disease resistance gene Pto (Li et al., 2002) and in some lesion-mimic mutants, including bon1, bap1, ssi4, slh1, bir1, and cpr30 (Yang & Hua, 2004; Noutoshi et al., 2005; Yang et al., 2006, 2007; Zhou et al., 2008; Gao et al., 2009; Gou et al., 2009). By contrast, cell death in rice spl7 mutants requires high temperatures and UV solar radiation (Yamanouchi et al. 2002). In our study, cell death occurs in lsd1-3 only at low temperatures. Null mutations in EDS1 and PAD4 suppress the chilling-sensitive phenotype of lsd1-3. This finding is in agreement with the observation that pad4 and eds1 block lsd1-conditioned RCD triggered by long photoperiods, high light, photorespiratory conditions, or SA (Jabs et al., 1996; Rusterucci et al., 2001; Mateo et al., 2004). Therefore, a similar mechanism is likely utilized by LSD1-regulated cell death machinery in response to different environmental cues.

In summary, our data indicate that low temperature is another trigger of the cell death caused by the imbalance of ROS in lsd1 mutants, and that the cell death phenotype requires EDS1 and PAD4. Thus, LSD1 plays an important role in ROS responses to repress cell death under various environmental stimuli.

Acknowledgements

We thank Dr Jian Hua for her critical reading of the manuscript. We are grateful to Drs J. Dangl, X Dong, B. Staskawicz, and ABRC for providing seeds. This work was supported by the National Natural Science Foundation of China (grant nos 30670181 and 90817007), the National Key Basic Research Program of China (no. 2009CB119100), the Ministry of Science and Technology of China (no. 2007AA021402), Research Fund for the Doctoral Program of Higher Education of China (No. 20070019002), and the Program for New Century Excellent Talents in University (no. NCET-05-0124).

Ancillary