Differential leaf flooding resilience in Arabidopsis thaliana is controlled by ethylene signaling-activated and age-dependent phosphorylation of ORESARA1

Abstract

The phytohormone ethylene is a major regulator of plant adaptive responses to flooding. In flooded plant tissues, ethylene quickly increases to high concentrations owing to its low solubility and diffusion rates in water. Ethylene accumulation in submerged plant tissues makes it a reliable cue for triggering flood acclimation responses, including metabolic adjustments to cope with flood-induced hypoxia. However, persistent ethylene accumulation also accelerates leaf senescence. Stress-induced senescence hampers photosynthetic capacity and stress recovery. In submerged Arabidopsis, senescence follows a strict age-dependent pattern starting with the older leaves. Although mechanisms underlying ethylene-mediated senescence have been uncovered, it is unclear how submerged plants avoid indiscriminate breakdown of leaves despite high systemic ethylene accumulation. We demonstrate that although submergence triggers leaf-age-independent activation of ethylene signaling via EIN3 in Arabidopsis, senescence is initiated only in old leaves. EIN3 stabilization also leads to overall transcript and protein accumulation of the senescence-promoting transcription factor ORESARA1 (ORE1) in both old and young leaves during submergence. However, leaf-age-dependent senescence can be explained by ORE1 protein activation via phosphorylation specifically in old leaves, independent of the previously identified age-dependent control of ORE1 via miR164. A systematic analysis of the roles of the major flooding stress cues and signaling pathways shows that only the combination of ethylene and darkness is sufficient to mimic submergence-induced senescence involving ORE1 accumulation and phosphorylation. Hypoxia, most often associated with flooding stress in plants, appears to have no role in these processes. Our results reveal a mechanism by which plants regulate the speed and pattern of senescence during environmental stresses such as flooding. Age-dependent ORE1 activity ensures that older, expendable leaves are dismantled first, thus prolonging the life of younger leaves and meristematic tissues that are vital to whole-plant survival.

Keywords: abiotic stress; ethylene; flooding; hypoxia; senescence.

Figure 1

Ethylene perception during submergence is systemic but mediates age-dependent leaf death (A) Arabidopsis wild-type (accession Col-0) plants submerged in darkness show age-dependent leaf death. Images show representative plants submerged for the duration indicated in the top left of each image. Numbers in the first image indicate leaf numbering according to age. Scale bar corresponds to 1 cm. (B) Immunoblot analyses showing that EIN3-GFP accumulates within 3 h of submergence or ethylene treatment in both old (leaf 3) and young (leaf 7) leaves of transgenic 35S:EIN3-GFP ein3eil1 plants. Samples were run on the same gel; the vertical line indicates where samples were cropped out. The large subunit of Rubisco (RbcL) served as a loading control. (C and D) Quantification of leaf death across three age categories of an Arabidopsis rosette. Age categories are indicated by leaf number (#) as in (A). Age-dependent leaf death observed in wild-type plants is lost in ethylene-insensitive ein3eil1 and ein2-5 mutants. P values indicate the effect of leaf age on the proportion of dead leaves per genotype, determined by a two-way ANOVA (leaf age × submergence duration). n = 5–6 plants per time point. (E) The effect of different flooding cues on leaf yellowing in wild-type Arabidopsis plants. Yellowing is indicated by the median hue of old (leaf 3) and young (leaf 7) leaves after 4 days of exposure to each treatment. Ethylene induces age-dependent leaf yellowing, and this process is slowed by hypoxia. Images below each bar show representative plants from each treatment. Asterisks indicate differences between old (leaf 3) and young (leaf 7) leaves (paired t-test), and different letters indicate significant differences between treatments (two-way ANOVA and Tukey’s post-hoc test). n = 7 plants per treatment. Treatment combinations are indicated, where normal (normoxia) or low oxygen (hypoxia) was combined with (ethylene) or without (air) ethylene gas in the presence (light) or absence (dark) of light. (F–H) Age-dependent leaf death is not lost in pco124, erfVII, or prt6-1 mutants, which have impaired oxygen sensing. Age categories are indicated by leaf number (#) as in (A). P values indicate the effect of leaf age on the proportion of dead leaves per genotype determined by a two-way ANOVA (leaf age × submergence duration). n = 5–6 plants per time point.

Figure 2

Submergence-induced senescence is primarily controlled by the ethylene-responsive NAC transcription factor ORE1 (A) Chlorophyll content of old and young leaves of Col-0 and ore1-1 plants before and after 5 days of ethylene treatment. n = 5 leaves per sample. (B)ORE1 mRNA abundance increases in both old and young leaves after 1 day of ethylene treatment. n = 4 leaves per sample, each consisting of 2 old or young leaves from different plants pooled together. Expression levels were normalized to those in old leaves of non-submerged plants. (C)ore1-1 mutants show reduced yellowing of old leaves after 4 days of submergence. Representative images show Col-0 and ore1-1 plants at the indicated time points. Scale bar corresponds to 1 cm. (D) Chlorophyll content of old and young leaves of Col-0 and ore1-1 before and after 4 days of submergence. n = 6 leaves per sample. (E) Chlorophyll content of Col-0, ore1-1, and two independent pORE1:ORE1-HA ore1-1 lines before and after 3 days of submergence. n = 6 leaves per sample. (F) Dead leaves per Col-0 and ore1-1 plant during recovery from 4 days of submergence. Leaves were scored as dead or alive at each of the indicated time points, n = 17–21 plants per genotype. These same plants were phenotyped for supplemental Figure 2E and 2H. (G) Ion leakage of Col-0 and ore1-1 before and after 6 days of submergence. n = 3 pools of 5 old or young leaves from different plants per sample. (H) Total living rosette area of Col-0 and ore1-1 plants before and after 4 days of submergence (sub) and after 13 days of recovery. Images of plants were categorized into dead, senescing, and healthy pixels using PlantCV. Senescing and healthy pixels were combined for each plant and converted to an area in cm². n = 16–20 plants per sample. (I) Seed yield of Col-0 and ore1-1 plants under control conditions and of plants that were submerged for 6 days. n = 15 plants per group. Different letters indicate significant differences between groups (two-way ANOVA and Tukey’s post-hoc test). Asterisks indicate significant differences between Col-0 and ore1-1 per time point.

Figure 3

ORE1 is induced in an age-independent manner during flooding stress (A)ORE1 mRNA abundance in whole rosettes before and after darkness and dark submergence. Asterisks indicate significant differences compared with untreated plants (one-way ANOVA and Dunnett’s post-hoc test). Expression was normalized to that of untreated plants. n = 3, each sample consists of one rosette. (B)ORE1 mRNA abundance in old and young leaves of Col-0 and ein3eil1 before and after 4 days of submergence. Different letters indicate significant differences between groups (two-way ANOVA and Tukey’s post-hoc test). Expression was normalized to that of non-submerged old leaves of Col-0. Three biological replicates were analyzed. ORE1 mRNA was not detected in one of the non-submerged Col-0 and ein3eil1 young leaf samples; each sample consists of two leaves pooled together from different plants. (C) Immunoblots showing ORE1-HA protein abundance in old and young leaves before and after 1 and 3 days of submergence using an antibody against HA. Each pORE1:ORE1-HA ore1-1 sample consists of five old or young leaves pooled together from different plants. Proteins of the Col-0 sample were extracted from one whole rosette. The large subunit of Rubisco (RbcL) served as a loading control. (D) mRNA abundance of the ORE1 target gene BFN1 in old and young leaves of Col-0 and ein3eil1 before and after 4 days of submergence. Different letters indicate significant differences between groups (two-way ANOVA and Tukey’s post-hoc test). Expression was normalized to that of non-submerged old leaves of Col-0. Three biological replicates were analyzed. BFN1 mRNA was not detected in one of the submerged Col-0 and ein3eil1 young leaf samples; each sample consists of two leaves pooled together from different plants.

Figure 4

ORE1 target activation is age independent (A) Leaf samples of Col-0 and ore1-1 were harvested before submergence, after 4 days of submergence, and after 6 h of recovery. The number of differentially expressed genes (DEGs) that show a genotype-dependent response to 4 days of submergence is greater in old leaves than in young leaves. None showed a genotype-dependent effect in their response to submergence followed by recovery. Most (428/720) DEGs that showed a genotype-specific response to post-submergence recovery showed the opposite pattern during the submergence phase. (B) Electrophoretic mobility shift assay (EMSA) showing in vitro binding of recombinant ORE1-GST to the promoters of MC9, ANAC010, DPD1, and CV. From left to right in each image: lane 1, labeled probe (5′-DY682-labeled double-stranded oligonucleotides); lane 2, labeled probe plus ORE1-GST protein; lane 3, labeled probe, ORE1-GST protein, and competitor (unlabeled oligonucleotide containing an ORE1 binding site; 200× molar access). Arrows indicate retarded bands (bound oligo) and non-bound DNA probes (free oligo). (C) ChIP–qPCR showing in vivo binding of ORE1 to the promoters of MC9, ANAC010, DPD1, and CV. Asterisks indicate significant enrichment relative to the negative control (AT2G22180) (one-way ANOVA and Dunnett’s post-hoc test). Chromatin was extracted from immunoprecipitated samples of whole pORE1:ORE1-HA rosettes submerged for 1 day, n = 3.

Figure 5

ORE1 phosphorylation during flooding is age dependent (A)pORE1:ORE1-HA ore1-1 protein samples from submerged old leaves move more slowly through a Phos-tag gel than samples from young leaves, indicating age-specific phosphorylation of ORE1. Five old or young leaves were pooled together from different plants per pORE1:ORE1-HA ore1-1 sample, and the Col-0 sample was from one whole rosette. (B) Representative images of Col-0, ore1-1, 35S:ORE1, and 35S:ORE1Δ17 plants after 5 days of submergence followed by 1 day of recovery. (C) Chlorophyll content of Col-0, ore1-1, 35S:ORE1, and 35S:ORE1Δ17 plants before and immediately after 5 days of submergence. (D–F) Expression of ORE1, MC9, and BFN1 in Col-0, ore1-1, 35S:ORE1, and 35S:ORE1Δ17 before and after 4 days of submergence. Expression was normalized to that of non-submerged old leaves of Col-0. Two old or young leaves from different plants were pooled together per sample. Different letters indicate significant differences among groups (two-way ANOVA and Tukey’s post-hoc test). Error bars indicate SEM.

Figure 6

Ethylene controls leaf-age-dependent ORE1 phosphorylation (A) Immunoblots showing ORE1 accumulation in old and young leaves after submergence or 1 day of ethylene treatment in darkness. Five old or young leaves from different plants were pooled together per pORE1:ORE1-HA ore1-1 sample. The Col-0 sample was from one whole rosette. Stain-free imaging of the protein gel was used as a loading control. (B) ORE1-HA from old leaves treated with ethylene in darkness for 1 day moves more slowly through a Phos-tag gel than ORE1-HA from young leaves. Samples are the same as those run on the non-Phos-tag gel in (A). Ponceau staining of the large subunit of Rubisco was used as a loading control. (C) Shoot phenotypes in response to submergence or ethylene in light or dark conditions. Representative images of Col-0 and ein3eil1 plants immediately after 5 days of the indicated treatments are shown. Scale bars correspond to 1 cm. (D) Chlorophyll content of old and young leaves after treatments with different combinations of ethylene and submergence (sub) in light and darkness. Asterisks indicate significant differences from the chlorophyll levels before treatment (one-way ANOVA and Dunnett’s test), and error bars indicate SEM. n = 10 per sample from 2 independent experiments; circles and triangles indicate experimental replicates.

Figure 7

A model for ethylene-mediated sequential leaf senescence in flooded plants Upon submergence, ethylene rapidly accumulates throughout the plant. This age-independent accumulation of ethylene induces the age-independent accumulation of ORE1 mRNA and protein via EIN3 stabilization. Leaf-age-dependent senescence is triggered by ethylene via ORE1 phosphorylation and activation specifically in old leaves via an unknown mechanism. This age-dependent phosphorylation of ORE1 ensures that it induces senescence in old leaves; the oldest leaves are thus broken down first and the youngest leaves and meristem last.