Reversal of senescence by N resupply to N-starved Arabidopsis thaliana: transcriptomic and metabolomic consequences

Abstract

Leaf senescence is a developmentally controlled process, which is additionally modulated by a number of adverse environmental conditions. Nitrogen shortage is a well-known trigger of precocious senescence in many plant species including crops, generally limiting biomass and seed yield. However, leaf senescence induced by nitrogen starvation may be reversed when nitrogen is resupplied at the onset of senescence. Here, the transcriptomic, hormonal, and global metabolic rearrangements occurring during nitrogen resupply-induced reversal of senescence in Arabidopsis thaliana were analysed. The changes induced by senescence were essentially in keeping with those previously described; however, these could, by and large, be reversed. The data thus indicate that plants undergoing senescence retain the capacity to sense and respond to the availability of nitrogen nutrition. The combined data are discussed in the context of the reversibility of the senescence programme and the evolutionary benefit afforded thereby. Future prospects for understanding and manipulating this process in both Arabidopsis and crop plants are postulated.

Keywords: Arabidopsis; gene expression; metabolomics; nitrogen limitation; senescence; transcriptome..

Fig. 1.

Experimental set-up. (A) Schematic representation of the experimental set-up: Arabidopsis thaliana plants were grown hydroponically in complete Hoagland liquid medium (+N medium). Plants were transferred to nitrogen-free medium (–N medium) 19 d after sowing (DAS). After 4 d and 7 d of growth on –N medium, subsets of plants were transferred back to complete medium (N resupply) for 3h or 3 d, respectively. (B) Chlorophyll content in leaves no. 1 and 2 of N-starved (–N), N-resupplied plants, and plants not starved for N (+N). Numbers on the y-axis indicate relative chlorophyll content. (C) Photosystem II (PSII) efficiency (F _v/F _m) in leaves no. 1 and 2 of N-starved (–N), N-resupplied plants, and plants not starved for N (+N). The chlorophyll fluorescence measurements were performed after 30min of dark adaptation. Measurements were performed on the total surface area of leaves. Data in (B) and (C) represent means ±SD (n=10). (D) Representative leaves and their PSII efficiency (F _v /F _m). For each day (DAS), pictures on the left represent the PSII efficiency in a chlorophyll fluorescence false-colour image, with a photograph of the same leaf on the right-hand side (see colour scale). Mean values of F _v/F _m and standard deviations are displayed below each image (n ≥10 plants).

Fig. 2.

N depletion-induced and -repressed genes. (A) Schematic presentation of the sampling strategy. (B) and (C) Venn diagrams showing the numbers of significantly (B) up-regulated and (C) down-regulated genes that are uniquely or commonly regulated during nitrogen depletion. Numbers in parentheses represent numbers of SAGs (in B) and SDGs (in C) in each group. Thick arrows in (A) indicate samples that were compared for data shown in (B) and (C).

Fig. 3.

N resupply-induced and -repressed genes. (A) Schematic presentation of the sampling strategy. (B) and (C) Venn diagrams showing the numbers of significantly (B) up-regulated and (C) down-regulated genes that are uniquely or commonly regulated upon N resupply. Numbers in parentheses represent numbers of SDGs (in B) and SAGs (in C) in each group. Thick arrows in (A) indicate samples that were compared for data shown in (B) and (C).

Fig. 4.

Number and function of genes commonly and differentially expressed upon N resupply, Botrytis cinerea infection, and dark-induced senescence. The numbers of genes are shown in the Venn diagrams.

Fig. 5.

N depletion- and N resupply-responsive transcripts grouped according to temporal expression profiles. Differentially expressed genes obtained from Affymetrix ATH1 microarray-based transcriptome studies were divided into seven distinct significant temporal profiles, using STEM software (Ernst and Bar-Joseph, 2006). Each of the profiles is represented as a different plot, with mean expression ratios (log₂) for each of the assigned transcripts at each time point. The presence of TFs in each profile is indicated.

Fig. 6.

Regulatory sequences associated with groups of co-expressed genes. Significantly over-represented (P<0.01) motifs from JASPAR and ATCOECIS as well as motifs discovered de novo using MotifSuite and MEMESuite are shown for the seven STEM clusters. Note that the de novo detected motifs were also subjected to an enrichment test and only motifs present in multiple clusters are included in this image (a full overview is given in Supplementary Table S7 available at JXB online). In particular, motifs that contain the light-responsive G-box and ABRE sequences occur frequently.

Fig. 7.

Primary metabolite profiling of N deficiency-induced leaf senescence and its reversal by N resupply. Hierarchical average linkage clustering of all detected primary metabolites. For every metabolite, the metabolic content of the control sample harvested at 19 DAS was considered as 1 and the metabolic content of all other samples at any given time point (+N, –N, and N resupply) normalized to that. Metabolic ratios: blue, minimum (between 0 and 1); red, maximum (between 1 and 2); see also Supplementary Table S9 available at JXB online.