Ran-GTP/-GDP-dependent nuclear accumulation of NONEXPRESSOR OF PATHOGENESIS-RELATED GENES1 and TGACG-BINDING FACTOR2 controls salicylic acid-induced leaf senescence

NPR1 is a functional receptor of salicylic acid (SA), and the Ran small G-protein-dependent nuclear accumulation of NPR1 and TGA2 proteins constitutes an important regulatory node in SA-induced leaf senescence.

Abstract

Leaf senescence is the final stage of leaf development and can be triggered by various external factors, such as hormones and light deprivation. In this study, we demonstrate that the overexpression of the GTP-bound form of Arabidopsis (Arabidopsis thaliana) Ran1 (a Ras-related nuclear small G-protein, AtRan1) efficiently promotes age-dependent and dark-triggered leaf senescence, while Ran-GDP has the opposite effect. Transcriptome analysis comparing AtRan1-GDP- and AtRan1-GTP-overexpressing transgenic plants (Ran1T27Nox and Ran1G22Vox, respectively) revealed that differentially expressed genes (DEGs) related to the senescence-promoting hormones salicylic acid (SA), jasmonic acid, abscisic acid, and ethylene (ET) were significantly upregulated in dark-triggered senescing leaves of Ran1G22Vox, indicating that these hormones are actively involved in Ran-GTP/-GDP-dependent, dark-triggered leaf senescence. Bioinformatic analysis of the promoter regions of DEGs identified diverse consensus motifs, including the bZIP motif, a common binding site for TGACG-BINDING FACTOR (TGA) transcription factors. Interestingly, TGA2 and its interactor, NONEXPRESSOR OF PATHOGENESIS-RELATED GENES1 (NPR1), which are two positive transcriptional regulators of SA signaling, differed in their extent of accumulation in the nucleus versus cytoplasm of Ran1T27Nox and Ran1G22Vox plants. Moreover, SA-induced, Ran-GTP-/-GDP-dependent functions of NPR1 included genome-wide global transcriptional reprogramming of genes involved in cell death, aging, and chloroplast organization. Furthermore, the expression of AtRan1-GTP in SA signaling-defective npr1 and SA biosynthesis-deficient SA-induction deficient2 genetic backgrounds abolished the effects of AtRan1-GTP, thus retarding age-promoted leaf senescence. However, ET-induced leaf senescence was not mediated by Ran machinery-dependent nuclear shuttling of ETHYLENE-INSENSITIVE3 and ETHYLENE-INSENSITIVE3-LIKE1 proteins. We conclude that Ran-GTP/-GDP-dependent nuclear accumulation of NPR1 and TGA2 represents another regulatory node for SA-induced leaf senescence.

Introduction

Ras-related nuclear protein (Ran) is a member of the Ras small GTPase superfamily; members of this family play essential roles in regulating nucleocytoplasmic transport of signaling molecules during interphase (Sorokin et al., 2007; Lee et al., 2008) and in controlling mitotic cell cycle progression through the regulation of mitotic spindle formation, chromosome segregation, and post-mitotic nuclear envelope (NE) re-assembly (Bamba et al., 2002; Silverman-Gavrila and Wilde, 2006; Bao et al., 2018; Boudhraa et al., 2020; Drutovic et al., 2020). Like other small GTP-binding proteins, Ran switches between GDP- and GTP-binding states due to interactions with nuclear-localized REGULATOR OF CHROMOSOME CONDENSATION1 (RCC1), a Ran guanine nucleotide exchanging factor (RanGEF), and RanGAP, a cytoplasmic Ran GTPase-activating protein (Bischoff and Ponstingl, 1991; Bischoff et al., 1994). The unequal distribution of Ran effector proteins across the NE generates an asymmetric gradient of cytoplasmic Ran-GDP and nuclear Ran-GTP, which acts as a major driving force giving directionality to nuclear transport of nuclear localization signal (NLS)-containing cargo proteins (Lee et al., 2008).

The intracellular gradient of Ran-GTP and Ran cycling between -GDP and -GTP-bound states is essential for Ran function. Depletion of RCC1 or RanGAP by RNAi resulted in chromosome misalignment and aberrant chromosome segregation in Caenorhabditis elegans (Bamba et al., 2002). In addition, RCC1-dependent generation and activation of Ran-GTP accelerated cell cycle and DNA repair, inhibiting DNA-damage-induced cell senescence (Cekan et al., 2016). Ran bound to Mog1, a guanine nucleotide release factor, lost its ability to bind to RCC1 and load GTP, while acetylation-elicited Ran binding to RCC1 promoted high levels of Ran-GTP, which is essential for chromosome segregation in mitosis (Bao et al., 2018). Ran GTPase is required for chromosome alignment and segregation during mitosis, and injection of RanT24N, a dominant negative allele of GDP-locked Ran (Kornbluth et al., 1994), prevented correct targeting of Aurora A kinase and the kinesins KLP3A and KLP61F to mitotic spindles and disrupted spindle assembly (Silverman-Gavrila and Wilde, 2006). In Xenopus egg extracts, the addition of a GTPase-deficient mutant (Ran Q69L, GTP-locked) stabilized microtubule asters and inhibited nuclear assembly while GDP-bound RanT24N arrested nuclear growth after disappearance of the aster (Zhang et al., 1999).

In plants, several Ran proteins and Ran-interacting proteins including Ran-binding proteins (RanBPs), RanGAP, and nucleoporins have been reported to be involved in various cellular and physiological responses, such as regulation of hormone sensitivities, mitotic progression, cold/drought/salt tolerance, female gametophyte and seed development, stomatal closure, and resistance to pathogens. For example, antisense expression of an Arabidopsis (Arabidopsis thaliana) Ran-BP, AtRanBP1c, rendered transgenic roots hypersensitive to auxins and altered auxin-induced root growth and development by arresting mitotic progress (Kim et al., 2001). Co-related to this observation, the overexpression of rice (Oryza sativa) RAN1 and wheat RAN1 altered auxin sensitivity and mitotic progression, indicating active involvement of RAN1 in the regulation of cold resistance, root growth, reduced apical dominance, and flowering (Wang et al., 2006b; Xu and Cai, 2014). In addition to these observations, the ectopic expression of rice RAN2 rendered transgenic plants hypersensitive to salinity and osmotic stress (Zang et al., 2010). Moreover, Arabidopsis RAN1 mediated seed development by affecting the onset of endosperm cellularization (Liu et al., 2014). Arabidopsis RanGAP1 lacks the SUMOylated C-terminal domain of vertebrate RanGAP, but instead contains a plant-specific N-terminal WPP domain that directs plant-specific targeting of the protein to the NE in interphase and the cell plate in mitosis (Jeong et al., 2005). Supporting the roles of Arabidopsis RanGAP proteins in cell cycle progression, a null double mutant of RanGAP1 and RanGAP2 arrested at interphase, predominantly after the first mitotic division following meiosis, and thus produced defective and lethal female gametophytes because of its lack of GAP activity but not WPP-dependent subcellular targeting (Rodrigo-Peiris et al., 2011; Boruc et al., 2015).

Leaf senescence is a genetically programmed process that appears at the last stage of leaf development, allowing plants to efficiently remobilize nutrients from leaves to other new growing sinks (Kim et al., 2018). Because of its profound effects on gene expression and metabolism, senescence involves substantial multilayered genetic and metabolic reprogramming including chlorophyll degradation, protein degradation, reallocation of nutrients, an increase in reactive oxygen species (ROS), enhanced programmed cell death (PCD)/necrosis, membrane ion leakage, and differential expression of numerous genes, including senescence-associated genes (SAGs; Woo et al., 2013; Havé et al., 2017; Kim et al., 2018). Hormones are endogenous factors that regulate the onset and progression of leaf senescence, mostly by relaying environmental changes to cell machinery. Cytokinins and auxins delay senescence, whereas ethylene (ET), jasmonic acid (JA), abscisic acid (ABA), and gibberellin (GA) accelerate leaf senescence (Jibran et al., 2013).

Salicylic acid (SA) plays essential roles in plant immune responses via transcription factor-mediated rapid and massive transcriptional reprogramming upon recognition of pathogens (Chen et al., 2020). SA also has been reported to promote leaf senescence (Zhang et al., 2013; Wang et al., 2021). NONEXPRESSOR OF PATHOGENESIS-RELATED GENES1 (NPR1) is a SA receptor, which in response to SA-triggered cellular redox changes converts from cytoplasmic oligomers in the cytoplasm to functional monomers in the nucleus (Mou et al., 2003; Tada et al., 2008). In addition to redox-induced conformational changes of NPR1, diverse post-translational modifications such as phosphorylation, ubiquitination/de-ubiquitination, and SUMOylation are tightly associated with SA-induced systemic acquired resistance (SAR) by regulating NPR1 stability/activity and its nucleocytoplasmic partitioning (Chen et al., 2021). In the nucleus, NPR1 boosts immune responses through a TGACG-BINDING FACTOR (TGA)- and WRKY-mediated transcriptional cascade, leading to the massive induction of defense genes including pathogen-related (PR) genes (Li et al., 2004; Wang et al., 2006a; Jin et al., 2018). In contrast, NPR3/NPR4 (NPR1 paralogs) play opposite roles in the transcriptional regulation of SA-induced defense gene expression by functioning as E3 ligases that promote NPR1 degradation (Ding et al., 2018). Recently, it was demonstrated that positive transcriptional feedback acting on the NPR1–WRKY46–WRKY6 signaling cascade efficiently promoted SA-elicited leaf senescence in Arabidopsis (Zhang et al., 2021). MPK6-mediated transcriptional activation of WRKY6 and Trx h5 also co-regulated the expression of NPR1 and subsequent NPR1 monomerization and nuclear accumulation in SA-induced detached leaf senescence (Chai et al., 2014).

Although Ran has been thoroughly investigated in animals and yeast cells, there is still a lack of knowledge about Ran-GTP/-GDP-dependent differential regulation of Ran functions in plant growth and development. In this study, we generated transgenic plants overexpressing GTP- or GDP-bound forms of Arabidopsis Ran1 and demonstrated that the overexpression of AtRan1-GTP efficiently promoted age-dependent and dark-triggered leaf senescence while Ran-GDP retarded senescence. We also provide evidence that positive regulators of SA signaling, such as TGA2 and NPR1, accumulate differentially in the nucleus of these plants, with opposite expression patterns of senescence-related genes and resulting leaf senescence.

Results

Generation and confirmation of the GTP-/GDP-bound status of site-directed AtRan1 mutant proteins

G19V and T24N mutations of Ran were made in a mammalian cell system based on analogous residues that affect either the guanyl nucleotide state or interactions with the regulatory proteins of Ras (Marshall, 1993; Kornbluth et al., 1994; Lounsbury et al., 1996). We took a similar approach and designed and generated various point-mutated constructs of Arabidopsis Ran1 (AtRan1) at equivalent sites to Ran G19V and Ran T24N of mammalian Ran, namely Ran1G22V (Ran-GTP) and Ran1T27N (Ran-GDP), respectively. We also generated a C-terminal deletion form of Ran1, which was named dCRan1, to study Ran-GTP/-GDP-mediated plant growth and responses (Figure 1A). The deletion of the DEDDDL motif in the C-terminal extension (ΔDEDDDL Ran) destabilized the GDP-bound form of Ran while ΔDEDDDL Ran pulled down RanBP1 to a similar extent as Ran wild-type (Richards et al., 1995).

Figure 1.

Confirmation of the GTP-/GDP-bound status of AtRan1 mutant proteins. A, Schematic diagram showing a variety of mutants of AtRan1 used in this study. B, Co-IP assays of transiently expressed AtRanBP1c-YFP with various mutated forms of AtRan1-MYC. C, In planta BiFC interaction tests of AtRanBP1c-nYFP with various mutated forms of AtRan1-cCFP. DAPI was used to stain nuclei. Scale bar = 20 µm. D, Co-IP assays of HA-tagged NTF2A with mutated forms of AtRan1-GFP. E, In planta BiFC interaction tests of nYFP-tagged NTF2A with mutated forms of AtRan1-cCFP. Scale bar = 20 µm.

Next, we confirmed the GTP or GDP affinity of the mutated proteins by testing their interactions with AtRanBP1c or NTF2A. Previously, it was demonstrated that the Arabidopsis Ran-BP, AtRanBP1c, bound preferentially to the GTP-bound conformation of a pea (Pisum sativum) Ran1, PsRan1, over the GDP-bound conformation (Kim et al., 2001). Moreover, Ran-GDP specifically bound to its nuclear receptor NTF2 to move into the nucleus (Ribbeck et al., 1998). NTF2A is an Arabidopsis homolog of yeast NTF2 and can functionally replace NTF2 in yeast (Zhao et al., 2006). To evaluate if Ran1G22V is a bona fide Ran-GTP, we initially examined interactions of YFP-fused AtRanBP1c with various mutated forms of Myc-tagged AtRan1 by transiently expressing these fusion constructs in Nicotiana benthamiana leaves and performing co-immunoprecipitation (co-IP) assays. As expected, AtRanBP1c successfully co-immunoprecipitated with Myc-tagged Ran1WT, Ran1G22V, and dCRan1 while Ran1T27N failed to interact with AtRanBP1c (Figure 1B). These differential interactions of Ran proteins with AtRanBP1c were further confirmed by demonstrating that the co-expression of the N-terminal portion of YFP-tagged RanBP1c (p35S::RanBP1c-nYFP) and C-terminal portion of CFP-tagged Ran1WT, Ran1G22V, and dCRan1 (p35S::Ran1WT-cCFP, p35S::Ran1G22V-cCFP, or p35S::dCRan1-cCFP, respectively) resulted in a GFP fluorescence signal in the nuclei of N. benthamiana epidermal cells (Figure 1C). In contrast, the transient co-expression of p35S::RanBP1c-nYFP and p35S::Ran1T27N-cCFP did not produce any GFP signal. By the same token, we found that only Ran1WT and the Ran1T27N interacted with NTF2A, as revealed by bimolecular fluorescence complementation (BiFC) and Co-IP assay (Figure 1, D and E). In contrast, neither Ran1G22V nor dCRan interacted with NTF2A. Based on these data, we conclude that Ran1G22V and dCRan1 preferentially bind GTP, while Ran1T27N preferentially binds GDP.

Overexpression of AtRan1-GTP promotes age-dependent and dark-triggered leaf senescence while Ran-GDP retards senescence

Ran expression has been shown to be associated with cell death in animals and yeasts as mentioned previously, mostly due to its roles in regulating mitosis and the cell cycle. Nuclear Ran-GTP level has been shown to be diminished during the early stages of apoptosis, along with the inactivation of the nuclear transport machinery (Wong et al., 2009). In addition, increased RCC1 expression, which increased cellular Ran-GTP levels, accelerated cell cycle progression and DNA damage repair, inhibiting cell cycle arrest and DNA-damage-induced senescence in a human hTERT-RPE1 cell line (Cekan et al., 2016). Considering the possible roles of GTP- or GDP-dependent AtRan1 in Arabidopsis senescence, we generated transgenic plant lines expressing GTP- or GDP-locked AtRan1 fused to EGFP, namely p35S::AtRan1WT-EGFP (Ran1WTox), p35S::AtRan1T27N-EGFP (Ran1T27Nox), p35S::AtRan1G22V-EGFP (Ran1G22Vox), and p35S::AtdCRan1-EGFP (dCRan1ox). We found that all three independent transgenic lines of Ran1T27ox, Ran1G22Vox, and dCRan1ox expressed similar levels of transgenic proteins to each other, except that Ran1WTox plants showed a relatively high expression of transgene protein (Supplemental Figure S1).

Ran1WTox plants produced fewer leaves than other transgenic lines. Nonetheless, the developmental timing of leaf emergence for other transgenic lines was similar, indicating that the leaves we used for examining age-dependent senescence can be considered to be at the same or similar developmental ages (Supplemental Figure S2A). The vegetative to reproductive phase transition, named as bolting or floral evocation, occurs prior to age-dependent leaf senescence in Arabidopsis. It is well known that age-dependent leaf senescence and flowering time are coordinately controlled by the circadian clock which regulates several senescence-related gene expression patterns (Kim et al., 2018). We demonstrated that the flowering time of Ran1WTox, Ran1G22V, and dCRan1ox transgenic lines, determined by measuring the number of rosette leaves or days after germination at bolting, was slightly earlier while Ran1T27N showed no significant difference with that of Col-0 plants (Supplemental Figure S2, B and C). Therefore, we used Ran1WTox-1, Ran1T27Nox-1, Ran1G22Vox-1, and dCRan1ox-1 lines for further analysis. Interestingly, the ectopic expression of GTP-bound forms of AtRan1, such as in Ran1G22Vox and dCRan1ox plants, accelerated age-dependent leaf senescence as determined by the observation that 5-week-old Ran-GTP-overexpressing plants showed visible yellow-colored third and fourth leaves while the leaves of Col-0, Ran1WTox, and Ran1T27Nox plants were still green (Figure 2A). Consistent with this observation, more cell death occurred and was visualized by trypan blue (TB) staining in the leaves of Ran1G22Vox and dCRan1ox plants. Furthermore, diaminobenzidine (DAB) staining detected prominent ROS accumulation in these Ran-GTP overexpressing plants (Figure 2A). In addition, these Ran-GTP-overexpressing plants had already developed mature chloroplasts in younger leaves as assessed by reddish autofluorescence. Moreover, they progressed rapidly to senescing leaves as determined by blue and purple autofluorescence in older leaves, indicating accelerated age-dependent leaf senescence in the Ran-GTP overexpressors (Figure 2B). This developmental progress was slower in Col-0, Ran1WTox, and Ran-GDP overexpressing Ran1T27Nox plants. Supporting the above findings, Ran-GTP overexpressors lost chlorophylls faster; the third/fourth leaves of Ran1G22Vox and dCRan1ox plants at DAG 49 had only 10% or 42% chlorophylls of the leaves at DAG 20, respectively (Figure 2C). This is comparable with Col-0 plants that lost 45% and maintained 55% of their chlorophylls at DAG 49. In contrast, Ran1WTox and Ran1T27Nox plants broke down chlorophylls more slowly than Col-0 plants. Similarly, membrane ion leakage assays revealed a significant loss of ions in Ran1G22Vox and dCRan1ox plants, indicating rapid cell death (Figure 2D). SAG12, SEN4, and AtNAP1, which are widely used as senescence-associated markers, are specifically induced in senescing leaves (Pontier et al., 1999; Guo and Gan, 2006; Jiang et al., 2014). When we compared the expression of these genes in Col-0 plants, we found that these genes were dramatically upregulated in Ran-GTP-overexpressing Ran1G22Vox and dCRan1ox plants, while they were slightly upregulated in leaves of Ran-GDP-overexpressing Ran1T27Nox plants. In contrast, CAB1 and RBCS1A, two genes involved in active photosynthesis, were downregulated in Ran-GTP overexpressors and upregulated in Ran-GDP overexpressors (Figure 2E). Our observation that the ectopic expression of GTP-bound forms of AtRan1-accelerated age-dependent leaf senescence was consistently confirmed in other independent transgenic lines of Ran1ox plants (Supplemental Figure S3).

Figure 2.

Age-dependent leaf senescence of AtRan1ox plants. A, Pictures showing aerial phenotype (scale bar = 2 cm) of 5-week-old AtRan1ox transgenic plants, including TB and DAB staining of transgenic leaves (scale bar = 200 µm). Aerial views of plants were taken using Canon EOS 500D camera and the images were placed on black-field background using the MS Paint program. B, Autofluorescence images of leaves taken from 5-week-old AtRan1ox transgenic plants. C–E, Relative chlorophyll content (C) and degree of ion leakage (D) measured at indicated times, and RT-qPCR gene expression analysis of senescence marker genes (E) performed using the third and fourth rosette leaves of 5-week-old AtRan1ox plants. DAG: days after germination. Leaves were taken from more than 20 plants for each time point of plants (C, D). RT-qPCR analysis was performed using three biological samples. Bar graphs represent means ± sd, and asterisks indicate statistical differences from either DAG20 plants (C, D) or the Col-0 control (E) at *P <0.05 and **P <0.01 (Student’s t test).

Light deprivation in the form of severe shading and darkening of leaves also accelerates senescence and darkness is widely used to trigger synchronous senescence in detached leaves (Weaver and Amasino, 2001; Liebsch and Keech, 2016). Similar to the age-dependent phenotypes, we found that detached leaves of Ran1G22Vox and dCRan1ox plants exhibited an accelerated dark-induced senescence phenotype as ascertained by their intense staining with TB and DAB (Figure 3A and Supplemental Figure S4A), significantly reduced chlorophyll content (Figure 3B and Supplemental Figure S4B), increased ion leakage (Figure 3C), and further downregulation of RBCS1A and CAB1 together with upregulation of SAG12 and AtNAP1 (Figure 3, D–G). Collectively, these results indicate that AtRan1 differentially regulates leaf senescence and that the overexpression of Ran-GTP promotes both age-dependent and dark-triggered leaf senescence while Ran-GDP antagonistically retards senescence.

Figure 3.

Dark-triggered leaf senescence of AtRan1ox plants. The fifth and sixth rosette leaves of 4-week-old AtRan1ox transgenic plants were placed in the dark for 4 d. Seen above are (A) visible leaf phenotype (scale bar = 500 µm) and the results of TB and DAB staining (scale bar = 200 µm), B, total chlorophyll content, C, degree of ion leakage. Leaves were taken from more than 25 plants for each line of plant. D–G, Relative gene expression of RBCS1A (D), CAB1 (E), SAG12 (F), and AtNAP1 (G) before and after dark incubation. RT-qPCR analysis was performed using three biological samples. Bar graphs represent means ± sd, and asterisks indicate statistical differences from the 0-day control before and after incubation in the dark or between dark-incubated AtRan1 transgenic groups. Significant interaction of Ran-GTP/-GDP expression with dark-triggered leaf senescence was analyzed by two-way ANOVA. *P <0.05 and **P <0.01, as determined by Tukey post hoc test. n.s., not significant.

AtRan1-GTP/-GDP transgenic plants show opposite transcriptional expression of hormone-regulated and senescence-related genes

Next, to gain molecular insights into Ran-GTP/-GDP-dependent regulation of leaf senescence, we performed RNA-seq analysis and compared the dark-induced differential gene expression profiles of Ran1G22Vox and Ran1T27Nox plants. Comparison of the transcriptomes of leaves from these plants resulted in the identification of 3,298 differentially expressed genes (DEGs) in Ran1G22Vox and 2,495 DEGs in Ran1T27Nox, with an 851 gene overlap between these two DEG sets (Figure 4A; two-fold change, P <0.05). In addition to these overlapping DEGs, some genes that are not differentially expressed could show differential regulation either in Ran1G22Vox or Ran1T27Nox. Based on this assumption, we extended our investigation and retrieved 2,559 DEGs that showed a more than two-fold change in gene expression between Ran1G22Vox and Ran1T27Nox during dark-triggered leaf senescence. We classified these genes into eight groups based on their expression patterns (Groups A–D relatively upregulated and Group E–H relatively downregulated in senescence-experiencing Ran1G22Vox) and subsequently analyzed their functions based on Gene Ontology terms (Figure 4B and Supplemental Table S1).

Figure 4.

RNA-seq analysis of dark-induced differential gene expression profiles comparing Ran1-GDP-overexpressing AtRan1T27Nox and Ran1-GTP-overexpressing AtRan1G22Vox transgenic leaves. A, Venn diagram showing overlap in DEGs in dark-triggered senescing leaves of AtRan1T27Nox and AtRan1G22Vox plants. DEGs were identified based on the criterion of a more than two-fold change in expression at an FDR-adjusted P < 0.05 before (D0) and after dark incubation (D3). B, Heatmap and GO analysis of AtRan1T27Nox- and AtRan1G22Vox-dependently regulated DEGs. About 2,559 dark-triggered DEGs that showed more than a two-fold difference in gene expression between Ran1T27Nox and Ran1G22Vox were retrieved from (A). These significantly regulated DEGs were then sorted into eight groups (A–H) depending on their regulation pattern and were assigned functional GO term classifications. Group A: upregulated in both AtRan1G22Vox and AtRan1T27Nox, with a further increase in AtRan1G22Vox. Group B: upregulated in AtRan1G22Vox and no significant change in AtRan1T27Nox. Group C: no significant change in AtRan1G22Vox and downregulated in AtRan1T27Nox. Group D: downregulated in AtRan1G22Vox and further suppressed in AtRan1T27Nox. Group E: upregulated in both AtRan1G22Vox and AtRan1T27Nox, with a further increase in AtRan1T27Nox. Group F: no significant change in AtRan1G22Vox and upregulated in AtRan1T27Nox. Group G: downregulated in AtRan1G22Vox and no significant change in AtRan1T27Nox. Group H: downregulated in both AtRan1G22Vox and AtRan1T27Nox, with further suppression in AtRan1G22Vox.

In line with the senescence-promoting phenotypes of Ran1G22Vox leaves, genes involved in stress, aging, leaf senescence, cell death, autophagy, and chlorophyll catabolism were highly upregulated and enriched in Groups A–C. On the contrary, genes involved in photosynthesis, chlorophyll and pigment metabolism, protein folding, and DNA repair were relatively suppressed and enriched in Groups F–H. Furthermore, we found that genes related to the senescence-promoting hormones SA, JA, ABA, and ET were significantly upregulated in dark-triggered senescing leaves of Ran1G22Vox, which indicates that signaling transduction pathways mediated by these hormones are actively involved in Ran-GTP/-GDP-dependent dark-triggered leaf senescence. In fact, SA, BR, ABA, and ET-responsive genes were also found among groups of genes downregulated in Ran1G22Vox or relatively upregulated in Ran1T27Nox plants (Figure 4B, Group E). However, as listed in Supplemental Table S2, many of these genes were negatively involved in senescence-promoting ET signaling pathways. For example, ARGOS (At3g59900) and its homologue OSR1 (At2g41230) physically interact with RTE1, an ET receptor interacting protein, and reduce plant sensitivity to ET, leading to modulation of early ET signaling and improved drought tolerance (Shi et al., 2015). In addition, the overexpression of RAP2.6L (At5g13330) delayed waterlogging-induced premature senescence, possibly through modulation of the ABA signaling pathway (Liu et al., 2012).

Overexpression of AtRan1-GTP facilitates nuclear accumulation of TGA2 and NPR1 transcription factors, which may be the causal factor for Ran1-GTP-promoted age-dependent and dark-triggered leaf senescence

Plant hormones exert many of their physiological roles by adopting stimulus-dependent nucleocytoplasmic partitioning of transcription factors or regulatory proteins (Lee et al., 2008). Ran has pivotal roles in nucleocytoplasmic transportation of cargo proteins, with the Ran-GTP/-GDP gradient across the nuclear membrane directing their movement. In this regard, it is crucial to know which regulatory proteins are transported into or out of the nucleus in a Ran-GDP/-GTP-dependent manner to relay hormonal signals related to age- and dark-induced leaf senescence. As an initial step to identify these regulatory proteins, we retrieved 1,000 base pairs upstream of translational start site from genes belonging to Groups A–H (Figure 4B) and analyzed them using the MEME analysis tool to identify statistically overrepresented consensus motifs (Machanick and Bailey, 2011). This analysis led us to identify diverse promoter consensus motifs including the bZIP motif, which is a common binding site for TGAs, and the TCP motif, which is a binding site for transcription factors belong to the TEOSINTE BRANCHED 1, CYCLOIDEA, and PCF1 (TCPs) families (Table 1; Jakoby et al., 2002; Li et al., 2018). It is noteworthy that both TGAs and TCPs are binding partners of the transcription factor NPR1 (Fan and Dong, 2002; Li et al., 2018). NPR1 is a well-known receptor of SA and contains a functional bipartite NLS (Kinkema et al., 2000; Wu et al., 2012). SA-mediated NPR1 monomerization and phosphorylation-dependent nuclear localization have been reported to be critically involved in regulation of SA-mediated leaf senescence and systemic immunity (Tada et al., 2008; Chai et al., 2014; Lee et al., 2015). Therefore, we proceeded to focus on SA signaling and investigated Ran-GTP/-GDP-dependent spatial dynamics of TGAs and NPR1. We hypothesized that the Ran-GDP/-GTP-dependent subcellular distribution of SA-regulated proteins is responsible for the observed differences in senescence phenotypes between AtRan1-overexpressing plants.

Table 1.

Representative cis-motifs found in promoters of genes categorized in Figure 4B

Motif	Cis-acting element identified in this study	Binding transcription factors	References
Relatively upregulated genes in Ran1G22Vox (A–D)
ABI3VP1	A(A/C)AA(A/C)AAAAA	ABI3
bZIP	GCCACGTG	TGA2, TGA3, TGA5	Jakoby et al. (2002)
AP2/EREEP	(A/G)(A/G)CG(G/A)CG (G/A)(A/C)G	RAP2-6, ERF113	Xie et al. (2019)
NAC	GAAGAAGA(A/C)(G/A)	ANAC042, ANAC083	Nuruzzaman et al. (2013)
Relatively downregulated genes in Ran1G22Vox (E–H)
ABI3VP1	AAAAAAAAA	ABI3
C2H2	GG(A/T)GGAG	SRS5, SEU	Englbrecht et al. (2004)
Trihelix	(A/C)CG(A/G)CGACG	PEX1, LRX2	Nagano et al. (2001)
TCP	(A/T)GGCCCA	TCP8, TCP14, TCP15	Li et al. (2018)

TGA transcription factors in Arabidopsis are divided into five subclades (clade I: TGA1 and TGA4; clade II: TGA2, TGA5, and TGA6; clade III: TGA3 and TGA7; clade IV: TGA9 and TGA10; and clade V: TGA8) with highly overlapping roles in plant defense, stress responses, and/or development (Gatz, 2013). We assessed subcellular accumulation of clades I and II TGAs because of their crucial roles in SA biosynthesis, signaling, and SA-induced transcriptional reprogramming of genes (Zhang et al., 2003; Jin et al., 2018; Sun et al., 2018). When monitoring the Ran-GTP/-GDP-dependent spatial dynamics of transiently expressed RFP-fused TGAs in N. benthamiana in the presence of co-expressed GTP- or GDP-locked AtRan1-EGFP, we realized that a relatively higher ratio of the GDP-bound form of AtRan1-EGFP (Ran1T27N) accumulated in the nucleus while more of the GTP-bound form (Ran1G22V, dCRan1) accumulated in the cytosol (Figure 5A and Supplemental Figure S5, A and B). Interestingly, TGA2 also accumulated in different cellular locations depending on the co-expression of Ran1-GTP or -GDP. Transiently expressed TGA2, in the absence of AtRan1, was localized in both the subcellular area of the cell with about 55% of TGA2 distributed in the nucleus and 45% in the cytoplasm (N-to-C ratio of 1.22% or 122%; Figure 5A and Supplemental Figure S5, A and C). However, the overexpression of the GTP-bound form of Ran1G22V or dCAtRan1 resulted in a significant increase in the level of TGA2 in the nucleus, as revealed by the high fluorescence intensity of TGA2-RFP in the nuclear area (Figure 5A and Supplemental Figure S5C). This preferential accumulation of TGA2 in the nucleus was effectively reversed by the co-expression of the GDP-bound form of Ran1 (Ran1T27N), so that a relatively higher amount of TGA2 was observed in the cytoplasm.

Figure 5.

AtRan1-GTP/-GDP-dependent nucleocytoplasmic accumulation of the transcription factor TGA2. A, Subcellular distribution of TGA2-RFP protein transiently co-expressed with AtRan1-GFP in N. benthamiana. Data show confocal microscope images. Fluorescent signal intensities of TGA2-RFP and AtRan1-GFP were determined along a line drawn on the confocal images using LSM Image Browser 4.0 software and are presented in Supplementary Figure S5. Scale bar = 20 µm. B, Immunoblot analysis of nucleus (N)/cytoplasm (C)-fractionated TGA2-myc proteins transiently co-expressed with AtRan1-GFP. T: total protein extract. Nuclear protein extracts for Ran1-GFP were 10× concentrated compared with cytoplasmic proteins. Histone H3 was used as a nuclear marker. C, Quantified nuclear-cytoplasmic distribution of TGA2-myc and AtRan1-GFP proteins. Blots from (B) were quantitated using ImageJ software. Data show relative protein accumulation in the N and C fractions compared with total protein extracts, which were set to 100%. These experiments were repeated three times. Bar graphs represent means ± sd and asterisks on bracketed samples represent statistical differences between the two compared samples at *P <0.05 and **P <0.01 (Student’s t test).

We further confirmed the Ran1-GTP-/-GDP-dependent differential accumulation of TGA2 by performing immunoblot analysis against total proteins obtained from nuclear and cytosolic fractions. Wild-type AtRan1 showed a distribution of 99% in the cytoplasm and 1% in the nucleus (Figure 5, B and C). Again, we found that a relatively higher amount of Ran1-GDP (Ran1T27N) accumulated preferentially in the nucleus (7.8%) while relatively more Ran1-GTP (Ran1G22V or dCRan1) was found in the cytosol (nuclear levels decreased compared with Ran1T27N). Regarding TGA2 distribution in the presence or absence of AtRan1, 45% of TGA2-myc was found in the cytosolic fraction and 55% in the nucleus when AtRan1 was not co-expressed (Figure 5, B and C). Consistent with our previous results, the co-expression of Ran1-GDP (Ran1T27N) led to predominant TGA2-myc localization in the cytoplasm while Ran1-GTPs (Ran1G22V or dCRan1) promoted TGA2-myc accumulation in the nucleus. These results indicated that the subcellular distribution of TGA2 was differentially regulated by GTP-/GDP-bound AtRan1, and that the overexpression of AtRan1-GTP promoted the nuclear accumulation of TGA2 while the overexpression of AtRan1-GDP supported its translocation to the cytoplasm. Surprisingly, subcellular accumulation of other TGA proteins such as TGA1, TGA4, TGA5, and TGA6 was not regulated by GTP- or GDP-bound AtRan1 overexpression (Supplemental Figure S6). These findings suggest that the Ran machinery is specific for TGA2-mediated cellular processes.

NPR1 is the transcriptional co-activator of TGA2, and the formation of SA-induced transcriptional reprogramming by HAC histone acetylase–NPR1–TGA complexes activates massive expression of pathogenesis-related genes (PRs) and the resulting immune responses in plants (Fan and Dong, 2002; Jin et al., 2018). Co-transformation of NPR1-RFP together with GTP- or GDP-locked AtRan1-GFP into N. benthamiana leaves revealed that co-expression with AtRan1-GTP (Ran1G22V or dCRan1) led to accumulation of NPR1–RFP predominantly in the nucleus (Figure 6A and Supplemental Figure S7). In contrast, upon the co-transformation of GDP-bound Ran1T27N, the signal of nucleus-localized RFP-NPR1 decreased significantly, and higher accumulation of this protein was observed in the cytosol. Consistent with these confocal observations, we detected a similar distribution pattern of NPR1-myc when we performed western blot assays of proteins in the nuclear and cytoplasmic fractions (Figure 6, B and C). In brief, NPR1-myc not co-transformed with AtRan1 showed a roughly equal distribution between the cytoplasm and nuclear fraction, with approximately 45% NPR1-myc in the nucleus and 55% in the cytoplasm. The nuclear level of NPR1-myc increased to 80% in the presence of the wild-type form of AtRan1-GFP, which implied that AtRan1 is involved in nuclear localization of NPR1. This nuclear localization of NPR1 was further reinforced by the co-expression of GTP-bound AtRan1G22V and dCRan1 with about 90% of NPR1-myc found in the nuclear fraction. In contrast, co-expression with GDP-bound AtRan1T27N resulted in the opposite outcomes; more than 85% of NPR1-myc was found in the cytosolic fraction. In fact, cytosolic accumulation was higher than that observed for NPR1-myc accumulation without AtRan1 co-transformation. Collectively, these results indicate that Ran differentially regulates the subcellular distribution of NPR1, and AtRan1-GTP promotes the nuclear accumulation of NPR1, similar to the distribution patterns observed for NPR1’s binding partner TGA2.

Figure 6.

AtRan1-GTP/-GDP-dependent nucleocytoplasmic accumulation of the transcription factor NPR1. A, Subcellular distribution of NPR1-RFP protein transiently co-expressed with AtRan1-GFP in N. benthamiana. Confocal microscope images are shown. Fluorescent signal intensities of NPR1-RFP and AtRan1-GFP were measured as described in Figure 5A and are presented in Supplemental Figure S7. Scale bar = 20 µm. B, Immunoblot analysis of nucleus (N)/cytoplasm (C)-fractionated NPR1-myc protein transiently co-expressed with AtRan1-GFP. T, total protein extract. Nuclear protein extracts for Ran1-GFP were 10× concentrated compared with cytoplasmic proteins. Histone H3 was used as a nuclear marker. C, Quantified nuclear-cytoplasmic distribution of NPR1-myc and AtRan1-GFP proteins. Blots from (B) were quantitated and normalized as described in Figure 5C. D, E, Effect of SA on Ran1-GTP/-GDP-dependent subcellular accumulation of NPR1 protein. Nicotiana benthamiana leaves transiently co-expressing NPR1-myc and AtRan1-GFP proteins were incubated in the presence and absence of 0.1-mM SA for 1 h. After nucleocytoplasmic fractionation, immunoblot analysis (D) and quantification of proteins (E) were performed as described in B and C. Bar graphs represent mean ± sd, and asterisks on bracketed samples represent statistical differences between the two compared samples at *P <0.05 and **P <0.01 (Student’s t test). Experiments in (C) and (E) were repeated three times.

In the cytosol, NPR1 exists mainly as oligomers. Upon pathogen infection or SA treatment, it is reduced to a monomeric form by thioredoxins and translocates to the nucleus to activate downstream targets (Mou et al., 2003; Tada et al., 2008). As described previously, NPR1-myc not co-expressed with AtRan1 showed a roughly equal cytosolic versus nuclear accumulation pattern (55% in the cytosol and 45% in the nucleus; Figure 6, D and E). Consistent with previous reports, SA effectively directed NPR1 to the nucleus so that the level of nuclear NPR1-myc increased significantly up to 75%. However, the co-expression of Ran1-GDP (AtRan1T27N) attenuated this effect of SA with 60% of NPR1 observed in the nucleus. In contrast, the co-expression of Ran1-GTP (AtRan1G22V) resulted in more than 95% accumulation of NPR1 in the nucleus regardless of exogenous SA application. Interestingly, it seemed that SA also affected Ran1 protein distribution itself depending on its GDP- or GTP-bound status. Exogenous treatment with SA increased the level of nuclear Ran1-GDP from 10% to 30% while the amount of nuclear Ran1-GTP that accumulated in the nucleus did not change significantly. Similar trends in nucleocytoplasmic accumulation were observed for SA-triggered Ran1-GDP/-GTP and NPR1-myc in the presence or absence of AtRan1 in N. benthamiana leaves transiently expressing those proteins (Supplemental Figure S8). In summary, we provided clear evidence that the Ran nuclear transport machinery regulates the subcellular localization of TGA2 and NPR1. We propose that the differential distribution pattern of TGA2 and NPR1 in different Ran-GDP/-GTP overexpressing transgenic plant backgrounds is the causal factor in Ran-GTP-modulated senescence.

SA-induced, Ran-GTP-/-GDP-dependent nuclear accumulation of NPR1 and TGA2 results in genome-wide global transcriptional reprogramming including that of genes involved in cell death, aging, and photosynthesis

To further explore possible SA-induced and AtRan1-mediated NPR1 target genes at a genome-wide level, we retrieved two sets of publicly available transcriptome profile data sets to compare the overlap between these genes and AtRan1-GDP/-GTP-dependently regulated genes. One of the public sets was DEGs obtained by comparing the gene expression profiles of leaves from 4-week-old Col-0 plants and those of npr1 plants, and the other set comprised genes differentially expressed between INA (a SA analog)- versus mock-treated Col-0 plants (GSE101567, Jin et al., 2018). We identified 342 overlapping genes (Figure 7A and Supplemental Table S3), and subsequently classified them into four groups based on their differential expression patterns incurred by INA treatment, npr1 genetic background, and AtRan1-GDP/-GTP-dependent dark-triggered senescence (Figure 7B, Groups A–D and Supplemental Table S3). In brief, genes in Groups A and B were upregulated by INA treatment and downregulated in npr1 plants and were either relatively downregulated (Group A) or upregulated (Group B) in dark-incubated senescing Ran1G22Vox plants. In comparison, genes in Groups C and D were downregulated by INA treatment and upregulated in npr1 plants and were either relatively downregulated (Group C) or upregulated (Group D) in dark-incubated senescing Ran1G22Vox plants. We demonstrated that the overexpression of AtRan1-GTP (Ran1G22V and dCRan1) facilitated nuclear accumulation of NPR1 and TGA2 transcription factors and suggested that this might be the cause of Ran1-GTP-promoted age-dependent and dark-triggered leaf senescence. Based on this hypothesis, we expected that a group of genes with regulatory pattern similar to INA-treated Col-0 plants and senescing Ran1G22Vox plants, such as genes in Groups B and C, would be potential SA-induced and AtRan1-mediated NPR1 target genes. In facts, Group B genes, which were INA-induced and relatively upregulated in senescing Ran1G22Vox plants, included genes involved in stress responses, cell death, aging, and cell wall organization. For example, genes for SA methyltransferase (FAMT) and senescence-promoting factors such as MKK9, SAG21, SEN4, and SAG113 were found among Group B genes (Supplemental Table S4). By the same token, genes in Group C, which were INA-repressed and also further suppressed in senescing Ran1G22Vox plants, were involved in an SA-induced negative feedback loop regulating systemic immunity, photosynthesis, chloroplast organization, and chlorophyll biosynthesis.

Figure 7.

Transcriptome analysis of SA-induced and AtRan1-regulated NPR1 target genes. A, Venn diagram showing overlapping DEGs among dark-triggered senescing leaves of AtRan1T27Nox and AtRan1G22Vox plants, Col-0 and npr1, and mock- and INA (a SA analog)-treated Col-0 plants. Genes with a two-fold or greater change in expression were analyzed further. B, Heatmap-based classification (Groups A–D) and functional GO term assignment of the 342 overlapping DEGs taken from (A). Group A: INA-induced and npr1-suppressed, with relatively higher expression in senescing AtRan1T27Nox leaves. Group B: INA-induced and npr1-suppressed, with relatively higher expression in senescing AtRan1G22Vox leaves. Group C: INA-repressed and upregulated in npr1 leaves with relatively higher expression in senescing AtRan1T27Nox leaves. Group D: INA-repressed and upregulated in npr1 leaves with relatively higher expression in senescing AtRan1G22Vox leaves.

To validate the results of RNA-seq analysis, we examined the expression of MKK9, SAG21, and SEN4 by performing reverse transcription quantitative polymerase chain reaction (RT-qPCR) and demonstrated that these genes were upregulated in senescing leaves of Ran1-GTP-overexpressing Ran1G22Vox and dCRan1ox plants while Ran1-GDP-overexpressing Ran1G22Vox plants showed a limited increase in the expression of these genes compared with Col-0 and Ran1WTox plants (Figure 8, A–C). This Ran-GDP/-GTP-dependent differential regulation of senescence-promoting genes was further confirmed by testing effector-induced transient expression of MKK9 promoter-, SEN4- promoter, or SAG21 promoter-driven luciferase reporters (pMKK9::Luc, pSEN4::Luc and pSAG21::Luc, respectively) in N. benthamiana leaves. Consistent with the RT-qPCR results, the co-expression of Ran1-GTP-locked Ran1G22V and dCRan1 together with NPR1 and TGA2 greatly enhanced the expression of promoter-fused luciferase reporters, while the opposite was observed for Ran1-GDP-locked Ran1T27N (Figure 8, D–F). Contrarily, the co-expression of TGA5 was less effective in these transcriptional regulations of reporter genes. Collectively, these results indicate that SA-induced, Ran-GTP-/-GDP-dependent nuclear accumulation of NPR1 and TGA2 induces genome-wide global transcriptional reprogramming including that of genes involved in cell death, aging, and chloroplast organization, resulting in age-dependent and dark-triggered leaf senescence.

Figure 8.

AtRan1-GTP/-GDP-dependent transcriptional regulation of diverse SA-responsive, senescence-related genes. A–C, RT-qPCR analysis of the gene expression of MKK9 (A), SEN4 (B), and SAG21 (C) performed using the fifth and sixth rosette leaves of 5-week-old AtRan1ox plants before and after dark incubation for 4 d. Transcript accumulation of genes was normalized to that of Col-0 or each transgenic line before dark treatment, which was set to 1. Significant interaction of Ran-GTP/-GDP expression with dark-triggered expression of senescence-related genes was detected by two-way ANOVA. *P <0.05 and **P <0.01, as determined by Tukey’s post hoc test. n.s., not significant. Experiments were repeated three times. Bar graphs represent mean ± sd. D–F, Transactivation tests of pMKK9::Luc (D), pSEN4::Luc (E), and pSAG21::Luc reporters (F) by different AtRan1 effector proteins. Luciferase activity was normalized to GUS activity of the same transfected samples. Bar graphs represent mean ± sd, and asterisks indicate statistical differences from a reporter control expressing p35S::GUS together with pMKK9::Luc, pSEN4::Luc, or pSAG21::Luc reporter at *P <0.05 and **P <0.01 (Student’s t test). Experiments were repeated three times.

AtRan1-mediated control of leaf senescence is tightly linked to SA-triggered NPR1 functions, rather than ET-regulated EIN3/EIL1 functions

Thus far, we have demonstrated that SA-induced and Ran-GTP-dependent nuclear accumulation of NPR1 and TGA2 results in enhancement of leaf senescence via global transcriptional reprogramming. To genetically evaluate if there is a causal relationship between SA signaling and Ran-GDP/-GTP-dependent regulation of leaf senescence, we generated lines of transgenic plants expressing GTP- or GDP-locked AtRan1-EGFP under the genetic background of the SA signaling-defective mutant npr1 (Cao et al., 1997), SA biosynthesis-deficient mutant sid2 (mutated in SA-biosynthetic ICS1 gene; Nawrath and Métraux, 1999), or ET signaling-defective mutant ein3eil1 (Alonso et al., 2003). We showed that the expression level of different Ran transgene proteins is relatively similar in different genetic backgrounds except that Ran1WT is expressed relatively highly (Supplemental Figure S9). Investigating the effects of these genetic backgrounds on Ran-GDP/-GTP-dependent leaf senescence revealed that the expression of GTP-bound Ran1G22V or dCRan1 in the non-npr1 Col-0 wild-type genetic background dramatically enhanced age-dependent leaf senescence as expected, as determined by greater loss of chlorophyll content (0.42 mg·g^-1 FW or 0.52 mg·g^-1 FW, respectively) than Col-0 control plants (1.92 mg·g^-1 FW; Figure 9, A and B). However, the expression of these Ran1-GTP forms in npr1 plants severely attenuated this Ran1 effect, so that a high level of chlorophyll (1.75–1.77 mg·g^-1 FW) remained in the aging leaves of Ran1G22Vox/npr1 and dCRan1ox/npr1 plants. In contrast, the expression of the GDP-bound form of Ran1 (Ran1T27N) in Col-0 wild-type did not result in any notable differences in chlorophyll amounts compared with Col-0 control plants, nor were there significant changes when this mutant protein was expressed in npr1 plants. This NPR1-dependent differential effect of Ran1-GTP and -GDP was further confirmed by our observation that the ectopic expression of Ran1-GTP in Col-0 plants facilitated a massive loss of ions from leaves, but this effect was effectively negated in npr1 plants (Supplemental Figure S10A). Similar to these npr1 effects on Ran-GDP/-GTP-regulated age-dependent leaf senescence, the expression of Ran1-GTP in the SA biosynthesis-deficient sid2 mutant successfully abolished the effects of Ran1-GTP, thereby retarding age-promoted leaf senescence (Figure 9, A and C and Supplemental Figure S10B). These data imply that SA-induced and age-dependent leaf senescence are tightly regulated by Ran1-mediated NPR1 functions.

Figure 9.

Genetic verification of AtRan1-GTP/-GDP-dependent leaf senescence due to NPR1-mediated SA signaling. A, Pictures showing aerial phenotypes of 5-week-old AtRan1ox transgenic plants, ectopically expressed in Col-0, npr1, sid2, or ein3ei1l mutant backgrounds. Scale bar = 1 cm. B–D, Total chlorophyll content in the third and fourth rosette leaves of plants overexpressing different forms of AtRan1 in npr1 (B), sid2 (C), and ein3eil1 (D) genetic backgrounds. Leaves were taken from more than 20 plants for each line of plant. E, Pictures showing the 4-d dark-triggered senescing phenotype of fifth and sixth rosette leaves, taken from 4-week-old AtRan1ox transgenic plants ectopically overexpressing different forms of AtRan1 in Col-0, npr1, sid2, or ein3ei1l mutant backgrounds. Scale bar = 1 cm. F–H, Total chlorophyll content of leaves presented in E overexpressing different forms of AtRan1 in npr1 (F), sid2 (G), and ein3eil1 (H) genetic backgrounds Leaves were taken from more than 25 plants for each line of plant. Bar graphs represent mean ± sd, and asterisks on bracketed samples represent statistical differences between the two compared samples at *P <0.05 and **P <0.01 (Student’s t test). ns, not significant.

We showed in an earlier experiment that genes related to senescence-promoting ET were significantly enriched in the dark-triggered senescing leaves of Ran1G22Vox (Figure 4B). Unexpectedly, we found that Ran1-GTP-driven chlorophyll and ion losses were observed regardless of expression of this protein in Col-0 or ein3eil1 plants, which suggests that age-driven and ET-induced leaf senescence may be not directly mediated by Ran-GTP/-GDP-dependent nucleocytoplasmic shuttling of EIN3 and EIL1 proteins (Figure 9, A and D and Supplemental Figure S10C). Supporting our speculation that AtRan1-mediated control of leaf senescence is tightly regulated by SA-triggered NPR1 functions but not by ET-regulated EIN3/EIL1 functions, detached leaves of Ran1G22Vox and dCRan1ox plants exhibited an accelerated dark-induced senescence phenotype in Col-0 plants as evidenced by significantly reduced chlorophyll content and increased ion leakage after 4 d of dark incubation (Figure 9, E–H and Supplemental Figures S10, D–F). However, the expression of these Ran1-GTP forms in npr1 or sid2 plants severely attenuated the Ran1 effects, while expression in ein3eil1 plants could not suppress the Ran1-GTP-driven promotion of dark-induced leaf senescence. In reality, more than 97% of either EIN3 (Supplemental Figure S11A) or EIL1 (Supplemental Figure S11B) was already accumulated in the nucleus regardless of the co-expression with GTP- or GDP-bound AtRan1. EIN3 physically interacts with NPR1 in senescing leaves to synergistically promote the expression of the SAG ORE1 and SAG29 (Wang et al., 2021). We tested in N. benthamiana leaves whether the co-expression of EIN3 or EIL1 with NPR1 influences the transient expression of SAG29 promoter-driven luciferase reporters (pSAG29::Luc) in a Ran-GDP and Ran-GTP-dependent manner. As expected, the transient expression of NPR1 or EIN3 alone significantly increased the reporter expression of SAG29 gene (Supplemental Figure S11C). The co-expression of NPR1 and EIN3 further enhanced the reporter expression; nonetheless, this increase in gene expression was observed in a similar level regardless of the co-expression of Ran-GDP or Ran-GTP. The similar pattern of SAG29 regulation was also found in case of the co-expression with the EIL1. These results imply that ET-induced leaf senescence is not mediated by Ran machinery-dependent nuclear shuttling of EIN3 and EIL1 proteins.

Discussion

Ran is the master regulator of nucleocytoplasmic transportation of NLS-containing cargo proteins through nuclear pore complexes, and the unequal distribution of cytoplasmic Ran-GDP and nuclear Ran-GTP gives directionality to nuclear import/export processes. Proteins with NLS are targeted and transported into the nucleus with the help of RanGDP-binding karyopherins such as importins and nucleoporins, and proteins with nuclear export signal (NES) are excluded from the nucleus by the action of other karyopherin members such as RanGTP-binding CRM1/exportin (Lee et al., 2008; Merkle, 2011).

We showed that wild-type or mutated forms of AtRan1 did not directly interact with NPR1 and TGA2 (Supplemental Figure S12). Nonetheless, we demonstrated that the overexpression of AtRan1-GTP resulted in predominant nuclear accumulation of TGA2 and NPR1 regardless of the presence of SA, thus facilitating premature promotion of age-dependent and dark-triggered leaf senescence. Regarding the regulatory functions of Ran-GTP and Ran-GDP, the co-expression of a GTPase-deficient, Ran-GTP-locked RanQ69L in Hela cells led to predominantly nuclear accumulation of an NES-exposed and cytoplasm-destined mutant of superoxide dismutase 1 (SOD1), SOD1^G85R (Zhong et al., 2017), consistent with our findings. By the same token, the co-transfection of dominant negative RanT24N (GDP-bound form) into breast cancer cell lines efficiently abolished nuclear transport of endogenous TβRI, a type I TGF-β receptor, while the transfection of a constitutively active Ran mutant, RanF35A, enabled nuclear transport of the TβRI protein (Chandra et al., 2012). Nonetheless, several other studies have reported contradictory findings for Ran-GTP and Ran-GDP. For example, the co-expression of GTP-bound RanQ69L in human glioblastoma U87MG cells prevented nuclear accumulation of PTEN, a NLS-containing tumor suppressor phosphatase (Gil et al., 2006). In this regard, it was argued that not only the Ran-GDP/-GTP gradient across the NE but also the GTPase activity of Ran is important for control of nucleocytoplasmic transport of proteins and mRNAs. Addition of nonhydrolyzable GTP to digitonin-permeablized HeLa cells effectively inhibited both protein and snRNP import (Palacios et al., 1996). In the same report, it was proven that both RanGTP-locked RanQ69L and RanGDP-locked RanT24N blocked nuclear import of NLS-fused proteins. Moreover, mutations affecting the Ran-GTP/-GDP cycle such as Rna1p (RanGAP) and Prp20p (RanGEF), or the overproduction of dominant negative Ran-GTP (RanG21V) inhibited the nuclear export of the ribosomal protein L25 in yeast cells, as evidenced by the predominantly nuclear accumulation of L25-GFP (Hurt et al., 1999). Ran-mediated nucleocytoplasmic partitioning of effector proteins can be regulated either through import control or export control. We demonstrated that a GTP-bound form of AtRan1 (Ran1G22V, dCRan1) accumulated at relatively high levels in the cytoplasm while the GDP-bound form (Ran1T27N) accumulated relatively more in the nucleus. It is not clear at this point what in our Ran-GTP or -GDP-overexpressing plants was modulated to facilitate differential regulation of leaf senescence. Was it caused by changes in the intrinsic Ran gradient, GTPase activity, or both? Further studies are needed to clarify the mechanism of Ran-GTP/-GDP-dependent nucleocytoplasmic regulation of hormone effector proteins.

Previous reports demonstrated that several TGAs, especially class II TGAs such as TGA2, TGA5, and TGA6, work redundantly as pathogen-induced SAR, JA-, and ET-dependent defenses against necrotrophic pathogens, as well as UV-B-incurred abiotic stress (Zhang et al., 2003; Zander et al., 2010; Herrera-Vásquez et al., 2021). In particular, TGA2 and TGA3 were strongly recruited to PR1 promoters harboring an activating sequence-1 (as-1) cis element in an SA- and NPR1-dependent manner to function as transcriptional activators (Fan and Dong, 2002; Johnson et al., 2003). In this regard, our findings are interesting in that only NPR1 and TGA2, but not other TGAs, affected effector-induced transient expression of senescence-related genes such as MKK9, SEN4, and SAG21 as well as dark-triggered leaf senescence in a Ran-GTP/-GDP-dependent manner. We speculate that it may be because TGA2 had the strongest interaction with NPR1 while TGA5 and TGA6 have weaker affinity (Zhou et al., 2000). TGA3 was shown to be expressed primarily in a subset of young seedlings and degraded in mature tissues, while TGA2 was persistently detectable in mature leaves as well as in seedlings (Pontier et al., 2002), suggesting that TGA3 is not involved in senescence. In N. benthamiana, TGA2.2 is the main component of as-1-binding factor-1 (ASF-1), and the expression of a mutated form of TGA2.2 unable to form heterodimers with the endogenous pool of TGA factors led to reduced SA-inducibility of the “immediate early” gene, Nt103 (Thurow et al., 2005). At this time, we cannot rule out the possibility that there are no binding partners of other TGAs in the N. benthamiana leaves in which we performed these experiments.

Although most NPR1 studies have focused on its role in SA-induced plant immune responses, some studies have reported a causal relationship between NPR1-mediated SA signaling pathways and leaf senescence (Yoshimoto et al., 2009; Chai et al., 2014; Zhao et al., 2016). Interestingly, accumulation of SA or constitutive SA signaling resulted in premature developmental leaf senescence, but not dark- or carbon starvation-induced leaf senescence (Yoshimoto et al., 2009; Zhao et al., 2016). Indeed, previous transcriptomic and genetic studies suggested that SA signaling was mostly involved in natural, but not dark-induced leaf senescence (van der Graaff et al., 2006). In contrast, SA treatment or the ectopic accumulation of SA in a dominant negative mutant of maize (Zea mays, inbred line B73) MAPK, ZmMEK1^KR, facilitated the dark-induced senescence process in Arabidopsis and maize (Li et al., 2016). Plants use two independent routes to produce SA, the isochorismate synthase (ICS) and phenylalanine ammonia-lyase (PAL) pathways (Huang et al., 2020). We demonstrated that Ran machinery was involved in both age-dependent and dark-triggered leaf senescence under our experimental conditions. Ran1G22Vox plants had elevated expression levels of SA-biosynthesis genes such as PAL1, PAL4, PBS3, and EDS5 compared with Col-0 plants of the same developmental age (Gene Expression Omnibus [GEO] data set GSE173375). These results imply that Ran-GTPox plants not only have active NPR1/TGA2-mediated SA signaling but also effectively increase endogenous SA levels through ICS1 and PAL pathways to synergistically induce both age-dependent and dark-triggered leaf aging.

We found that a variety of genes related to senescence-regulating hormones, including SA and ET, were differentially regulated in dark-triggered senescing leaves, which indicates that signaling transduction by these hormones is actively involved in Ran-GTP/-GDP-dependent dark-triggered leaf senescence. Contrary to our expectations, the expression of Ran1-GTP in ein3eil1 plants was not effective at interfering with the Ran1-GTP-driven promotion of dark-induced leaf senescence, while the expression of Ran1-GTP in npr1 or sid2 plants severely attenuated dark-driven leaf aging. Recently, it was reported that NPR1 binding to EIN3 synergistically promoted the expression of the SAGs ORE1 and SAG29 (Wang et al., 2021). It is probable that SA-directed NPR1 affects leaf senescence through several parallel pathways, including EIN3-dependent regulation of ORE1 and SAG29 genes and EIN3-independent, maybe TGA2-dependent, regulation of senescence genes. In support of this idea, SA can induce early responding genes through either NPR1-dependent or -independent pathways, with some involvement of TGA2, TGA5, and TGA6-activated transcriptional regulation (Blanco et al., 2009). In addition, autophagy-defective mutants exhibited premature leaf senescence and excessive PCD that was SA signaling dependent but did not require intact JA or ET signaling pathways (Yoshimoto et al., 2009).

Leaf senescence is a genetically controlled dismantling program that involves substantial metabolic and genetic reprogramming and whose timing is affected by hormone-relaying developmental and environmental signals: different hormones and regulatory proteins have diverse time-dependent roles throughout the progression of age- and dark-induced leaf senescence (Jibran et al., 2013; Liebsch and Keech, 2016). Recently, utilization of multi-omics approaches including transcriptomics and proteomics has successfully resolved the complicated process of leaf senescence, replacing a single-component-based molecular understanding of mechanisms. In this study, we focused on the effects of Ran1 on NPR1/TGA2 localization in SA-induced leaf senescence, based on our observation that the TGA-binding bZIP cis-motif was overrepresented among Ran1-regulated genes. Further identification of hormone-regulatory proteins and clarification of how the AtRan1 G-protein controls nucleocytoplasmic shuttling of these proteins would add another level to our understanding of age-dependent and dark-triggered leaf senescence.

Materials and methods

Plant materials and growth conditions

Wild-type Arabidopsis (A. thaliana; Columbia-0, Col-0), SA-related mutants npr1-1 and sid2-1, the ET-insensitive mutant ein3eil1, and variants of AtRan1-overexpressing transgenic plants were used in this study. Seeds of npr1-1 and sid2-1 were kindly provided by Dr Yoo-Sun Noh (Seoul National University, South Korea). Seeds of ein3eil1 were provided by Dr Rashmi Sasidharan (Utrecht University, The Netherlands). Following 2 d of cold stratification, seeds were grown in pots containing Sunshine No. 5 soil (Sun Gro Horticulture, USA) in an environmentally controlled growth room under a 16-h light (100–150 µmol m⁻² s⁻¹) and 8-h dark cycle at 23°C–25°C with 80%–85% humidity.

Plasmid construction and generation of transgenic plants

To generate transgenic plants ectopically expressing variants of AtRan1-EGFP, cDNAs encoding either full-length or truncated forms of AtRan1 were PCR-amplified (Supplemental Table S5) and cloned into the CaMV35S promoter-driven EGFP-expressing pB7FWG2 (Karimi et al., 2002) binary vector using Gateway technology (Invitrogen, USA). The QuickChange site-directed mutagenesis kit was used to create point mutations (AtRan1G22V and AtRan1T27N; Agilent Technologies, USA). Subsequently, Agrobacterium tumefaciens GV3101 carrying each construct was used to transform Col-0, npr1, sid2, or ein3eil1 genetic lines using the floral dipping method (Structure of expression constructs in Supplemental Table S6). The expression of EGFP- tagged transgenic proteins was confirmed by western blot analysis using anti-GFP antibodies as a probe.

Co-IP and BiFC assays in N. benthamiana

To test protein interactions in planta using co-IP assays as described in Figure 1, cDNAs encoding variants of AtRan1, the full-length coding region of AtRanBP1c, and NTF2A were PCR-amplified (Supplemental Table S5) and cloned into pGWB20 in fusion with MYC, pB7YWG2 in fusion with YFP, and pGWB15 in fusion with HA, respectively, as previously described (Karimi et al., 2002; Kim et al., 2020; Supplemental Table S6). Next, A. tumefaciens GV3101 containing p35S::RanBP1c–YFP and a variant construct of AtRan1-MYC was co-infiltrated through the underside of 2- to 4-week-old N. benthamiana leaves. Similarly, A. tumefaciens GV3101 containing p35S::HA-NTF2A and a variant construct of AtRan1-GFP was co-infiltrated into N. benthamiana leaves. Total proteins were then extracted from infiltrated leaves after 36–48 h of incubation, and immunoprecipitation was performed using Protein G magnetic beads (Bio-Rad, USA) conjugated either to anti-GFP or anti-HA antibodies (Abcam, USA). Co-immunoprecipitated proteins were then separated by 12% (w/v) sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), and the presence of Ran1-MYC or Ran1-EGFP proteins was analyzed by immunoblotting with anti-MYC (Abcam, USA) or anti-GFP antibodies (Santa Cruz Biotechnology, USA). The immunoblot images were taken using Image Quant LAS 400 (GE Healthcare Life Science, USA).

To examine in planta interaction using BiFC assays, Agrobacterium containing wild-type or mutant AtRan1 fused with the C-terminal half of CFP (cCFP) in pPZP312-cCFP, together with RanBP1c or NTF2A fused with the N-terminal half of YFP (nYFP) in pPZP312-nYFP were co-infiltrated into N. benthamiana leaves as described previously (Kim et al., 2020). After 36–48 h of incubation, epidermal cell layers were observed for fluorescence using a Zeiss LSM 710 confocal laser scanning microscope. Diamidino-2-phenylindole (DAPI) was used to stain nuclei. Co-transformed samples were excited at 488 nm to collect GFP signal at 490–530 nm or excited at 405 nm to detect DAPI signal at 436–486 nm.

Luciferase reporter assays

To examine effector-induced transient transcription of the MKK9 promoter-, SEN4 promoter-, SAG21 promoter-, or SAG29 promoter-driven luciferase reporters (pMKK9::Luc, pSEN4::Luc, pSAG21::Luc, and pSAG29::Luc, respectively) in planta, genomic DNA fragments covering 1,398, 1,356, 1,490, or 2,054 bp upstream of the MKK9 (At1G73500), SEN4 (At4G30270), SAG21 (At4G02380), or SAG29 (At5G13170) transcription start sites, respectively, were PCR-amplified (Supplemental Table S5) and cloned into the Gateway-compatible binary vector pBGWL7 in fusion with the full-length coding region of the luciferase gene as described previously (Kim et al., 2020; Supplemental Table S6). Next, Agrobacterium carrying one of these reporter constructs, together with Agrobacterium containing mutant forms of Ran1 and other effectors were co-infiltrated along with a transfection control (p35S::GUS) into N. benthamiana leaves. After 36–48 h of incubation, luciferase activity in transfected leaves was measured using a GloMax 20/20 luminometer (Promega, USA). Luciferase activity was normalized to GUS activity of the same transfected samples.

Transient expression of proteins and determination of their subcellular distribution in N. benthamiana

To determine subcellular localization of transiently expressed proteins in N. benthamiana, Agrobacterium expressing CaMV35S promoter-driven wild-type or mutated forms of Ran1 fused to EGFP at the C-terminus were co-infiltrated into N. benthamiana leaves with Agrobacterium containing a recombinant binary vector ectopically expressing TGA2- and NPR1-RFP (see Supplemental Tables S5 and S6 for gene amplification and binary vectors used). After 36–48 h of incubation, epidermal cell layers were observed for fluorescence using a Zeiss LSM 710 confocal laser scanning microscope. DAPI was used to stain nuclei. The signal intensities of GFP, RFP, and DAPI in indicated areas were quantitatively determined using LSM Image Browser 4.0 software (Carl Zeiss, Germany). RFP was excited with 488 nm and 543 nm laser lines to detect the light emission at 568–608 nm.

To further investigate whether subcellular accumulation of the proteins being evaluated was influenced by different Ran1 genetic backgrounds, Agrobacterium expressing a variant of Ran1-EGFP was co-infiltrated into N. benthamiana leaves together with Agrobacterium containing a recombinant binary vector ectopically expressing TGA1-, TGA2-, TGA4-, TGA5-, TGA6-, NPR1-, EIN3-, or EIL1-MYC (see Supplemental Table S6). Nuclear-cytoplasmic fractions were prepared as described previously (Wang et al., 2011), except that leaves of the transfected N. benthamiana were used for fractionation and subsequent immunodetection of proteins. Subcellular distribution of Ran1-EGFP or the MYC-tagged TGAs, NPR1, EIN3, and EIL1 was determined by western blot analysis using anti-GFP or anti-MYC antibodies as a probe after separation of proteins by 12% (w/v) SDS–PAGE. Histone H3 was used as a nuclear marker. The immunoblot images were taken using Image Quant LAS 400 (GE Healthcare Life Science, USA).

Leaf senescence assays

To examine age-dependent and dark-induced leaf senescence, the third and fourth rosette leaves of 5-week-old Arabidopsis plants (for age-dependent leaf senescence) or the fifth and sixth leaves of 4-week-old plants (for dark-induced leaf senescence) were prepared, and their chlorophyll content, membrane ion leakage, autofluorescence, and transcript expression of senescence marker genes were examined as described previously (Kim et al., 2020). Leaves were stained with TB or DAB to assess cell death or the presence of ROS such as hydrogen peroxide and superoxide, respectively.

RT-qPCR analysis

Total RNA was isolated from leaves using a plant RNA extraction kit (iNtRON Biotechnology, Korea) and reverse transcribed to first-strand cDNA using the ReverTraAce qPCR RT Master Mix Kit (Toyobo, Japan). RT-qPCR was performed using the SYBR green method (Applied Biosystems, USA) and the primers listed in Supplemental Table S7. The expression of each transcript was normalized to that of the UBC1 control in each sample.

RNA-sequencing and data analysis

To determine genes differentially expressed between Ran-GTP- and Ran-GDP-overexpressing plants, total RNA was obtained from the leaves (the 5th and 6th leaves) of 4-week-old Col-0, p35S::Ran1T27N-EGFP/Col-0 (Ran1T27Nox), and p35S::Ran1G22V-EGFP/Col-0 (Ran1G22Vox) before and after 3 d of incubation in the dark. RNA libraries were constructed from these RNA samples using the SENSE 3′ mRNA-Seq Library Prep Kit (Lexogen Inc., Austria) according to the manufacturer’s instructions. High-throughput sequencing was performed using NextSeq. 550 (Illumina Inc., USA). Next, we retrieved the published transcriptome profile data of genes differentially expressed between the leaves of Col-0 and npr1-1 plants or Col-0 leaves before and after treatment with the SA analog INA (a GEO data set GSE101567; Jin et al., 2018). We compared these DEGs with the Ran-regulated gene sets (deposited as GEO data set GSE173375) to identify genes that overlapped with SA-responsive genes.

Transcriptome analysis was performed as described previously (Kim et al., 2020). In brief, DEGs were identified using R (version 3.5.2) and the edgeR package was used for normalization and correction within arrays. Statistic P-value was below 0.05, and expression levels 1 < log2(fold change) <−1 were used to identify significantly regulated DEGs. Heatmaps were generated using the “gplots” package of R (https://cran.r-project.org/web/packages/gplots/gplots.pdf). Venn diagrams were constructed using the Venny tool (https://bioinfogp.cnb.csic.es/tools/venny/). Gene ontology (GO) annotation and enrichment of GO terms were conducted using resources from The Gene Ontology Consortium (http://www.geneontology.org/) and a P <0.05 was used to identify significantly enriched GO categories.

Cis-motif predictions in promoters

Multiple EM for Motif Elicitation (MEME; http://meme.sdsc.edu) analysis tool was used to search for cis-motifs (6–10 bps) conserved in the promoter regions 1,000 bp upstream of the start codons of genes that were differentially regulated between Ran1T27Nox and Ran1G22Vox plants.

Statistical analysis

Twenty to thirty leaves were collected for each experiment. Three biological replicates were included in each experiment, and the significance of differences between treatments or groups was assessed using Student’s t test or two-way ANOVA analysis.

Data availability

All data associated with the paper are available in this manuscript. Novel materials used and described in the paper are available for noncommercial research purposes from soohwan@yonsei.ac.kr.

Accession numbers

Global gene expression data set mentioned in this article can be downloaded from the GEO under accession numbers GSE173375 and GSE101567.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Western blot analysis of AtRan1-GFP expression in independent lines of AtRan1-overexpressing transgenic plants.

Supplemental Figure S2. Leaf emergence and floral evocation in independent lines of AtRan1-overexpressing transgenic plants.

Supplemental Figure S3. Age-dependent leaf senescence in independent lines of AtRan1-overexpressing transgenic plants.

Supplemental Figure S4. Dark-triggered leaf senescence in independent lines of AtRan1-overexpressing transgenic plants.

Supplemental Figure S5. Plot profiles and measurement of AtRan1-GTP/-GDP-dependent nucleocytoplasmic accumulation of TGA2.

Supplemental Figure S6. Immunoblot analysis of nucleus (N)/cytoplasm (C)-fractionated TGAs-myc protein transiently co-expressed with different forms of AtRan1-GFP proteins.

Supplemental Figure S7. Plot profiles and measurement of AtRan1-GTP/-GDP-dependent nucleocytoplasmic accumulation of NPR1.

Supplemental Figure S8. Effect of SA on Ran1-GTP/-GDP-dependent subcellular accumulation of NPR1.

Supplemental Figure S9. Western blot analysis of AtRan1-GFP expression in AtRan1-overexpressing transgenic plants transformed into different genetic backgrounds.

Supplemental Figure S10. AtRan1-GTP/-GDP-dependent leaf senescence due to NPR1-mediated SA signaling as determined by the ion leakage assay.

Supplemental Figure S11. AtRan1-GTP/-GDP-independent nucleocytoplasmic accumulation of EIN3/EIL1 and the resulting transient expression of pSAG29::Luc reporter.

Supplemental Figure S12. Analysis of interaction of NPR1 and TGA2 with AtRan1 proteins.

Supplemental Table S1. A list of differentially regulated genes grouped (Figure 4B, Groups A–H) according to their expressions in dark-triggered senescing leaves of AtRan1T27Nox and AtRan1G22Vox plants.

Supplemental Table S2. A list of representative genes belonging to Group E in Figure 4.

Supplemental Table S3. A list of differentially-regulated genes grouped (Figure 7B, Groups A–D) according to their differential expressions in Col-0 versus Col-0 + INA, Col-0 versus npr1, and Ran1T27Nox versus Ran1G22Vox undergoing dark-induced senescence.

Supplemental Table S4. Representative overlapping senescence-related genes which are differentially regulated in Col-0 versus Col-0 + INA, Col-0 versus npr1, and Ran1T27Nox versus Ran1G22Vox undergoing dark-induced senescence.

Supplemental Table S5. Primers used in the amplification of cDNA, promoter or site-directed mutagenized genomic DNA.

Supplemental Table S6. Structure of expression constructs generated in this study.

Supplemental Table S7. Primers used in RT-qPCR analysis.

 

G.P. conceived the research plan and performed most of the experiments including Figures 1–3, 6, 8, and 9. D.-M.S. analyzed the RNAseq data and generated Figures 4 and 7 and Table 1. Y.K. provided the technical assistance to G.P. and D.-M.S. S.-H.K. conceived and supervised the research project and wrote the article with contributions of all the authors.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Soo-Hwan Kim (soohwan@yonsei.ac.kr).