- © 2019 American Society of Plant Biologists. All rights reserved.
Abstract
Chloroplast-to-nucleus retrograde signaling is essential for the coupled expression of photosynthesis-associated nuclear genes (PhANGs) and plastid genes (PhAPGs) to ensure the functional status of chloroplasts (Cp) in plants. Although various signaling components involved in the process have been identified in Arabidopsis (Arabidopsis thaliana), the biological relevance of such coordination remains an enigma. Here, we show that the uncoupled expression of PhANGs and PhAPGs contributes to the cell death in the lesion simulating disease1 (lsd1) mutant of Arabidopsis. A daylength-dependent increase of salicylic acid (SA) appears to rapidly up-regulate a gene encoding SIGMA FACTOR BINDING PROTEIN1 (SIB1), a transcriptional coregulator, in lsd1 before the onset of cell death. The dual targeting of SIB1 to the nucleus and the Cps leads to a simultaneous up-regulation of PhANGs and down-regulation of PhAPGs. Consequently, this disrupts the stoichiometry of photosynthetic proteins, especially in PSII, resulting in the generation of the highly reactive species singlet oxygen (1O2) in Cps. Accordingly, inactivation of the nuclear-encoded Cp protein EXECUTER1, a putative 1O2 sensor, significantly attenuates the lsd1-conferred cell death. Together, these results provide a pathway from the SA- to the 1O2-signaling pathway, which are intertwined via the uncoupled expression of PhANGs and PhAPGs, contributing to the lesion-mimicking cell death in lsd1.
INTRODUCTION
Plastids are semiautonomous subcellular organelles whose biogenesis and functions largely depend on coordinated nuclear gene expression. Plastid-generated retrograde signals mediate this coordination via a process called retrograde signaling (RS), which involves biogenic and operational signaling (Pogson et al., 2008; Jarvis and López-Juez, 2013; Chan et al., 2016; Kleine and Leister, 2016). Biogenic signaling coordinates the expression of photosynthesis-associated nuclear genes (PhANGs) during chloroplast (Cp) biogenesis; Operational signaling regulates the expression of nuclear genes, which are mostly associated with stress/defense responses, in response to environmental fluctuation. Several Cp-derived signals, including tetrapyrrole intermediates (Mochizuki et al., 2001; Larkin et al., 2003; Strand et al., 2003), reactive oxygen species (ROS; Wagner et al., 2004; Lee et al., 2007; Ramel et al., 2013), redox signals (Escoubas et al., 1995; Baier and Dietz, 2005), β-cyclocitral (Ramel et al., 2012), metabolites (Estavillo et al., 2011; Xiao et al., 2012), calcium (Guo et al., 2016), and chloroplast-to-nucleus shuttling proteins (Sun et al., 2011; Isemer et al., 2012b), participate in RS pathways. Disturbance of Cp biogenesis by genetic mutations or by inhibitors such as norflurazon (an inhibitor of carotenogenesis) and lincomycin (LIN, an inhibitor of plastid translation) coordinately reduces the expression of a set of PhANGs and photosynthesis-associated plastid genes (PhAPGs; Susek et al., 1993; Pesaresi et al., 2006; Woodson et al., 2013). This awareness led to the discovery of several genetic loci, such as GENOMES UNCOUPLED (GUN), whose inactivation impairs the coordinated expression of PhANGs (Susek et al., 1993; Mochizuki et al., 2001; Larkin et al., 2003; Koussevitzky et al., 2007). However, to date, the biological relevance of this impaired coordination and/or uncoupled expression of photosynthesis-associated genes distributed to nucleus and plastid genomes on plant development and/or stress responses remains unclear.
Programmed cell death (PCD) is a genetically regulated cellular suicide initiated during developmental or caused by biotic or abiotic stresses (Pennell and Lamb, 1997; Huysmans et al., 2017). Indeed, the identification and the analysis of many lesion mimic mutants displaying spontaneous hypersensitive response-like cell death raised our understanding of the mechanisms underlying PCD. RS also influences PCD via Cp-generated pro-death signals, particularly ROS-derived RS (Wagner et al., 2004; Lee et al., 2007; Kim et al., 2012). The photosynthetic electron transport chain (PETC) is a major site of ROS production. When the absorption of light energy exceeds the maximum capacity of the PETC for CO2 assimilation, singlet oxygen (1O2) and hydrogen peroxide (H2O2)/superoxide anion (O2●−) are generated by PSII and PSI, respectively (Apel and Hirt, 2004; Laloi et al., 2004). Moreover, the uncontrolled accumulation of free tetrapyrrole molecules including protoporphyrin IX, protochlorophyllide, and chlorophyll catabolites, all of which act as potent photosensitizers upon light absorption, leads to the generation of 1O2 (Tanaka et al., 2003; Wagner et al., 2004; Lee et al., 2007; Woodson et al., 2015). Previous studies on the Arabidopsis fluorescent (flu) mutant, which conditionally generates 1O2 in Cps upon a dark-to-light shift, reported that 1O2 rapidly induces changes in the nuclear gene expression and cell death (Meskauskiene and Apel, 2002; op den Camp et al., 2003). Subsequent genetic screens revealed that the nuclear-encoded Cp proteins EXECUTER1 (EX1) and EX2 mediate 1O2 signaling in the flu mutant, indicating that 1O2-triggered cell death is genetically controlled rather than directly resulting from the cytotoxicity of 1O2 (Wagner et al., 2004; Lee et al., 2007; Kim et al., 2012).
Along with the enhanced levels of ROS, accumulation of the plant defense hormone salicylic acid (SA) also facilitates PCD, suggesting a probable crosstalk between ROS- and SA-mediated signaling pathways (Herrera-Vásquez et al., 2015). Recent studies have also demonstrated that Cps deploy stromules, stroma-filled tubular extensions, to coordinate the pro-PCD signals, ROS and SA, during effector-triggered immunity (Caplan et al., 2015; Kumar et al., 2018). In addition, it was suggested that the nuclear-encoded plastid protein WHIRLY1 senses the Cp-synthesized SA, facilitating its translocation from the Cps to the nucleus to mediate retrograde immune signaling (Desveaux et al., 2004; Isemer et al., 2012a, 2012b; Carella et al., 2015). The vast majority of SA is synthesized through the isochorismate (ICS) pathway in Cps (Garcion et al., 2008). SA plays a role in increasing ROS levels by inhibiting the activities of ROS-scavenging enzymes (Chen et al., 1993; Durner and Klessig, 1995) and by inducing stomatal closure, which in turn induces photorespiration by lowering intracellular CO2 concentrations and results in concomitant ROS production (Mateo et al., 2006). Conversely, elevated levels of ROS from Cps also lead to SA accumulation (Ochsenbein et al., 2006; Lv et al., 2015). In Arabidopsis, this positive feedback loop between ROS and SA is known to be negatively regulated by LESION STIMULATING DISEASE1 (LSD1), a cysteine-2/cysteine-2-class zinc-finger protein (Dietrich et al., 1997; Aviv et al., 2002; Kaminaka et al., 2006). Hence, LSD1 negatively regulates cell death and basal defense responses (Dietrich et al., 1994, 1997). Interestingly, the lsd1 mutant develops uncontrolled cell death in a light-dependent manner (Dietrich et al., 1994). The excess excitation energy (EEE)-induced hyper-reduction of the plastoquinone (PQ) pool in Cps, which causes the photorespiratory burst of ROS, triggers lsd1-conferred cell death by inducing SA-dependent signaling pathways through its key signaling components, PHYTOALEXIN DEFICIENT4 (PAD4) and ENHANCED DISEASE SUSCEPTIBILITY1 (EDS1; Mateo et al., 2004; Mühlenbock et al., 2008; Karpiński et al., 2013). Consequently, dynamic crosstalk between Cp-derived ROS and SA is important in mediating light-dependent cell death. However, so far there is very little evidence underlying the molecular link between ROS and SA (Lv et al., 2015).
In this study, by exploiting the lesion mimic lsd1 mutant, we reveal that the uncoupled expression of PhANGs and PhAPGs plays an important role in downstream and upstream events of the SA and ROS signaling pathways, respectively. We show that SA-dependent transcriptional induction of SIGMA FACTOR BINDING PROTEIN1 (SIB1) appears to alter the expression of both PhANGs and PhAPGs upon its dual targeting to the nucleus and Cps, which results in the subsequent activation of 1O2-triggered and EX1-dependent RS that contributes to the lsd1-conferred cell death.
RESULTS
Daylength Determines the Timing of Runaway Cell Death in lsd1 Mutants
The lsd1 mutant plants develop uncontrolled cell death, also called runaway cell death (RCD), under both long-day (LD, 16-h light/8-h dark cycles) and continuous light (CL; 24 h of light) but not under short-day (SD, 8-h light/16-h dark cycles) conditions (Dietrich et al., 1994; Senda and Ogawa, 2004). We re-evaluated this daylength-dependent RCD to examine the global gene expression changes in the lsd1 mutant before the onset of RCD. For this, we used an in vitro culture system to accurately determine the time of emergence of RCD. The RCD phenotype was observed under both LD and CL (Figures 1A and 1B), coinciding with previous reports (Dietrich et al., 1994; Senda and Ogawa, 2004). When grown under CL, the first and the second leaves of the lsd1 mutant started to show RCD at ∼19–20 d after seed imbibition (Figures 1A and 1B). Similar phenotypes were observed in the lsd1 mutant grown under LD, but the onset of RCD started ∼6 d later than under CL (Figures 1A and 1B). Unexpectedly, the RCD phenotype was also observed in the lsd1 mutant grown under SD, previously defined as a permissive condition (Dietrich et al., 1994; Mühlenbock et al., 2008). The visible RCD phenotype became clear after 44 d under SD condition (Figures 1A and 1B). These results suggest that the timing of emergence of the lsd1-conferred RCD (hereafter lsd1 RCD) is proportional to the daylength (length of light per day).
Daylength Determines the Timing of RCD in lsd1 Mutants.
(A) Phenotypes of wild-type and lsd1 plants grown under CL, LD (16-h light / 8-h dark), or SD (8-h light/16-h dark) conditions.
(B) Quantitative analysis of leaf RCD in wild-type and lsd1 plants grown under CL, LD, or SD conditions. Results represent the mean of three independent measurements. For each measurement, at least 20 plants were analyzed. Error bars indicate sd.
The degree of RCD in the leaves of the lsd1 mutant grown under CL was visualized with trypan blue (TB), which selectively stains dead cells. The TB staining confirmed the observed onset of RCD in the lsd1 mutant plants 20 d after seed imbibition as indicated by the emergence of TB-stained areas, whereas no clear TB staining was detected in wild-type plants (Supplemental Figure 1A). In addition, the RCD was also quantified by measuring the cellular ion leakage and the maximum photochemical efficiency of PSII (Fv/Fm). Significant increases in ion leakage (Supplemental Figure 1B) and decreases in Fv/Fm were consistently confirmed in the mutant plants (Supplemental Figure 1C).
lsd1 Mutation Causes Substantial Up-Regulation of Immune-Related Genes
The visible onset of RCD in lsd1 mutant plants under CL condition started at around nine d after imbibition. What are the molecular mechanisms behind this process? To answer this question, we tried to gain insight into the molecular basis of lsd1 RCD by performing RNA sequencing (RNA-Seq) and comparing the changes in global gene expression between wild-type and lsd1 mutant plants grown under CL condition. Total RNA was extracted from 17- and 19-d-old plants to identify a group of genes that may be involved in the lsd1 RCD. Compared to wild type, 17- and 19-d-old lsd1 mutant plants showed a down-regulation of 70 and 73 genes (at least twofold), respectively, of which 22 genes overlapped (Supplemental Figure 2A; Supplemental Data Set 1).
The results of gene ontology (GO) term enrichment analyses indicated that 121 different genes down-regulated in the lsd1 mutant are mainly involved in photosynthesis, electron transport chain, and oxidation-reduction processes (Supplemental Data Set 2). By contrast, many genes appeared to be up-regulated in the lsd1 mutant compared to wild type. In 17- and 19-d-old lsd1 mutant plants, 395 and 514 genes are up-regulated (at least twofold), respectively (Supplemental Figures 2B and 2C; Supplemental Data Set 3). Nearly 72% (285 of the 395) of genes up-regulated at 17 d overlapped with genes up-regulated at 19 d (Supplemental Figure 2B; Supplemental Data Set 3), indicating that transcriptional reprogramming occurred in lsd1 mutant before the onset of RCD. The RNA-Seq results of selected genes, including PATHOGENESIS-RELATED1 (PR1), PR2, MITOCHONDRIAL PHOSPHATE TRANSPORTER2, ALTERNATIVE OXIDASE1D (AOX1D), SENESCENCE-ASSOCIATED GENE13, ATP BINDING CASSETTE G40 (ABCG40), BON ASSOCIATION PROTEIN2 (BAP2), and IMMUNE-ASSOCIATED NUCLEOTIDE BINDING PROTEIN7, were validated using reverse transcription quantitative PCR (RT-qPCR). Although no significant differences were observed in 15-d-old plants, these genes were clearly up-regulated in the 17- and 19-d-old lsd1 mutant plants compared to wild type, coinciding with the RNA-Seq results (Supplemental Figure 2D).
The GO term enrichment analysis may provide further insight into the processes affected by the up-regulation of the genes in lsd1. The analysis (P value < 0.05) of the 624 different genes up-regulated in the lsd1 mutant revealed that nearly 31% (196 of the 624) of the genes were annotated to the top 20 significantly enriched GO terms in the Biological Process (BP) ontology after eliminating the redundancy. Of the identified GO terms, “response to bacterium” (P value = 2.36E-35), “immune system process” (P value = 2.72E-18), “regulation of defense response” (P value = 1.01E-13), “PCD” (P value = 7.05E-12), and “SA-mediated signaling pathway” (P value = 8.99E-09) were the most significantly overrepresented (Supplemental Figure 2E; Supplemental Data Set 4). Among these genes, we found 54 genes associated with protein phosphorylation-mediated cellular signaling networks (Shiu and Bleecker, 2001; Matsushima and Miyashita, 2012). These proteins include cysteine-rich receptor-like protein kinases, mitogen-activated protein kinases, leucine-rich repeat kinase family proteins, and wall-associated kinases (Supplemental Data Set 4). In addition, almost 22% (16 of 72) of the Arabidopsis WRKY transcription factor (TF) genes were up-regulated in the lsd1 mutant (Supplemental Data Set 3). Seven WRKY TFs, including WRKY18, WRKY25, WRKY33, WRKY40, WRKY46, WRKY53, and WRKY70, function as potential hubs in the formation of the WRKY regulatory circuits (Mouna et al., 2015). Notably, all of these WRKYs except WRKY25 were significantly up-regulated in lsd1. Indeed, 218 (35%) of the 624 lsd1-dependent up-regulated genes were identified as target genes of WRKY18, WRKY33, and/or WRKY40 (Supplemental Data Set 5; Birkenbihl et al., 2017), suggesting a possible role of the WRKY regulatory network in priming the lsd1 RCD.
Uncoupled Expression of Photosynthesis-Associated Genes Before the Onset of RCD
Not only were a lot of genes involved in immune and defense processes differentially regulated in the lsd1 mutant after the onset of RCD but also PhANGs (five genes) and PhAPGs (17 genes) were notably up- (Supplemental Data Set 3) and down-regulated (Supplemental Data Set 1), respectively. The whole sets of PhANGs and PhAPGs were retrieved from the pathways (map00195 and map00196) of the Kyoto Encyclopedia of Genes and Genomes database (http://www.genome.jp/kegg/). Their expression patterns (up- and down-regulation in the lsd1 mutant in comparison with the wild type) were clustered, and z-scores were visualized in heatmaps (Supplemental Figures 3A and 3B). These data sets comprise genes mainly encoding for proteins constituting light-harvesting antennae, PSII, PSI, cytochrome b6f, and ATP synthase (Supplemental Table 1). The heatmaps show that the overall expression levels are markedly higher for PhANGs and lower for PhAPGs in the lsd1 mutant than in wild type (Supplemental Figures 3A and 3B). Moreover, the difference was more apparent in 17-d-old plants, suggesting that the uncoupled expression occurred before the onset of RCD. Among the 47 genes (39 PhANGs and eight PhAPGs) that were significantly differentially expressed (P value < 0.05), 35 PhANGs were clearly up-regulated and eight PhAPGs were down-regulated in 17-d-old lsd1 mutant plants (Figures 2A; Supplemental Data Set 6). Only four PhANGs, including PHOTOSYNTHETIC NDH SUBCOMPLEX L2 (PNSL2), PNSL3, CONSERVED ONLY IN THE GREEN LINEAGE160, and PSBP-LIKE PROTEIN1, were down-regulated in the lsd1 mutant (Figure 2A; Supplemental Data Set 6). Interestingly, those 35 PhANGs up-regulated in the lsd1 mutant included 13 light-harvesting chlorophyll a/b binding protein (LHCB) and four LHCA genes encoding LHCAs in PSII and PSI, respectively (Figure 2A; Supplemental Data Set 6). While most of the up-regulated PhANGs were involved in light harvesting, the eight PhAPGs down-regulated in the lsd1 mutant encode proteins that function in photochemical quenching of the absorbed light energy, i.e. conversion of light energy to chemical energy. These eight PhAPGs encode PsbA/D1, PsbB/CP47, and PsbE/Cytochrome b559 in PSII; PetG and PetN in cytochrome b6f; PsaA and PsaB in PSI; and AtpH, a subunit of the ATP (adenosine triphosphate) synthase (Figure 2A). Importantly, the changes in expression of the PhANGs and PhAPGs were correlated with their protein abundance (Figure 2B). It is also noted that the uncoupled expression of PhANGs and PhAPGs was observed in not only old but also new emerging leaves being green and healthy from 19-d-old lsd1 mutants, implying no distinct effect of emerging leaves on the uncoupled transcript abundances of PhANGs and PhAPGs in lsd1 (Supplemental Figure 4). Taken together, these observations suggest that the lsd1 mutation led to the uncoupled expression of PhANGs and PhAPGs before the emergence of the visible cell death.
Uncoupled Expression of PhANGs and PhAPGs Before the Onset of the Isd1 RCD.
(A) Genome-wide expression analysis revealed contrasting expression patterns of PhANGs and PhAPGs between both 17- and 19-d-old lsd1 mutant and wild-type plants grown under CL; these patterns are illustrated by the heatmap. The colors of the heatmap represent the z-scores ranging from green (z-score of −1.5) through black to red (z-score of 1.5).
(B) Levels of LHCBs (encoded by LHCB genes in a subset of PhANGs); PSII reaction center D1, D2, and CP43 proteins (encoded by psbA, psbB, and psbC genes, respectively, in a subset of PhAPGs); and AtpF protein (encoded by atpF gene in a subset of PhAPGs) in wild type and lsd1 plants grown under CL at the indicated time points. Denaturing gels stained with Coomassie Brilliant Blue were used as the loading control. CBB: Coomassie Brilliant Blue; DAI: days after imbibition; WT: wild type.
lsd1 Is Not a gun Mutant
Because of the gun-like phenotype (Susek et al., 1993), i.e. uncoupled expression of PhANGs and PhAPGs, it is conceivable that LSD1 could act as a downstream component of the GUN-mediated RS pathway. Although earlier genetic approaches have not revealed LSD1 in this signaling pathway, we examined its potential role in the GUN-mediated RS pathway. Along with wild type and gun1, lsd1 mutants were germinated on Murashige & Skoog (MS) medium supplemented with LIN, a plastid translation inhibitor leading to the repression of PhANGs expression via the GUN-mediated RS. As expected, the significant derepression of PhANGs such as LHCB1.1 and LHCB2.1 was confirmed in the gun1 mutant seedlings because of the impaired RS. By contrast, both wild type and lsd1 mutant were found to repress the expression of LHCBs (Supplemental Figure 5), indicating that lsd1 does not belong to the gun mutants. Moreover, unlike most gun mutants that display their phenotype during early seedling development (Susek et al., 1993; Mochizuki et al., 2001; Larkin et al., 2003), LSD1 is likely involved in the operational RS because the concurrent up-regulation of stress/defense-related genes and the uncoupled expression of PhANGs and PhAPGs observed in lsd1 mutant plants appear during a later developmental stage.
SA Synthesized Via the Cp ICS Pathway Primes lsd1 RCD
The results of the GO term enrichment analyses revealed a substantial number of defense/immune-related genes (Supplemental Figure 2E; Supplemental Data Set 4) that are up-regulated in the lsd1 mutant before the emergence of the visible cell death. Because SA is one of the major defense hormones in plants and a key regulator in the expression of defense/immune-related genes (Delaney et al., 1994; Ding et al., 2018), we further compared the 624 lsd1-dependent up-regulated genes with the SA-responsive genes (204 genes, Supplemental Data Set 7) obtained from a previously published microarray data (GSE61059; Zhou et al., 2015). Eighty-seven genes up-regulated in 17-d-old lsd1 mutant plants also appeared in the list of SA-responsive genes (Figure 3A; Supplemental Data Set 8), implying the possible up-regulation of SA in the lsd1 mutant plants grown under CL condition. Thus, we quantified the amount of free SA and the expression level of ISOCHORISMATE SYNTHASE1 (ICS1), which encodes a key enzyme involved in SA synthesis, in the lsd1 mutant and in wild type. Both free SA and ICS1 transcripts were seemingly increased in the 16-d-old lsd1 mutant, while 14-d-old lsd1 mutant showed levels of both free SA and ICS1 transcripts comparable to those of wild type (Figures 3B and 3C). In addition, the lsd1 mutant also rapidly accumulated PR1 and PR2 proteins, two SA-related marker proteins, in accordance with the accumulation of free SA (Figure 3D). The rapid increase of the cellular SA content following d 14 after seed imbibition suggests that SA accumulation and up-regulation of SA-responsive genes precede lsd1 RCD.
SA Accumulation and SA Signaling Prime lsd1 RCD.
(A) Venn diagram showing the numbers of uncommon and shared genes between lsd1-dependent up-regulated genes and SA-responsive genes.
(B) and (C) Endogenous free SA levels (B) and ICS1 transcript levels (C) were examined in wild-type and lsd1 plants grown under CL at the indicated time points. The results represent the means of three independent biological replicates. ACT2 was used as an internal control for (C). Error bars indicate sd. Lowercase letters in (B) indicate statistically significant differences between mean values (P < 0.05, one-way analysis of variance with posthoc Tukey’s Honest Significant Difference test). DAI: days after imbibition; FW: fresh weight; WT: wild type.
(D) PR1 and PR2 protein levels in wild-type and lsd1 plants grown under CL. RBCL was used as loading control.
(E) and (F) Impact of the inactivation of key SA signaling components and SA depletion on the lsd1-conferred RCD. lsd1 eds1 (l/eds1), lsd1 pad4 (l/pad4), and lsd1 npr1 (l/npr1) double mutants, as well as lsd1 transgenic plants (l/cpN) overexpressing the cpNahG (bacterial SA-hydrolyzing enzyme NahG fused with a Cp transit peptide of small subunit of RUBISCO) were grown on MS medium under CL. RCD phenotype (E) and maximum photochemical efficiency of PSII (Fv/Fm) (F) were examined in 26-d-old plants. The representative images are shown at the same scale. For the measurement of Fv/Fm, 10 leaves per genotype were used for each measurement. Data represent the means from three independent measurements. Error bars indicate sd. Lowercase letters in (B) and (F) indicate statistically significant differences between mean values (P < 0.05, one-way analysis of variance with posthoc Tukey’s Honest Significant Difference test).
The requirement of SA signaling in priming RCD in the lsd1 mutant grown under CL condition was further substantiated by inactivating key SA signaling components (Figures 3E and 3F). Double mutants were created crossing lsd1 with mutants impacted in the function of either of these components, including PAD4, EDS1, and NONEXPRESSER OF PR GENES1 (NPR1; Rustérucci et al., 2001; Aviv et al., 2002), or alternatively expressing the bacterial SA hydroxylase (NahG) fused to the Cp transit peptide of the small subunit of RUBISCO under the control of the CaMV 35S promoter. The Cp-targeted NahG (cpNahG) hydrolyzes SA synthesized via the ICS pathway. All analyzed lsd1 eds1, lsd1 pad4, and lsd1 npr1 double mutants as well as the lsd1 cpNahG transgenic plants completely abolished RCD, which was accompanied with the decline of Fv/Fm (Figures 3E and 3F).
Rapid and Transient Up-Regulation of the Transcriptional Coregulator SIB1 in Response to SA
Given that the lsd1 RCD phenotype is completely abrogated by the inactivation of those SA signaling components and by the depletion of SA (Figure 3E), we speculated that SA-responsive TFs and/or transcriptional coregulators may also function in the regulation of PhANGs and PhAPGs. Hence, we first focused on the 87 genes that were up-regulated both in the lsd1 mutant and in response to SA (Figure 3A; Supplemental Data Set 8) and identified seven TFs and four transcriptional coregulators, respectively (Supplemental Data Set 8). Among the proteins encoded by these 11 genes, the transcriptional coactivator SIB1 (also known as valine–glutamine [VQ]23) attracted our attention because of its dual targeting to both the nucleus and the Cps (Lai et al., 2011). SIB1 is a nuclear-encoded protein and a member of the plant-specific VQ (FxxxVQxxTG; x, any amino acid) motif-containing protein family (Xie et al., 2010; Lai et al., 2011). Previous reports demonstrated that SIB1 interacts with the WRKY33 TF in the nucleus (Lai et al., 2011) and the RNA polymerase δ-factor SIG1 in the Cps (Morikawa et al., 2002). It was also shown that SIB1 positively regulates the SA- and JA-mediated expression of nuclear-encoded defense genes after infection of the bacterial pathogen Pseudomonas syringae (Xie et al., 2010) and the necrotrophic fungal pathogen Botrytis cinerea (Lai et al., 2011). By contrast, the expression of plastid-encoded genes, such as psaA and psaB, are repressed by SIB1 (Xie et al., 2010). Because the lsd1 mutant shows the down-regulation of PhAPGs including psaA and psaB but up-regulation of SA-responsive genes along with PhANGs before the onset of RCD (Figures 2A and 3A), we reasoned that SIB1 might play a role as a transcriptional coregulator in both nucleus and Cps in priming the lsd1 RCD.
To examine the protein abundance and the subcellular localization of SIB1 in response to SA, we generated transgenic Arabidopsis plants expressing SIB1 fused to the GREEN FLUORESCENT PROTEIN (GFP) reporter gene driven by the SIB1 promoter. Five-d-old transgenic plants grown on MS medium were transferred to fresh MS medium containing 1.0 mM SA. The GFP signal was determined by confocal microscopy 6 h after the SA treatment. Under normal growth conditions, the expression of SIB1-GFP was nearly absent, as verified by the lack of GFP signal by confocal microscopy and by immunoblot analysis using an anti-GFP antibody (Figures 4A and 4B). However, exogenous application of SA caused the rapid accumulation of SIB1-GFP proteins in the nucleus (Figure 4A). Although the GFP signal in the Cps was not as strong as that in the nucleus (Figure 4A), the Cp form of mature SIB1-GFP (37 kD) lacking its N-terminal plastid transit peptide was clearly detected by immunoblot analysis (Figure 4B). An earlier study also showed that SIB1 is localized in both nucleus and Cps in transgenic Arabidopsis plants overexpressing SIB1 under the control of the CaMV 35S promoter (Lai et al., 2011).
SIB1-GFP Fusion Proteins Are Dually Targeted to the Nucleus and Cps in Response to SA.
(A) Subcellular localization of SIB1-GFP in response to SA. Five-d-old transgenic wild-type plants harboring pSIB1:SIB1-GFP were initially grown on MS medium under CL. The plants were subsequently transferred to MS medium containing 1.0 mM SA. GFP fluorescent signals in guard cells (top) and mesophyll cells (bottom) were detected by confocal microscopy 6 h after SA treatment. Chl.: Chlorophyll.
(B) Both the nuclear and Cp forms of the SIB1-GFP proteins were detected by immunoblot analysis using anti-GFP antibody at the indicated time points after 1.0 mM SA treatment. PR1 and UGPase were used as a positive control of the SA response and as a loading control, respectively. CBB: Coomassie Brilliant Blue; Nu: nucleus.
SA-Mediated Up-Regulation of SIB1 Leads to Uncoupled Expression of PhANGs and PhAPGs
We further examined the role of SIB1 as a transcriptional coregulator in both the nucleus and the Cps. Because SIB1 was rapidly up-regulated by SA (Figure 4B), the effect of SA on the expression of PhANGs and PhAPGs was first determined by RT-qPCR. Fifteen-d-old wild-type plants grown on MS medium under CL were sprayed with either a 0.5 mM solution of SA or a mock solution. Then, leaf tissues were harvested from the wild-type plants at 6 h after each treatment. As expected, the strong up-regulation of SIB1 and PR1 genes were observed in wild-type plants after SA treatment (Supplemental Figure 6A). The RT-qPCR results also showed that SA treatment resulted in clearly induced expression of selected PhANGs such as LHCB1.1 and LHCB2.2 compared to the mock-treated wild-type plants. By contrast, the expression of PhAPGs including psbA and psbB were significantly repressed by SA treatment (Supplemental Figure 6A), indicating that SA leads to the uncoupled expression of PhANGs and PhAPGs.
Next, to determine whether this SA-mediated uncoupled expression of photosynthesis-associated genes is implicated in the up-regulation of SIB1, we generated transgenic sib1 plants overexpressing the GFP-fused SIB1 under the control of the CaMV 35S promoter (sib1 oxSIB1). Subsequently, the effect of the SIB1 overexpression on the expression of PhANGs and PhAPGs was examined in sib1 oxSIB1 compared to sib1 and wild-type plants grown under CL conditions. Unlike in sib1, the uncoupled expression of PhANGs and PhAPGs was clearly observed in sib1 oxSIB1 plants (Supplemental Figure 6B). Taken together, these results suggest that SA plays an important role in altering the expression of both PhANGs and PhAPGs through SIB1, which acts as a transcriptional coregulator upon its dual targeting to both the nucleus and the Cps.
SIB1 Acts as a Positive Regulator of lsd1 RCD
Because SIB1 is localized to both the nucleus and the Cps; acts as a transcriptional regulator in both cellular compartments; and is up-regulated after exposure to SA, we were interested if it plays a role in mediating lsd1 RCD. To explore a possible genetic interaction between SIB1 and LSD1, lsd1 sib1 and lsd1 sib2 double knock-out mutants were generated. Like SIB1, SIB2 is a member of the VQ motif-containing protein family and functions redundantly with SIB1 in increasing resistance against B. cinerea (Lai et al., 2011). Interestingly, the RCD phenotype was significantly compromised in lsd1 sib1 but not in lsd1 sib2 in comparison to lsd1 (Figure 5A). The results of ion leakage were consistent with the macroscopic phenotypes (Figure 5B), suggesting that SIB1 functions in lsd1 RCD. The rapid up-regulation of SIB1 but not SIB2 in the lsd1 mutant plants before the onset of RCD further supports this notion (Supplemental Figure 7). Unlike in lsd1 cpNahG transgenic plants and lsd1 mutants lacking key SA signaling components, in which RCD was completely abrogated, the lsd1 sib1 mutant seemed to considerably attenuate RCD but not absolutely (Figures 3E and 5A). This finding is in agreement with the slightly increased ion leakage observed at 30 d after seed imbibition. PR1 and PR2 proteins accumulated also in lsd1 sib1 (Figure 5C), but at obviously lower levels than in lsd1.
SIB1-Deficient lsd1 Mutant Significantly Compromises lsd1 RCD and the Uncoupled Expression of PhANGs and PhAPGs.
(A) and (B) The RCD (A) and the related ion leakage (B) were examined at the indicated time points in wild type, lsd1, lsd1 sib1, lsd1 sib2, and lsd1 cpNahG plants grown under CL. WT: wild type.
(C) PR1 and PR2 protein levels were determined in 24-d-old plants. l/s1: lsd1 sib1; l/s2: lsd1 sib2; l/cpN: lsd1 cpNahG. WT: wild type.
(D) Expression levels of immune-responsive genes (IAN7 and WRKY70), PhANGs (LHCB1.1 and LHCB2.2), and PhAPGs (psbA and psbB) were analyzed by RT-qPCR. ACT2 was used as an internal standard. The data represent the means of three independent biological replicates. Error bars indicate sd. Asterisks indicate significant differences between mean values by Student’s t-test (P < 0.05). WT: wild type.
Next, to decipher the role of SIB1 in lsd1-dependent transcriptional reprogramming, we selected six in lsd1 differentially regulated genes and analyzed their relative abundances by RT-qPCR. These genes can be classified into immune-related genes (IAN7 and WRKY70), PhANGs (LHCB1.1 and LHCB2.2), and PhAPGs (psbA and psbB). Indeed, the expression of these genes largely relied on SIB1, as evidenced in that the lack of SIB1 in the lsd1 sib1 double mutant restored the expression levels of these genes to those of wild type (Figure 5D). Notably, we found that the level of free SA in lsd1 sib1 was even slightly higher than that in lsd1 (Supplemental Figure 8), and that inactivation of the key SA signaling components NONEXPRESSER OF PR GENES 1 (NPR1) and EDS1 almost completely suppressed the up-regulation of SIB1 in lsd1 mutant plants (Supplemental Figure 9A), indicating that SIB1 acts as a downstream component of SA. Consistent with this notion, the SA-induced SIB1-mediated up-regulation of immune-related genes and uncoupled expression of PhANGs and PhAPGs observed in lsd1 mutant plants were also suppressed in lsd1 npr1 and lsd1 eds1 mutant plants as well as in lsd1 cpNahG transgenic plants (Figure 5D; Supplemental Figure 9B).
Next, we performed a complementation analysis by expressing SIB1-GFP driven by the native SIB1 promoter in the lsd1 sib1 double mutant. The SIB1-GFP expression fully restored the lsd1 RCD in the lsd1 sib1 mutant (Figure 6A), suggesting that the SIB1-GFP fusion protein is biologically active and that SIB1 is involved in lsd1 RCD. The confocal microscopy (Figure 6B) and immunoblot analyses (Figure 6C) indicated that the accumulation of SIB1-GFP and its dual targeting depend on the rapid accumulation of free SA. The concurrent accumulation of SIB1 and PR1 proteins further showed that SIB1 functions downstream of SA (Figure 6C). Neither GFP fluorescence nor SIB1-GFP protein was detected in sib1 SIB1-GFP transgenic plants (Figures 6B and 6C).
The SIB1-GFP Fusion Protein Restores the RCD in lsd1 sib1.
(A) GFP-fused SIB1 genomic DNA including the promoter region (SIB1-GFP) was used for complementation assays. The images were taken from 26-d-old wild-type, lsd1, lsd1 sib1, lsd1 sib1 SIB1-GFP and sib1 SIB1-GFP plants grown on MS agar medium under CL. WT: wild type.
(B) Subcellular localization of SIB1-GFP in 16-d-old sib1 SIB1-GFP and lsd1 sib1 SIB1-GFP plants. N: nuclear-localized SIB1; Chl.: Chlorophyll.
(C) Immunoblot analysis showing the protein abundances of SIB1-GFP and PR1 in sib1 SIB1-GFP and lsd1 sib1 SIB1-GFP at the indicated time points. Coomassie Brilliant Blue (CBB) staining of the gel was performed after SDS-PAGE, and the amount of protein loaded for each sample is presented. DAI: day after imbibition; Nu: nuclear-localized SIB1.
Although SIB1 appears to play a role in the expression of both nuclear- and plastid-encoded genes (Figure 5D), whether the emergence of RCD requires both forms of SIB1 (Nu and Cp) was unclear. To this end, we deployed an identical strategy (Lai et al., 2011) to verify which form of SIB1 or both is required for lsd1 RCD. We generated lsd1 sib1 transgenic lines expressing modified versions of SIB1 fused with GFP driven by the native SIB1 promoter: These lines included a SIB1ΔPTP lacking the part of the N-terminal plastid transit peptide (14 amino acids, Met-1 to Leu-14) and a SIB1NLS harboring an inactivated nuclear localization signal (NLS) via substitution of all the Lys residues with Ala between Lys-16 and Lys-32 as described in Lai et al. (2011). It is important to note that the CpSIB1 is a dual-targeting protein but not a shuttling protein from the Cps to the nucleus because its NLS (Lys-16 to Lys-32) is located within the plastid transit peptide (Met-1 to Ser-54). Thus, the cleavage of the plastid transit peptide concurrently removes the NLS upon its import into the Cps (Lai et al., 2011). In contrast with the results observed with SIB1-GFP, lsd1 sib1 transgenic plants expressing either SIB1ΔPTP or SIB1NLS failed to restore the RCD phenotype (Supplemental Figure 10A) as well as the expression of immune-related genes (Supplemental Figure 10B). However, SIB1ΔPTP and SIB1NLS largely restored the expression levels of PhANGs and PhAPGs, respectively, alleviating their uncoupled expression (Supplemental Figure 10B). These results indicate that the effect of SIB1 on the transcriptional regulation of photosynthesis-associated genes in both Cps and nucleus is necessary to contribute to the lsd1 RCD.
The positive role of SIB1 in lsd1 RCD was further substantiated by the inactivation of the WRKY33 TF in lsd1. It was previously shown that SIB1 interacts with WRKY33 and stimulates the DNA binding activity of WRKY33 (Lai et al., 2011; Cheng et al., 2012). Therefore, a similar phenotype of lsd1 wrky33 to that of lsd1 sib1 in the context of lsd1 RCD was expected. Indeed, the RCD phenotype was compromised in lsd1 wrky33 (Figure 7A). However, unlike the immune-related genes whose expression were drastically down-regulated in lsd1 wrky33 compared to lsd1, no noticeable difference in the expression of both PhANGs and PhAPGs was found between lsd1 wrky33 and lsd1 (Figure 7B), suggesting that WRKY33 is not involved in the expression of photosynthesis-associated genes and therefore SIB1 may interact with other nuclear TFs involved in the expression of PhANGs. As both proteins are localized in the nucleus (Figure 7C), the physical interaction between SIB1 and WRKY33 was confirmed via bimolecular fluorescence complementation (BiFC) and co-immunoprecipitation (Co-IP) analyses in Nicotiana benthamiana leaves (Figures 7D and 7E), which is consistent with the previous report (Lai et al., 2011).
WRKY33-SIB1 Interaction Contributes to lsd1 RCD But Not for Uncoupled Expression of Photosynthesis-Associated Genes.
(A) Phenotypes of wild-type, lsd1, and lsd1 wrky33 (lsd1/w33) plants grown under CL. Arrows indicate the appearance of RCD. WT: wild type.
(B) Expression levels of immune-related genes (IAN7 and PR1), PhANGs (LHCB1.1 and LHCB2.2), and PhAPGs (psbA and psbB) in CL-grown wild type, lsd1, and lsd1/w33 were analyzed using RT-qPCR. ACT2 was used as an internal standard. The data represent the means of three independent biological replicates and error bars indicate sd. Lowercase letters indicate statistically significant differences between mean values at each of the indicated time points (P < 0.05, one-way analysis of variance with posthoc Tukey’s Honest Significant Difference test). WT: wild type.
(C) Nuclear localizations of GFP-tagged SIB1, WRKY33 (W33), and LSD1 in N. benthamiana leaves. Representative images are shown at the same scale. Chl: Chlorophyll.
(D) and (E) In vivo interaction between SIB1 and W33 by BiFC (D) and Co-IP (E) analyses upon transient coexpression in N. benthamiana leaves. LSD1 was used as a negative control as it does not interact with WRKY33. DAPI was used to stain the nucleus. For the BiFC analysis in (D), results were reproduced in at least two independent experiments using three or more N. benthamiana leaves in each experiment, and representative enlarged images of nucleus are shown at the same scale. Chl: Chlorophyll.
Nuclear SIB1 May Act as a Positive Regulator of GLK1 and GLK2 Activities
Golden2-like (GLK) TFs are found to regulate Cp development in a cell-autonomous manner (Waters et al., 2008). GLK1 and GLK2, functional equivalents, regulate the expression of PhANGs including LHCBs (Waters et al., 2009). Moreover, multiple lines of evidence indicate that GLKs are also involved in immune responses (Savitch et al., 2007; Murmu et al., 2014; Han et al., 2016; Wang et al., 2017; Townsend et al., 2018). In this regard, we hypothesized that, upon their translocation into the nucleus (Figure 8A), SIB1 may stimulate the activity of the GLKs toward LHCBs expression via physical interaction. Indeed, the results of the BiFC and Co-IP analyses confirmed their interactions in N. benthamiana leaves after coexpression (Figures 8B and 8C).
Nuclear SIB1 Interacts with the GLK1 and GLK2 TFs and Enhances GLK1 Transcriptional Activity.
(A) Nuclear localizations of GFP-tagged GLK1 and GLK2 in N. benthamiana leaves. Chl: chlorophyll.
(B) Interactions of SIB1 with GLK1 and GLK2 in the BiFC assays. YFP fluorescence was observed in the nucleus when the C-terminal part of YFP tagged with SIB1 (SIB1-YFPC) was coexpressed with the N-terminal part of the YFP tagged with GLK1 (GLK1-YFPN) or GLK2 (GLK2-YFPN) in N. benthamiana leaves. DAPI was used to stain the nucleus. Results in (A) and (B) were reproduced in at least two independent experiments using three or more N. benthamiana leaves in each experiment, and representative enlarged images of nucleus are shown.
(C) Co-IP of SIB1-GFP with GLK1 or GLK2 fused with MYC tag upon transient coexpression in N. benthamiana leaves.
(D) and (E) Antagonistic effect of PhyB and PIF actions on the lsd1-conferred RCD. Wild-type, lsd1, lsd1 phyB, and lsd1 pif1 pif3 pif4 pif5 (pifq) plants were grown under CL. The RCD phenotype (D) and Fv/Fm (E) were examined at the indicated time points. The representative images are shown at the same scale. Arrows in (D) indicate the appearance of leaf RCD. For the measurement of Fv/Fm, 10 leaves per genotype were used for each measurement. Data represent the means from three independent measurements. Error bars indicate sd. WT: wild type.
(F) The relative transcript levels of GLKs (GLK1 and GLK2), PhANGs (LHCB1.1, LHCB2.2, and LHCB2.3), and immune-related genes (IAN7, WRKY70, PR1, and PR2) in CL-grown wild type, lsd1, lsd1 phyB, and lsd1 pifq at the indicated time points, were analyzed by RT-qPCR. ACT2 was used as an internal standard. The data represent the means of three independent biological replicates and error bars indicate sd. Lowercase letters indicate statistically significant differences between mean values at each of the indicated time points (P < 0.05, one-way analysis of variance with posthoc Tukey’s Honest Significant Difference test). WT: wild type.
(G) ChIP-RT-qPCR assays were performed to examine the effect of SIB1 on GLK1 binding to the promoter regions of GLK1 target genes (LHCB1.4, LHCB6, and HEMA1). Arabidopsis leaf protoplasts were isolated from sib1 (− SIB) and sib1 oxSIB1 (+ SIB1) plants to overexpress GLK1-4xMYC. ChIP assay was performed using MYC antibody. The fold enrichment was calculated as described in Methods. Data represent the means from two independent ChIP assays, and error bars indicate sd. Asterisks denote statistically significant differences by Student’s t test (*P < 0.01; **P < 0.05) from mean value of sib1 (− SIB1).
The concurrent inactivation of both GLK1 and GLK2 was found to arrest Cp development (Waters et al., 2008), which may indirectly affect the lsd1 RCD. We therefore decided to inactivate upstream regulators of GLKs, such as the photoreceptor phytochrome B (PhyB) and PHYTOCHROME-INTERACTING FACTOR (PIF)-class bHLH TFs in the lsd1 mutant background. While the expression of GLK1 is repressed in the dark by PIFs, light-activated Phys lead to the derepression of GLK1 expression by directing PIF degradation (Martín et al., 2016). Interestingly, stressed Cps under excess light conditions appear to inhibit this light-driven derepression of GLK1 expression via the GUN1-mediated RS pathway, resulting in the repression of GLK1 expression in the presence of light (Martín et al., 2016).
This antagonistic action of GUN1-mediated RS to the expression of GLK1 may provide an adaptable system toward fluctuating light: Reduced levels of photosystem apparatus due to the repression of GLK1 may reduce the levels of ROS under light stress. This assumption is in agreement with a recent discovery that plants inhibit the light-dependent photosynthetic activity under photo-oxidative stress conditions to reduce the levels of ROS produced by the photosystems (Ling and Jarvis, 2015). This finding also suggests that, as environmental sensors, Cps can override the development-related intracellular signaling network by activating stress-related RS pathways. Nevertheless, based on the proposed positive and negative role of Phys and PIFs on GLKs expression, respectively, we investigated whether inactivation of such regulators can alter lsd1 RCD. When a quartet of PIFs (PIFQ; PIF1, 3, 4, and 5), negative regulators of GLK1 expression, was genetically inactivated, the lsd1 RCD was drastically enhanced, while the mutation of PhyB, a positive regulator of GLK1 expression, led to the attenuation of the lsd1 RCD (Figures 8D and 8E). The transcript abundances of selected immune-related genes (IAN7, WRKY70, PR1, and PR2) coincided with the RCD phenotypes (Figure 8F). The antagonistic effect of the PhyB and PIFs actions on the expression of GLK1 and GLK2 was confirmed by RT-qPCR (Figure 8F). The expression of LHCBs including LHCB1.1, LHCB2.2, and LHCB2.3 depended greatly on the transcript abundances of the GLKs in lsd1 phyB and lsd1 pifq (Figure 8F). These results strongly suggest that upon nuclear localization in response to the increased levels of SA, SIB1 may act as a positive transcriptional coregulator for GLK1 and GLK2, potentiating the expression of PhANGs. Indeed, chromatin immunoprecipitation (ChIP)-RT-qPCR assays using Arabidopsis leaf protoplasts confirmed that SIB1 positively affects the binding activity of GLK1 to the promoters of its known target genes (Waters et al., 2009) such as PhANGs (LHCB1.4 and LHCB6) and a chlorophyll synthesis gene (HEMA1; Figure 8G).
Uncoupled Expression of Photosynthesis-Associated Genes Disrupts Cp ROS Homeostasis
We hypothesized that the uncoupled expression of photosynthesis-associated genes, particularly the up-regulation of genes encoding the PSII LHCBs (encoded by PhANGs) and the down-regulation of genes encoding the PSII core proteins D1, D2, and CP43 (encoded by PhAPGs; Figure 2A), would disrupt the stoichiometry of LHCBs to PSII core proteins. This disturbance of the stoichiometry would result in that the absorbed light energy surpasses the photochemical efficiency of PSII, thereby inducing photoinhibition in PSII. If this assumption is correct, then the lsd1 mutant may generate 1O2 by the photoinhibited PSII via energy transfer from the chlorophylls to molecular oxygen in a light-dependent manner (Krieger-Liszkay et al., 2008; Kim and Apel, 2013). As a reminder, Cp-generated 1O2 triggers RS (op den Camp et al., 2003; Wagner et al., 2004; Kim and Apel, 2013) and 1O2-responsive nuclear genes have been identified (op den Camp et al., 2003; Lee et al., 2007; Kim and Apel, 2013; Dogra et al., 2017).
If the uncoupled expression of photosynthesis-associated genes increases the levels of 1O2 due to the disruption of PSII stoichiometry (Figure 2B), 1O2-responsive genes (SORGs) could be up-regulated in lsd1, along with the SA-responsive genes. Therefore, we compared the 624 genes up-regulated in the lsd1 mutant plants (Supplemental Data Set 3) with the SORGs (168 genes) that were identified to be up-regulated in the flu mutant plants through the EX1-mediated RS (Dogra et al., 2017). The comparative analysis indicated that out of the 168 SORGs, 34 and 47 were significantly up-regulated relative to wild type in the 17-and 19-d-old lsd1 plants, respectively (Figures 9A and 9B; Supplemental Data Set 9). This result suggests that the 1O2 level may increase in the Cps of 17-d-old lsd1 mutant plants, or even earlier. Indeed, we measured substantial levels of 1O2 in the Cps of 17-d-old lsd1 plants as evidenced by the strong fluorescence of the Singlet Oxygen Sensor Green (SOSG; Figures 9C and 9D). SOSG is a highly selective detection reagent for 1O2 (Flors et al., 2006).
Accumulation of 1O2 Results in the Activation of EX1-Mediated Cell Death in lsd1.
(A) Venn diagram showing the uncommon and shared genes between genes up-regulated in lsd1 and SORGs.
(B) Heatmap showing the differential expression of those 52 shared genes. The colors of the heatmap represent the z-scores ranging from green (z-score of −1.5) through black to red (z-score of 1.5). SIB1 is indicated by an arrow. WT: wild type.
(C) Representative confocal images of SOSG fluorescence in leaf mesophyll cells of 17-d-old wild-type and lsd1 plants grown under CL. Images were taken using the same microscope settings. Chl.: chlorophyll; WT: wild type.
(D) The fluorescence intensity of SOSG as shown in (C) was calculated using the ImageJ software and normalized to Cp size. Ten Cps per genotype were used for each measurement. Data are from three independent measurements. Error bars indicate sd. Asterisk indicates significant differences between mean values by Student’s t-test (P < 0.01). RFI: relative fluorescence intensity; n.d.: non-detectable; WT: wild type.
(E) and (F) The relative transcript levels of SIB1 (E) and ICS1 (F) in 16-d-old wild-type, lsd1, lsd1 ex1 (l/ex1), and lsd1 cpNahG (l/cpN) plants were determined by RT-qPCR. ACT2 was used as an internal standard. Data represent the means from three independent biological replicates. Lowercase letters indicate statistically significant differences between mean values at each of the indicated time points (P < 0.05, one-way analysis of variance with posthoc Tukey’s Honest Significant Difference test). WT: wild type.
(G) Phenotype was collected from 22- and 24-d-old wild-type, lsd1, and ex1 lsd1 plants grown on MS medium under CL. WT: wild type.
(H) and (I) Fv/Fm (H) and ion leakage (I) were analyzed in wild-type, lsd1, and lsd1 ex1 plants at the time indicated. Ten leaves per genotype were used for each measurement. Data are from three independent measurements. Error bars indicate sd. WT: wild type.
To ensure that the 1O2 production is caused by the uncoupled expression of PhANGs and PhAPGs, we also examined the 1O2 production in CL-grown sib1 oxSIB1 plants because SIB1 overexpression led to the uncoupled expression of PhANGs and PhAPGs (Supplemental Figure 6B). The fluorescence of SOSG was strongly enhanced in sib1 oxSIB1 plants relative to wild-type or sib1 plants (Supplemental Figure 11A). Consistent with the SOSG result, the selected SORGs were also significantly up-regulated in sib1 oxSIB1 plants (Supplemental Figure 11B). These results strongly suggest that the SIB1-mediated uncoupled expression of PhANGs and PhAPGs potentiates the 1O2 production. Interestingly, although no visible cell death was detected in sib1 oxSIB1 plants by TB staining (data not shown), the sib1 oxSIB1 plants displayed a pale green leaf phenotype (Supplemental Figure 11C), implying impaired Cp function. The Fv/Fm levels were indeed decreased in sib1 oxSIB1 plants as compared to wild-type and sib1 plants (Supplemental Figure 11C).
The nuclear-encoded Cp EX1 protein has been implicated in mediating 1O2-triggered chloroplast-to-nucleus RS that primes PCD in the Arabidopsis flu mutant plants (Wagner et al., 2004; Lee et al., 2007; Kim et al., 2012). Considering the increased levels of 1O2 in lsd1 before the emergence of cell death, a possible role of EX1 in contributing to lsd1 RCD was examined using lsd1 ex1 double mutant plants grown under CL conditions. Unlike in lsd1 cpNahG in which the SIB1 and ICS1 genes are fully repressed, the levels of these transcripts were comparable between lsd1 and lsd1 ex1 (Figures 9E and 9F), implying that SA signaling remained active in the lsd1 ex1 double mutant plants. Despite the active SA signaling, the onset of RCD was found to be considerably impaired in lsd1 ex1 compared to lsd1. The strong RCD phenotype was observed in 22-d-old lsd1 mutant plants with a concurrent decline of the Fv/Fm (Figure 9H). An increased ion leakage due to the cell death was apparent in 20-d-old lsd1 mutant (Figure 9G). By contrast, no clear phenotypic changes were observed in the double mutant lsd1 ex1 until 22 d, but gradually decreasing Fv/Fm and slightly increased ion leakage were measured at 24 d (Figures 9H and 9I). Taken together, these results suggest that the rapid increase of SA and the subsequent SIB1-related uncoupled expression of photosynthesis-associated genes is likely to activate 1O2-triggered EX1-dependent RS that may contribute to the lsd1 RCD.
DISCUSSION
It has been shown that lsd1 mutant plants display an RCD phenotype implicated in the incapability to restrict the spread of systemic cell death after a local cell death initiated by various environmental stimuli such as a shift in growing condition from SD to LD, high light, cold, UV-C irradiation, red light, hypoxia, and invasion of pathogens. (Dietrich et al., 1994; Jabs et al., 1996; Torres et al., 2005; Mühlenbock et al., 2007, 2008; Huang et al., 2010; Karpiński et al., 2013; Chai et al., 2015; Rusaczonek et al., 2015). This lsd1 RCD is largely dependent on EDS1 and PAD4, which positively regulate the production of SA and ethylene, and a key regulator of SA-mediated systemic acquired resistance, NPR1 (Rustérucci et al., 2001; Aviv et al., 2002; Mühlenbock et al., 2008; Roberts et al., 2013). Therefore, LSD1 is considered as a pivotal regulator of cell death and basal defense response toward biotic and abiotic stresses. Here, we further demonstrated that daylength, increased SA content and its cognate signaling pathways, uncoupled expression of PhANGs and PhAPGs, and Cp-generated ROS chronically direct the RCD phenotype in the lsd1 mutant plants grown under nonstressful growth conditions (Figure 1). Given that SA plays a vital role in priming lsd1 RCD in the absence of external stimuli (Figure 3E) and that daylength alters the timing of the lsd1 RCD (Figures 1A and 1B), it is reasonable to assume that the daylength may also determine the timing of SA accumulation in lsd1. In fact, Zheng et al. (2015) had recently shown that the basal SA level in Arabidopsis plants is regulated by a core circadian oscillator, which is evidenced by the observation that the expression of ICS1 appears to show circadian oscillation. Moreover, considering that the circadian period is shortened with leaf age, which is further accelerated by extended daylength (Kim et al., 2016), the combination of foliar aging and daylength may lead to the gradual accumulation of SA wherein LSD1 may antagonistically regulate the process to evade inappropriately induced immune responses such as cell death.
The transcriptional coregulator SIB1 gene was rapidly up-regulated in response to the elevated levels of SA and then posttranslationally targeted to both the nucleus and Cps (Figure 4; Supplemental Figure 6A). This SA-mediated dual targeting of SIB1 to both subcellular compartments leads to a concomitant up-regulation of PhANGs and down-regulation of PhAPGs (Supplemental Figures 6A and 6B). Given that the lsd1 sib1 double mutant largely restores the transcript levels of PhANGs and PhAPGs to nearly wild-type levels and significantly compromises lsd1 RCD (Figure 5D), it is conceivable that the uncoupled expression of these photosynthesis-associated genes may contribute to lsd1 RCD. The positive role of SIB1 toward lsd1 RCD was further substantiated through the inactivation of the WRKY33 TF that was previously found to interact with SIB1 in the nucleus. The SIB1-WRKY33 interaction is part of the plant defense response toward the necrotrophic fungal pathogen B. cinerea (Lai et al., 2011). As a transcriptional coregulator, SIB1 stimulates the DNA binding activity of WRKY33, potentiating the expression of its downstream target genes involved in immune responses (Xie et al., 2010; Lai et al., 2011). However, the transcript abundance of the examined LHCBs remained unchanged in lsd1 wrky33 relative to lsd1 (Figure 7B), indicating that SIB1 may regulate other TFs functioning in the expression of PhANGs. In agreement with this hypothesis, our analyses demonstrated that SIB1 interacts with both TFs GLK1 and GLK2 (Figures 8B and 8C), whose function has been implicated in the positive regulation of the expression of PhANGs as well as the Cp biogenesis in a cell-autonomous manner (Waters et al., 2008, 2009). SIB1 stimulates GLK1 binding to the promoter regions of its target genes including LHCBs (Figure 8G). Because GLK1 and GLK2 function redundantly and the glk1 glk2 double mutant demonstrated a significantly impaired Cp biogenesis with reduced chlorophyll content (Fitter et al., 2002; Waters et al., 2008), the effect of the inactivation of upstream regulators of GLK1 and GLK2 on the expression of LHCBs and lsd1 RCD was examined. GLK1 and GLK2 are regulated via the light-dependent function of the phytochromes and their interacting factors (PIFs). Consistent with a previous report regarding the functions of Phys and PIF TFs in the context of the expression of PhANGs (Martín et al., 2016), our results indicate that PIFs repress the expression of GLKs in lsd1 whereas PhyB induces their expression (Figure 8F). Moreover, the mutations in PhyB and the PIFs significantly affect the lsd1 RCD (Figures 8D and 8E). These results somewhat support our notion that the uncoupled expression of PhANGs to PhAPGs contributes to lsd1 RCD. However, these results cannot provide an answer regarding the biological relevance of the SIB1-GLK1/2 interaction upon release of SA. Therefore, this requires further investigation. In addition to this question, despite that the genetic and molecular evidences strongly suggest that SIB1 may act as a positive transcriptional coregulator for the GLKs, another question is how SIB1 interacts with these functionally distinct TFs, i.e. WRKY33 and GLKs.
It has to be noted that loss of SIB1 function in the lsd1 sib1 double mutant significantly attenuated but did not completely abolish the RCD (Figures 5A and 5B). Because the SIB1 expression depends on NPR1 (Xie et al., 2010), a bona-fide SA receptor (Wu et al., 2012), and because no RCD phenotype was detectable in the lsd1 npr1 mutant (Figures 3E and 3F), it is reasonable to assume that NPR1 might activate SIB1-independent signaling pathway(s), which probably contribute to the lsd1 RCD (Figure 10). Indeed, we found that an additional VQ gene, VQ10, was highly up-regulated in the lsd1 mutant (Supplemental Data Set 3). Like SIB1, VQ10 belongs to a group of SA-responsive genes (Supplemental Data Set 8) and interacts with WRKY33, WRKY25, and WRKY26 (Cheng et al., 2012). Given that several WRKYs were up-regulated in the lsd1 mutant before the onset of RCD (Supplemental Data Set 3), SIB1 and VQ10 may coordinately or independently regulate the DNA binding activity of those WRKYs. Therefore, the role of VQ10 in lsd1 RCD still needs to be addressed.
Simplified Model Elucidating the Pathways Involved in lsd1 RCD.
(A) In the lsd1 mutant, the daylength-dependent cell death is linked to the rapid accumulation of SA, synthesized via the Cp-established ICS pathway. SA-mediated NPR1 activation leads to the expression of SIB1.
(B) Upon translation, SIB1 proteins are targeted to both the nucleus (nuclear SIB1) and Cps (CpSIB1), where they may act as transcriptional coregulators for targeted photosynthesis-associated genes, in the up- and down-regulation of PhANGs and PhAPGs, respectively. CpSIB1 is known to reduce a subset of PhAPGs via the interaction with SIG1 (plastid RNA polymerase; Xie et al., 2010) and nuclear SIB1 may activate GLK TFs to express PhANGs. Besides, WRKY33-SIB1 interaction seems to mainly involve the expression of immune-related genes. NuSIB1: nuclear SIB1.
(C) The uncoupled expression of PhAPGs and PhANGs may disrupt the stoichiometry of light-harvesting antenna complex to PSII core proteins. This perturbation in PSII homeostasis eventually leads to the generation of 1O2 by PSII (Figures 9A to 9D; Supplemental Figures 11A and 11B). Because 1O2 and other ROS affect the redox status of the PQ pool in the PETC (Karpinska et al., 2000; Krieger-Liszkay et al., 2008; Kruk and Szymańska, 2012), the disrupted PSII stoichiometry may also instigate the redox signals, which contribute to cell death (Mühlenbock et al., 2008). The nuclear-encoded Cp protein EX1, a putative 1O2 sensor, senses the increased levels of 1O2 and mediates a 1O2-triggered genetically controlled cell death program (Lee et al., 2007; Kim et al., 2012) via retrograde signaling. NPR1-dependent but SIB1-independent transcriptional reprogramming might also participate in the regulation of cell death in lsd1.
Remarkably, the majority of the PhANGs encoding LHCBs of PSII (LHCBs) were up-regulated in the lsd1 mutant whereas the PhAPGs encoding PSII core proteins, such as D1, D2, and CP43, were down-regulated (Figure 2A). This impaired stoichiometry of LHCBs to PSII core proteins (Figure 2B) may augment the photoinhibition of PSII that usually occurs under excess light conditions, leading to the generation of 1O2 (Apel and Hirt, 2004). Consistent with this notion, a mutation in the CHLOROPHYLL A/B BINDING PROTEIN ORGANELLE SPECIFIC gene, also known as the CHLOROPLAST SIGNAL RECOGNITION PARTICLE43, which results in a lessening of light harvesting capacity in PSII due to the impaired assembly of LHCBs into PSII, also compromises the excess light-induced RCD in lsd1 mutant by an increase in nonphotochemical quenching (Mateo et al., 2004). The attenuated RCD phenotype by loss of EX1, a putative 1O2 sensor (Figure 9G), further strengthens the assumption that the uncoupled expression of LHCBs and PSII core proteins result in the activation of 1O2-triggered and EX1-dependent RS, which was shown to elicit PCD in the Arabidopsis flu mutant (Wagner et al., 2004; Lee et al., 2007; Kim et al., 2012). However, given that RCD was considerably but not completely attenuated in the lsd1 ex1 mutant (Figures 9G to 9I), the uncoupled expression may generate not only 1O2 but also other Cp-derived signal(s) involved in RCD. In agreement with this, the Cp redox changes, i.e. excess-excitation-energy– or red-light–induced hyper-reduction of the PQ pool are reportedly involved in the induction of RCD in the lsd1 mutant through the ROS-, SA-, and ethylene-mediated multiple signaling pathways under the control of EDS1 and PAD4 (Mühlenbock et al., 2008; Chai et al., 2015). Moreover, because SA signaling is still active in lsd1 ex1 (Figures 9E and 9F), it seems that EX1-independent RS pathways contribute to the lsd1 RCD.
Inactivation of a crucial SA biosynthesis enzyme ICS1, which catalyzes the conversion of chorismate to isochorismate, also compromises the lsd1 RCD (Li et al., 2013). Isochorismate is a precursor of SA as well as of phylloquinone (Vitamin K1), which functions as an electron carrier from PSI to the iron-sulfur cluster (Sigfridsson et al., 1995) and plays an important role in the regulation of state transitions (Gawroński et al., 2013). As a positive factor of lsd1 RCD, the hyper-reduction of the PQ pool can also trigger a signal that leads to the state transitions through phosphorylation–dephosphorylation of photosystem and light harvesting complex proteins, resulting in the alteration of gene expression both in the Cps and the nucleus (Rochaix, 2013). Therefore, not only isochorismate-derived SA but also phylloquinone may participate in contributing to the lsd1 RCD.
Consistent with our model (Figure 10), SIB1 overexpression results in the uncoupled expression of PhANGs and PhAPGs along with the enhanced production of 1O2 even in the wild-type background (Supplemental Figures 6B, 11A, and 11B). Although SIB1 overexpression results in an impaired Cp function in wild type (Supplemental Figure 11C), which is likely due to the 1O2 production, no RCD phenotype was observed. This might be due to the presence of LSD1 that may counteract SIB1-mediated stress responses to repress cell death. Indeed, it has been reported that LSD1 suppresses ROS burst by directly or indirectly activating ROS scavenger enzymes in the Cps, the cytoplasm as well as the extracellular space (Jabs et al., 1996; Kliebenstein et al., 1999; Mateo et al., 2004; Mühlenbock et al., 2008; Li et al., 2013). LSD1 also inhibits the EDS1- and PAD4-dependent production of SA and ethylene that lead to cell death (Mühlenbock et al., 2008). Moreover, a recent interactome analysis demonstrated that nucleocytoplasmic LSD1 protein forms complexes with other proteins involved in multiple molecular pathways (Czarnocka et al., 2017). This LSD1 interactome is largely dependent on the cellular redox status, as shown its dynamic changes in response to oxidative stress (Czarnocka et al., 2017). LSD1 may negatively regulate a set of proteins involved in cell death through direct interaction. In fact, previous studies demonstrated that LSD1 interacts with a BASIC LEUCINE ZIPPER10 TF and a METACASPASE1, which act as positive regulators of cell death and basal defense response in lsd1 mutant (Kaminaka et al., 2006; Coll et al., 2010). In addition, because LSD1 has been reported to function as a transcriptional regulator (Kaminaka et al., 2006; Czarnocka et al., 2017), LSD1-dependent transcriptional reprogramming might also participate in the suppression of cell death in response to stress.
In this study, we provided a hierarchical pathway from SA-primed immune response to 1O2-triggered RS, contributing to the lesion-mimic phenotype in lsd1. The physical interactions between the transcriptional coregulator SIB1 (downstream component of SA) and the GLK TFs are likely to enhance the expression of PhANGs, in contrast with the down-regulated PhAPGs, which consequently leads to the increased levels of 1O2 in the Cps, presumably through alteration of the stoichiometry of in the PSII apparatus. The putative 1O2 sensor EX1 subsequently mediates 1O2 signaling, which was previously shown to prime the genetically controlled cell death. This finding needs to be further studied to elucidate the underlying mechanism governing up- and down-regulation of the PhANGs and PhAPGs in the lsd1 mutant, and to determine whether such molecular repertoire plays a role in mediating cell death in response to various natural stresses.
METHODS
Plant Materials and Growth Conditions
All Arabidopsis (Arabidopsis thaliana) seeds used in this study were derived from Columbia-0 ecotype and were harvested from plants grown under CL condition (100 µmol·m−2·s−1 of light from Cool White Fluorescent bulbs) at 22 ± 2°C. Arabidopsis mutant seeds of lsd1-2 (SALK_042687), npr1 (SALK_204100), pad4 (SALK_206548), sib1-4 (SM_3.30596), sib2-1 (SM_3.16236), ex1 (SALK_002088), wrky33-1 (SALK_006603), phyB-9, and pifq (pif1-1 pif3-7 pif4-2 pif5-3) were obtained from the Nottingham Arabidopsis Stock Centre. eds1-2 were reported in Bartsch et al. (2006). The transgenic cpNahG line overexpressing the GFP-tagged Cp-localized bacterial SA hydrolase under the control of CaMV 35S promoter was described in Fragnière et al. (2011). All double and pifq lsd1 quintuple mutants as well as cpNahG lsd1 were created by crossing the homozygous plants. The homozygous cpNahG was selected based on Basta resistance and the genotypes of all mutants confirmed by PCR-based analysis. Primer sequences for PCR are listed in Supplemental Table 2.
Seeds were surface-sterilized by soaking in 1.6% hypochlorite solution for 10 min, followed by washing five times with sterile water. Seeds were then plated on MS medium (Duchefa Biochemie) containing 0.65% (w/v) agar (Duchefa Biochemie). After a 3-d stratification at 4°C in darkness, seeds were placed in a growth chamber (CU-41L4; Percival Scientific) under CL, 16-h light/8-h dark (LD), or 8-h light/16-h dark (SD) conditions. The light intensity was maintained at 100 μmol·m−2·s−1 at 22°C ± 2°C. For SA treatment, 5-d-old seedlings grown under CL were transferred to MS medium containing 1.0 mM SA (Sigma-Aldrich).
Vector Construction and Generation of Transgenic Plants
The stop-codon–less genomic SIB1 DNA containing the 1.8-kb promoter region and the stop-codon–less SIB1 coding sequence (CDS) were cloned into a pDONR221 Gateway vector (Thermo Fisher Scientific) through the Gateway BP reaction (Thermo Fisher Scientific) and subsequently recombined into the Gateway-compatible plant binary vectors pGWB504 and pGWB505, respectively (Nakagawa et al., 2007) for C-terminal fusion with sGFP through the Gateway LR reaction (Thermo Fisher Scientific). The same procedure was performed for constructing the different versions of SIB1 (SIB1ΔPTP and SIB1NLS) with their CDSs being modified according to Lai et al. (2011). The generated vectors were transformed by electroporation into the Agrobacterium tumefaciens strain GV3101. Arabidopsis transgenic plants were generated using Agrobacterium-mediated transformation using the floral dip procedure (Clough and Bent, 1998), and homozygous transgenic plants were selected on MS medium containing 50 mg/L Hygromycin (Thermo Fisher Scientific).
RNA Extraction and RT-qPCR
Total RNA was extracted using the Spectrum Plant Total RNA Kit (Sigma-Aldrich) and spectrophotometrically quantified at 260 nm with the NanoDrop 2000 (Thermo Fisher Scientific). Total RNA (1 μg) was reverse-transcribed using the PrimeScript RT Reagent Kit (Takara) according to the manufacturer’s recommendations. The RT-qPCR was performed with iTaq Universal SYBR Green PCR master mix (Bio-Rad) on a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems). Relative transcript level was calculated by the ddCt method (Livak and Schmittgen, 2001) and normalized to the ACTIN2 (At3g18780) or UBQ10 (At4g05320) gene transcript levels. The sequences of the primers used in this study are listed in Supplemental Table 2.
Protein Extraction and Immunoblot Analysis
Total protein was extracted from 100 mg of plant tissue with an extraction buffer (20 mM HEPES [pH7.4], 2 mM EDTA, 2 mM EGTA, 25 mM NaF, 1 mM Na3VO4, 50 mM Glycerophophate, 10% [v/v] Glycerol, 100 mM sodium chloride (NaCl) and 0.5% [v/v] Triton X-100) containing a protease inhibitor tablet (Sigma-Aldrich) and then quantified using a Pierce BCA protein assay kit (Thermo Fisher Scientific). The extracted total protein was then separated on 10% SDS-PAGE gels and blotted onto Immun-Blot PVDF membrane (Bio-Rad). PR1, PR2, LHCB1, LHCB2, LHCB3, LHCB4, D1, D2, CP43, and AtpF proteins were immunochemically detected using rabbit anti-PR1 (1:5,000; Agrisera) and rabbit anti-PR2 (1:5,000; Agrisera), rabbit anti-LHCB1 (1:4,000; Agrisera), rabbit anti-LHCB2 (1:7,000; Agrisera), rabbit anti-LHCB3 (1:4,000; Agrisera), rabbit anti-LHCB4 (1:7,000; Agrisera), rabbit anti-D1 (1:10,000; Agrisera), rabbit anti-D2 (1:10,000; Agrisera), rabbit anti-CP43 (1:5,000; Agrisera), and rabbit anti-AtpF (1:5,000; Agrisera) antibodies, respectively. SIB1-GFP fusion protein was detected using a mouse anti-GFP monoclonal antibody (1:5,000; Roche). The RUBISCO large subunit (RBCL) and UDP-glucose pyrophosphorylase (UGPase) proteins, detected with Rabbit anti-RBCL (1:10,000; Agrisera) and rabbit anti-UGPase (1:3,000; Agrisera) antibodies, were used as loading controls.
Subcellular Localization and Confocal Laser-Scanning Microscopy
For subcellular localization of SIB1, WRKY33, GLK1, GLK2, and LSD1 proteins, the pDONR/Zeo entry vectors (Thermo Fisher Scientific) containing their CDSs were recombined into the destination vectors pGWB505 for C-terminal fusion with sGFP through the Gateway LR reaction (Thermo Fisher Scientific). These constructs were then transformed into A. tumefaciens strain GV3101 and transiently expressed in Nicotiana benthamiana leaves. The GFP, chlorophyll, and 4′, 6-diamidino-2-phenylindole (DAPI) fluorescence signals were detected by confocal laser-scanning microscopy analysis using a TCS SP8 (Leica Microsystems) 30 h after infiltration. All the images were acquired and processed using Leica LAS AF Lite software, version 2.6.3 (Leica Microsystems).
BiFC Assay
BiFC assays were conducted using a split-yellow fluorescent protein (YFP) system in N. benthamiana leaves as described in Lu et al. (2010). Briefly, the pDONR/Zeo entry vectors containing the CDS of SIB1, WRKY33, GLK1, GLK2, and LSD1 were recombined into the split-YFP vectors (pGTQL1221 or 1211) through Gateway LR reaction (Thermo Fisher Scientific). For the assay, A. tumefaciens mixtures carrying the appropriate BiFC constructs were infiltrated with a 1-mL syringe without a needle into the abaxial side of 4-week-old N. benthamiana leaves. After 48 h of infection, the absence or presence of YFP signal was imaged using a TCS SP8 (Leica Microsystems).
Co-IP Assay
For Co-IP assays, the pDONR/Zeo entry vector containing the stop-codon–less CDSs of SIB1, WRKY33, GLK1, GLK2, and LSD1 were recombined into the destination vector pGWB505 for C-terminal fusion with sGFP, pGWB620 for C-terminal fusion with 10xMyc tag, or pGWB617 for C-terminal fusion with 4xMYC through Gateway LR reaction (Thermo Fisher Scientific) to create p35S:SIB1-sGFP, p35S:LSD1-sGFP, p35S:WRKY33-10xMYC, p35S:GLK1-4xMYC, and p35S:GLK2-4xMYC constructs. For the p35S:GLK1-4xMYC and p35S:GLK2-4xMYC constructs, a flexible linker DNA encoding Gly-Gly-Ser-Gly-Gly-Ser was added between MYC and GLK1 or GLK2 CDSs to increase conformational flexibility of the fusion proteins as described in Tokumaru et al. (2017). The different combinations of selected vectors were coexpressed in 4-week-old N. benthamiana leaves after Agrobacterium infection. Total protein was isolated using an IP buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 5 mM EDTA, 10% glycerol, 1% Nonidet P-40, 1% deoxycholate, 0.1% SDS, and 1× cOmplete protease inhibitor cocktail [Roche]). After protein extraction, 15 μL of GFP-Trap magnetic agarose beads (GFP-TrapMA, Chromotek) was added to 20 mg of the total protein extract, and the mixture incubated for 2 h at 4°C by vertical rotation. The beads were washed five times with the washing buffer (10 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, and 1× cOmplete protease inhibitor cocktail) and then eluted with 2× SDS protein sample buffer for 20 min at 70°C. The eluates were loaded into 10% SDS-PAGE gels and the interaction between coexpressed proteins determined by immunoblot analyses using a mouse anti-GFP monoclonal antibody (1:5,000; Roche) and a mouse anti-MYC monoclonal antibody (1:10,000; Cell Signaling Technology), respectively.
ChIP-RT-qPCR Assays
The ChIP assays using Arabidopsis leaf protoplasts were performed as described in Yoo et al. (2007) and Lee et al. (2017) with minor modifications. Briefly, 1 mg of pSAT6 vector (Tzfira et al., 2005) containing p35S:GLK1-4xMYC DNA was transfected into leaf protoplasts isolated from 3-week-old sib1 and sib1 oxSIB1 plants grown on soil under LD conditions using the PEG-mediated transfection method. After incubating the protoplasts at room temperature for 16 h under dim light condition, the protoplast chromatins were crosslinked by 1% formaldehyde in 1× PBS buffer (pH 7.4) for 10 min and quenched with 0.1 M Glc for 5 min. Subsequently, the protoplasts were lysed, and the chromatins were sheared by sonication into a major size of ∼500 bp. The lysates were precleared by incubation with 50 μL Protein-A agarose beads/Salmon sperm DNA (Millipore) for 1 h at 4°C and then incubated with anti-MYC monoclonal antibody (Cell Signaling Technology) at 4°C overnight. In parallel, ChIP assays were performed without antibody to determine nonspecific binding. The beads were washed according to Lee et al. (2017). After eluting the immunocomplexes by elution buffer (1% [w/v] SDS and 0.1 M NaHCO3), the bound DNA fragments were recovered and purified according to Lee et al. (2017). RT-qPCR was performed on bound and input DNAs. The sequences of primers for each gene are listed in Supplemental Table 2. The amount of DNA precipitated by anti-MYC antibody was calculated in comparison with the respective input DNA used for each ChIP. Then, the fold enrichment was calculated by normalizing against the corresponding control sample (without antibody). ACT7 (At5g09810) was used as a negative control.
SA Measurements
The endogenous SA levels in plant samples were measured by ultra performance liquid chromatography-tandem mass spectrometer. Briefly, the plant tissue was homogenized to a fine powder in liquid nitrogen using a mortar and pestle. Approximately 25 mg of the fine powder was mixed with 2 ng of d4-SA (internal standard; Sigma-Aldrich) and 500 μL extraction solvent (69.9% [v/v] methanol and 0.1% [v/v] formic acid). After 1 h of shaking (1,200 rpm) at 4°C, the homogenate was centrifuged at 18,000 g for 15 min at 4°C. The supernatant (300 μL) was transferred to a fresh high-performance liquid chromatography vial (Waters) containing 300 μL H2O and vortexed for a few seconds before injection. Chromatographic separations were conducted on a BEH C18 column (2.1 mm × 150 mm, 1.7 μm particle diameter; Waters) at 45°C using an ACQUITY UPLC I-class system (Waters) equipped with an ACQUITY Sample Manager (Waters) and an ACQUITY Binary Solvent Manager (Waters). Formic acid (0.1%, v/v) and acetonitrile with 0.1% (v/v) formic acid were used as mobile phases A and B, respectively. The elution profile after injection of each sample was: 0–5 min, 20% to 60% B; 5–9 min, 90% B; 9–14 min, 20% B. The flow rate of mobile phase was 300 μL/min and the injection volume 50 µL. The eluate was monitored by an online mass spectrometry using TripleTOF 5600+ (AB Sciex) set to high-resolution multiple reaction monitoring in negative electrospray mode. SA concentrations were determined by comparing the peak area of the product ion of 93.04 dissociated from precursor ion of 137.0244 (SA) with that of product ion of 97.03 dissociated from precursor ion of 141.0484 (d4-SA).
Determinations of Photochemical Efficiency and Maximum Chlorophyll Fluorescence
The maximum photochemical efficiency of PSII (Fv/Fm) determined with a FluorCam system (FC800-C/1010GFP; Photon Systems Instruments) containing a charge-coupled device camera and an irradiation system according to the instrument manufacturer’s instructions.
Determination of Cell Death
Plant tissues were submerged in TB staining solution (10 g phenol, 10 mL glycerol, 10 mL lactic acid, 0.02 g TB, and 10 mL H2O) diluted with ethanol 1:2 (v/v) and boiled for 2 min. After 16 h of incubation at room temperature, nonspecific staining was removed with destaining solution (250 g chloral hydrate and 100 mL H2O). Plant tissues were then stored in 50% (v/v) glycerol for taking images. To determine electrolyte leakage, first or second leaves were harvested at the indicated time points and transferred to a 15-mL tube containing 6-mL deionized water. After 6 h of incubation at room temperature, conductivity of the solution was measured with an Orion Star A212 conductivity meter (Thermo Fisher Scientific). For each measurement, six leaves per genotype were used, and the experiment was repeated three times.
Imaging SOSG Fluorescence
To detect the accumulation of 1O2 in leaf mesophyll cells, the first pair of leaves from each genotype grown under CL condition was immersed in a solution of 260 μM SOSG (Thermo Fisher Scientific, Molecular Probes) in 50 mM phosphate buffer (pH 7.4). Leaves were vacuum-infiltrated for 2 min and then imaged using a TCS SP8 (Leica Microsystems). 1O2-activated SOSG was visualized with an excitation wavelength of 488 nm and an emission wavelength of 530 nm. At least 10 leaves from each genotype were monitored and representative images were shown.
RNA-Seq Library Construction and Data Analysis
Three biological replicates of 17- and 19-d-old wild type and lsd1 grown under a CL condition (100 μmol·m−2·s−1) were used for RNA extraction. Total RNA extracted using the RNeasy Plant Mini Kit (Qiagen) was subjected to on-column DNase digestion with RNase-free DNase Set (Qiagen) according to the manufacturer’s instruction. The purity of RNA was verified with a Nano Photometer Spectrophotometer (IMPLEN). Qubit RNA Assay Kit in Qubit 2.0 Fluorometer (Life Technologies) was used to measure RNA concentration, and RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies) was used to evaluate RNA integrity. Only RNA samples that passed the quality control were further used for RNA-Seq analyses. RNA-Seq libraries were constructed using the NEBNext Ultra Directional RNA Library Prep Kit for Illumina (New England Biolabs) following manufacturer’s instructions. RNA-Seq libraries were sequenced on an Illumina HiSeq 2500 sequencer to generate 100-bp paired-end reads. SolexaQA and cutadapt were used to remove low quality regions and adapter sequences from the raw reads. The resulting clean reads were mapped to the Arabidopsis Columbia-0 reference genome (TAIR10) using Tophat with default parameters. The read counts of the genes were calculated by htseq-count, and edgeR, a package in R language, was used to identify differentially expressed genes (DEG) using fold change > 2 and false discovery rate < 0.05 as significance cutoffs. Heatmaps showing gene expression patterns of selected genes were generated using MultiExperiment Viewer (MeV4.9.0) software (Saeed et al., 2006). In brief, normalized read counts of selected genes were retrieved from edgeR results. Hierarchical clustering as performed by Pearson Correlation and average linkage clustering.
GO Enrichment Analysis
GO enrichment analysis of DEGs was performed with a public web tool gprofiler (http://biit.cs.ut.ee/gprofiler) to determine the significantly enriched GO terms in the data set of biological processes in Arabidopsis with a significance of P value < 0.05. DEGs were applied to hierarchical sorting and filtering (best per parent [moderate]), and the top 20 GO terms with the lowest P values were selected.
Accession Numbers
Sequence information of the genes studied in this article can be found in the Arabidopsis TAIR database (https://www.arabidopsis.org) under the following accession numbers: ABCG40 (At1g15520), ACT2 (At3g18780), ACT7 (At5g09810), AOX1D (At1g32350), BAP2 (At2g45760), EDS1 (At3g48090), ERF105 (At5g51190), EX1 (At4g33630), GLK1 (At2g20570), GLK2 (At5g44190), HEMA1 (At1g58290), IAN7 (At1g33950), ICS1 (At1g74710), LHCB1.1 (At1g29920), LHCB1.4 (At2g34430), LHCB2.1 (At2g05100), LHCB2.2 (At2g05070), LHCB2.3 (At3g27690), LHCB6 (At1g15820), LSD1 (At4g20380), MPT2 (At3g48850), NPR1 (At1g64280), PAD4 (At3g52430), PHYB (At2g18790), PIF1 (At2g20180), PIF3 (At1g09530), PIF4 (At2g43010), PIF5 (At3g59060), PR1 (At2g14610), PR2 (At3g57260), SAG13 (At2g29350), SIB1 (At3g56710), SIB2 (At2g41180), STZ (At1g27730), WRKY33, (At2g38470), WRKY70 (At3g56400), and ZAT12 (At5g59820).
Supplemental Data
Supplemental Figure 1. Quantitative measurement of cell death in lsd1.
Supplemental Figure 2. RNA-Seq analysis revealed that many stress-responsive genes are up-regulated in the lsd1 mutant grown under CL.
Supplemental Figure 3. Expression patterns of the whole set of photosynthesis-associated genes in wild-type and lsd1 plants.
Supplemental Figure 4. Expression levels of PhANGs and PhAPGs in old and new emerging leaves from wild-type and lsd1 plants.
Supplemental Figure 5. The lsd1 mutant does not show a gun-like mutant phenotype on LIN.
Supplemental Figure 6. Effect of the SA-mediated up-regulation of SIB1 on the expression of photosynthesis-associated genes.
Supplemental Figure 7. Expression levels of SIB1 and SIB2 in wild-type and lsd1 plants.
Supplemental Figure 8. Endogenous free SA levels in wild-type, lsd1, and lsd1 sib1 plants.
Supplemental Figure 9. Inactivation of the key SA-signaling components NPR1 and EDS1 significantly suppresses the induction of SIB1 expression and the uncoupled expression of PhANGs and PhAPGs in the lsd1 mutant.
Supplemental Figure 10. Both nuclear- and Cp-localized SIB1 proteins are necessary for the proper function of SIB1 to mediate RCD in lsd1.
Supplemental Figure 11. Overexpressing SIB1 significantly enhances the production of singlet oxygen.
Supplemental Table 1. List of PhANGs (54 genes) and PhAPGs (29 genes) in Arabidopsis.
Supplemental Table 2. List of primer sets used in this study.
Supplemental Data Set 1. List of genes (121) down-regulated in lsd1 compared to wild type.
Supplemental Data Set 2. List of lsd1-dependent down-regulated genes annotated to GO terms for BP.
Supplemental Data Set 3. List of genes (624) up-regulated in lsd1 compared to wild type.
Supplemental Data Set 4. List of lsd1-dependent up-regulated genes annotated to the top 20 GO terms for BP.
Supplemental Data Set 5. List of WRKY18, 33, and 40 target genes, which are up-regulated in lsd1 compared to wild type.
Supplemental Data Set 6. List of PhANGs and PhAPGs showing a significant differential expression (P value < 0.05) between 17-d-old wild-type and lsd1 mutant plants.
Supplemental Data Set 7. List of SA-responsive genes (204) obtained from Zhou et al. (2015).
Supplemental Data Set 8. List of SA-responsive genes up-regulated in 17-d-old lsd1 compared to wild type.
Supplemental Data Set 9. List of 1O2-responsive genes up-regulated in lsd1 compared to wild type.
Dive Curated Terms
The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:
Acknowledgments
We thank the Core Facility of Genomics, Shanghai Center for Plant Stress Biology for carrying out RNA-Seq. We thank Rosa Lozano-Durán, Nuria Sánchez Coll, and Junghee Lee for critical comments on the article. This research was supported by the Chinese Academy of Sciences (100-Talents Program to C.K. and R.Q.L., and the Strategic Priority Research Program XDB27040102 to C.K.) and by the National Natural Science Foundation of China (NSFC grant 31570264 to C.K.).
AUTHOR CONTRIBUTIONS
R.Q.L., Z.L., M.L., K.P.L., and C.K. designed the research; R.Q.L., Z.L., M.L., and K.P.L. conducted the experiments; R.Q.L., Z.L., M.L., V.D., S.L., R.Y.L., K.P.L., and C.K. analyzed the data; K.P.L. and C.K. wrote the article; all authors reviewed and edited the article.
Footnotes
↵1 These authors contributed equally to this work.
- Received October 26, 2018.
- Revised December 18, 2018.
- Accepted December 27, 2018.
- Published January 3, 2019.
References
