CCCH protein-PvCCCH69 acted as a repressor for leaf senescence through suppressing ABA-signaling pathway

CCCH is a subfamily of zinc finger proteins involved in plant growth, development, and stresses response. The function of CCCH in regulating leaf senescence, especially its roles in abscisic acid (ABA)-mediated leaf senescence is largely unknown. The objective of this study was to determine functions and mechanisms of CCCH gene in regulating leaf senescence in switchgrass (Panicum virgatum). A CCCH gene, PvCCCH69 (PvC3H69), was cloned from switchgrass. Overexpressing PvC3H69 in rice suppressed both natural senescence with leaf aging and dark-induced leaf senescence. Endogenous ABA content, ABA biosynthesis genes (NCED3, NCED5, and AAO3), and ABA signaling-related genes (SnRKs, ABI5, and ABF2/3/4) exhibited significantly lower levels in senescencing leaves of PvC3H69-OE plants than those in WT plants. PvC3H69-suppression of leaf senescence was associated with transcriptional upregulation of genes mainly involved in the light-dependent process of photosynthesis, including light-harvesting complex proteins, PSI proteins, and PSII proteins and downregulation of ABA biosynthesis and signaling genes and senescence-associated genes. PvC3H69 could act as a repressor for leaf senescence via upregulating photosynthetic proteins and repressing ABA synthesis and ABA signaling pathways.


Introduction
Natural or stress-induced leaf senescence adversely affects photosynthetic capacity and plant productivity [1][2][3] . Leaf senescence development is regulated at multiple levels, involving molecular, transcriptional, posttranscriptional, and metabolic processes [3][4][5] . At the transcriptional level, CCCH zinc finger proteins with three Cys and one His residues as the conserved motif that function as RNA-binding proteins and regulate RNA metabolism have been found to act as key regulators of leaf senescence in Arabidopsis thaliana and rice (Oryza. Sativa). CCCH genes, AtKHZ1 and AtKHZ2, can accelerate leaf senescence when overexpressed in Arabidopsis 6 .
OsDOS and OsTZF1 were found to be negative regulators of leaf senescence in rice [7][8][9] . Despite the knowledge of the involvement of CCCH genes in leaf senescence, the upstream and downstream regulatory mechanisms of CCCHs controlling leaf senescence remain largely unknown.
Several CCCH genes have been found to interact with abscisic acid (ABA), a well-known stress hormone, in the regulation of seed germination, plant growth, and stress responses. A CCCH gene cloned from Arabidopsis, AtTZF1, was reported to upregulate the expression of ABA-response genes, RD29A and COR15A, for regulating plant growth and stress responses 4 . Other CCCH genes in Arabidopsis, including AtC3H49 (AtTZF3) and AtC3H20 (AtTZF2) was inducible by ABA and also feed forward to upregulate ABA-response genes, such as RD29B during drought stress 10 . AtTZF4/5/6 genes are found to be upregulated under ABA treatment and in turn enhance ABA biosynthesis and signaling by activating NCED9 and ABI5 during seed germination 11 . Furthermore, AtTZF4 (SOMNUS) can be directly targeted by ABI5 and ABI3 in the process of seed germination 12,13 . AtTZF5 can interact with two ABA-response proteins, MARD1 and RD21A, during drought responses in Arabidopsis 14 . In rice, OsC3H47 can also be induced by ABA and promote drought tolerance, but the regulatory mechanism has not been reported 15 . Some CCCH proteins from other species, such as GhZFP1 in cotton and IbC3H18 in sweet potato can also interact with ABA downstream regulators 16,17 . ABA is well-known to play crucial roles in inducing or accelerating leaf senescence, involving activation of ABA biosynthesis genes (i.e., NCED9, ABA2, AAO3) or ABA signaling genes (i.e., ABI5, ABF2, ABF3) [18][19][20][21][22] . However, whether and how CCCH may interact with ABA regulating leaf senescence are not well understood.
Switchgrass is a perennial warm-season bunchgrass producing feathery delicate flowers from July to September, which is widely used as ornamental grasses in warm climate regions. We have previously cloned a CCCH gene in switchgrass, PvC3H69, which is phylogenetically homologous to OsTZF1; the expression of this gene can be induced by ABA whereas its overexpression resulted in delaying leaf senescence in rice 8,9,23,24 . Here, we hypothesized that PvC3H69 may serve as a repressor for leaf senescence, which could be activated by ABA accumulation at the onset of leaf senescence. The objectives of this study were to characterize the function of PvC3H69 in leaf senescence and determine whether PvC3H69 participate ABA-mediated leaf senescence and how it may regulate the ABA-mediated leaf senescence with a goal to clarify the regulatory mechanism of PvC3H69 and ABA signaling in leaf senescence. The regulatory mechanisms of PvC3H69 were examined by endogenous ABA detection and exogenous application of ABA to find the main signaling pathway in plants overexpressing PvC3H69.

Results
PvC3H69 as a nuclear-localized protein without transactivation function The Subcellular localization of PvC3H69 was studied by fusing it with a GFP tag. From Fig. 1a, it clearly showed that GFP signal of PvC3H69-GFP was merged with the DAPI stained nuclear, while the GFP control was dispersed in nuclear and cytosol, indicating that PvC3H69 was exclusively localized in the nuclear.
To test whether PvC3H69 had transcriptional activity, we performed yeast-based transactivation assay. As shown in Fig. 1b, PvC3H69 and the negative control (GUS) fused with the GAL4 DNA-binding domain (GAL4-DB) did not activate the reporter gene in the yeast system, while the positive control of a known transcription factor, PvC3H72, had transactivity 25 . Furthermore, we carried out in planta transcriptional activity assay by transient coexpressing 35S::PvC3H69-GAL4-DB (effector), GAL4(4x):: GUS (reporter), and 35S::LUC (internal control) in Arabidopsis protoplasts, with 35S::PvC3H72-GAL4-DB and 35S::GAL4-DB as the positive and empty vector controls, respectively (Fig. 1c). The positive control of PvC3H72 activated the expression of the GUS reporter gene, while PvC3H69 did not activate or suppress the expression of the GUS reporter gene (Fig. 1d). Taken together, PvC3H69 showed no transcriptional activity in neither yeast-based nor in planta assay.

Suppression of dark-induced leaf senescence in switchgrass and rice by overexpression of PvC3H69
To understand the functions of PvC3H69 regulating leaf senescence, we overexpressed the gene under driven of maize ubiquitin promoter in switchgrass and rice by Agrobacterium-mediated transformation, with transformation being confirmed by PCR and GUS staining ( Fig. S1a-d). Overexpressing PvC3H69 (abbreviated as C3H69-OE hereafter) resulted in transgenic plants of both switchgrass and rice retained more green leaves in each tiller and lower level of leaf senescence compared to their respective wild-type (WT) plants (Fig. 2a, b). Leaf photochemical efficiency (Fv/Fm) and net photosynthetic rate (Pn) of C3H69-OE lines were significantly higher than that of WT. Chlorophyll (Chl) content of the second to fourth leaves showed significantly higher levels in C3H69-OE lines than that in WT plants (Fig. 2c, d). Under dark conditions, leaves of WT plants exhibited severe leaf senescence with 90% loss of Chl content, whereas Chl content of transgenic lines maintained at 50% of the control level; Fv/Fm of transgenic lines also were significantly higher than that of WT ( Fig. 3a-h). Those results demonstrated that PvC3H69 played a positive role in suppressing dark-induced leaf senescence.

Suppression of ABA synthesis and signaling by overexpression of PvC3H69
Exogenous application of 5 or 20 µM ABA significantly accelerated leaf senescence in rice WT plants, as shown in Fig. 4a, b. In contrast, C3H69-OE-Oz lines of rice were less sensitive to ABA treatment. Leaf Chl content and Fv/ Fm of overexpression lines were significantly higher than those of WT plants. Expression levels of three ABAresponsive genes (OsLIP9, OsLEA3, and OsRAB16A) increased 120 to 5000 times in WT plants, but only increased~4 to 25 times in C3H69-OE-Oz transgenic lines of rice when treated with 20 µM ABA (Fig. 4c). These results demonstrated that PvC3H69 suppressed ABA-accelerated leaf senescence in rice by downplaying ABA responses.
We further measured expression levels of four ABA biosynthesis genes (OsABA2, OsZEP, OsNCED3, OsNCED5) and ABA content in leaves of WT and C3H69-OE-Oz lines. As shown in Fig. 6, significantly lower expression levels of OsNCED3 and OsNCED5 were detected in C3H69-OE-Oz lines than in WT, and ABA content in C3H69-OE-Oz lines was also significantly lower than that in WT (Fig. 6a, b), indicating that PvC3H69 suppressed ABA biosynthesis.

Global gene expression analysis of ABA-and senescencerelated genes in PvC3H69overexpression transgenic plants
Comparative analysis of transcriptome of C3H69-OE-Oz and WT leaves exposed to dark for 10 d to induce leaf senescence and those leaves collected prior to dark exposure (0 d) were performed by using Illumina-HiSeq™ 4000. Principle component analysis (PCA) of transcriptomic data showed that differentially expressed genes (DEGs) in WT leaves at 0 d (WT-0), C3H69-OE-Oz leaves at 0 d (C3H69-OE-Oz -0), WT leaves at 10 d of dark treatment (WT-10), and C3H69-OE leaves at 10 d of dark treatment (C3H69-OE-Oz -10) were clustered in separate groups, with DEGs in C3H69-OE-Oz -0 and   Gene Set Enrichment Analysis (GSEA) of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were carried out to understand biological functions of DEGs related to dark-induced leaf senescence due to overexpressing PvC3H69. Comparative analysis of DEGs between WT-10 and C3H69-OE-Oz -10 found PvC3H69-regulated genes and their functions related to the suppression of dark-induced leaf senescence. The most enriched GO term pathways of PvC3H69-regulated genes were "photosynthesis", "thylakoid", "signaling", "macromolecular complex", "response to hormone", "structural molecule activity", and "signal transducer activity" (Fig. S4a and Table S2). In particular, the most enriched KEGG pathways were related to hormones and photosynthetic metabolism including "photosynthesis", "carbon fixation in photosynthetic organisms", "alpha-Linolenic acid metabolism", "diterpenoid biosynthesis", "flavonoid biosynthesis", "tryptophan metabolism", "carotenoid biosynthesis", "Linoleic acid metabolism", and "plant hormone signaling pathway" (Fig. S4b and Table S3). The results indicated that PvC3H69 could mainly regulate photosystems and hormone metabolism and signaling to delay dark-induced senescence.
The expression of selected DEGs was confirmed with qRT-PCR analysis, including four ABA signaling-related genes (SnRK2, ABI5, NAC103, and NAC58), three other hormonal signaling pathways (ARR, ERF103, and ORE1), and two chlorophyll degradation genes (SGR and NOL). The relative expression of these nine genes in qRT-PCR analysis corresponded to those generated from the RNA-Seq, supporting the reliability of RNA-Seq results (Fig. S5).

Discussion
The CCCH-type zinc finger family includes multiple genes, and PvC3H69 cloned from switchgrass in our study was found to be a homology of ZmC3H38 and OsTZF1 24 , but exhibited unique characteristics from the homologs in maize and rice. PvC3H69 was located in nucleus, but had no transcriptional activity (Fig. 1a-d), suggesting it is not a transcriptional factor. Other CCCH proteins in rice and Arabidopsis with functions in DNA binding, RNA binding, and mRNA turnover or silencing can shuttle between the nucleus and cytoplasm foci under different stresses 9,26 . TZF1 proteins have been found to directly interact with stress regulators, such as RD21A and PR5 or the mRNA of Ank-β 9,16 . However, whether PvC3H69 has mRNA or protein binding function deserves further investigation.
Although CCCH genes are known to regulate leaf senescence, different genes in the CCCH family were found to have distinct functions with AtKHZ1 and AtKHZ2 being characterized as a positive regulator inducing leaf senescence in Arabidopsis 6 and OsDOS and OsTZF1 as a negative regulator of leaf senescence in rice [7][8][9] . In our study, PvC3H69 acted as a negative regulator, which suppressed both natural senescence associated with leaf aging and dark-induced leaf senescence, as manifested by the maintenance of greater chlorophyll content, photochemical efficiency, and net photosynthetic rate in plants overexpressing PvC3H69 (Figs. 2 and 3). Furthermore, transcriptomic analysis of plants overexpressing PvC3H69 found that a large number of genes involved in the lightdependent process of photosynthesis, including lightharvesting complex proteins (Lhca2, Lhcb1, Lhcb4, Lhcb5, Lhcb), PSI proteins (PsaH, PsaK, PsaN, PsaO) and  (Fig. 7c). The combined physiological and transcriptomic data provided strong evidence that PvC3H69 acted as a repressor for leaf senescence, which helped to sustain or maintain photosynthesis by enhancing lightharvesting and photochemical capacity in both PS I and PS II of photosynthesis.
The underlying molecular mechanisms and key regulatory pathways for CCCH suppressing leaf senescence are yet to be fully understood. The delay in stress-induced leaf senescence by OsTZF1 has been associated with regulating stress-related genes (i.e., AK112082, JAZ1, Ferritin, MT-type1, ChaC-like) 8,9 . In our study, DEGs by PvC3H69 were highly enriched in hormone metabolism  and signaling pathways, with most genes in ABA biosynthesis and signaling being downregulated significantly by overexpressing PvC3H69 (Fig. 7b). Moreover, transgenic rice plants overexpressing PvC3H69 exhibited lower sensitivity of leaf senescence to ABA treatment compared with WT plants (Fig. 4a, b). These results indicated that function of PvC3H69 is involved in ABA.
ABA-induced leaf senescence involves activation of ABA biosynthesis genes, such as NCEDs, ABA2, AAO3, and ABA signaling genes, such as ABI5, ABF2, and ABF3 [18][19][20][21][22] . OsNECD3 and OsNECD5 are involved in xanthophyll cleavage for ABA biosynthesis 27,28 . The relatively lower content of ABA in C3H69-OE plants could be resulted from the downregulated transcript level of OsNECD3 and OsNECD5. Consistently, global genes expression analysis showed that nine genes in ABA biosynthesis, including OsNCED3 and OsAAO3 were downregulated, indicating that PvC3H69 could be a negative regulator for ABA biosynthesis (Fig. 6a, b). In ABA signaling, PP2C-SnRK-ABF regulatory model is considered as the core pathway that is required for ABA-triggered Chl degradation 21,22 . In this model, PYL as a receptor can accept ABA signal and form PYL-ABA bound complexity, which leads to PP2C inactivated and released the repressed SnRKs (mainly SnRK2s); the activated SnRKs can phosphorylate ABA-response binding factors (ABFs) 29,30 . SnRK2 genes are plant-specific serine/threonine kinases involving in plant responses to abiotic stresses and ABA-dependent plant development 31,32 . In rice, there are 10 SnRK proteins in SnRk2.0 family and are designated as SAPKs (stress-activated protein kinase). Among them, OsSAPK8, OsSAPK9, and OsSAPK10, were also activated by ABA 33 . OsSAPK1 can be upregulated by osmotic stressors and it has been reported that OsSAPK1 can be directly regulated by OsNAC2 through an ABAdependent pathway 34 . Ectopic expression of SAPK6 (OSRK1) in tobacco confers reducing ABA sensitivity 35,36 . In our study, three OsSAPKs, OsSAPK1, OsSAPK6, and OsSAPK10 were downregulated by overexpressing PvC3H69. All these results suggested that PvC3H69 could interrupt both ABA biosynthesis and signaling, thereby suppressing leaf senescence in C3H69-OE plants.
Further evaluation of the relative transcript changes of ABA-biosynthesis and ABA-signaling genes in leaves in C3H69-OE lines and WT exposed to dark found that PvC3H69 caused downregulation of all genes in the ABA signaling pathway in a greater magnitude compared to those genes for ABA biosynthesis, suggesting ABA signaling could be more sensitive to negative transcriptional control of PvC3H69. In addition, ABI5 and ABF2/3/4 are key phosphorylating substrates for SnRKs which belong to basic leucine zipper (bZIP) TFs and AREB binding TFs, respectively, and positively affect leaf senescence by directly anchoring the promoter of SAGs including SGR, NYC1, and ABR 5,22 . The reduced transcript level of SnRKs, ABI5, and ABF2/3/4 in C3H69-OE plants treated with ABA (Fig. 5) indicated that PvC3H69 could negatively regulate ABA-induced leaf senescence mainly through PP2C-SnRK-ABF signaling pathway.
In summary, overexpression of PvC3H69 in rice or its native plants resulted in a stay-green phenotype, strongly suggesting that PvC3H69 was a negative regulator in leaf senescence. PvC3H69 could facilitate the stay-green phenotype or delayed leaf senescence mainly by upregulating light-dependent process of photosynthesis, including light-harvesting complex proteins, PSI proteins, and PSII proteins and repressing ABA biosynthesis and signaling genes and senescence-associated genes such as NCED3/5, AAO3, SnRK1/6/10, ABF2/3/4, ABI5, SGR, and NYC1 (Fig. 9). PvC3H69 suppression of ABA-mediated leaf senescence with leaf aging or induced by darkness was mainly through regulating PP2C-SnRK-ABF signaling Fig. 9 Proposed pathways of PvC3H69 regulating ABA-mediated leaf senescence: PvC3H69 could repress the expression of NCED3/5 and AAO3, and then repressed ABA synthesis. In ABA transduction signaling, PvC3h69 could also repress the expression of SnRK1/6/10 and further decrease the phosphorylationof SnRK1/6/10 on ABF2/3/4 and ABI5; and then further surpressed the expression of SGR1/NYC1 and PSI/PSII protein degradation, then delay the leaf senescence pathway. Future research could identify its upstream regulatory factors and further confirm the functions of PvC3H69 in leaf senescence induced by other abiotic stress, such as heat, drought, and salinity in order to improve plant tolerance to diverse environmental stresses.

Gene cloning and vector construction
The full-length gene of PvC3H69 (Phytozome accession no.: Pavir.J04795.1), with 394 amino acids without intron was amplified from the gDNA of a lowland ecotype 'Alamo' switchgrass 37 . The gene was firstly cloned into the Gateway entry vector pENTR/D (Invitrogen). It was subcloned into p2GWF7.0 38 , pGBKT7 (Invitrogen), and pVT1629 37 through LR reaction (Invitrogen). The primers used for PvC3H69 cloning are listed in Table S1.

Observation of subcellular localization of PvC3H69-GFP
The PvC3H69 was subcloned into a modified gatewaycompatible P2GWF7.0 vector to put PvC3H69 in fusion with GFP. By polyethylene glycol (PEG)-mediated Arabidopsis protoplast transformation 39 , the PvC3H69-GFP fusion gene was overexpressed in Arabidopsis protoplasts. DAPI was used to stain the nucleus, and the GFP signals were detected under a Zeiss LSM 780 laser scanning confocal microscope (Carl Zeiss SAS, Jena, Germany).

Transactivation assay
PvC3H69 was subcloned into the BD vector pGBKT7 to fuse PvC3H69 with the DNA-binding domain of GAL4. The pGBKT7-PvC3H69 and the control vector pGBKT7-GUS (UiDA gene) were then transformed into the yeast strain Y2HGold (Clonetech), separately. The pGBKT7-PvC3H72 was used as a positive control 25 . The transformed positive clones grown well on SD/-Trp were then grown on plates containing SD/-Trp-Ade-His and SD/-Trp-Ade-His + 25 mM 3-AT for auto-transactivation assay.
For the transcriptional activity assay of PvC3H69 in plant cells, PvC3H69 was cloned into the 35S promoterdroven pZB370 vector to fuse with the yeast GAL4 DNAbinding domain (GAL4BD) as effector (pZB369-PvC3H69), while the vector without the target gene was used as the negative control. As a positive control, PvC3H72 has been reported to be a transcriptional activator 25 . The internal control vector was pZB371-Luciferase under driven of 35S promoter as well. The reporter vector (pZB370-GUS) was constituted of four copies of GAL4 DNA-binding sites (GAL4(4x)-D1-3(4x)) to drive the GUS (UidA) reporter gene. Three plasmids (effector, reporter, and internal control) were cotransferred into Arabidopsis protoplasts at the ratio of 5:4:1. The transcriptional ability was assessed by the GUS/ LUC ratio. Three biological replicates were included for each combination.

Plant transformation and verification
Switchgrass genetic transformation followed the protocol described in Xu et al. 37 . Embryogenic calluses of a selected line 'HR8' from switchgrass lowland ecotype 'Alamo' were infected with Agrobacterium tumefaciens strain 'AGL1' harboring the binary vector pVT1629-PvC3H69 with the target gene under driven of the maize ubiquitin promoter, and selected the putative transgenic lines on 50 mg L −1 hygromycin (Sigma). GUS staining and regular PCR for the presence of the T-DNA fragment of transgenic lines were the same as reported previously 25 .
Nipponbare (Oryza sativa japonica cvNipponbare) was used in this study. The transformation system was referred to Toki et al. 40 . Leaves of transgenic lines were stained with GUS solution. DNA was extracted from the C3H69-OE plants for PCR detection. RNA was extracted from the leaves of transgenic lines as a template for determining the transcriptional level of PvC3H69.

Plant growth conditions and dark treatment
Switchgrass transgenic lines and WT were grown in the green house with temperatures set at 28°C/22°C, day/ night with a 14-h light/10-h darkness. The plants were watered twice a week. In order to induce leaf senescence, the middle 1/3 part of whole detached full-expand leaves from 3-month-old seedlings were cut into 3-cm segments and placed in a dark room with air temperature controlled at 28°C. Leaf samples were collected at the time points of 0, 10, and 15 day for phenotypic and physiological analysis. Five-month-old plants were used for phenotypic and physiological analysis of natural senescence.
T3 seeds of WT and transgenic rice 'Nipponbare' were sterilized and germinated on 1/2 Murashige and Skoog (MS) medium, which were transferred to bucket filled with IRRI (International Rice Research Institute) nutrient solution in a growth chamber controlled at 30°C during the day and 25°C at night with 16-h light/8-h of darkness. The middle 1/3 part of whole detached full-expand leaves of 6-week-old plants were cut into about 3-cm fragments for dark treatment with air temperature controlled at 25°C. Leaf fragments were collected at the time points of 0 d (day), 10 d, and 13 d for further phenotype and physiological index measurement. For ABA treatment, leaf fragments (same cut with the dark treatment) of WT and transgenic rice were soaked in different concentration ABA solutions (0, 5, and 20 µM) under darkness. Leaf samples were collected at the time points of 0, 7, and 10 d for phenotypic, physiological analysis. Samples at 7 day were used for qRT-PCR analysis.

Physiological analysis of transgenic switchgrass and rice
Leaf net photosynthetic rate (Pn) was measured with a LI-6400 system (LI-COR Inc., Lincoln, NE) equipped with a standard 2 × 3 cm 2 leaf chamber with light-emitting diodes as a light source. The measurements were taken at the PAR of 800 µmol m −2 s −1 and flow rate of 500 µmol s −1 . The block temperature was set to 25°C for optimal temperature. For leaf photochemical efficiency (Fv/Fm), leaves were put in a 30 min dark-adaptation period and measured using a fluorescence induction monitor (OPTI-Sciences, Hudson, USA) as the ratio of variable (Fv) to maximum (Fm) fluorescence. Chlorophyll content (Chl) was measured by extracting chlorophyll from 0.1 g fresh leaves in 10 ml dimethyl sulfoxide (DMSO) under dark for 4 d and measuring the absorbance at 663 and 645 nm. The blades were dried in an oven at 80°C for dry weight. Chl content was calculated using the formula described in Arnon 41 .
For endogenous ABA content analysis, leaves of 4week-old rice seedlings were used to measure the endogenous ABA content. ABA was extracted from 500 mg frozen leaf powder according to the method reported by Krishnan et al. 42 . The samples were suspended with extraction buffer (methanol:water:acetic acid, 80:19:1, v/v/ v) and shake for 12 h at 4°C, then centrifuged at 14,000 rpm for 20 min. The supernatant was collected in a new tube, pellet were reextracted with 500 μl extraction buffer, shaken for 4 h at 4°C under darkness, then centrifuged at 14,000 rpm for 20 min at 4°C. Two tubes of supernatant were mixed and dried using centrifugal vacuum concentrator and then dissolved in 300 µL methanol. The resulted supernatant ABA concentration was determined by high-performance liquid chromatograph using SCIEX-6500Qtrap mass spectrometer (HPLC-MS/MS; Aglient1290, Agilent, USA).

Rice transcriptome analysis and qRT-PCR analysis
For transcriptome analysis, total RNA was extracted from the leaves of 4-week-old rice WT and overexpressing PvC3H69 (C3H69-OE) plants exposed to dark treatment for 10 d and prior to dark treatment (0 d) according to the manufacturer's instructions using RNA extract kit (Invitrogen, Carlsbad, CA, USA). mRNA was enriched by Oligo (dT) beads. Then the enriched mRNA was fragmented into short fragments and reverse into transcripted into cDNA. cDNA were purified with QiaQuick PCR extraction kit (Qiagen, Venlo, The Netherlands). Three independent biological replicates were conducted for the WT and C3H69-OE plants. The differently expressed genes in WT and C3H69-OE plants were classified functionally using the biological process category of Rice Gene Ontology (ftp:// ftp.ensemblgenomes.org/pub/plants/release-39). Significant interactors were determined using a two-sample analysis (t test).
For qRT-PCR analysis, total RNA extraction, firststrand cDNA synthesis, PCR reaction, and data analysis were the same as our previously reported by Xie et al. 25 . Primers for qRT-PCR are listed in Table S1.

Statistical analysis
Leaf senescence (including dark-induced and ageinduced) and ABA treatment effects and variations among WT and transgenic plants for physiological parameters and gene expression levels were analyzed using SAS v9.2 (SAS Institute, Cary, NC, USA). Mean data were separated Fisher's protected LSD at the probability of 0.05.