The C2H2‐type zinc finger protein ZOS202 regulates leaf senescence in rice by targeting senescence‐associated transcription factors and ROS homoeostasis

SUMMARY

Leaf senescence is a tightly regulated developmental process orchestrated by multiple transcription factor (TF) families. Although C2H2‐type zinc finger proteins are known to participate in various aspects of plant growth and abiotic stress responses, their specific role in regulating leaf senescence remains poorly understood. Here, we show that knockout mutation of the C2H2‐type zinc finger TF gene ZOS202 (LOC_Os02g02424) resulted in delayed leaf senescence under both dark‐induced and natural conditions, while overexpression lines exhibit accelerated leaf senescence, as indicated by reactive oxygen species (ROS) overaccumulation, chloroplast degradation, and leaf cell death. As a dual‐function transcription factor whose expression peaks at senescence onset, ZOS202 directly activates senescence‐associated TF genes OsWRKY42, OsWRKY53, and ONAC096, while repressing peroxidase genes OsPRX113, OsPRX114, and OsPRX122, thereby disrupting ROS homeostasis and promoting senescence. Collectively, our findings uncover ZOS202 as a positive regulator of leaf senescence that acts by coordinating chlorophyll breakdown and ROS accumulation, providing promising strategies for fine‐tuning leaf senescence in rice breeding.

Significance Statement

The C2H2‐type zinc finger TF family constitutes one of the largest and most functionally diverse TF families in plant genomes, yet investigations into their regulatory functions during leaf senescence remain relatively limited. We address this knowledge gap by identifying ZOS202 as a positive senescence regulator, which directly activates multiple senescence‐promoting TFs and represses peroxidase‐encoding genes, orchestrating the coordinated progression of chloroplast degradation and ROS accumulation.

INTRODUCTION

Rice (Oryza sativa L.) serves as the staple food for more than half of the global population. However, diverse biotic and abiotic stresses frequently trigger premature leaf senescence, which severely impairs rice yield and grain quality (Sakuraba et al., 2020; Swift et al., 2026; Xu et al., 2023; Yang et al., 2026; Zheng et al., 2024). As the terminal stage of leaf development, leaf senescence is characterized by the breakdown of the photosynthetic apparatus, accumulation of reactive oxygen species (ROS), and remobilization of nutrients to grains and young leaves (Buelbuel et al., 2023; Guo et al., 2025; Yang et al., 2026). Notably, the onset and progression of leaf senescence are precisely modulated by the coordination of endogenous physiological dynamics and external environmental cues. Fine‐tuning the balance between vegetative growth and leaf senescence is an effective strategy to maximize rice yield and improve grain quality (Woo et al., 2019; Zhang et al., 2024).

A wide range of transcription factors have been identified as regulators of leaf senescence in rice, with the majority belonging to NAC and WRKY transcription factor families (Li et al., 2023; Zhang et al., 2023, 2024). Functional studies have demonstrated that OsNAP, ONAC054, ONAC096, OsWRKY53, and OsWRKY42 act as positive regulators of leaf senescence, whereas OsNAC109 and ONAC106 function as negative regulators (Han et al., 2014; Kang et al., 2019; Li et al., 2021; Liang et al., 2014; Sakuraba et al., 2015, 2020; Xie et al., 2021). Nevertheless, only a limited number of C2H2‐type transcription factors have been reported to participate in leaf senescence.

C2H2 zinc finger proteins constitute one of the largest transcription factor families and play critical roles in plant growth, developmental processes, as well as response to biotic and abiotic stresses (Lu et al., 2024). The C2H2 protein DST regulates cytokinin homeostasis in the shoot apical meristem by modulating OsCKX2 expression, thereby influencing reproductive organ formation (Guo et al., 2020). A single nucleotide polymorphism in the promoter of Bsr‐d1, which encodes a C2H2 zinc finger protein, reduces its expression, inhibits H₂O₂ degradation, and enhances disease resistance (Li et al., 2017). Notably, LS1 is the sole C2H2 protein currently identified as a negative regulator of leaf senescence in rice (Zhang et al., 2022). Therefore, dissecting the biological roles of C2H2 family members in leaf senescence will advance our comprehension of the intricate regulatory networks of leaf senescence.

In this study, we demonstrate that knockout of the C2H2‐type zinc finger TF gene ZOS202 delays leaf senescence under both natural and dark‐induced conditions, while its overexpression accelerates senescence by promoting chloroplast degradation and ROS accumulation. Mechanistically, ZOS202 regulates rice leaf senescence by directly activating senescence‐associated transcription factors OsWRKY42, OsWRKY53, and ONAC096, while repressing the ROS scavenging genes OsPRX113, OsPRX114, and OsPRX122. Our work uncovers a novel regulatory node in the plant leaf senescence network, expands the functional understanding of C2H2‐type zinc finger proteins in developmental senescence, and provides a potential molecular target for genetic improvement of leaf senescence traits in rice.

RESULTS

ZOS202 positively regulates leaf senescence

Dark treatment efficiently induces leaf senescence (Buelbuel et al., 2023; Liang et al., 2014; Sakuraba et al., 2020). Screening of CRISPR/Cas9‐mediated knockout mutants revealed that two independent zos202 lines (z‐1 and z‐2) showed delayed leaf senescence under dark treatment (Figure 1A–C; Figure S1A). Before dark treatment, no significant differences in chlorophyll content, SPAD values, and relative electrolytic leakage (REL) were observed between ZH11 and zos202 mutants. After 6 days of dark treatment, zos202 mutants retained significantly higher chlorophyll content and SPAD values, and lower REL than wild‐type ZH11 (Figure 1D–F). Transmission electron microscopy observations revealed that the leaves of both zos202 mutants and ZH11 had structurally intact chloroplasts with well‐organized thylakoid grana before dark treatment. Following dark exposure, the chloroplasts remained better structural integrity in the zos202 mutants compared to ZH11 (Figure 1G). Under field conditions, zos202 mutants showed no obvious difference from ZH11 at the seedling stage (35 Days Post Sowing/DPS; Figure S1B–F) or heading stage (75 DPS). However, at full maturity (120 DPS), zos202 mutants displayed significantly delayed leaf senescence, as reflected by higher chlorophyll content, SPAD values, net photosynthetic rate, and lower REL compared with ZH11 (Figure 1H–M).

Figure 1.

Loss of function of ZOS202 delays leaf senescence in rice.

(A) Leaf segments from 2‐month‐old ZH11 and zos202 mutants before and after 6 days of DT. Scale bar, 1 cm.

(B, C) Two‐week‐old ZH11 and zos202 mutants before and after 6 days of DT, which were photographed as whole plant (B) or scanned with detached leaves (C). Scale bars: (B) 3 cm; (C) 6 mm.

(D–F) Chlorophyll content (D), SPAD values (E), REL (F) of leaves from ZH11 and zos202 mutants before and after 6 days of DT. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

(G) Chloroplast ultrastructure of leaves from ZH11 and zos202 mutants before and after 6 days of DT. CP, chloroplast; G, granum; OG, osmiophilic granule. Scale bar, 2 μm.

(H, I) Plant phenotypes of ZH11 and zos202 mutants at 75 and 120 DPS which were photographed as whole plant (H) or scanned with detached leaves (I). Scale bars: (H) 10 cm; (I) 1.5 cm.

(J–M) Chlorophyll content (J), SPAD values (K), relative electrolytic leakage (L), net photosynthetic rate (M) of flag leaves from ZH11 and zos202 mutants at 75 and 120 DPS. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

Conversely, two ZOS202 overexpression lines (Z‐OE1 and Z‐OE2; Figure S2A) exhibited accelerated leaf senescence. Mild leaf yellowing was observed starting at the tillering stage (50 DPS; Figure S2B), with significantly enhanced senescence at 65 DPS and 100 DPS (Figure 2A,B). Consistent with the senescence phenotype, Z‐OE lines showed reduced chlorophyll content, SPAD values, and net photosynthetic rate, together with elevated REL compared with control KT plants from tillering to maturity (Figure 2C–F; Figure S2C–F). Under dark treatment, Z‐OE lines showed further accelerated senescence, with lower chlorophyll content and SPAD values, and higher REL (Figure 2G–L). Ultrastructurally, chloroplasts in dark‐treated Z‐OE lines showed severe degradation and loosely stacked thylakoid grana relative to KT (Figure 2M).

Figure 2.

ZOS202 overexpression accelerates rice leaf senescence.

(A, B) Plant phenotypes of KT and Z‐OE1/2 transgenic lines at 65 and 100 DPS which were photographed as whole plant (A) or scanned with detached leaves (B). Scale bars: (A) 10 cm; (B) 1.5 cm.

(C–F) Chlorophyll content (C), SPAD values (D), relative electrolytic leakage (E), net photosynthetic rate (F) of flag leaves from KT and Z‐OE1/2 transgenic lines at 65 and 100 DPS. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

(G) Leaf segments from 2‐month‐old KT and Z‐OE1/2 transgenic lines after 5 days of DT. Scale bar, 1 cm.

(H, I) Two‐week‐old KT and Z‐OE1/2 transgenic lines before and after 5 days of DT, which were photographed as whole plant (H) or scanned with detached leaves (I). Scale bars: (H) 3 cm; (I) 6 mm.

(J–L) Chlorophyll content (J), SPAD values (K), relative electrolytic leakage (L) of leaves from KT and Z‐OE1/2 transgenic lines before and after 5 days of DT. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

(M) Chloroplast ultrastructure of leaves from KT and Z‐OE1/2 transgenic lines before and after 5 days of DT. CP, chloroplast; G, granum; OG, osmiophilic granule. Scale bar, 2 μm.

Expression of the senescence‐promoting transcription factors (OsNAP, ONAC300), chlorophyll catabolic genes (OsSGR, OsPAO), and senescence‐associated genes (Osl2, OsSH36) was upregulated in Z‐OE lines and downregulated in zos202 mutants (Figure S3A–F). No significant differences in heading date were observed in either zos202 mutants or Z‐OE lines (Figure S4A,G). Z‐OE lines showed significantly reduced plant height, tiller number per plant, panicle length, grain weight, and grain yield per plant compared to wild type (KT), probably due to the premature senescence (Figure S4B–F). However, no significant changes were observed in the zos202 mutants compared to wild type (ZH11) (Figure S4H–L). These genetic and physiological data establish ZOS202 as a positive regulator of leaf senescence that accelerates chlorophyll degradation and chloroplast dismantling, thereby shortening leaf photosynthetic duration without affecting flowering time.

ZOS202 promotes ROS accumulation

Leaf senescence is tightly linked to the excessive ROS accumulation, which accelerates senescence and promotes cell death (Woo et al., 2019). We therefore first examined ROS levels in leaves before and after dark treatment using 3,3′‐diaminobenzidine (DAB), nitroblue tetrazolium chloride (NBT), and H₂DCFDA staining, indicating markedly lower ROS levels in zos202 mutants compared with ZH11 (Figure 3A–C). Biochemical measurements further confirmed that zos202 mutants accumulated significantly lower levels of hydrogen peroxide (H₂O₂), superoxide radical anion (${\mathrm{O}}_2^{·-}$), and malondialdehyde (MDA) after dark treatment (Figure 3D–F). Similarly, ROS levels in flag leaves at the mature stage were lower in zos202 mutants than ZH11, consistent with reduced cell death and cell membrane damage in zos202 mutants as shown by trypan blue (TB) staining and MDA content (Figure S5A–D).

Figure 3.

ZOS202 promotes ROS accumulation in leaves during dark‐induced senescence.

(A, B) DAB (A) and NBT (B) staining of leaves from ZH11 and zos202 mutants before and after 6 days of DT. Scale bar, 1.5 mm.

(C) H₂DCFDA staining of leaves from ZH11 and zos202 mutants before and after 6 days of DT. Red, chlorophyll; green, oxidized H₂DCFDA. Scale bar, 300 μm.

(D–F), MDA content (D), ${\mathrm{O}}_2^{·-}$ content (E), H₂O₂ content (F) of leaves from ZH11 and zos202 mutants before and after 6 days of DT. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

(G, H) DAB (G) and NBT (H) staining of leaves from KT and Z‐OE1/2 transgenic lines before and after 5 days of DT. Scale bar, 1.5 mm.

(I) H₂DCFDA staining of KT and Z‐OE1/2 transgenic lines before and after 5 days of DT. Red, chlorophyll; green, oxidized H₂DCFDA. Scale bar, 300 μm.

(J–L) MDA content (J), ${\mathrm{O}}_2^{·-}$ content (K), H₂O₂ content (L) in leaves of KT and Z‐OE1/2 transgenic lines before and after 5 d of DT. Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

In contrast, Z‐OE lines accumulated significantly higher ROS levels in seedlings after dark treatment and in flag leaves at maturity, leading to aggravated cell membrane damage and cell death compared with KT (Figure 3G–L; Figure S5E–H). Notably, H₂O₂ levels were already elevated in Z‐OE lines even before dark treatment (Figure 3G,I,L). Collectively, these results demonstrate that ZOS202 promotes leaf senescence by elevating ROS accumulation.

ZOS202 encodes a nuclear‐localized transcriptional regulator with elevated expression during senescence transition and SA treatment

To characterize ZOS202 as a transcription factor, we transiently expressed pCAMBIA2300‐35S:ZOS202‐eGFP in rice protoplasts and confirmed that ZOS202 localizes exclusively in the nucleus (Figure 4A). GAL4‐dependent transcriptional activity assays in rice protoplasts further revealed that ZOS202 functions as a transcriptional activator and potentially as a transcriptional repressor (Figure 4B; Figure S6A).

Figure 4.

ZOS202 encodes a nuclear‐localized transcriptional regulator with elevated expression during senescence transition and SA treatment.

(A) Subcellular localization of ZOS202‐eGFP fusion proteins in rice sheath protoplasts. OsbZIP52‐RFP fusion proteins served as a nuclear marker, with eGFP as the control. Scale bar, 10 μm.

(B) GAL4‐dependent transactivation assay of ZOS202 in rice protoplasts. The left part is a schematic illustration of the effector and reporter constructs. Values are mean ± SD (n = 3 biologically independent samples). **P < 0.01 (Student's t‐test).

(C) Expression analysis of ZOS202 in different tissues of KT at 65 DPS under natural long‐day field conditions. Values are mean ± SD (n = 3 biologically independent samples).

(D, E) Temporal expression analysis of ZOS202 and OsNAP during natural senescence. Flag leaves of field‐grown KT were harvested at 7‐day intervals from 14 DPS to 112 DPS. Values are mean ± SD (n = 3 biologically independent samples).

(F, G) Expression analysis of ZOS202 (F) and OsNAP (G) in different leaf ages of field‐grown KT at 50 DPS. L1 to L5 represent the flag leaf to the fifth leaf of rice plant from top down, respectively. Values are mean ± SD (n = 3 biologically independent samples).

(H, I) Expression analysis of ZOS202 and OsNAP in different segments of the same leaf from field‐grown KT at 50 DPS. a–e correspond to the five truncated sections from leaf base to tip of the third leaf, respectively. Values are mean ± SD (n = 3 biologically independent samples).

(J) Expression patterns of ZOS202 in response to phytohormone treatments. Detached leaf disks from 2‐week‐old wild‐type plants were floated on 3 mM MES buffer (pH 5.8) containing individual phytohormones at the indicated concentrations. Control, 3 mM MES buffer alone; SA, salicylic acid; ABA, abscisic acid; ACC, 1‐aminocyclopropane‐1‐carboxylic acid; MeJA, methyl jasmonate; Values are mean ± SD (n = 3 biologically independent samples). *P < 0.05; **P < 0.01 (Student's t‐test).

To explore the expression pattern of ZOS202, we first analyzed the transcript levels of ZOS202 in different tissues of rice plant at heading stage by RT‐qPCR and found that ZOS202 was most highly expressed in flag leaves (Figure 4C). Expression data from the Rice Expression Profile Database (RiceXPro) show that ZOS202 transcript levels first increase and then decline during leaf development (Figure S6B). This expression pattern was validated by RT‐qPCR profiling of upper leaves from seedling (14 DPS) to mature (112 DPS) stages (Figure 4D). Interestingly, the peak of ZOS202 expression precedes the senescence marker gene OsNAP, which increases progressively throughout leaf development (Figure 4E; Figure S6C). We then examined ZOS202 expression along leaf age and position gradients. At the tillering stage (50 DPS), ZOS202 transcript abundance increased then decreased in older leaves (L1 to L5), whereas OsNAP increased continuously (Figure 4F,G). Within individual leaves, ZOS202 showed higher expression in the middle regions than at the tip or base, while OsNAP increased from base to tip (Figure 4H,I).

To identify upstream hormonal signals that modulate ZOS202 expression, we analyzed its transcript levels in response to four senescence‐associated phytohormones: abscisic acid (ABA), the ethylene precursor 1‐aminocyclopropane‐1‐carboxylic acid (ACC), methyl jasmonate (MeJA), and salicylic acid (SA). SA treatment induced the most rapid and robust increase in ZOS202 transcript levels, while ABA, ACC, and MeJA exerted only mild or negligible effects (Figure 4J).

Together, these findings reveal that ZOS202 is a nuclear‐localized transcription factor whose expression peaks at the senescence transition and is strongly induced by SA.

ZOS202 regulates genes involved in ROS homeostasis and photosynthesis

To uncover the gene regulatory network controlled by ZOS202, we performed RNA‐seq on flag leaves of KT and Z‐OE lines at 80 DPS. For the nine RNA‐seq libraries, the proportion of clean reads that mapped to the reference genome was above 95% (Table S2). In total, 2209 genes were commonly upregulated and 2048 genes were commonly downregulated in Z‐OE1 and Z‐OE2 relative to KT (Figure S7A,B; Tables S3 and S4). KEGG enrichment analysis showed that upregulated genes were strongly enriched in defense‐related secondary metabolism and redox homeostasis, whereas downregulated genes were predominantly associated with photosynthesis and carbon metabolism (Figure S7C,D). These transcriptomic data indicate that ZOS202 coordinately regulates ROS homeostasis and photosynthesis during leaf senescence. By suppressing photosynthesis‐related gene expression and activating oxidative stress responses, ZOS202 drives the transition from anabolism to catabolism, which in turn facilitates leaf senescence.

ZOS202 directly targets multiple transcription factors related to ageing

To identify direct targets of ZOS202, we performed DNA affinity purification sequencing (DAP‐seq). For the DAP‐seq libraries, the mapping rate of clean reads exceeded 97%, and we identified 6277 ZOS202‐binding peaks distributed across all 12 rice chromosomes (Figure S8A; Table S5). Binding sites were strongly enriched near transcription start sites (TSSs), with 45.91% in intergenic regions and 20.65% in promoter regions (Figure S8B,C; Table S6). MEME‐ChIP identified five high‐confidence ZOS202‐binding motifs (Figure S8D).

By integrating DAP‐seq and RNA‐seq datasets (Tables S3 and S7), we identified 118 genes that were promoter‐bound by ZOS202 and upregulated in Z‐OE lines (Figure 5A; Table S8). GO enrichment highlighted terms related to DNA‐binding transcription factor activity and kinase activity (Figure 5B). Among these, three known senescence‐promoting TFs, OsWRKY42, OsWRKY53, and ONAC096, were selected for validation (Figure S9).

Figure 5.

ZOS202 directly activates OsWRKY42, OsWRKY53, and ONAC096 transcription.

(A) Venn diagram showing the identification of shared genes between DAP‐seq target genes (≤3000 bp upstream of the transcription start site) and upregulated DEGs in Z‐OE lines.

(B) GO enrichment analysis of the 118 ZOS202‐targeted upregulated genes.

(C–E) Expression analysis of OsWRKY42 (C), OsWRKY53 (D), ONAC096 (E) in leaves of KT, Z‐OE1/2, ZH11, and zos202‐1/2 lines after 48 h of DT by RT‐qPCR. Values are mean ± SD (n = 3 biologically independent samples). *P < 0.05; **P < 0.01; ns, not significant (Student's t‐test).

(F) Detection of ZOS202 binding to the promoters of OsWRKY42, OsWRKY53, and ONAC096 by DAP‐seq. DAP‐1 and DAP‐2: two independent biological replications. Shaded area: peak position; red bar: peak and motif binding site.

(G) Electrophoretic mobility shift assays (EMSAs) showing that ZOS202 binds to the promoters of OsWRKY42, OsWRKY53, and ONAC096. GST was used as a negative control; competition for binding was performed with 5× or 20× unlabeled oligos. Plus and minus symbols represent the presence and the absence of components, respectively.

(H) Transactivation analysis of OsWRKY42, OsWRKY53, and ONAC096 promoters by ZOS202 in rice protoplasts. Protoplasts co‐transformed with the 35S empty vector plus reporter plasmid were used as controls. The left part is a schematic illustration of the effector and reporter constructs. Values are mean ± SD (n = 4 biologically independent samples). **P < 0.01 (Student's t‐test).

RT‐qPCR confirmed that these three TFs were upregulated in Z‐OE lines and downregulated in zos202 mutants after dark treatment (Figure 5C–E). DAP‐seq revealed ZOS202‐binding peaks in their promoters (Figure 5F), and EMSA verified direct binding of ZOS202 to these promoter regions (Figure 5G). Dual‐luciferase assays in rice protoplasts further demonstrated that ZOS202 transcriptionally activates the promoters of OsWRKY42, OsWRKY53, and ONAC096 (Figure 5H). Thus, ZOS202 directly activates multiple senescence‐promoting transcription factors, positioning it upstream of a transcriptional cascade that amplifies leaf senescence signal.

ZOS202 regulates ROS homeostasis by repressing peroxidase activity

Given the dual function of ZOS202, we next intersected DAP‐seq promoter‐bound targets with RNA‐seq downregulated genes (Tables S4 and S7), identifying 117 overlapping genes (Figure 6A; Table S9). GO enrichment revealed significant enrichment for peroxidase activity, including PRX113, PRX114, and PRX122 (Figure 6B). These three peroxidase genes were selected for functional validation.

Figure 6.

ZOS202 directly represses peroxidase genes to regulate ROS‐mediated leaf senescence.

(A) Venn diagram showing the identification of shared genes between DAP‐seq target genes (≤3000 bp upstream of the transcription start site) and downregulated DEGs in Z‐OE lines.

(B) GO enrichment analysis of the 117 ZOS202‐targeted downregulated genes.

(C–E) Expression analysis of PRX113 (C), PRX114 (D), PRX122 (E) in leaves of KT, Z‐OE1/2, ZH11 and zos202‐1/2 lines after 48 h of DT by RT‐qPCR. Values are mean ± SD (n = 3 biologically independent samples). *P < 0.05; **P < 0.01; ns, not significant (Student's t‐test).

(F) Transactivation analysis of PRX113, PRX114 and PRX122 promoters by ZOS202 in rice protoplasts. Protoplasts co‐transformed with the 35S empty vector plus reporter plasmid were used as controls. Upper: schematic illustration of the effector and reporter constructs. Values are mean ± SD (n = 4 biologically independent samples). *P < 0.05; **P < 0.01 (Student's t‐test).

(G) Electrophoretic mobility shift assays (EMSAs) showing that ZOS202 binds to the promoters of PRX113, PRX114, and PRX122. GST was used as a negative control; competition for binding was performed with 5× or 20× unlabeled oligos. Plus and minus symbols represent the presence and the absence of components, respectively.

(H) Phenotypes of ZH11 and zos202 mutants treated with 20 μM melatonin, 40 mM H₂O₂, or a combination of 40 mM H₂O₂ and 20 μM MT. Scale bar, 3 cm.

(I–M) Chlorophyll content (I), relative electrolytic leakage (J), MDA content (K), H₂O₂ content (L), peroxidase (POD) activity (M) of leaves from ZH11 and zos202 mutants as shown in (H). Values are mean ± SD (n = 3 biologically independent samples). Different lowercase letters indicate significant differences (P < 0.05, one‐way ANOVA with Tukey's test).

RT‐qPCR showed that PRX113, PRX114, and PRX122 were downregulated in Z‐OE lines and upregulated in zos202 mutants after dark treatment (Figure 6C–E). ZOS202‐binding peaks were detected in their promoters (Figure S10), and direct binding was confirmed by EMSA (Figure 6G). Dual‐luciferase assays demonstrated that ZOS202 represses the promoter activities of PRX113, PRX114, and PRX122 (Figure 6F). These results indicate that ZOS202 directly represses genes encoding ROS‐scavenging peroxidases.

To verify the role of ZOS202 in ROS‐dependent senescence, we treated ZH11 and zos202 mutants with 20 μM melatonin (MT, a well‐characterized ROS scavenger), 40 mM H₂O₂, or a combination of 40 mM H₂O₂ and 20 μM MT. In the absence of H₂O₂, leaf senescence was comparable between ZH11 and zos202 mutants. After 5 days of H₂O₂ treatment, zos202 mutants showed significantly alleviated senescence compared with ZH11. Specifically, the mutants retained higher chlorophyll content and peroxidase activity, and accumulated lower levels of REL, MDA, and H₂O₂ (Figure 6H–M). Melatonin strongly alleviated H₂O₂‐induced senescence in both genotypes and fully abolished the senescence difference between ZH11 and zos202 mutants. By contrast, Z‐OE lines were hypersensitive to H₂O₂ relative to KT, and melatonin partially rescued this phenotype (Figure S11A–F).

Together, these data demonstrate that ZOS202 directly represses a set of peroxidase genes, thereby compromising enzymatic ROS detoxification. The enhanced tolerance to oxidative stress in zos202 mutants can be attributed to their elevated peroxidase activity.

DISCUSSION

Leaf senescence is a highly regulated developmental process that directly determines crop yield and quality, driven by a complex transcriptional regulatory network (Dai et al., 2026; Guo et al., 2025; Ma et al., 2025). Among the transcription factors (TFs) implicated in this process, NAC, WRKY, and MYB family members have been extensively characterized (Li et al., 2023; Lu et al., 2025; Ma et al., 2025; Sasi et al., 2022; Yu et al., 2021; Zheng et al., 2024). C2H2 zinc finger proteins represent one of the largest TF families with well‐established roles in plant growth and stress responses, yet only a few members have been functionally linked to leaf senescence (Bonchuk & Georgiev, 2024; Lu et al., 2024). Here, we identified ZOS202, a C2H2 zinc finger protein, as a positive regulator of leaf senescence. We demonstrated that ZOS202 directly activates multiple senescence‐associated TFs and represses peroxidase genes, leading to chloroplast degradation and ROS accumulation. Our findings establish a key role for ZOS202 and provide new insights into how C2H2 zinc finger protein regulates leaf senescence.

An intriguing finding of this study is that ZOS202 functions as a dual‐role transcription factor, directly activating senescence‐promoting TF genes such as OsWRKY42, OsWRKY53, and ONAC096 while simultaneously repressing ROS‐scavenging peroxidase genes such as PRX113, PRX114, and PRX122. This dual functionality has also been reported for other C2H2 zinc finger proteins: the rice C2H2 protein DST represses NAL1 to modulate leaf width but activates OsNR1.2 to improve nitrogen‐use efficiency, and the citrus C2H2 protein CitZAT4 coordinately activates and represses distinct carotenoid biosynthetic genes to regulate peel pigmentation (Han et al., 2022; Sun et al., 2025; You et al., 2022). However, the mechanistic basis by which a single C2H2 transcription factor discriminates between activation and repression targets remains poorly understood.

Several studies have suggested that transcription factors may function as activators or repressors by binding to distinct motifs in the promoters of downstream target genes (Chakravarthy et al., 2003; Dong et al., 2025). To investigate whether ZOS202 achieves opposing regulation through distinct cis‐elements, we first compared the binding motifs enriched in the promoters of activated targets (OsWRKY42, OsWRKY53, and ONAC096) and repressed targets (PRX113, PRX114, and PRX122). The promoters of activated TF genes (OsWRKY42, OsWRKY53, and ONAC096) were preferentially associated with Motif 5, whereas the promoters of repressed peroxidase genes (PRX113, PRX114, and PRX122) were enriched for Motif 1 (Table S10). We then systematically evaluated the motif distribution between the 118 upregulated DAP‐seq targets and the 117 downregulated DAP‐seq targets (Figure S12; Table S11). Among the upregulated DAP‐seq targets, 127 motifs were identified, with Motif 1 accounting for 42.52% and Motif 5 accounting for 29.92% (Figure S12A). Meanwhile, among the downregulated DAP‐seq targets, 121 motifs were identified, with Motif 1 accounting for 48.76% and Motif 5 accounting for 25.62% (Figure S12B). Although there is a slight increase of Motif 1 in repressed genes than in activated genes (48.76% versus 42.52%) and a slight decrease of Motif 5 in repressed genes than in activated genes (25.62% versus 29.92%), we cannot conclude that the different binding motifs mediate the opposing regulatory activities of ZOS202 to different downstream targets.

It has also been proposed that dual‐function transcription factors confer regulatory specificity via differential protein–protein interactions. The maize transcription factor ZmMYB127 was recently shown to switch between activation and repression by interacting with distinct protein partners (Shi et al., 2026). Hence, identifying the interaction partners of ZOS202 is necessary to fully elucidate the mechanism underlying its dual functionality.

Leaf senescence is intricately regulated by phytohormones, among which salicylic acid (SA) is recognized as a key promoter of senescence (Niu et al., 2020; Zhu et al., 2025). We found that ZOS202 expression was most strongly and rapidly induced by SA compared with other senescence‐promoting hormones, suggesting that ZOS202 acts downstream of SA signaling to promote leaf senescence (Figure 4J). Notably, the spatiotemporal expression pattern of ZOS202 is highly consistent with that of OsWRKY45, which is a central regulator of the SA signaling pathway (Ichimaru et al., 2022; Ueno et al., 2015); both genes peak at the early reproductive stage and decline thereafter (Figure S6B,D). Therefore, future studies on the interplay between ZOS202 and SA signaling in the leaf senescence pathway may offer novel insights into the role of ZOS202.

Numerous transcription factors, particularly from the NAC and WRKY families, have been identified as key regulators of rice leaf senescence (Sasi et al., 2022). Well‐characterized senescence‐associated NAC proteins include OsNAP, ONAC054, ONAC096, OsNAC2, ONAC011, OsNAC109, ONAC106, and ONAC300, while prominent WRKY members include OsWRKY5, OsWRKY10, OsWRKY23, OsWRKY42, and OsWRKY53 (Lee & Masclaux‐Daubresse, 2021). However, the upstream regulatory hierarchy and functional crosstalk among these transcription factors remain largely unclear. Here, we identified ZOS202 as an upstream regulator that directly targets and activates several key senescence‐promoting transcription factors, including OsWRKY42, OsWRKY53, and ONAC096 (Figure 5). These findings place ZOS202 at a pivotal regulatory node that initiates and coordinates a transcriptional cascade during leaf senescence. Future investigations will be needed to further resolve the detailed hierarchical regulatory network among these TFs.

In summary, our findings characterize ZOS202, a nuclear‐localized dual‐function C2H2 transcription factor, as a key positive regulator of rice leaf senescence. ZOS202 promotes senescence through two parallel molecular mechanisms: directly activating a suite of downstream senescence‐associated transcription factors and repressing peroxidase genes, thereby impairing ROS scavenging and disrupting redox homeostasis (Figure 7). Although no yield benefit was achieved by the delayed leaf senescence in zos202 mutants in our current study, future studies on the grain quality and mineral nutrients content will be necessary to fully uncover the potentials of ZOS202 in breeding. Meanwhile, fine‐tuning the expression of ZOS202 may also achieve differential effects on the timing and progression of leaf senescence, providing breeding targets in the future.

Figure 7.

Proposed model for ZOS202‐mediated regulation of rice leaf senescence.

ZOS202 acts as a positive regulator of leaf senescence by promoting chloroplast degradation and disrupting ROS homeostasis through activating senescence‐associated transcription factors and repressing peroxidase‐encoding genes. ZOS202 expression declines as senescence progresses, likely via feedback repression to fine‐tune leaf senescence. On the other hand, knockout of ZOS202 reduces ROS levels and chloroplast degradation, thereby delaying leaf senescence. Arrows indicate promotion, blunt‐ended arrows indicate inhibition, and thicker lines represent enhanced regulation.

MATERIALS AND METHODS

Plant materials and growth conditions

The rice (Oryza sativa L.) varieties Kitaake (KT) and Zhonghua 11 (ZH11) were used in this study. The ZOS202‐knockout mutants were obtained from the ZH11 background CRISPR/Cas9 mutant library of Biogle GeneTech. Two ZOS202‐overexpressing lines were generated in the Kitaake background. ZH11 and zos202 mutants were cultured under natural field conditions at the Chinese Rice Research Institute located in Hangzhou, Zhejiang Province, China. Meanwhile, KT and the Z‐OE lines were grown under natural field conditions in Changchun, Jilin Province, China. For dark induced senescence (DIS) of detached leaves, leaf segments were kept in 3 mM MES buffer (pH 5.8) without light at 28°C. For H₂O₂ treatment, 10‐day‐old seedlings were transferred to hydroponic culture solution containing 20 μM melatonin, 40 mM H₂O₂, or a combination of 40 mM H₂O₂ and 20 μM MT. All seedlings were grown in a walk‐in plant growth room (WIPGC‐CP162, Fujian Jiupo Biotechnology Co., Ltd.) under full‐spectrum white fluorescent light with a photon flux density of 300–350 μmol m⁻² sec⁻¹, under a 12‐h light (30°C)/12‐h dark (28°C) photoperiod cycle. The relative humidity in the greenhouse was maintained at 60%.

Vector construction for generation of transgenic rice

The full‐length coding region of ZOS202 was amplified and cloned into pCAMBIA2301‐ACTIN1 to generate the pCAMBIA2301‐ACTIN1:ZOS202 overexpression construct. These constructs were confirmed by sequencing and transformed by Agrobacterium tumefaciens‐mediated transformation into KT (Liu et al., 2007). All primers used to generate the constructs are listed in Table S1.

Measurement of chlorophyll content and net photosynthetic rate

To assess chlorophyll content, SPAD values were recorded using a SPAD‐502 chlorophyll meter. For each leaf, three readings were taken: one at the midpoint and two at positions 40 mm from the midpoint on either side. Chlorophyll concentration was determined according to previously established methods (Qiu et al., 2019). Briefly, leaf tissues were cut into small pieces and incubated in 10 mL of 95% ethanol in the dark for 24 h. The absorbance of the extracted solution was measured at 645, 470, and 663 nm using a UV spectrophotometer. Net photosynthetic rates were measured using a LI‐6400 portable photosynthesis system (LI‐COR Biosciences, Lincoln, NE, USA). All samples were analyzed in triplicate.

Transmission electron microscopy

For transmission electron microscopy, leaf samples were fixed in 2.5% glutaraldehyde in phosphate buffer, followed by standard sample preparation procedures. Transmission electron microscopy was performed following the protocol described previously (Liu et al., 2018). The procedure consisted of the following steps: Samples were sectioned into small pieces, fixed in 2.5% glutaraldehyde at 4°C for a minimum of 4 h, post‐fixed in 1% osmium tetroxide for 2 h, stained with uranyl acetate, dehydrated through a graded ethanol series, and embedded in Spurr's medium. Observations were carried out using a Hitachi H‐7650 transmission electron microscope.

Histochemical analysis

Nitro blue tetrazolium (NBT) staining, 3,3′‐diaminobenzidine (DAB) staining, and the 2′,7′‐dichlorofluorescein diacetate (H₂DCFDA) assay were performed according to previously established protocols (Han et al., 2012). Briefly, leaf samples were immersed in 0.05% (w/v) NBT solution or 0.1% (w/v) DAB solution, followed by incubation at 37°C for 16 h in the dark. After staining, the leaves were transferred to 95% ethanol to elute chlorophyll and subsequently transferred to 70% glycerol for imaging. For the H₂DCFDA assay, leaf samples were incubated with 10 mM H₂DCFDA dissolved in 10 mM Tris–HCl buffer (pH 7.2) for 30 min at room temperature. Fluorescence signals of the samples were observed and imaged using a confocal laser scanning microscope (Leica STELLARIS 5). Trypan blue (TB) staining was carried out following the procedure outlined in a prior study (Yin et al., 2000). Specifically, leaves were submerged in a lactic acid‐phenol‐Trypan blue solution (composed of 25% lactic acid, 23% water‐saturated phenol, 0.25% Trypan blue, and 25% glycerol) and boiled for 10 min. After incubation in the dark overnight, the stained leaf samples were decolorized in 25 mg/mL chloral hydrate solution for 3 consecutive days.

Measurement of relative electrolytic leakage, H₂O₂ , ${\mathrm{O}}_2^{·-}$, MDA, and POD

Relative electrolytic leakage (REL) was evaluated by cutting leaves into 1‐cm segments and submerging them in 10 mL of deionized water in test tubes, followed by incubation at room temperature for 12 h. The initial conductivity (R1) of the solution was measured using a DDSJ‐308A conductivity meter. The test tubes were then placed in boiling water for 30 min and then cooled naturally to room temperature, after which the total conductivity (R2) was measured. REL was calculated as the percentage of R1 relative to R2. The levels of hydrogen peroxide (H₂O₂) and superoxide anion (${\mathrm{O}}_2^{·-}$) were measured using specific assay kits (Grace Biotech, Suzhou, China). The malondialdehyde (MDA) content was determined following the methods described in Moradi and Ismail (2007). Peroxidase (POD) activities were assessed according to previously established protocols (Ashraf et al., 2017).

Phytohormone treatments

Leaves from 2‐week‐old wild‐type plants were excised and floated on 3 mM MES buffer (pH 5.8) containing each phytohormone at the indicated concentrations: abscisic acid (ABA; 20 or 100 μM), methyl jasmonate (MeJA; 20 or 100 μM), salicylic acid (SA; 100 or 500 μM), or 1‐aminocyclopropane‐1‐carboxylic acid (ACC; 100 or 500 μM). Control leaves were floated on hormone‐free MES buffer. All treatments were performed under continuous white light at 28°C for 24 h, as described previously (Sakuraba et al., 2020).

RNA extraction and RT‐qPCR

Total RNA was extracted using Trizol Reagent (Invitrogen). About 2 μg total RNA was used to synthesize the single‐strand cDNA with a reverse transcription kit (TransScript® One‐Step gDNA Removal and cDNA Synthesis SuperMix). RT‐qPCR was performed using TransStart® Green qPCR SuperMix on a Bio‐Rad CFX Opus 96 system. Three technical replicates were carried out for each gene. Primers used are listed in Table S1.

Subcellular localization

To investigate the subcellular localization of ZOS202, the full‐length coding region of ZOS202 was fused in frame to the N terminus of the enhanced GFP (eGFP) coding sequence under the control of CaMV 35S promoter in the pCAMBIA2300‐35S:eGFP vector. Plasmids were extracted and purified using DP118 EndoFree Midi II (Tiangen, Beijing, China). The construct was transfected into rice protoplasts as described previously (Zhang et al., 2011). Fluorescence imaging of the protoplasts was carried out using a confocal laser microscope (Leica STELLARIS 5). Primers used are listed in Table S1.

RNA‐seq assays

Total RNAs were extracted from flag leaf with three biological replicates of the KT, ZOS202‐OE1, and ZOS202‐OE2 strains respectively at 80 DPS grown in field and RNA sequencing was performed by Beijing Biomarker Technologies Co., LTD. The DESeq2 software was employed to identify differentially expressed genes with a false discovery rate (FDR) of less than 0.01 and |Log₂FC| greater than ≥1. BMKCloud (www.biocloud.net) was utilized for the generation of venn plots, heat maps, as well as GO and KEGG analyses.

DNA affinity purification sequencing analysis

The DAP‐seq experiment was conducted following the methodology of previous studies (Bartlett et al., 2017). Genomic DNA (gDNA) was extracted and purified from Kitaake rice plants, fragmented, and then ligated with short DNA sequencing adapters to construct a DAP‐seq library. This work was carried out by Lan Jing Ke Xin Technology Co., Ltd. (Hebei, China). The transcription factor ZOS202 was prepared through in vitro expression, linked to affinity tags (Halo, GST), and co‐constructed into an expression vector, followed by purification using Magne HaloTag magnetic beads. The gDNA library was added to ZOS202 bound to the affinity tags, and unbound DNA was washed away. The bound fraction was eluted, amplified with PCR primers, indexed adapters were introduced, and the DNA was sequenced. By aligning the sequencing data to the reference genome, the target genes of ZOS202 and the recognition motifs were identified using the enriched sites (peaks).

GAL4‐dependent chimeric transactivation assay

The GAL4‐dependent chimeric transactivation assay was performed in rice protoplasts as described previously (Guo et al., 2013). The effectors were constructed by fusing the full‐length coding region of ZOS202 to the GAL4 DNA‐binding domain. The reporter plasmid GAL4‐LUC includes five repeats of the yeast GAL4 protein binding site and the minimal TATA region of the CaMV 35S promoter located upstream of the firefly LUC gene (fLUC). The reference plasmid pPTRL, including a Renilla LUC gene (rLUC) under the control of CaMV 35S promoter, was used as an internal control. LUC activity was measured using the Dual‐Luciferase® Reporter Assay System (Vazyme, DL101‐01, China) with a LumiPro luminometer (YPH‐Bio, Beijing, China). The results were reported as the ratio between the activity of firefly LUC (fLUC) and Renilla LUC (rLUC). Primers used for constructing the vectors are listed in Table S1.

The transient dual‐LUC reporter system

For LUC reporter constructs, the promoters of ONAC096, OsWRKY42, OsWRKY53, OsPRX113, OsPRX114, and OsPRX122, amplified from genomic DNA were inserted into the linearized pGreen0800‐LUC vector sequentially by recombination using the ClonExpress II One Step Cloning Kit (Vazyme). The effector was constructed by fusing the full‐length coding region of ZOS202 to the pCAMBIA2300‐35S vector, with the empty pCAMBIA2300‐35S vector as the control. Different combinations of vectors were transfected into rice protoplasts according to the methods described previously (Zhang et al., 2011). LUC activity was quantified as described above. Primers used for constructing the vectors are listed in Table S1.

Electrophoretic mobility shift assay (EMSA)

To verify a direct target of ZOS202 in vitro, the full‐length CDS of ZOS202 was cloned into the pGEX‐4T‐2 vector to generate GST fusion proteins. The resulting GST‐ZOS202 construct and the empty GST vector were transformed into the Escherichia coli strain BL21 (DE3) for protein expression, and the expressed proteins were purified using GST Resin. Both 6‐FAM 5′‐end‐labeled and unlabeled probes were synthesized by Sangon Bio‐Tech Company (Shanghai, China). Electrophoretic mobility shift assays (EMSA) were conducted as previously described (Gao et al., 2019). Fluorescence signals in the gel were visualized using a Tanon 5200Multi fluorescence imaging system (Tanon, Shanghai, China). The primers used for EMSA are listed in Table S1.

AUTHOR CONTRIBUTIONS

ZZ, ZL, and XD conceived and designed the experiments. ZZ and WZ analyzed the data and wrote the manuscript; WZ and ZG jointly performed the experiments with the assistance of SL, YW, JZ, SG, JY, YB, YL, BZ, SH, and WJ; CC revised the manuscript; WZ and ZG contributed equally to this work.

CONFLICT OF INTEREST

The authors declare that they have no conflict of interest.

Supporting information

Figure S1. Identification of ZOS202 knockout mutants.

Figure S2. Identification of ZOS202 overexpression transgenic lines.

Figure S3. Validation of the expression of senescence‐related genes in zos202 mutants, OE lines their WT under dark treatment by RT‐qPCR.

Figure S4. Agronomic traits of zos202 mutants and ZOS202 overexpression lines.

Figure S5. ZOS202 promotes ROS accumulation and cell death in leaves at mature stage under field‐grown conditions.

Figure S6. Transcriptional repression activity analysis of ZOS202, spatiotemporal expression profiles of ZOS202, OsNAP, and OsWRKY45 from RiceXPro.

Figure S7. RNA‐seq analysis of Z‐OE lines and KT.

Figure S8. Identification of the genome‐wide direct targets of ZOS202 by DAP‐seq.

Figure S9. Expression cluster analysis of genes enriched in GO term “transcription factor activity”.

Figure S10. Detection of ZOS202 binding to the promoters of PRX113, PRX114 and PRX122 by DAP‐seq.

Figure S11. ZOS202‐overexpressing lines are more sensitive to H₂O₂.

Figure S12. Distribution of motifs in ZOS202‐upregulated and ‐downregulated genes.

Table S1. Primers used in this study.

Table S2. Summary of RNA‐seq reads alignment statistics.

Table S3. Commonly upregulated DEGs in Z‐OE1 and Z‐OE2 compared to KT.

Table S4. Commonly downregulated DEGs in Z‐OE1 and Z‐OE2 compared to KT.

Table S5. Summary of DAP‐seq reads alignment statistics.

Table S6. ZOS202‐bound genes identified by DAP‐seq.

Table S7. ZOS202‐bound genes (≤3000 bp upstream of transcription start site) identified by DAP‐seq.

Table S8. Overlapping genes between DAP‐seq target genes (≤3000 bp upstream of the transcription start site) and upregulated DEGs in Z‐OE lines.

Table S9. Overlapping genes between DAP‐seq target genes (≤3000 bp upstream of the transcription start site) and downregulated DEGs in Z‐OE lines.

Table S10. Motifs identified in the promoter of OsWRKY42, OsWRKY53, ONAC096, PRX113, PRX114, and PRX122 by DAP‐seq.

Table S11. Motifs identified in the promoters (≤3 kb) of genes that were both DAP‐seq targets and RNA‐seq DEGs (up‐ and downregulated).

DATA AVAILABILITY STATEMENT

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive in the National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA039535) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa.