The OsUVR8-OsNAC3-OsERF117 signaling module mediates metabolic acclimation and climate adaptation in rice Authors: Wenxin Liu, Dongxiao Zhou, Jianying Xing, Shangyong Xue, Tian-Jing Wang, Peng Guo, Yutong Liu, Zheng-Yi Xu Published: 2026 ABSTRACT Ultraviolet-B (UV-B; 280 to 315 nanometers) radiation increasingly threatens crop productivity, yet the genetic basis of plant adaptation remains poorly understood. We delineate a UV-B signaling module in rice that links photoreceptor activation to transcriptional reprogramming and metabolic acclimation. The AP2/ERF transcription factor OsERF117 acts as a central regulator, directly activating flavonoid and melatonin biosynthetic genes to drive photoprotective metabolite accumulation and enhance UV-B stress tolerance. Promoter variation in OsERF117 defines 10 haplotypes across 4093 rice accessions, with high-expression haplotypes enriched in high-UV-B regions and correlated with adaptive divergence. OsERF117 is transcriptionally activated by OsNAC3, with a cis-regulatory SNP at an OsNAC3-binding site modulating responsiveness and contributing to subspecies diversification. Genetic and biochemical evidence supports a model in which UV-B–activated OsUVR8 promotes OsNAC3 activity and antagonizes OsCOP1-mediated ubiquitination and degradation in rice. This OsUVR8-OsCOP1-OsNAC3-OsERF117 module reveals how UV-B perception drives regulatory and metabolic diversification, offering targets for breeding UV-B–resilient crops. INTRODUCTION The ultraviolet-B (UV-B; 280 to 315 nm) radiation is a persistent and intensifying constraint on crop productivity, causing DNA, protein, and lipid damage that disrupts plant growth and development (1–3). Climate change is expected to heighten this threat: Altered stratospheric temperatures and atmospheric circulation are projected to slow ozone recovery and create regionally heterogeneous increases in surface UV-B, particularly across low-latitude agricultural zones (4, 5). These changes directly jeopardize tropical and subtropical cropping systems, where plants already contend with heat, drought, and other compounding stresses. In rice, elevated UV-B provokes a stress syndrome resembling heat and drought responses—including stomatal closure, leaf thickening, enhanced flavonoid accumulation, and reduced photosynthetic performance (6, 7). Although protective, these responses impose metabolic and physiological costs, and field and controlled-environment studies report associated yield losses of 5 to 25% (8). Together, these converging pressures highlight a critical gap in understanding how cereals perceive, transduce, and adapt to UV-B. Defining the genetic, molecular, and metabolic mechanisms that underlie UV-B acclimation is therefore essential for developing climate-resilient crop varieties and safeguarding global food security. Decades of research in the model dicot Arabidopsis thaliana have defined a robust molecular framework for UV-B perception. In its ground state, UV RESISTANCE LOCUS 8 (UVR8) forms an inactive homodimer; absorption of UV-B photons triggers charge neutralization within interfacial tryptophan clusters, promoting rapid monomerization (9–11). Monomeric UVR8 engages the WD40 domain of CONSTITUTIVELY PHOTOMORPHOGENIC 1 (COP1) with high affinity, displacing COP1 from the CUL4-DDB1-COP1-SPA E3 ligase and reducing COP1-mediated ubiquitination of central transcriptional regulators, including the bZIP factor ELONGATED HYPOCOTYL 5 (HY5) and its homologs (12, 13). In parallel, UVR8-COP1 complex formation contributes to the destabilization of bHLH PHYTOCHROME-INTERACTING FACTOR (PIF) repressors, while UVR8 independently interacts with additional transcription factor families—such as BRI1EMS-SUPPRESSOR 1 (BES1), BES1-INTERACTING MYC-LIKE 1 (BIM1), MYB DOMAIN PROTEIN 13 (MYB13), MYB73, MYB77, and WRKY DNA BINDING PROTEIN 36 (WRKY36)—to modulate their DNA binding activities (14–17). These combined mechanisms orchestrate broad transcriptional reprogramming that underpins UV-B–induced photomorphogenesis and acclimation. Feedback regulation is conferred by the WD40-repeat proteins REPRESSOR OF UV-B PHOTOMORPHOGENESIS1 (RUP1) and RUP2, which promote UVR8 redimerization and restoration of the inactive ground state (18, 19). However, recent work indicates that the UVR8-COP1 relationship in rice differs from this canonical model; specifically, OsUVR8a and OsUVR8b accumulate in the nucleus independently of UV-B and OsCOP1, and OsCOP1 acts as a negative regulator of rice UV-B responses (20). Although UVR8 and related UV-B signaling components are widely conserved across the algae to angiosperms (2, 21–23), their functional relationships and regulatory outputs may differ between species. Nevertheless, current mechanistic understanding is based largely on studies in A. thaliana, leaving key questions unresolved in crops (1, 22, 24). How are UV-B cues integrated into crop-specific transcriptional networks that govern growth and metabolic adaptation? Which natural polymorphisms modulate UV-B responsiveness across rice subspecies adapted to distinct latitudinal environments? In addition, critically, how can these mechanisms be harnessed to improve UV-B resilience without impairing yield under multifactorial climate stress? Here, we delineate a UV-B signaling module in rice that links OsUVR8 perception to stabilization of the NAC transcription factor OsNAC3—through antagonism of OsCOP1-mediated degradation—and subsequent activation of the AP2/ERF regulator OsERF117. This OsUVR8-OsCOP1-OsNAC3-OsERF117 module coordinates flavonoid, melatonin, and γ-aminobutyrate [γ-aminobutyric acid (GABA)] metabolism to drive acclimation, with promoter haplotypes shaping adaptive variation across subspecies. Our findings reveal how environmental light cues sculpt heritable regulatory architectures, providing a molecular blueprint for breeding UV-B–resilient crops in an era of climatic volatility. RESULTS OsERF117 positively regulates UV-B–induced photomorphogenesis and confers enhanced tolerance to environmental stress To identify genetic loci involved in UV-B–induced photomorphogenesis, we conducted a mutant screen using an activation-tagging mutant pool generated in the rice cultivar Kitaake (25) under short-day conditions (10-hour light/14-hour dark photoperiod). The screen was performed in both white light (WL) and WL supplemented with narrowband UV-B light at 30 μW cm−2 (equivalent to ∼4.5 μmol m−2 s−1) during the daytime. As controls, we generated osuvr8a osuvr8b double mutants (osuvr8ab) and an osrup mutant via CRISPR-Cas9–mediated gene editing in the Kitaake background (fig. S1A). Two independent osuvr8ab lines showed no notable changes in seedling height in response to UV-B, reflecting impaired UV-B–induced growth inhibition (Fig. 1A). In contrast, two independent osrup mutant lines displayed enhanced sensitivity to UV-B, as evidenced by increased seedling growth inhibition (Fig. 1A). Moreover, under broadband UV-B irradiation (500 μW cm−2, equivalent to ∼75 μmol m−2 s−1), osuvr8ab seedlings displayed reduced survival, whereas osrup mutants showed increased survival rates (fig. S1, B and C). The AC37 line, exhibiting impaired growth inhibition, was identified through thermal asymmetric interlaced polymerase chain reaction (TAIL-PCR) to have a transferred DNA (T-DNA) insertion within the first exon of OsKitaake11g045700, which encodes OsERF117 (Fig. 1, A to D). This insertion led to a complete loss of OsERF117 transcript accumulation in the AC37 line, while neighboring genes remained unaffected (Fig. 1E). OsERF117 is a member of the AP2 transcription factor subfamily, characterized by a highly conserved N-terminal DNA binding domain and a C-terminal variable region with transcriptional activation activity (fig. S2, A to C). It predominantly localizes to the nucleus (fig. S2D). Promoter-GUS assays and quantitative reverse transcription–PCR (RT-qPCR) revealed that OsERF117 is broadly expressed across multiple tissues, including seeds, roots, internodes, young panicles, leaf sheaths, and leaves (fig. S2, E and F). Notably, OsERF117 transcript levels are rapidly induced following UV-B exposure, suggesting a role in early stress-responsive signaling (fig. S2G). Two independent loss-of-function mutants of OsERF117, generated using the CRISPR-Cas9 system, also exhibited impaired growth inhibition, which was rescued in the complementation lines (Fig. 1, F and G, and fig. S3, A to C). Furthermore, the oserf117 mutants displayed reduced survival rates under broadband UV-B exposure (Fig. 1, H and I); however, OsERF117 overexpression lines showed increased seedling growth inhibition during photomorphogenesis and UV-B stress tolerance (fig. S3, D to I). Notably, oserf117 mutants showed reduced plant height, tiller number, panicle length, and root biomass accumulation compared to both the wild-type Kitaake and complementation lines (fig. S4, A to H). Together, these results suggest that OsERF117 plays a crucial role in UV-B–induced photomorphogenesis, UV-B stress tolerance, and overall plant growth. OsERF117 positively regulates UV-B–triggered growth inhibition and enhances UV-B tolerance. (A) Seedling phenotypes of wild-type (Kitaake), AC37, osuvr8ab-1, and osrup-1 mutants under WL and WL supplemented with narrowband UV-B. Scale bar, 1 cm. (B) Relative inhibition rate of seedling height in (A). n = 60 seedlings per line. (C) Mapping of the T-DNA insertion site in the AC37 mutant by TAIL-PCR. (D) Genomic organization of 10-kb flanking the T-DNA insertion site in AC37. (E) Semiquantitative RT-PCR analysis of five selected gene transcripts in Kitaake (wild-type) and AC37 plants. The primer sets are indicated by blue arrows in (D). OsACT2 was used as a loading control. (F and G) Seedling phenotypes (F) and relative inhibition rate of seedling height (G) of oserf117-1 and oserf117-2 mutants and complemented lines (Com) under WL and WL supplemented with narrowband UV-B. n = 60 seedlings per line. Scale bar, 1 cm. (H) Phenotypes of 2-week-old Kitaake, oserf117-1 and oserf117-2 mutants, and complemented lines following broadband UV-B treatment and recovery. (I) Survival rates of Kitaake, oserf117-1 and oserf117-2 mutants, and complemented lines under broadband UV-B (n = 8). Scale bar, 6 cm. In (B), (G), and (I), data are means ± SD, and different lowercase letters indicate statistically significant differences (P < 0.05) by one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. OsERF117 orchestrates a UV-B–responsive transcriptional and metabolic network To define the transcriptional network governed by OsERF117 under UV-B stress, we integrated DAP-seq binding profiles with RNA sequencing (RNA-seq) datasets generated from Kitaake, oserf117-1, and osuvr8ab-1 seedlings after broadband UV-B treatment (0 hours of versus 6 hours of UV-B exposure). Differential expression analysis identified genes that were UV-B responsive in wild type but misregulated in both mutants, revealing a core set of UV-B–induced genes coregulated by OsERF117 and OsUVR8 (fig. S5A and table S1). The DAP-seq assay showed strong reproducibility across biological replicates, with 44,487 overlapping peaks identified between two independent runs (fig. S5B). A heatmap of normalized signal intensities confirmed that peak enrichment is sharply centered within promoter regions (fig. S5C). Genomic distribution analysis further revealed that OsERF117 binding is predominantly associated with promoters (49.69%), with smaller fractions mapping to transcription termination sites and to intergenic or intronic regions (fig. S5D), consistent with its role as a transcriptional activator. This integrative approach revealed a strong convergence between transcriptional responses and direct DNA binding regulation by OsERF117. RNA-seq identified 533 genes that were UV-B responsive in Kitaake but misregulated in both oserf117-1 and osuvr8ab-1 mutants, indicating joint control by OsERF117 and OsUVR8. DAP-seq analysis, encompassing 26781 OsERF117-bound loci (table S2), further delineated its direct genomic targets. Intersection of these datasets yielded 317 high-confidence genes coregulated by OsERF117 and OsUVR8 (Fig. 2A). Gene Ontology enrichment of this set highlighted processes associated with UV-B adaptation and secondary metabolism, including flavonoid biosynthetic process, chalcone biosynthetic process, positive regulation of anthocyanin metabolic process, and response to UV-B (Fig. 2B). These categories underscore a coordinated transcriptional program linking UV-B perception to flavonoid biosynthesis and photoprotective metabolism. Among these is the flavonoid-biosynthetic enzyme OsCHS1 (LOC_Os11g32650), which encodes a chalcone synthase that channels p-coumaroyl-CoA into naringenin chalcone, thereby committing flux into the flavonoid pathway. In rice, CHS marks the first dedicated step in flavonoid production and may act as a bottleneck under UV-B stress (20), when flavonoid accumulation is up-regulated to protect photosynthetic tissues from oxidative damage. A further component of this regulatory module is uridine diphosphate–glycosyltransferase OsUGT88C2 (LOC_Os01g53390), a homolog of the malvidin 3-O-galactoside synthase OsUGT88C3. OsUGT88C2 catalyzes a key modification in flavone maturation and acts as a rheostat for UV-B acclimation, fine-tuning the accumulation of glycosylated flavones that buffer excess radiation and alleviate UV-B stress in photosynthetic tissues. In parallel, OsERF117 directly engages OsGABA-T (LOC_Os02g02210), encoding γ-aminobutyrate transaminase, which channels stress-induced γ-aminobutyrate (GABA) through the GABA shunt to form succinic semialdehyde (24). OsGABA-T mitigates the accumulation of reactive oxygen intermediates and suppresses cell death under UV-B stress. OsASMT15 (LOC_Os10g02880), encoding N-acetylserotonin O-methyltransferase (ASMT), completes this photoprotective circuit. As the terminal enzyme in melatonin biosynthesis, OsASMT15 elevates endogenous melatonin levels, reinforcing antioxidant capacity and enhancing tolerance to UV-B–induced oxidative stress. Notably, RNA-seq heatmaps showed strong UV-B induction of these genes in Kitaake but not in oserf117-1 mutants (Fig. 2C), underscoring their dependence on OsERF117 activity. To validate the RNA-seq results, we performed RT-qPCR analysis of four representative target genes, OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15. Consistent with the transcriptome data, all four genes were strongly induced by UV-B in Kitaake, whereas their induction was markedly reduced in oserf117-1 and osuvr8ab-1 mutants (fig. S5E). OsERF117 orchestrates a UV-B–responsive transcriptional and metabolic network. (A) Venn diagram showing the overlap between UV-B–responsive differentially expressed genes identified by RNA-seq and OsERF117-bound genes identified by DAP-seq. (B) Gene Ontology enrichment analysis of OsERF117-regulated genes under UV-B, P values were calculated using Fisher’s exact test. (C) Heatmap showing transcript abundance of four UV-B–responsive genes in Kitaake and oserf117-1 mutants. Color scale denotes log2 (fold change). (D) Diagrams showing OsERF117 binding to the promoters of OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15, consistently detected in two independent DAP-seq replicates. (E) ChIP-qPCR validation of OsERF117 binding to the promoters of OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15 in Kitaake and OsERF117pro::FLAG-OsERF117 Plants. Data are mean ± SD (n = 3). Different letters indicate significant differences determined by two-way ANOVA followed by Tukey’s multiple comparisons test (P < 0.05). (F) Hierarchical clustering of flavonoid abundance in Kitaake and oserf117-1 leaves under WL and WL + broadband UV-B conditions. Z-scores indicate relative abundance across samples. Promoter analysis pinpointed potential OsERF117 target recognition sites, with a predominant GCCGCC motif being common to multiple UV-B–responsive gene promoters (fig. S5F), as validated by electrophoretic mobility shift assays (EMSAs) (fig. S5G). Dendrogram analysis of the DAP-seq data confirmed that OsERF117 directly binds the promoters of OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15, providing genome-wide evidence of sequence-specific DNA binding (Fig. 2D). Chromatin immunoprecipitation (ChIP)–qPCR in OsERF117-complemented plants confirmed in vivo promoter occupancy at these loci (Fig. 2E), and luciferase reporter assays established OsERF117 as a potent transcriptional activator (fig. S5H). Metabolomic profiling further illuminated the functional consequences of OsERF117 regulation. Principal components analysis of metabolic data clearly separated genotypes along the first axis and light treatments along the second (fig. S6A and tables S3 and S4), highlighting distinct metabolic states. Under UV-B exposure, wild-type leaves accumulated a suite of photoprotective flavonoids and tannins. Flavonoids are central to photoprotection, acting as both antioxidants and UV filters. Compounds such as quercetin exhibit high antioxidant efficacy and superior capacity to dissipate absorbed UV energy; hesperetin directly scavenges reactive oxygen species (ROS) and absorbs UV-B radiation; and proanthocyanidins, as potent free-radical quenchers, collectively prevent oxidative and cellular damage under UV-B irradiation (26–28). In contrast, oserf117-1 mutants failed to induce this protective response (Fig. 2F). To further validate the metabolomic results and assess whether these metabolic changes are associated with oxidative stress protection, we performed targeted quantification of representative metabolites and ROS-staining assays. Under broadband UV-B treatment, Kitaake seedlings showed increased accumulation of proanthocyanidins, quercetin, melatonin, and GABA, whereas these UV-B–induced increases were markedly attenuated in oserf117-1 and osuvr8ab-1 mutants. By contrast, osrup-1 seedlings accumulated higher levels of these metabolites under UV-B (fig. S6, B to E). We further examined ROS accumulation using nitrotetrazolium blue chloride (NBT) and diaminobenzidine (DAB) staining, which detect H2O2 and O2−, respectively. Following UV-B treatment, oserf117-1 and osuvr8ab-1 seedlings showed stronger DAB and NBT staining than Kitaake, whereas osrup-1 showed weaker staining (fig. S6F). These results provide targeted biochemical and physiological validation that OsERF117- and OsUVR8-dependent metabolic activation contributes to UV-B–induced antioxidant protection. Because OsERF117 also regulates UV-B–induced seedling growth inhibition, we further examined OsERF117-dependent transcriptional responses under narrowband UV-B (table S5). RNA-seq analysis of Kitaake and oserf117-1 seedlings revealed that genes misregulated in oserf117-1 were enriched for Gene Ontology terms associated with photomorphogenesis and brassinosteroid homeostasis (fig. S7, A and B). We therefore examined four representative genes involved in light-regulated seedling development and hormone-dependent elongation growth: OsHY5a/OsbZIP1, OsPIL11, CYP90D2/D2, and OsGA20ox1. OsHY5a/OsbZIP1 is a rice HY5 homolog involved in skoto- and photomorphogenesis, whereas OsPIL11 is a PIF-like regulator associated with skotomorphogenic growth. CYP90D2/D2 encodes a brassinosteroid-biosynthetic enzyme, and OsGA20ox1 encodes a gibberellin 20-oxidase that promotes early seedling growth. Under UV-B, OsHY5a was induced, whereas OsPIL11, CYP90D2, and OsGA20ox1 were repressed in Kitaake seedlings. By contrast, these UV-B–dependent transcriptional changes were strongly attenuated in oserf117-1, as confirmed by RT-qPCR (fig. S7, C and D). Together, these results indicate that OsERF117 coordinates two interconnected outputs of UV-B signaling: photoprotective metabolic acclimation, which supports UV-B stress tolerance, and transcriptional modulation of light- and hormone-related growth regulators, which contributes to UV-B–induced photomorphogenesis and seedling growth inhibition. Haplotype variation in OsERF117 drives rice adaptation to UV-B environments To investigate the role of OsERF117 in rice adaptation to natural UV-B habitats, we analyzed 3990 RiceVarMap v2.0 accessions with documented geographic origins, together with 103 germplasms we newly sequenced, yielding a panel of 4093 accessions. Because the OsERF117 coding sequence lacks synonymous mutations, we focused on promoter polymorphisms and identified 10 distinct haplotypes. Phylogenetic analysis revealed two major groups: group A, enriched in indica with minor contributions from japonica, aus, and intermediates, and group B, largely japonica with smaller fractions of other subspecies (Fig. 3, A and B, and table S6). Geographic distribution further showed that group A haplotypes dominate in regions with high UV-B intensity, whereas group B haplotypes are enriched in areas with lower UV-B exposure, suggesting that promoter variation in OsERF117 underlies rice adaptation to diverse UV-B environments (Fig. 3C and table S7). Under UV-B irradiation, the expression level of OsERF117 was significantly elevated in group A compared to group B. This result is consistent with the phenotypic data, which showed that group A exhibited stronger induced photomorphogenic responses (Fig. 3, D and E). To test this hypothesis, we developed near-isogenic lines (NILs) between indica 9311 (group A) and japonica Nipponbare (NIP) (group B). NIL-OsERF1179311 showed stronger growth inhibition, higher UV-B–induced OsERF117 expression, and higher UV-B stress tolerance than NIL-OsERF117NIP, indicating that natural haplotype variation in OsERF117 modulates divergent UV-B responses in rice (Fig. 3, F to J). To dissect subspecies-specific functions, two independent oserf117 CRISPR mutants were generated in both ZH11 and 9311 backgrounds. Under narrowband UV-B, relative growth was reduced more in 9311 (52%) than in ZH11 (62%), consistent with enhanced UV-B sensitivity in the indica background. Disruption of OsERF117 conferred UV-B insensitivity (∼81 to 83% relative growth) in both genetic contexts, with a stronger rescue in 9311 (fig. S8, A to C). Under broadband UV-B stress followed by recovery, clear survival phenotypes were observed in both genetic backgrounds. Wild-type ZH11 and 9311 seedlings showed substantially higher survival rates than their respective oserf117 mutants, demonstrating that loss of OsERF117 compromises UV-B stress tolerance in both subspecies. Notably, the magnitude of the survival defect differed between backgrounds, with the 9311 background showing a stronger reduction in survival than ZH11 (fig. S8, D and E). These results indicate that OsERF117 contributes to UV-B stress tolerance across genetic backgrounds while also reinforcing the divergent UV-B responsiveness between indica and japonica rice. Haplotype variation in OsERF117 drives rice adaptation to UV-B environments. (A) Phylogenetic relationships among the 10 OsERF117 promoter haplotypes, grouped into group A (Hap2, Hap3, Hap6, and Hap8) and group B (Hap1, Hap4, Hap5, Hap7, Hap9, and Hap10). (B) Haplotype network of OsERF117 from RiceVarMap v2.0. Circle size reflects the number of accessions within each haplotype, and colors denote subspecies identity. (C) Geographic distribution of 4093 rice accessions classified into group A and group B. Pie charts indicate the proportion of group A (green) and group B (white) genotypes at each location. The map is overlaid with global UV-B radiation intensity (red, high; blue, low). (D) Relative OsERF117 expression under broadband UV-B treatment in group A and group B accessions. Group A, n = 89 accessions; group B, n = 50 accessions; RT-qPCR was performed with three biological replicates. (E) UV-B–induced seedling growth inhibition in representative group A and group B accessions. (F and G) Seedling phenotypes (F) and relative inhibition of seedling height (G) for NIL-OsERF117NIP and NIL-OsERF1179311 under WL and WL supplemented with narrowband UV-B (n = 30 seedlings per line). Scale bar, 1 cm. (H) RT-qPCR analysis of OsERF117 expression in NIL-OsERF117NIP and NIL-OsERF1179311 in response to broadband UV-B. OsUBQ10 served as the internal control. Data are mean ± SD (n = 3). Different letters indicate significant differences determined by two-way ANOVA followed by Tukey’s multiple comparisons test (P < 0.05). (I) Phenotypes of 2-week-old NIL-OsERF117NIP and NIL-OsERF1179311 seedlings before and after broadband UV-B treatment and recovery. Scale bar, 3 cm. (J) Survival rates of NIL-OsERF117NIP and NIL-OsERF1179311 under broadband UV-B (n = 8). In (D), (E), (G), and (J), data are mean ± SD. Student’s t test was used for significance analysis. **P < 0.01. To investigate the transcriptional control of OsERF117 under UV-B, we performed yeast one-hybrid (Y1H) screening using the OsERF117 promoter as bait, which identified seven putative regulators (table S8). To further define potential regulatory elements, we analyzed the OsERF117 promoter with PlantPAN4.0, revealing several candidate cis-elements (table S9). By integrating Y1H data with PlantPAN4.0 predictions, we pinpointed OsNAC3 as a strong candidate binding factor (Fig. 4A and fig. S9A). Functional assays demonstrated that these cis-elements mediate OsERF117 promoter activation (fig. S9, B to D). Notably, a single nucleotide variation at −1774 (C in group B versus A in group A) within the second NAC-binding site markedly enhanced OsNAC3-dependent promoter activation, resulting in substantially higher OsNAC3-mediated activation of the group A promoter compared with that of group B (Fig. 4B and fig. S9E). EMSA and ChIP-qPCR confirmed stronger OsNAC3 binding to the group A element (Fig. 4, C and D, and fig. S9F). To establish in vivo function, we generated two independent osnac3 CRISPR-Cas9 mutants (fig. S9G). Both mutants showed attenuated growth inhibition under narrowband UV-B and reduced survival under broadband UV-B (Fig. 4, E to H). Moreover, UV-B induction of OsERF117 was markedly compromised in osnac3 plants (fig. S9H). Consistent with OsNAC3 acting upstream of OsERF117 in this pathway, targeted metabolite analysis showed that UV-B–induced accumulation of proanthocyanidins, quercetin, melatonin, and GABA was also reduced in osnac3-1 and osnac3-2 compared with Kitaake (fig. S10, A to D). Moreover, DAB and NBT staining revealed stronger ROS accumulation in osnac3 mutants after UV-B treatment, similar to the phenotypes observed in oserf117-1 and osuvr8ab-1 mutants (fig. S10E). These data support the conclusion that OsNAC3 is required for full activation of the OsERF117-dependent photoprotective metabolic response under UV-B. Together, these results establish OsNAC3 as a key trans-acting regulator of OsERF117, driving natural variation in rice adaptation to UV-B conditions. OsNAC3 transcriptionally activates OsERF117 to regulate UV-B responses. (A) Y1H assay showing specific binding of OsNAC3 to the OsERF117 promoter, assessed by yeast growth on SD/-Leu medium supplemented with aureobasidin A (700 ng/ml; AbA). (B) Transactivation assays in rice protoplasts indicate that coexpression of OsNAC3 significantly enhances LUC activity driven by the OsERF117 promoter from group A and group B under narrowband UV-B. AtUBQ10pro::GUS served as an internal control. Data are mean ± SD (n = 3). Different letters indicate significant differences determined by two-way ANOVA followed by Tukey’s multiple comparisons test (P < 0.05). (C) EMSA showing specific binding of recombinant His-OsNAC3 to biotin-labeled DNA probes corresponding to the OsERF117 promoter from group A and group B. His-GFP was used as a negative control. (D) ChIP-qPCR analysis in 2-week-old transgenic 9311 and NIP plants expressing FLAG-OsNAC3 from its native promoter (OsNAC3pro::FLAG-OsNAC3), revealing strong enrichment of OsNAC3 at the OsERF117 promoter, with markedly higher occupancy in the 9311 background. The OsUBQ10 locus served as a negative control. Data are mean ± SD (n = 3). (E and F) Seedling phenotypes (E) and relative inhibition of seedling height (F) of Kitaake, osnac3-1, osnac3-2, osuvr8ab-1, and osrup-1 under WL and WL supplemented with narrowband UV-B. n = 60 seedlings per line. Scale bar, 1 cm. (G) Representative phenotypes of 2-week-old seedlings of Kitaake, osnac3-1, osnac3-2, osuvr8ab-1, and osrup-1 before and after broadband UV-B treatment and recovery. (H) Survival rates of the same lines under broadband UV-B (n = 8). Scale bar, 3.5 cm. In (D), (F), and (H), data are mean ± SD; different lowercase letters indicate statistically significant differences (P < 0.05, one-way ANOVA followed by Tukey’s multiple comparisons test). OsUVR8 facilitates OsNAC3 activity under UV-B To identify the protein complex associated with OsNAC3 under UV-B stress, we conducted immunoprecipitation followed by mass spectrometry (IP-MS) using FLAG-tagged OsNAC3 expressed in rice seedlings, with or without UV-B irradiation. Compared to control immunoprecipitates, OsNAC3 pull-downs revealed a distinct enrichment of interacting proteins, including key components of the UV-B signaling pathway and stress-responsive transcriptional regulators. Notably, OsCOP1 preferentially interacted with OsNAC3 under dark conditions (table S10), while OsUVR8 specifically associated with OsNAC3 under UV-B exposure (table S11). These findings indicate that OsNAC3 acts as a context-dependent hub, linking UV-B perception to transcriptional regulation. To validate the OsNAC3-OsUVR8 interaction identified by IP-MS, we performed coimmunoprecipitation assays in rice protoplasts coexpressing FLAG-OsNAC3 and hemagglutinin (HA)–OsUVR8a/b. Under normal conditions, HA-OsUVR8a/b coimmunoprecipitated with FLAG-OsNAC3, indicating a constitutive association (Fig. 5A). Upon UV-B exposure, the interaction was enhanced, and HA-OsUVR8a/b appeared as a monomer rather than the dimeric form observed in darkness, suggesting UV-B–induced conformational changes (Fig. 5A). We further examined the OsNAC3-OsUVR8 interaction in vitro using pull-down assays with recombinant glutathione S-transferase (GST)–OsUVR8 and His-OsNAC3 proteins. A basal level of interaction was detected under normal conditions, which was markedly enhanced upon UV-B exposure (fig. S11A). No binding was observed with GST alone, confirming the specificity of the interaction (fig. S11A). Domain-mapping analyses revealed that the N-terminal NAC domain of OsNAC3 mediated this interaction, as fragments lacking the NAC domain failed to associate with OsUVR8 (fig. S11, B and C). To gain structural insights, we modeled the OsNAC3-OsUVR8b complex using AlphaFold and ChimeraX. The predicted structure indicated that OsNAC3 engages OsUVR8b through its N-terminal NAC domain. Within OsNAC3, Tyr83, Asn85, Ser87, Arg88, Lys118, Pro128, Lys132, Arg141, and Ala154 formed specific interactions with Ile347, Glu191, Arg243, Asp138, Asp51, Asp86, Asp116, Thr365, Trp409, Asp105, and Arg209 in the β-propeller domain of OsUVR8b (fig. S12A). Specifically, the interactions include Tyr83 (OsNAC3)–Ile347 (OsUVR8b), Asn85 (OsNAC3)–Glu191 (OsUVR8b), Asn85 (OsNAC3)–Arg243 (OsUVR8b), Ser87 (OsNAC3)–Asp138 (OsUVR8b), Arg88 (OsNAC3)–Asp51 (OsUVR8b), Arg88 (OsNAC3)–Asp86 (OsUVR8b), Lys118 (OsNAC3)–Asp116 (OsUVR8b), Pro128 (OsNAC3)–Thr365 (OsUVR8b), Lys132 (OsNAC3)–Trp409 (OsUVR8b), Arg141 (OsNAC3)–Asp105 (OsUVR8b), and Ala154 (OsNAC3)–Arg209 (OsUVR8b) (fig. S12B). These predictions provide a mechanistic framework for experimental validation. Consistent with the structural model, GST-tagged OsUVR8b efficiently bound His-tagged OsNAC3 in vitro, and this interaction was abolished when key residues in the NAC or β-propeller domains were simultaneously mutated to alanine (fig. S12C). Bimolecular fluorescence complementation (BiFC) analysis was performed using nV-OsUVR8a, nV-OsUVR8b, nV-OsNAC3, and nV-OsNAC3ΔN together with cV-OsNAC3, cV-OsNAC3ΔN, or cV-OsCOP1. Strong nuclear fluorescence was observed upon coexpression of nV-OsUVR8a/cV-OsNAC3 and nV-OsUVR8b/cV-OsNAC3, whereas no detectable fluorescence was observed for nV-OsUVR8a/cV-OsNAC3ΔN or nV-OsUVR8b/cV-OsNAC3ΔN, indicating that the N-terminal NAC domain of OsNAC3 is required for interaction with OsUVR8, as verified by colocalization with the nuclear marker nuclear localization signal–red fluorescent protein (NLS-RFP) (Fig. 5B). These findings were further corroborated by luciferase complementation assays in Nicotiana benthamiana, where robust luminescence signals were detected upon coexpression of OsNAC3-nLUC and cLUC-OsUVR8, with signal intensity further enhanced by UV-B exposure (fig. S12, D and E). No luminescence was observed when the OsNAC3 construct lacking the NAC domain was cotransfected with either OsUVR8a or OsUVR8b, confirming the specificity of the interaction (fig. S12, D and E). Collectively, these in vitro, protoplast, and heterologous assays support the idea that OsUVR8 can associate with OsNAC3 and modulate its activity under UV-B. OsUVR8 interacts with and enhances OsNAC3 activity under UV-B. (A) Coimmunoprecipitation assays showing interactions between FLAG-OsNAC3 and HA-OsUVR8a/b in rice protoplasts under WL or narrowband UV-B. Protein complexes were immunoprecipitated with anti-FLAG beads and detected with anti-HA antibodies. OsUVR8a/b dimeric and monomeric forms were resolved under nonboiled conditions and are indicated by yellow and blue asterisks, respectively. (B) BiFC analysis of OsNAC3-OsUVR8a/b interactions in rice protoplasts. Nuclear yellow fluorescent protein (YFP) signals were detected for OsUVR8a/b with full-length OsNAC3 but not with OsNAC3ΔN or EV controls. NLS-RFP marks nuclei. Scale bar, 5 μm. (C to F) Transactivation assays in N. benthamiana leaves showing that OsUVR8a/b enhance OsNAC3-dependent activation of the OsERF117pro::LUC reporter under narrowband UV-B. 35Spro::REN was used as an internal control. The assay was performed with the following constructs under without UV-B (i to iii) and UV-B (iv to vi) conditions: (i and iv) EV, (ii and v) OsNAC3, and (iii and vi) OsNAC3 + OsUVR8a/b. The EV served as a negative control. (G and H) EMSAs showing that monomeric OsUVR8aW299A and OsUVR8bW294A enhance OsNAC3 binding to OsERF117 promoter probes; GST and His-GFP served as controls. (I) ChIP-qPCR analysis showing in vivo association of OsNAC3 with the OsERF117 promoter. Chromatin was isolated from 2-week-old seedlings expressing FLAG-OsNAC3 under its native promoter (OsNAC3pro:: FLAG-OsNAC3) in the Kitaake or osuvr8ab-1 background (OsNAC3pro:: FLAG-OsNAC3/osuvr8ab-1) grown under WL or WL + narrowband UV-B. Immunoprecipitation (IP) was performed with anti-FLAG antibody, and enrichment is shown relative to an intergenic control region. In (D), (F), and (I), values represent mean ± SD (n = 3). Different letters indicate significant differences [P < 0.05, Tukey’s test; two-way ANOVA for (D) and (F), one-way ANOVA for (I)]. To investigate whether OsUVR8 modulates OsNAC3 activity, we used the OsERF117pro::LUC reporter in conjunction with 35S::REN as the internal control. Various effector combinations, including empty vector (EV), OsNAC3, and OsNAC3 in combination with OsUVR8a or OsUVR8b, were cotransfected into N. benthamiana. As shown in Fig. 5 (C to F), no luminescence signals were detected in the EV + reporter control, regardless of UV-B treatment. In contrast, significant luminescence was observed upon coexpression of OsNAC3 with OsERF117pro::LUC, with signal intensity further enhanced upon UV-B exposure. Notably, cotransfection of OsNAC3 with either OsUVR8a or OsUVR8b resulted in markedly higher luminescence signals, which were further amplified under UV-B treatment. We next performed EMSAs to assess the impact of OsUVR8 on the DNA binding activity of OsNAC3. The monomeric forms of GST-OsUVR8aW299A and GST-OsUVR8bW294A markedly enhanced the binding of His-OsNAC3 to the OsERF117 probe in EMSA assays (Fig. 5, G and H, and fig. S13, A and B). To investigate the in vivo association of OsNAC3 with its target promoters and to assess the regulatory roles of UV-B light and OsUVR8, we performed ChIP-qPCR assays. FLAG-tagged OsNAC3 transgenic plants driven by its native promoter were crossed with osuvr8ab-1 to generate FLAG-OsNAC3/osuvr8ab-1 lines (fig. S13C). FLAG-OsNAC3 was found to associate with the OsERF117 promoter even in the absence of UV-B exposure. Notably, this association was further strengthened upon UV-B irradiation but was significantly reduced in the osuvr8ab-1 mutant background (Fig. 5I). To confirm the specificity of OsNAC3 binding, we performed ChIP-qPCR on the nontarget gene OsUBQ10 and an intergenic region. No significant enrichment was observed in these control regions, whereas OsNAC3 showed strong, UV-B–inducible occupancy at the OsERF117 promoter in an OsUVR8-dependent manner (fig. S13, D and E). To further explore how OsUVR8 enhances OsNAC3 DNA binding, we modeled both the OsNAC3-DNA binary complex and the OsUVR8b-OsNAC3-DNA ternary complex using a promoter fragment from OsERF117 containing the NAC consensus motif (CACG/CGTG). In the binary complex, OsNAC3 was predicted to contact DNA mainly through Lys82, Arg88, Lys105, and Pro127. In the ternary complex, the OsNAC3-DNA interaction surface was expanded, with additional residues including Thr101 and Lys132 contributing to DNA contact. Consistent with this, predicted aligned error (PAE) analysis indicated a better-defined OsNAC3-DNA interface in the presence of OsUVR8b. Together with the EMSA results, these data suggest that OsUVR8b enhances OsNAC3 function not only by stabilizing the protein but also by promoting a more DNA binding–competent state (fig. S14, A and B). To assess the in vivo relevance of the predicted OsUVR8-OsNAC3 interface, we generated transgenic rice expressing FLAG-OsNAC3 or FLAG-OsNAC3(8A), in which eight predicted OsUVR8-contacting residues were substituted with alanine. Both proteins accumulated in transgenic seedlings (fig. S14C). Compared with FLAG-OsNAC3 plants, FLAG-OsNAC3(8A) lines showed attenuated UV-B–induced growth inhibition and reduced survival after broadband UV-B stress, although both phenotypes remained stronger than those of Kitaake or osuvr8ab-1 controls (fig. S14, D to G). These results indicate that OsNAC3 exhibits basal occupancy at the OsERF117 promoter in vivo, whereas OsUVR8a/b is required for the full UV-B–induced enhancement of OsNAC3 promoter binding and transactivation activity. UV-B–activated OsUVR8 inhibits OsCOP1-mediated degradation of OsNAC3 To validate the interaction between OsNAC3 and OsCOP1, we first performed coimmunoprecipitation assays in rice protoplasts coexpressing FLAG-OsNAC3 and HA-OsCOP1, which showed specific coprecipitation of HA-OsCOP1 with FLAG-OsNAC3, confirming their association in planta (fig. S15A). We then examined this interaction using in vitro pull-down assays with recombinant FLAG-tagged OsCOP1 and GST-fused full-length OsNAC3, its N-terminal domain (OsNAC3-N), or its C-terminal domain (OsNAC3-C). OsCOP1 interacted with GST-OsNAC3 and GST-OsNAC3-N but not with GST-OsNAC3-C, indicating that the N-terminal NAC domain mediates the interaction (fig. S15B). To further substantiate these findings in vivo, we performed BiFC assays, fusing OsNAC3, an N-terminal deletion variant (OsNAC3ΔN), and OsCOP1 to the N- or C-terminal halves of Venus (nV-OsNAC3, nV-OsNAC3ΔN, and cV-OsCOP1). Coexpression of nV-OsNAC3 with cV-OsCOP1 produced strong fluorescence signals in nuclear speckles that colocalized with the nuclear marker NLS-RFP, whereas no signal was detected with nV-OsNAC3ΔN, nV-OsNAC3 with cV, or nV with cV-OsCOP1, demonstrating both specificity and N-terminal dependence of the interaction (fig. S15C). Last, luciferase complementation assays in N. benthamiana corroborated these results: Strong luminescence was observed upon coexpression of OsNAC3-nLUC and cLUC-OsCOP1, whereas deletion of the NAC domain abolished luminescence, underscoring the essential role of the N-terminal region in mediating the interaction (fig. S15D). To test the hypothesis that OsCOP1 functions as the E3 ligase responsible for ubiquitinating OsNAC3, we conducted ubiquitination assays using a bacterial reconstitution system (29). In this system, we coexpressed Arabidopsis UBA1-S (E1), AtUBC8-S (E2), OsCOP1-Myc (E3), MBP-OsNAC3-HA (substrate), and FLAG-tagged ubiquitin (Flag-ub). Immunoprecipitation (IP) with an anti-HA antibody revealed a characteristic laddering pattern of OsNAC3, which was observed only when all components were present. The absence of any one component resulted in a complete failure to detect ubiquitination of OsNAC3. Further analysis via IP-MS identified five robust ubiquitination sites (fig. S16, A and B). To further evaluate the OsCOP1-dependent degradation of OsNAC3, we generated OsCOP1-knockdown lines (OsCOP1-amiR) using an artificial microRNA system (30). Two specific 21-nucleotide reverse complements from the OsCOP1 coding sequence (CDS) were designed as amiRNA candidates (fig. S17A). OsCOP1 transcripts were markedly decreased in two independent OsCOP1-amiR lines (fig. S17B). Furthermore, field-grown OsCOP1-amiR plants exhibited markedly reduced heights (fig. S17, C and D), underscoring the OsCOP1’s critical role in regulating plant stature under field conditions. To investigate whether OsCOP1 mediates polyubiquitination of OsNAC3, FLAG-tagged OsNAC3, driven by the UBQ promoter, was crossed with OsCOP1-amiR-1-1 lines to generate UBQpro::FLAG-OsNAC3/OsCOP1-amiR-1-1 plants. Upon dark treatment, FLAG-OsNAC3 protein levels were markedly reduced in wild-type background, whereas treatment with the proteasome inhibitor MG132 effectively stabilized FLAG-OsNAC3, suggesting proteasome-dependent degradation (fig. S17E). In contrast, FLAG-OsNAC3 degradation was substantially attenuated in the OsCOP1-amiR-1-1 background (fig. S17E). Moreover, IP assays revealed that the level of polyubiquitinated FLAG-OsNAC3 was markedly decreased in OsCOP1-amiR-1-1 compared to the Kitaake background, supporting a role for OsCOP1 in promoting OsNAC3 ubiquitination and degradation in the dark (fig. S17F). To test whether these five lysine residues are functionally important for OsNAC3 regulation, we generated a ubiquitination-deficient variant in which all five IP-MS–identified lysines were substituted with arginine (OsNAC35KR). In semi–in vitro ubiquitination assays, GST-OsNAC35KR showed markedly reduced polyubiquitination relative to wild-type GST-OsNAC3 when incubated with total protein extracts from Kitaake seedlings, and ubiquitination of wild-type OsNAC3 was likewise reduced in extracts from the OsCOP1-amiR-1-1 background (fig. S18A). In vivo ubiquitination assays in rice protoplasts further showed that Myc-OsCOP1 efficiently promoted ubiquitination of FLAG-OsNAC3, whereas ubiquitination of FLAG-OsNAC35KR was strongly diminished (fig. S18B). Consistent with this, in cell-free degradation assays, GST-OsNAC35KR was more stable than wild-type GST-OsNAC3, and the difference between genotypes was largely attenuated in the OsCOP1-amiR-1-1 background (fig. S18C). Moreover, in transient transcription assays using the OsERF117pro::LUC reporter, OsNAC35KR exhibited greater reporter activation than wild-type OsNAC3 under both control and UV-B conditions (fig. S18D). Together, these results support that the five lysine residues identified by IP-MS are major functional sites contributing to COP1-dependent ubiquitination and negative regulation of OsNAC3 in vivo. We next investigated the effect of UV-B irradiation on OsNAC3 protein stability. FLAG-OsNAC3, driven by the UBQ promoter, was crossed into the osuvr8ab-1 background to generate UBQpro::FLAG-OsNAC3/osuvr8ab-1 plants. Immunoblot analysis revealed that FLAG-OsNAC3 levels were markedly elevated following UV-B treatment compared to dark conditions (Fig. 6A). This UV-B–induced stabilization was abolished in the osuvr8ab-1 double mutant (Fig. 6B), indicating a requirement for the OsUVR8 function. By contrast, FLAG-OsNAC3 degradation was substantially attenuated in the OsCOP1-amiR-1-1 background, irrespective of UV-B exposure (Fig. 6B). IP assays further demonstrated increased polyubiquitination of FLAG-OsNAC3 in osuvr8ab-1 mutants relative to wild-type background under UV-B conditions (Fig. 6C). Notably, pull-down assays showed that OsUVR8a and OsUVR8b compete with OsCOP1 for binding to the N-terminal region of OsNAC3, suggesting that OsUVR8 impedes OsCOP1 access to OsNAC3 (fig. S19, A and B). Together, these findings support a model in which UV-B–activated OsUVR8 promotes OsNAC3 stabilization, at least in part by antagonizing OsCOP1-mediated ubiquitination and degradation. In summary, our working model (Fig. 6D) proposes that UV-B–induced monomerization of OsUVR8 shields OsNAC3 from OsCOP1 while simultaneously enhancing its DNA binding activity, thereby enabling OsNAC3 accumulation and robust activation of OsERF117 transcription. OsERF117 then induces flavonoid, melatonin, and GABA-shunt biosynthetic pathways, producing photoprotective metabolites that alleviate UV-B stress. A naturally occurring promoter polymorphism at an OsNAC3-binding site (SNP −1774) further increases OsERF117 responsiveness in accessions from high-UV-B environments, contributing to subspecies-level UV-B adaptation. UV-B–activated OsUVR8 inhibits OsCOP1-mediated degradation of OsNAC3. (A) Immunoblot analysis of FLAG-OsNAC3 abundance in the UBQpro::FLAG-OsNAC3 line under darkness or narrowband UV-B illumination. OsNAC3 accumulation increases upon UV-B exposure. Histone H3 serves as a loading control. (B) Stability of FLAG-OsNAC3 in UBQpro::FLAG-OsNAC3, UBQpro::FLAG-OsNAC3/osuvr8ab-1, and UBQpro::FLAG-OsNAC3/OsCOP1-amiR-1-1 seedlings. Seedlings were pretreated with DMSO or 25 μM MG132 for 2 hours, followed by 100 μM CHX for 30 min in darkness, and subsequently transferred to continuous darkness or narrowband UV-B for the indicated times. Histone H3 serves as a loading control. (C) In vivo ubiquitination of OsNAC3. One-week-old dark-grown seedlings of the indicated genotypes were treated with MG132 for 4 hours under darkness or UV-B. Proteins were immunoprecipitated with anti-FLAG beads and immunoblotted with anti-FLAG, anti-ubiquitin, and anti-H3 antibodies. (D) Proposed model of UV-B–induced metabolic adaptation in rice. UV-B activates OsUVR8, which stabilizes OsNAC3 by antagonizing OsCOP1-mediated ubiquitination and enhancing its DNA binding activity. Stabilized OsNAC3 transcriptionally activates OsERF117, which in turn drives flavonoid, melatonin, and GABA-shunt biosynthetic pathways by inducing genes including OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15. Accumulation of these metabolites mitigates UV-B stress. A cis-regulatory SNP at an OsNAC3-binding site modulates OsERF117 expression level, contributing to local UV-B adaptation and subspecies diversification. h, hours. Created in BioRender. Liu, W. (2026) https://BioRender.com/w8yggf7. DISCUSSION Plants have evolved elegant photoregulatory networks to perceive and mitigate UV-B radiation, yet how these networks have been remodeled in crops to accommodate diverse light ecologies has remained unresolved. In Arabidopsis, the canonical dicot UV-B pathway is centered on UVR8 monomerization, UVR8-COP1 association, stabilization of HY5/HYH, destabilization of PIF repressors (31–33), and direct UVR8 interactions with transcription factors such as WRKY36, BES1/BIM1, and MYB family members that modulate DNA binding (1, 15–17). Together, these mechanisms couple UV-B perception to photomorphogenesis and photoprotective acclimation. Our study uncovers a functionally specialized UV-B signaling module in rice, in which UV-B–activated OsUVR8 enhances OsNAC3 activity and antagonizes OsCOP1-mediated ubiquitination and proteasomal degradation, thereby stabilizing OsNAC3 and promoting activation of the AP2/ERF regulator OsERF117. This OsUVR8-OsCOP1-OsNAC3-OsERF117 axis thus defines a crop-specific conduit between light perception and metabolic acclimation, coupling transcriptional reprogramming with physiological resilience. Mechanistically, our findings refine the interface between photoreception and proteostasis. Structural modeling and biochemical analyses indicate that OsUVR8 directly associates with the N-terminal NAC domain of OsNAC3 through a multivalent β-propeller interface. Alanine substitution of eight predicted OsUVR8-interacting residues in OsNAC3 partially compromised, but did not abolish, OsNAC3-mediated UV-B photomorphogenesis and stress tolerance in vivo, indicating that these residues are required for maximal OsNAC3 activity rather than basal function. Our data further suggest that OsNAC3 retains basal promoter-binding activity in the absence of UV-B, whereas UV-B activation strengthens the OsUVR8-OsNAC3 interaction and enhances OsNAC3 occupancy at the OsERF117 promoter in vivo. An intriguing feature of the AlphaFold3 models is that OsUVR8b association does not simply increase all predicted OsNAC3-DNA contacts. Instead, a flexible terminal segment positioned near DNA in the binary complex is displaced in the ternary complex, whereas additional contacts are predicted within the structured NAC-domain interface. This redistribution suggests that OsUVR8b may allosterically reorganize OsNAC3 into a more productive promoter-bound orientation rather than merely enlarging its DNA binding surface. This remodeling could influence DNA binding affinity, residence time, or sequence selectivity. However, because the flexible terminal region is modeled with lower confidence and AlphaFold3 does not resolve binding dynamics, direct structural and quantitative biophysical analyses will be required to distinguish among these possibilities. In parallel, UV-B–activated OsUVR8 antagonizes OsCOP1-mediated ubiquitination and degradation of OsNAC3, thereby promoting its accumulation and transcriptional output. Together, these results support a model in which OsUVR8 does not confer OsNAC3 promoter binding de novo but instead potentiates OsNAC3 activity under UV-B. Although our biochemical and genetic data support a monomerization-linked model, whether UV-B–induced OsUVR8 monomerization is itself strictly required for OsNAC3 interaction and activation remains to be established more rigorously in planta. Recent work has shown that the spatial regulation of UV-B signaling in rice differs from that in Arabidopsis: Nuclear accumulation of OsUVR8 is independent of both OsCOP1 and UV-B. This distinction is notable because, in dicots, UV-B–induced UVR8 nuclear accumulation is closely linked to signaling output, including COP1-dependent transcriptional reprogramming and direct inhibition of several transcription factors. In rice, by contrast, the constitutive nuclear localization of OsUVR8a/b and the weak OsUVR8-OsCOP1 interaction suggest divergence in downstream signaling architecture. Our data identify OsNAC3 as one such monocot effector: OsUVR8 not only protects OsNAC3 from OsCOP1-dependent turnover but also enhances its DNA binding competence, thereby converting UV-B perception into stronger activation of OsERF117. Thus, our results suggest that, in addition to conserved UVR8 signaling outputs, rice has a OsUVR8-OsNAC3-OsERF117 transcriptional branch that contributes to UV-B–induced metabolic acclimation. Supported by genetic analyses of rice osuvr8ab and OsCOP1-amiR lines, these findings illustrate how a conserved photoreceptor–E3 ligase interface can be adapted to regulate lineage-specific effectors and coordinate metabolic acclimation in a cereal crop. The OsNAC3-OsERF117 transcriptional module orchestrates a coherent metabolic program that underpins UV-B acclimation in rice. OsERF117 directly activates genes involved in flavonoid, melatonin, and GABA metabolism, including OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15, thereby coupling carbon and nitrogen flux to photoprotective scaffolds such as quercetin, proanthocyanidins, and melatonin. This coordinated regulation mitigates oxidative stress, maintains redox homeostasis, and shields chloroplasts from UV-B–induced photodamage, sustaining growth under environmental stress (34). Disruption of OsERF117 collapses this network, leading to hypersensitivity and impaired development, whereas its overexpression reinforces photoprotection without compromising growth—highlighting the module’s potential for breeding UV-B–resilient rice varieties. Evolutionary analyses further clarify how natural selection has shaped this signaling logic. Promoter polymorphisms at OsERF117 show a pronounced subspecies-biased distribution—group A haplotypes dominating indica and group B enriched in japonica, with only limited admixture between the two. A single −1774 C → A promoter substitution strengthens OsNAC3 binding and enhances UV-B responsiveness in low-latitude indica populations, consistent with adaptive divergence along historical UV gradients. This cis-regulatory variation, rather than coding-sequence change, accounts for much of the photoadaptive plasticity in rice. Together, these results illustrate how fine-scale promoter rewiring can repurpose a conserved light-sensing architecture into a geographically optimized regulatory circuit. Together, our findings delineate an integrative UV-B signaling framework that links photoreceptor perception with transcriptional control and adaptive metabolism. This module illuminates the evolutionary flexibility of plant light signaling and provides a framework for designing climate-resilient crops capable of thriving under intensifying solar regimes. MATERIALS AND METHODS Plant materials and growth conditions A whole-genome activation-tagging pool of Oryza sativa L. ssp. japonica cv. Kitaake, from which AC37 was identified in this study, was generated by Agrobacterium-mediated cocultivation of calli, as previously described (25). Loss-of-function mutants (oserf117, osuvr8ab, osrup, and osnac3) were generated using the CRISPR-Cas9 genome-editing system, and mutations were verified by Sanger sequencing of PCR amplicons. Complementation lines (OsERF117pro::FLAG-OsERF117; Com) were created in the oserf117-1 background by introducing a 3-kb OsERF117 promoter fragment fused to the corresponding genomic sequence. The 3-kb OsERF117 promoter was also cloned into pCAMBIA3301 to generate the OsERF117pro::GUS reporter construct. For overexpression lines, the coding sequences of OsERF117 and OsNAC3 were inserted into pCAMBIA1302 to produce UBQpro::FLAG-OsERF117, UBQpro::FLAG-OsNAC3, UBQpro::FLAG-OsNAC3(8A), and OsNAC3pro::FLAG-OsNAC3 transgenic plants, respectively. An OsCOP1-amiRNA sustained expression construct was generated using the rice pre-amiRNA528a backbone following established procedures (30). All constructs were introduced into Agrobacterium and transformed into the calli of different rice varieties by Wimibio (China). The following plant materials were generated by genetic crossing UBQpro::FLAG-OsNAC3/osuvr8ab-1, UBQpro::FLAG-OsNAC3/OsCOP1-amiR-1-1, and OsNAC3pro::FLAG-OsNAC3/osuvr8ab-1. A recombinant inbred line population derived from NIP × 9311 was constructed as previously described (35), from which NIL-OsERF117NIP and NIL-OsERF1179311 were selected. Field experiments were conducted under controlled conditions at experimental plots in Changchun, Jilin Province, China (125°41′N, 43°82′E). Phenotypic analysis of UV-B–induced photomorphogenesis and stress responses Rice seeds were surface-sterilized with sodium hypochlorite and germinated at 37°C for 3 days. Uniform seedlings were selected and transferred to petri dishes or 96-well plates and subsequently grown under a 10-hour light/14-hour dark photoperiod at 28°/25°C (day/night) under WL. UV-B treatments were conducted using Philips UV-B lamps (TL 20 W/01 for narrowband UV-B, 305 to 315 nm with a peak at 311 nm, and TL 40 W/12 for broadband UV-B, 290 to 315 nm). UV-B intensity was adjusted by varying the distance between the plants and the lamps and was measured using a UV-297 UV-B Light Meter (HANDY). For UV-B growth response assays, seedlings were grown under WL (∼200 μmol m−2 s−1) alone or WL supplemented with narrowband UV-B radiation (30 μW cm−2, ∼4.5 μmol m−2 s−1). After 10 days of treatment, seedling height was measured, and the relative inhibition rate was calculated as the ratio of seedling height under UV-B treatment to that under WL alone. For UV-B stress tolerance assays, 14-day-old seedlings grown under WL (∼280 μmol m−2 s−1) were exposed to WL supplemented with broadband UV-B radiation (500 μW cm−2, ∼75 μmol m−2 s−1) for 3 to 4 days. After treatment, seedlings were allowed to recover under normal WL conditions for 5 to 7 days, and survival rates were then determined. Plasmid construction Loss-of-function mutants of OsERF117, OsRUP, OsNAC3, OsUVR8a, and OsUVR8b were generated using the CRISPR-Cas9 system (36). Gene-specific single guide RNAs were designed with the online tool and individually cloned into the pYLCRISPR-Cas9Pubi-H binary vector. For complementation, the full-length OsERF117 coding sequence (CDS) with its native promoter was inserted into pCAMBIA1302 to generate the OsERF117pro::FLAG-OsERF117 construct. Similarly, the OsNAC3 coding sequence (CDS) driven by its native promoter was inserted into pCAMBIA1302 to generate the OsNAC3pro::FLAG-OsNAC3 construct. For subcellular localization, the OsERF117 CDS was cloned into the 326-nGFP vector under the cauliflower mosaic virus (CaMV) 35S promoter using Xba I and Bam HI to produce GFP-OsERF117. For Y1H assays, a 3-kb OsERF117 promoter fragment and the full-length OsNAC3 CDS were inserted into pAbAi and pGADT7, respectively, yielding pAbAi-OsERF117 and pGADT7-OsNAC3. For coimmunoprecipitation assays, FLAG-OsNAC3 was generated by cloning the OsNAC3 CDS downstream of the CaMV 35S promoter using Xba I and Bam HI into the 326-FLAG vector. HA-tagged OsUVR8a, OsUVR8b, and OsCOP1 constructs were produced by cloning the corresponding CDSs using Sal I and Bam HI into the 326-HA vector. For BiFC, the CDSs of OsUVR8a, OsUVR8b, OsNAC3 and an N-terminal deletion of OsNAC3 (OsNAC3ΔN) were fused in-frame to the N-terminal fragment of Venus (nV) in the 326-nYNEE vector, generating nV-OsUVR8a, nV-OsUVR8b, nV-OsNAC3, and nV-OsNAC3ΔN. Similarly, the CDSs of OsCOP1, OsNAC3, and OsNAC3ΔN were fused to the C-terminal Venus fragment (cV) in the 326-nYCHA vector, producing cV-OsCOP1, cV-OsNAC3, and cV-OsNAC3ΔN. All BiFC constructs were generated using Xba I and Bam HI sites. For GST pull-down assays, the CDSs of OsUVR8a, OsUVR8b, OsNAC3, the OsNAC3 DNA binding domain (OsNAC3-N), the OsNAC3 transactivation domain (OsNAC3-C), and an OsNAC3 variant containing eight alanine substitutions at key interaction sites [OsNAC3(8A)] were cloned into pGEX4T-1 using Bam HI and Eco RI. The CDSs of OsNAC3, OsUVR8a, OsUVR8b, and an OsUVR8b mutant carrying 11 alanine substitutions [OsUVR8b(11A)] were individually cloned into pET28a. The CDS of OsCOP1 was successfully cloned into the pMAL-c6T expression vector. For luciferase complementation assays, the CDSs of GFP, OsNAC3, and OsNAC3ΔN were cloned into pCAMBIA1300-nLUC, whereas the CDSs of GFP, OsUVR8a, OsUVR8b, and OsCOP1 were inserted into pCAMBIA1300-cLUC, using Kpn I and Sal I restriction sites. For the luciferase reporter assays, ∼3-kb promoter fragments of OsERF117 (group A), OsERF117 (group B), OsCHS1, OsUGT88C2, OsGABA-T, and OsASMT15 were cloned into the reporter vector using Bam HI and Nco I. HA-tagged OsERF117 and OsNAC3 were used as effectors. The AtUBQ10pro::GUS construct (37) was cotransfected for normalization. For promoter activity assays, the OsNAC3 CDS was inserted into pGreenII 62-SK, and the 3-kb OsERF117 promoter fragment was cloned into pGreenII 0800-LUC. All primers used for plasmid construction are listed in table S12. Y1H assay The Y1H assay was performed using the Y1HGold-pAbAi Yeast One-Hybrid System (Coolaber). To generate the promoter bait, a 3-kb fragment upstream of the OsERF117 ATG start codon was amplified by PCR and cloned into the pAbAi vector. For the effector construct, the coding sequence of OsNAC3 was amplified and inserted into pGADT7 to generate pGADT7-OsNAC3. Bait constructs were linearized with the appropriate restriction enzymes and integrated into the Y1HGold yeast genome. The basal sensitivity of the integrated strain to aureobasidin A (AbA) was assessed to determine the minimal inhibitory concentration before transformation with pGADT7-OsNAC3. Yeast transformants were selected on SD/-Leu medium supplemented with AbA (700 ng ml−1). The empty pGADT7 vector was introduced under identical conditions and served as the negative control. RNA preparation and quantitative RT-PCR analysis Total RNA was extracted using TriPure Isolation Reagent (Roche, 11667165001), and first-strand cDNA was synthesized using the cDNA Synthesis SuperMix (TransGen Biotech, AT301-03). RT-qPCR was performed as described previously (38). Ubiquitin10 (UBQ10) was used as the internal reference gene. Each target gene was analyzed using three biological replicates. Primer sequences used for RT-qPCR are provided in table S12. Electrophoretic mobility shift assay His-tagged proteins were purified in vitro using glutathione beads (LABLEAD, G10510) following the protocol of Wang et al. (38). Biotin-labeled DNA probes and unlabeled competitor probes were synthesized by Sangon Biotechnology. Probes were incubated with purified His-tagged proteins in EMSA binding buffer at 25°C for 20 min. Protein-DNA complexes were resolved on a 6% nondenaturing polyacrylamide gel electrophoresis (PAGE), transferred to a nylon membrane, and detected using the Chemiluminescent EMSA Kit (Beyotime, GS009). Primer sequences used for EMSA are listed in table S12. ChIP-qPCR assay ChIP assays were performed following the method of Liu et al. (25) Chromatin complexes were extracted from 15 g of 2-week-old seedlings of wild type, OsERF117pro::FLAG-OsERF117, OsNAC3pro::FLAG-OsNAC3-1 (9311), OsNAC3pro::FLAG-OsNAC3-3 (NIP), OsNAC3pro::FLAG-OsNAC3-5 (Kitaake), and OsNAC3pro::FLAG-OsNAC3-5/osuvr8ab-1 lines, with or without 3 hours of narrowband UV-B treatment before harvesting. DNA was sheared to ∼500-bp fragments using a Bioruptor Plus sonicator (Diagenode, UCD-300). Chromatin extracts were incubated with anti-FLAG antibodies (Sigma-Aldrich, F1804) under constant rotation at 4°C for 16 to 18 hours. After IP, cross-links were reversed, and DNA was purified. Enrichment of immunoprecipitated DNA was quantified by RT-qPCR, with OsUBQ10 used as a constitutive reference gene. Primer sequences are listed in table S12. Histochemical GUS staining and subcellular localization analysis Transgenic plants carrying the OsERF117pro::GUS construct were subjected to histochemical GUS staining. Tissues were incubated in staining buffer [100 mM Na3PO4, 5 mM K4Fe(CN)6, 5 mM K3Fe(CN)6, 10 mM EDTA, 1% X-Gluc, and 0.1% Triton X-100] and destained in ethanol/acetic acid (7:3). Representative images were captured from three independent transgenic lines using a digital single-lens reflex camera. Subcellular localization of OsERF117 was examined in rice protoplasts. Protoplasts isolated from leaf sheaths of 3-week-old seedlings were cotransfected with GFP-OsERF117 and NLS-RFP constructs via poly(ethylene glycol) (PEG)–mediated transfection. After incubation for 16 to 18 hours, fluorescence signals were observed using an Olympus fluorescence microscope. GUS activity assay GUS activity assays were performed as described previously (38). Briefly, effector plasmids encoding full-length or truncated OsERF117 fused in-frame with GD or GD–VP16 were cotransfected with the GAL4-mini35Spro::GUS reporter into rice protoplasts and incubated in the dark for 20 to 22 hours. GUS activity was quantified using an automatic microplate reader (Spark 10M). Data represent three independent biological experiments. DNA sample preparation and whole-genome resequencing A total of 103 rice accessions were obtained from the Rice Research Institute of Shenyang Agricultural University. Plants were grown under standard agronomic conditions, and young leaf tissue was collected and immediately snap-frozen in liquid nitrogen. Genomic DNA was extracted using the DNAsecure Plant Kit (Tiangen, DP320). Sequencing libraries were constructed with the BGI Optimal DNA Library Prep Kit (BGI, Shenzhen, China) following the manufacturer’s guidelines. Briefly, genomic DNA was fragmented, size-selected, and subjected to end repair, A-tailing, and adapter ligation, followed by PCR amplification and quality assessment. Single-stranded circular templates were then generated and amplified by rolling circle amplification to produce DNA nanoballs for sequencing. Libraries were sequenced on the DNBSEQ-T7 platform (BGI, Shenzhen, China) to generate 150-bp paired-end reads. To minimize technical artifacts, raw reads were filtered according to established criteria (39). Reads were removed if they contained ≥10% undetermined bases (N), >10 nucleotides of adapter sequence (≤10% mismatches allowed), >50% low-quality bases (Phred score < 5), or if they represented potential PCR duplicates. Haplotype and phylogenetic analysis The SNP dataset for the 3-kb promoter region of OsERF117 was compiled from two sources: publicly available data from RiceVarMap V2.0 (40) (http://ricevarmap.ncpgr.cn) and whole-genome resequencing of 103 rice accessions. A neighbor-joining phylogenetic tree was constructed with 1000 bootstrap replicates using MEGA12 and visualized with the interactive tool tvBOT (41). For haplotype analysis, a 3-kb promoter sequence of OsERF117 from 4093 accessions was used, and only haplotypes present in ≥10 accessions were included (40). Global UV-B radiation data were obtained from a previous study (42) and integrated using R. RNA-seq and transcriptome analysis Seedlings of O. sativa cv. Kitaake, oserf117-1, and osuvr8ab-1 were grown hydroponically in Yoshida’s nutrient solution for 14 days. Shoots were collected after exposure to 0 or 6 hours of broadband UV-B irradiation (500 μW cm−2). For each condition, three biological replicates were harvested, with each replicate comprising pooled shoot tissue from three seedlings. Total RNA was extracted using the RNeasy Plant Mini Kit (QIAGEN) and assessed for integrity and quantity with an Agilent 2100 Bioanalyzer (Agilent) and the Equalbit RNA BR Assay Kit. RNA-seq libraries were prepared and sequenced by BGI (Shenzhen, China). Raw reads were filtered with SOAPnuke to obtain high-quality clean reads (∼4.0 Gb per replicate). Clean reads were aligned to the rice reference genome using HISAT2. Gene-level quantification and differential expression analyses were performed using DESeq2, and genes with an adjusted P ≤ 0.05 and |log2(fold change)| ≥ 1 were considered significantly differentially expressed and used for downstream functional analyses. DAP-seq analysis DAP-seq was performed as previously described with minor modifications. Genomic DNA was extracted from wild-type (Kitaake) leaves, and sequencing libraries were prepared by SeqHealth (Wuhan, China). The coding sequence of OsERF117 was PCR-amplified, incorporating an SP6 promoter and a Kozak consensus sequence at the 5′ end and a 3 × FLAG tag at the 3′ end. In vitro translation was conducted using the TnT SP6 High-Yield Wheat Germ Protein Expression System (Promega) following the manufacturer’s protocol, with 6 μg of DNA template in a 50 μl reaction. After a 2-hour incubation at room temperature, the reaction mixture was combined with 20 μl of anti-FLAG magnetic beads (Merck Millipore), incubated for 1 hour, and washed three times. Genomic DNA from Kitaake was isolated using the Plant Genomic DNA Kit (Tiangen Biotech) and sheared to 300 to 500 bp by ultrasonication (20% amplitude, 2 s on/8 s off, 16 min, on ice). Fragmented DNA was ligated to Illumina adapters using the VAHTS Universal DNA Library Prep Kit for Illumina V3 (Vazyme) to generate the input library. Anti-FLAG beads bound to OsERF117 protein were incubated with 50 ng of input library, and affinity-purified DNA (“DAP”) was recovered by phenol-chloroform extraction. Both DAP and input libraries were PCR-amplified, purified, quantified, and sequenced on a DNBSEQ-T7 platform (MGI) using paired-end 150-bp (PE150) sequencing. Raw reads were quality-filtered with Fastp and aligned to the rice reference genome using Bowtie2 (v2.2.6). Read distribution was assessed with RSeQC (v2.6). Peak calling was performed using MACS2 (v2.1.1), and peak annotation and genomic distribution analyses were carried out with bedtools (v2.25.0). Conserved motifs within peak regions were identified using HOMER (v4.10). Metabolomic analysis Two-week-old seedlings of O. sativa cv. Kitaake and oserf117 were exposed to broadband UV-B radiation or maintained under control conditions for 6 hours, with three independent biological replicates per treatment. Flavonoid profiling was performed using an AB Sciex QTRAP 6500 LC-MS/MS platform (Metware Biotechnology Co. Ltd., Wuhan, China). Metabolite extraction and detection were conducted as previously described. Briefly, samples were freeze-dried using a lyophilizer (Scientz-100F) and ground to a fine powder (30 Hz, 1.5 min). Powdered samples were extracted in prechilled 70% methanol containing internal standards. Samples were vortexed for 30 s every 30 min for a total of 6 cycles, followed by centrifugation at 12,000 rpm for 3 min. The resulting supernatants were filtered through 0.22-μm membranes and transferred to autosampler vials for UPLC-MS/MS analysis. Chromatographic separation was performed on an Agilent SB-C18 column (2.1 mm by 100 mm, 1.8 μm) using a mobile phase consisting of water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) under a gradient elution program. Mass spectrometry was carried out on an ESI source operated in both positive and negative ion modes, with a source temperature of 500°C and an ion spray voltage of ±5500 V. Data acquisition was performed in multiple reaction monitoring mode using nitrogen as the collision gas, with declustering potentials and collision energies optimized for individual metabolites. Unsupervised principal components analysis was performed in R (www.r-project.org) using the prcomp function with unit-variance scaling. Metabolite abundances were normalized using z-scores. Variable importance in projection (VIP) values were obtained from orthogonal projections to latent structures discriminant analysis. Metabolites with VIP ≥ 1 and a fold change ≥ 1.5 or ≤ 0.67 between UV-B–treated and control samples in Kitaake or oserf117 were considered significantly accumulated. Metabolite extraction and detection Two-week-old seedlings of Kitaake, oserf117-1, osnac3-1, osuvr8ab-1, and osrup-1 were exposed to broadband UV-B radiation or maintained under control conditions for 6 hours, with three independent biological replicates per treatment. For measurement of proanthocyanidin, the content was determined according to a previously described method with minor modifications (43). Briefly, 0.1 to 0.15 g of sample was homogenized in 1 ml of 70% (v/v) aqueous acetone containing 0.1% (w/v) ascorbic acid and extracted at 4°C overnight. Following centrifugation, the supernatant was combined with diethyl ether and kept at −20°C in the dark for phase separation. The lower phase was then mixed with DMACA reagent and quantified as A643 g−1 fresh weight. For the determination of quercetin, 0.15 g of powder was extracted with 1 ml of methanol containing 1% HCl (v/v), followed by a 20-min ultrasound bath to ensure complete dissolution, and then kept at 4°C overnight in the dark. After centrifugation at 12,000 rpm for 10 min at 4°C, the supernatant was filtered through a 0.22-μm filter, and the filtrate was collected for high-performance liquid chromatography analysis using a Waters Alliance e2695 system (Waters Corporation, USA). For determination of melatonin and GABA, the samples were ground in liquid nitrogen and homogenized with phosphate-buffered saline, followed by ultrasonic disruption for 30 min. The homogenate was incubated at 4°C with shaking at 110 rpm for 2 hours and then centrifuged at 12,000 rpm for 10 min. The supernatant was filtered through a 0.22-μm filter membrane. Melatonin content was measured using a Plant Melatonin Quantitative ELISA Kit (mlbio, Shanghai, China), and GABA content was measured using a Plant GABA Quantitative ELISA Kit (mlbio, Shanghai, China). Both assays were performed according to the respective manufacturers’ instructions. The absorbance was recorded at 450 nm with a Multimode Plate Reader (BioTek Instruments, USA). DAB and NBT staining The NBT and DAB (Coolaber, Beijing, China) staining assays were performed according to the manufacturer’s instructions. Two-week-old seedlings were treated with or without broadband UV-B for 24 hours. For ROS detection, leaves were vacuum-infiltrated for 30 min and then submerged either in 10 mM potassium phosphate buffer (pH 7.8) containing 0.1% NBT and 10 mM NaN3 for 12 hours or in 0.1% DAB solution (pH 5.8) for 24 hours. To remove chlorophyll, stained leaves were transferred to 95% ethanol and incubated at 80°C for 15 min. DAB and NBT staining were used to detect H2O2 and O2− accumulation, respectively. Immunoprecipitation–mass spectrometry Approximately 3 g of leaf tissue from Kitaake and UBQpro::FLAG-OsNAC3 transgenic plants was harvested following narrowband UV-B (400 μW cm−2) exposure or dark incubation. Samples were flash-frozen in liquid nitrogen, finely ground, and homogenized in 30 ml of IP buffer [100 mM tris-HCl (pH 7.5), 150 mM NaCl, 2 mM dithiothreitol (DTT), 0.5% Triton X-100, 1× protease inhibitor cocktail, and 30 μM MG132]. Protein extracts were incubated with anti-FLAG M2 magnetic beads (Sigma-Aldrich, M8823) for 3 hours at 4°C under darkness or narrowband UV-B. Beads were washed four times with IP buffer, and bound protein complexes were eluted. Eluates were denatured by boiling and submitted to Suzhou Mass-elife Biotechnologies for mass spectrometric analysis. Protein digestion was carried out using the filter-aided sample preparation method (44). Peptides were separated by nanoflow liquid chromatography on a homemade C18 column (75 μm i.d. × 15 cm; 1.9 μm, 100 Å) using a 75-min linear gradient from 0 to 95% acetonitrile (0.1% formic acid) at a flow rate of 600 nl min−1. Mass spectrometry was performed on a Q Exactive HF-X instrument (Thermo Fisher Scientific) operating in data-dependent acquisition mode. The 25 most intense precursor ions (mass/charge ratio 340 to 1100) were isolated for higher-energy collisional dissociation using a normalized collision energy of 27%. MS1 and MS2 spectra were acquired at resolutions of 120,000 and 15,000, respectively. Raw data were processed with Proteome Discoverer 3.1 and searched against the Kitaake protein database (UniProt) using Sequest HT. Peptide and protein identifications were filtered to a false discovery rate < 1%. Co-immunoprecipitation Coimmunoprecipitation assays were performed as previously described (38). FLAG- and HA-tagged proteins were cotransfected into rice protoplasts and incubated for 16 hours at 22°C. After the 16-hour incubation, protoplasts were either kept under WL or exposed to narrowband UV-B (400 μW cm−2, 1 hour). Total protein was extracted using IP buffer and incubated with anti-FLAG magnetic beads (Sigma-Aldrich, M8823) under constant agitation at 4°C for 3 hours. Beads were washed three times, and bound proteins were eluted in SDS sample buffer. Proteins were detected by immunoblotting with anti-FLAG (Sigma-Aldrich, F3165) and anti-HA (Sigma-Aldrich, H6908) antibodies. OsUVR8 dimer/monomer transition assay The OsUVR8 dimer/monomer transition assay was performed as previously described (9), with minor modifications. Protoplasts were either kept under WL or exposed to narrowband UV-B (400 μW cm−2, 1 hour). Total proteins were extracted from rice protoplasts expressing HA-OsUVR8a, HA-OsUVR8b, HA-OsUVR8aW299A, or HA-OsUVR8bW294A using IP buffer. Protein extracts were resolved by SDS-PAGE without prior boiling. After electrophoresis, the gel was overlaid with a thin layer of SDS-PAGE running buffer to prevent dehydration and irradiated with narrowband UV-B (1000 μW cm−2) for 15 min. Proteins were then transferred to a polyvinylidene difluoride membrane and analyzed by immunoblotting with an anti-HA antibody. BiFC assay BiFC assays were conducted by cotransforming plasmid pairs into rice protoplasts along with NLS-RFP as a nuclear marker. Tested combinations included nV-OsUVR8a/cV-OsNAC3, nV-OsUVR8a/cV-OsNAC3ΔN, nV-OsUVR8b/cV-OsNAC3, nV-OsUVR8b/cV-OsNAC3ΔN, nV-OsNAC3/cV-OsCOP1, and nV-OsNAC3ΔN/cV-OsCOP1. After 16 hours of incubation, yellow fluorescent protein and RFP signals were visualized using fluorescence or confocal microscopy. Results were consistent in at least 50 cells per experiment. Luciferase complementation assay Luciferase complementation assays were performed as described previously (45), with minor modifications. Agrobacterium tumefaciens GV3101 cells carrying nLUC or cLUC fusion constructs were infiltrated into N. benthamiana leaves. Plants were incubated in the dark for 12 hours and then grown under normal conditions for 48 hours. Leaf samples were treated with 100 μM luciferin, and chemiluminescence was imaged using a Tanon 5200 system. To assess the effect of UV-B, half of each leaf was covered with aluminum foil (control), while the other half was exposed to narrowband UV-B (1000 μW cm−2) for 3 hours. The pCAMBIA1302-AtUBQ10pro::GUS was coinfiltrated as an internal control, and LUC/GUS ratios were used to quantify interactions. Protein expression and purification Proteins were expressed in Escherichia coli BL21 (DE3) as described (38). Cultures were induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at OD600 of 0.5 to 0.6 and grown overnight at 16°C. GST- and His-tagged proteins were purified using the respective affinity resins (TransGen Biotech, DP201-01 and DP101-01). MBP-FLAG-OsCOP1 was purified with amylose resin (NEB, E8021S), dialyzed into 1× TEV protease buffer, and cleaved with TEV protease (NEB, P8112S) at 30°C for 3 hours to release FLAG-OsCOP1. Pull-down assay Purified GST-, His-, or FLAG-tagged proteins were incubated with Glutathione–Sepharose beads (GE Healthcare) in 1 ml of pull-down buffer [100 mM tris-HCl (pH 7.5), 150 mM NaCl, 1 mM DTT, 1× protease inhibitor, 0.5% Triton X-100, and 30 μM MG132] at 4°C for 3 hours with rotation. Beads were washed extensively, eluted by boiling in SDS-PAGE loading buffer, separated by SDS-PAGE, and immunoblotted with anti-GST (TransGen Biotech, HT601-01) or anti-His (TransGen Biotech, HT501-01) antibodies. Transcription factor binding site enrichment analysis Upstream 3-kb promoter sequences of OsERF117 were extracted using in-house scripts. Transcription factor binding site scanning was performed using PlantPan v4.0 (46). Luciferase reporter assay Rice protoplasts were cotransfected with reporter and effector plasmids, with AtUBQ10pro::GUS as an internal control. After lysis in 1× cell lysis buffer and centrifugation, supernatants were assayed for firefly luciferase activity using the Dual-Luciferase Assay Kit (Promega). Relative luciferase activity was calculated as LUC/GUS. Protein sequence alignment and in silico 3D structure prediction Multiple sequence alignment of AtUVR8, OsUVR8a, and OsUVR8b was performed using Multalin13 and visualized with ESPript 3.2. The OsNAC3-OsUVR8b complex structure was predicted using AlphaFold (47), with model quality assessed via PAE. Key interface residues were identified by combining PAE analysis and were visualized in UCSF ChimeraX. Critical residues are highlighted in blue (OsNAC3) and red (OsUVR8b). Promoter activity analysis The pGreenII 0800-OsERF117pro::LUC reporter was cotransformed into A. tumefaciens GV3101 along with pGreenII 62-SK, pGreenII 62-SK-OsNAC3, pGreenII 62-SK-OsUVR8a, or pGreenII 62-SK-OsUVR8b. Leaves were agroinfiltrated and incubated for 2 days. Fluorescence imaging was performed with the Tanon 5200 system, and luciferase activity was quantified using the Dualucif Firefly and Renilla Assay Kit (UElandy, China). To test UV-B induction, half of each leaf was shielded with aluminum foil, while the other half was exposed to narrowband UV-B (1000 μW cm−2) for 3 hours. Protein abundance and degradation assays One-week-old dark-grown seedlings of UBQpro::FLAG-OsNAC3, UBQpro::FLAG-OsNAC3/osuvr8ab-1, and UBQpro::FLAG-OsNAC3/OsCOP1-amiR-1-1 were treated with dimethyl sulfoxide (DMSO) or 25 μM MG132 for 2 hours, followed by 100 μM cycloheximide (CHX) for 30 min in darkness. Seedlings were then incubated in continuous darkness or narrowband UV-B (400 μW cm−2) for indicated times. Total proteins were extracted with IP buffer, boiled in 1× SDS loading buffer at 98°C for 8 min, and subjected to immunoblotting with anti-FLAG (Sigma-Aldrich, F3165). Histone H3 (TransGen Biotech, HL102-01) was used as a loading control. For immunoblotting quantification, the band intensities were quantified using ImageJ. In vitro ubiquitination assay The in vitro ubiquitination assay was performed in E. coli BL21 (DE3) as previously described (29). Plasmids pCDFDuet-MBP-NAC3-HA-UBA1-S, pACYCDuet-COP1-MYC-UBC8-S, and pET-28a-FLAG-UBQ10 were cotransformed into E. coli BL21 (DE3). Cultures were induced with 0.5 mM IPTG at an OD600 of 0.4 to 0.6, incubated at 28°C for 10 to 12 hours, and subsequently stored at 4°C overnight. Bacterial pellets were lysed in IP buffer [100 mM tris-HCl (pH 7.5), 150 mM NaCl, 2 mM DTT, and 0.5% Triton X-100] supplemented with EDTA-free protease inhibitor cocktail (Roche) and 30 μM MG132. Total proteins were immunoprecipitated using anti-HA magnetic beads. Crude extracts were resolved by SDS-PAGE and analyzed by immunoblotting with anti-HA, anti-MYC, anti-FLAG, and anti-ubiquitin antibodies. Ubiquitination sites were determined by IP-MS (Suzhou Mass-elife Biotechnologies). In vivo ubiquitination assay Wild-type Kitaake, UBQpro::FLAG-OsNAC3, UBQpro::FLAG-OsNAC3/osuvr8ab-1, and UBQpro::FLAG-OsNAC3/OsCOP1-amiR-1-1 seedlings were grown under darkness or narrowband UV-B (400 μW cm−2) conditions and treated with 25 μM MG132 for 2 hours. This was followed by a 30-min treatment with 100 μM CHX and subsequent sample collection. Proteins were extracted using IP buffer supplemented with 30 μM MG132 and subjected to coimmunoprecipitation with anti-FLAG magnetic beads under darkness or narrowband UV-B for 3 hours. The beads were then washed three times, resuspended in 1× loading buffer, and boiled for 8 min. Immunoprecipitates were analyzed by immunoblotting using anti-FLAG, anti-ubiquitin, and anti-H3 antibodies. For in vivo ubiquitination assays in rice protoplasts, FLAG-OsNAC3 and FLAG-OsNAC35KR transiently coexpressed with Myc-COP1 proteins were cotransfected into rice protoplasts and incubated for 16 hours at 22°C and then treated with 25 μM MG132 for 2 hours. Total protein was extracted using IP buffer and incubated with anti-FLAG magnetic beads (Sigma-Aldrich, M8823) under constant agitation at 4°C for 3 hours. Beads were washed three times, and bound proteins were eluted in SDS sample buffer. Proteins were detected by immunoblotting with anti-ubiquitin, anti-FLAG, and anti-Myc (Sigma-Aldrich, H6908) antibodies. Cell-free protein degradation assays Cell-free protein degradation assays were performed as previously described (31). Total proteins were extracted from 7-day-old Kitaake and OsCOP1-amiR-1-1 grown under darkness or narrowband UV-B (400 μW cm−2) conditions in degradation buffer containing 25 mM tris-HCl (pH 7.5), 10 mM NaCl, 10 mM MgCl2, 4 mM phenylmethylsulfonyl fluoride (PMSF), 5 mM DTT, and 10 mM adenosine 5′-triphosphate (ATP). Total protein (500 μg) was incubated with 300 ng of recombinant protein at 22°C under the same conditions used for seedling growth, and the aliquots were harvested at different time points before boiling and immunoblotting. The proteasome inhibitor MG132 was selectively added as indicated. The samples were analyzed by Western blot using anti-GST and anti-H3 antibodies. Semi–in vitro ubiquitination assay Semi–in vitro ubiquitination assay was performed as previously described (48), with minor modifications. Purified GST-OsNAC3 recombinant protein was initially immobilized on Glutathione–Sepharose beads (GE Healthcare) by incubation at 4°C for 1 hour, followed by three sequential washes with 1× phosphate-buffered saline buffer. Total proteins were extracted from Kitaake and OsCOP1-amiR-1-1 plants using a protein extraction buffer [25 mM tris-HCl (pH 7.5), 10 mM NaCl, 10 mM MgCl2, 4 mM PMSF, 5 mM DTT, 10 mM ATP, and 50 μM MG132]. For the ubiquitination reaction, equal amounts of the protein-bound beads were incubated with equal amounts of the respective total protein extracts at 22°C for 5 hours. After incubation, the beads were washed four times, boiled in 1× SDS loading buffer, and analyzed by Western blotting using anti-ubiquitin and anti-GST antibodies.