NUA and ESD4 negatively regulate ABA signaling during seed germination

The phytohormone abscisic acid (ABA) plays important roles in plant growth, development and adaptative responses to abiotic stresses. SNF1-related protein kinase 2s (SnRK2) are key components that activate the ABA core signaling pathway. NUCLEAR PORE ANCHOR (NUA) is a component of the nuclear pore complex (NPC) that involves in deSUMOylation through physically interacting with the EARLY IN SHORT DAYS 4 (ESD4) SUMO protease. However, it is not clear how NUA functions with SnRK2 and ESD4 to regulate ABA signaling. In our study, we found that nua loss-of-function mutants exhibited pleiotropic ABA-hypersensitive phenotype. We also found that ABA-responsive genes remarkably up-regulated in nua by exogenous ABA. The nua snrk2.2 snrk2.3 triple mutant and nua abi5 double mutant partially rescued the ABA-hypersensitive phenotype of nua, thereby suggesting that NUA is epistatic to SnRK2s. Additionally, we observed that esd4-3 mutant was also ABA-hypersensitive. NUA and ESD4 were further demonstrated to physically interact with SnRK2s and negatively regulate ABA signaling by reducing SnRK2s stability. Taken together, our findings uncover a new regulatory mechanism that can modulate ABA signaling. Supplementary Information The online version contains supplementary material available at 10.1007/s44154-022-00062-1.


Introduction
The phytohormone ABA plays critical roles in regulating plant growth, development, and stress responses to abiotic stresses such as drought, salt, and cold (Cutler et al. 2010). The core ABA signaling pathway consists of three major components: PYRABACTIN RESISTANCE1 (PYR1)/PYR1-LIKE (PYL)/REGULATORY COMPO-NENTS OF ABA RECEPTORS (RCAR) ABA receptors, protein phosphatases 2C (PP2Cs), and SnRK2 protein kinases. In the presence of ABA, PYR1/PYL/PCAR bind with ABA and inhibit the activity of PP2Cs, thus relieving PP2C-mediated inhibition of SnRK2s. The SnRK2 kinases phosphorylate and regulate the activity of downstream components, such as ABI5 and MYB30 transcription factors, which lead to changes in gene expression (Fujii and Zhu 2009;Ma et al. 2009).
Since SnRK2s function as pivotal positive regulators of ABA signaling, they are tightly regulated to help plants respond to ABA and abiotic stresses. Up till now, a number of studies has focused on the role of SnRK2s phosphorylation modification. Disinhibition of PP2Cs is insufficient to activate SnRK2s (Cai et al. 2014; Open Access

Stress Biology
Handling editor: Huazhong Shi. *Correspondence: hrzhang@henau.edu.cn 1 College of Life sciences, Henan Agricultural University, Zhengzhou 450002, China Full list of author information is available at the end of the article et al. 2018a, b, c), as some kinases such as brassinosteroid insensitive 2 (BIN2), abiotic stress-responsive Raflike kinase (ARK) and Raf-like kinases (RAFs) are also required for SnRK2s activation (Saruhashi et al. 2015;Wang et al. 2018a, b, c;Lin et al. 2020;Takahashi et al. 2020). Although the role of SnRK2s' phosphorylation modification has been extensively studied, mechanisms that can regulate their stability remain elusive. Recently, the stability of SnRK2 proteins regulated via ubiquitylation has been studied. Arabidopsis phloem protein 2-B11 (AtPP2-B11), which is part of a SKP1/Cullin/F-box E3 ubiquitin ligase complex, negatively regulates plant responses to ABA through promoting the degradation of SnRK2.3 (Cheng et al. 2017). SnRK2.6 is ubiquitinated by HIGH OSMOTIC STRESS 15 (HOS15), another E3 ubiquitin ligase (Ali et al. 2019). Proteomic analysis also found that SnRK1.1 and SnRK2.4 were the targets of ubiquitination (Kim et al. 2013). Although ubiquitin proteasomemediated degradation of SnRK2s has been identified for regulating ABA responses, the underlying mechanism of how SnRK2s degradation is being triggered remains unknown. The post-translation modifications of SnRK2s for balancing ubiquitylation still needs to be investigated.
The nuclear pore complex (NPC) is a large multiprotein complex that is essential for macromolecular trafficking between the cytoplasm and nucleus in the eukaryotic cells. NUCLEAR PORE ANCHOR (NUA), the Arabidopsis homolog of vertebrate Tpr (Translocated Promoter Region), is a nuclear pore protein that is attached to the inner side of the NPCs. NUA physically interacts with EARLY IN SHORT DAYS 4 (ESD4), a small ubiquitin-related modifier (SUMO) protease, to regulate SUMO-homeostasis in Arabidopsis (Jacob et al. 2007;Xu et al. 2007). SUMO is an ubiquitin-like peptide that is conjugated to target proteins. SUMOylation is a type of post-translational modification, that regulates diverse biological processes including flowering, environmental stress, plant pathogen interactions, cell growth and development (Park et al. 2011). SUMOylation occurs in a series of biochemical steps: E1-activation, E2-conjugation, and E3-ligation. SUMO proteins are first processed by SUMO proteases that truncate a fragment of about ten amino acids long from the C terminus, leaving two glycine residues at the end of the mature SUMO protein.
SUMOylation is a reversible process, which is mediated by SUMO proteases, through their isopeptidase activity (Hickey et al. 2012;Yates et al. 2016). No study to date has investigated the function of NUA and its partner ESD4 SUMO proteases in the ABA signaling pathway. In our study, we found that the nua and esd4 mutants are hypersensitive to ABA during seed germination and seedling growth. Introduction of the snrk2.2 snrk2.3 or abi5 mutation can suppress the ABA hypersensitivity of nua, indicating that NUA acts upstream of SnRK2s in ABA signaling. Furthermore, we detected that NUA physically interacts with SnRK2.2, SnRK2.3, and SnRK2.6; regulating their stability. However, physical interactions between ESD4 and SnRK2s were found to be weaker than those in NUA-SnRK2s, suggesting that NUA can act as an adaptor that links ESD4 proteases and SnRK2s. Our findings thereby revealed a new regulatory mechanism that involves NUA, ESD4 and SnRK2 kinases to mediate ABA signaling.

NUA negatively regulates the ABA signaling pathway
To investigate the role of NUA in ABA signaling, two T-DNA insertion mutants nua-2 (SALK_069922) and nua-3 (SAIL_505_H11) were obtained. The developmental defects of nua-3 were observed to be more severe than those of nua-2 since nua-3 was a stronger loss-offunction NUA allele with T-DNA insertion in the exon (Xu et al. 2007).
We next investigated the effects of ABA treatment on nua mutants at the seed germination stage. Without ABA treatment, the green cotyledon ratios of nua-2 and nua-3 mutants were observed to be similar as the wild type (WT) at Day 5 after the end of vernalization. In the presence of ABA, however, the nua mutants were noted to be more sensitive to the treatment (Fig. 1a). With increasing concentrations of ABA in the MS medium, the cotyledon greening and green cotyledon ratios of nua at the seed germination stage exponentially reduced relative to the WT. (Fig. 1a, b).
We further analyzed the phenotype of nua mutants in response to ABA during the post germination stage. The 5-day-old seedlings on MS medium were transferred to MS medium containing different concentrations of ABA. Five days later, the primary root length elongation was measured. Our results showed that ABA inhibits primary root growth to a greater extent in nua-3 than in WT plants (Fig. 1d, e). In addition to affecting root growth, we also noticed that the shoot development of nua was hindered in the presence of ABA (significant growth inhibition and chlorotic) compared to the WT (Fig. 1c). Moreover, nua-3 was less sensitive to salt stress than the wild type, suggesting that NUA might also be involved in ABA-dependent salt-stress signal pathways (Fig. S1).
We also tested the relative expression levels of the ABA-responsive genes in nua mutant upon ABA treatment. The expressions of ABI5, RD29A and RD29B were remarkably induced by exogenous ABA treatment more than that in the WT (Fig. S2). Taken together, our results indicate that NUA is a negative regulator of ABA signaling.

NUA is epistatic to SnRK2s and ABI5 in ABA signaling
To further elucidate how NUA functions in the ABA signaling pathway, we crossed nua with genes involved in ABA signaling and analyzed their interactions. It has been known that snrk2.2 snrk2.3 and abi5 are insensitive to ABA during the seed germination stage (Fujii and Zhu 2009). We thus isolated nua-3 snrk2.2 snrk2.3 triple mutants and nua-3 abi5 double mutants for studying their responses to ABA during the cotyledon greening stage. In the absence of ABA, the cotyledon greening rates of all mutants were comparable to those of the WT. In the presence of ABA, nua-3 was hypersensitive to ABA, while snrk2.2 snrk2.3 and abi5 Fig. 1 The nua mutants are more sensitive to ABA than the wild type. A Comparisons of cotyledon greening on MS medium, or MS medium supplemented with different concentrations of ABA between wild type (WT) and two NUA loss-of-function mutant alleles. B Statistical analyses of green cotyledon ratios. The ratios were counted at Day 5 after the end of vernalization. At least 30 seeds were counted in each replicate. Data are shown as means ± SD of three independent experiments. ** P < 0.01 (Student's t-test). C The post-germination phenotypes of WT and nua mutants on MS containing ABA. Seeds were germinated on MS medium and grown for 4 d before being transferred to MS medium containing 80 μM ABA. The pictures were taken 7 days after transfer. D Root growth phenotypes of the WT, nua-2 and nua-3 on MS, or MS containing ABA. Seeds were germinated on MS medium and grown for 5 d before being transferred to MS medium or MS medium containing 20 μM ABA. The pictures were taken 5 days after transfer. E Statistical analysis of relative root growth of nua-3 and the WT in (D), the elongation root length of the WT and nua-3 on MS medium was set to 100%. Data are shown as means ± SD of three independent experiments. * P < 0.05 (Student's t-test) mutants were insensitive to ABA as described previously. nua-3 snrk2.2 snrk2.3 triple mutants and nua-3 abi5 double mutants exhibited similar ABA insensitive phenotype as snrk2.2 snrk2.3 and abi5 mutants. These results indicate that NUA might act upstream of SnRK2.2/2.3 and AIB5 (Fig. 2).
To test this hypothesis, we carried out yeast two-hybrid (Y2H) assays. NUA fused with AD domain (NUA-AD) and SnRK2.2/2.3/2.6 fused with BD domain (SnRK2.2/2.3/2.6-BD) were co-transformed into AH109 yeast cells. Their physical interactions were then determined by examining cell growth on selective media. All colonies developed normally on the medium without tryptophan and leucine (-Trp/-Leu). On -Trp/-Leu/-His 3AT and -Trp/-Leu/-His/-Ade plates, only SnRK2s appeared to have interaction with NUA whereas the negative controls failed to survive (Fig. 3a). These results indicated that NUA can directly interact with SnRK2.2, SnRK2.3 and SnRK2.6 in yeast cells. Because there are ten members of the SnRK2 protein kinase family in Arabidopsis, we also tested the interactions between NUA and other SnRK2 kinases, and no interactions were observed among other SnRK2 kinases (Fig. S3).
To determine whether NUA interacts with SnRK2s in planta, we performed split-luciferase complementation assays using Nicotiana benthamiana leaves. The coding sequences of NUA and SnRK2.2/2.3/2.6 were cloned inframe with those of nLUC and cLUC, respectively. The fusion plasmids were then transformed into Agrobacterium. LUC activity was then detected in the leaf area two days after transformation. Our results showed that NUA interacts with SnRK2.2/2.3/2.6 ( Fig. 3b). Together, our Y2H and LCI assays indicated that NUA can physically interact with SnRK2.2/2.3/2.6 to regulate ABA signaling.
ESD4 also interacts with SnRK2.2, SnRK2.3 and SnRK2.6 to negatively regulate ABA signaling It is known that NUA physically interacts with SUMO protease ESD4; and their loss-of-function mutants displayed similar phenotypes which include early flowering, stunted growth, mRNA accumulation in the nuclei, increased SUMO conjugates and decreased free SUMO (Xu et al. 2007). Based on these findings, we speculated that ESD4 may have similar roles in the ABA signaling as NUA. We thus obtained esd4-3 mutant to test this hypothesis. We first counted the green cotyledon ratios at Day 5 after the end of vernalization. The cotyledon greening rates of esd4-3 were similar with those of the WT under normal condition. However, when germinated on MS medium containing 0.2 μM or 0.5 μM ABA, esd4-3 showed greater sensitivity to ABA (Fig. 4a, b). The growth of esd4-3 seedling was also inhibited more by ABA compared to the WT (Fig. 4c). These data indicated that ESD4 can act as a negative regulator during seed germination and post-germination stages in response to ABA.
Moreover, we investigated whether ESD4 can interact with SnRK2s by performing split-LUC complementation assays using cLUC-NUA and SnRK2s-nLUC as positive controls. Our results showed that ESD4 might have interaction with SnRK2s in planta as LUC signals were detected to be weaker than those in the NUA-SnRK2s (Fig. 5a). Furthermore, our Y2H results showed that only SnRK2.2 can directly interact with ESD4 (Fig. 5b). These results together suggested that NUA can act as an adaptor that links ESD4 protease with SnRK2s, then ESD4 and NUA form a large complex with SnRK2s.

NUA and ESD4 affect the stability of SnRK2s
Previous research has reported that E3 ubiquitin ligases affect the stability of SnRK2s, while SUMOylation modification regulates protein stability (Miura et al. 2009;Zheng et al. 2012;Cheng et al. 2017). NUA has been reported to couple with ESD4 SUMO protease to participate SUMOylation (Xu et al. 2007). Since physical interactions were observed between NUA and SnRK2s in our study, we wanted to determine whether NUA and ESD4 can affect the stability of SnRK2s to mediate ABA signaling. We thus detected the protein levels of SnRK2s in 7-d-old nua-3 and esd4-3 seedlings that have been treated with 50 μM ABA at indicated time points. Our western blotting results showed that SnRK2 proteins level decrease more gradually in the nua-3 and esd4-3 mutants relative to WT. The proteins level of esd4-3 mutant treated with ABA was obviously higher than that of WT. In terms of contrast, the SnRK2s in nua-3 were more stable than WT only after 2-h ABA treatment (Fig. 6a). To test if ESD4 is more important than NUA, we investigated the SUMO conjugation pattern in nua-3 and esd4-3, compared with the wild type. The results showed that both nua-3 and esd4-3 mutants lead to an increase in SUMO conjugates and decrease in free SUMO. But esd4-3 mutants accumulated more SUMO conjugates and had less free SUMO than nua-3 mutant, indicating that ESD4 played a major role in deSU-MOylation, thus it is reasonable that SnRK2s in esd4-3 were more stable than nua-3 (Fig. 6a, b). To exclusive the affects of transcription, we detected the mRNA level of SnRK2s in the mutants upon 50 μM ABA treatment. Real-time quantitative PCR (RT-qPCR) analysis revealed that the expressions of SnRK2.2 and SnRK2.6 were induced by ABA, while the expression level of SnRK2.3 had no obvious changes. Moreover, there were no significant differences between WT and the mutants after ABA treatment (Fig. 6c). These results indicated that the changes of protein level (Fig. 6a) were not due to the mRNA level. Combined with the previous studies, we conclude that NUA and ESD4 can affect the stability of SnRK2s in the ABA signaling pathway in Arabidopsis.

Discussion
NUA is a large protein located on the nucleoplasmic side of the NPCs in Arabidopsis. NUA physically interacts with ESD4 SUMO protease to regulate SUMO-homeostasis, mRNA export and flowering (Xu et al. 2007). In this study, we uncovered a novel function of NUA acting as a negative regulator in ABA signaling. Fig. 4 The esd4-3 mutant is hypersensitive to ABA. a Phenotypic analyses of the wild type (WT) and esd4-3 mutant treated by 0.2 or 0.5 μM ABA. The images were taken at Day 5 after the end of vernalization. b Seedling greening rates of esd4-3 and WT plants. The ratios were counted at Day 5 after the end of vernalization. At least 30 seeds were counted in each replicate. Data are shown as means ± SD of three independent experiments. * P < 0.05 (Student's t-test). c The post-germination phenotype of esd4-3 treated with ABA. The WT and esd4-3 mutant seeds were germinated on MS medium or MS medium containing 0.2 μM ABA, and the pictures were taken 3 weeks after germination Although previous reports have shown that SUMOylation modification participates in ABA signaling (Lois et al. 2003;Miura et al. 2009;Zheng et al. 2012), no studies have yet investigated the functions of NUA and ESD4 in the ABA signaling pathway. Here, we revealed that both nua and esd4 mutants were hypersensitive to ABA in the germination and post-germination stages (Figs. 1,  4). The genetic interaction studies showed that NUA can act upstream of SRNK2s in ABA signaling (Fig. 2). The Y2H and LCI assays confirmed the physical interactions of NUA and SnRK2.2/2.3/2.6 (Fig. 3). The interactions between ESD4 and SnRK2s were observed to be weaker (Fig. 5a), so NUA might serve as an adaptor that links ESD4 SUMO protease and SnRK2s to negatively regulate SnRK2s stability. Based on these results, we speculate that SnRK2 kinases might be substrates of the ESD4 SUMO protease. SnRK2s have SUMO-binding ΨKXE motif(s) using SUMOplot (http:// www. abgent. com/ sumop lot) analyses and are nuclear localized ), which are consistent with previous evidence of most SUMOylated proteins are nucleoproteins containing at least one ΨKXE sequence (Miller et al. 2010).
Further experimental evidence is needed to confirm SnRK2s can be SUMOylated in Arabidopsis. It has been shown that the level of high molecular weight SUMO conjugates increased in both nua and esd4 mutants (Fig. 6b) (Murtas et al. 2003;Xu et al. 2007). It is possible that the level of SUMO-SnRK2 conjugates might also increase in nua and esd4 mutants.
It was previously reported that SUMOylation modulates protein stability. For example, SIZ1 (SAP and Miz) is a principal SUMO E3 ligase in Arabidopsis. SIZ1 negatively controls ABA signaling through the SUMOylation of ABI5 and MYB30 transcription factors. The conjugation of SUMO to ABI5 represses its activity but prevents ABI5 from undergoing degradation (Miura et al. 2009), while SUMOylation of MYB30 stabilizes MYB30 upon ABA treatment (Zheng et al. 2012). SIZ1 is also involved in aluminum (Al) resistance by SUMOylation of the zinc-finger transcription factor SENSITIVE TO PROTON RHIZOTOXICITY 1 (STOP1). The STOP1 protein levels were decreased in siz1 mutant (Fang et al. 2021;Mercier et al. 2021;Roy and Sadanandom 2021;Xu et al. 2021). Interestingly, the SUMOylation level of STOP1 is also regulated by ESD4 (Q. Fang et al. 2020). Our findings of NUA and ESD4 interacting with SnRK2s (Figs. 3, 5), combining with the protein stability results (Fig. 6a), suggested that NUA and its partner ESD4 SUMO protease can modulate SnRK2s stability mostly via affecting their SUMOylation levels.
Furthermore, it was reported that SUMOylation affects protein stability by regulating ubiquitination-mediated protein degradation (Lois et al. 2003;Wilkinson and Henley 2010;Fang et al. 2020;Mercier et al. 2021). The SUMO substrates, such as ABI5, MYB30 and STOP1 proteins described above, are also regulated by the ubiquitin-26S proteasome pathway Fig. 6 Protein abundance and transcription levels of SnRK2s in the nua-3 and esd4-3 mutants upon ABA treatment. a NUA and ESD4 positively regulate ABA-induced degradation of SnRK2s. Seeds were sown on 1/2 MS plates. 7-DAG seedlings were treated with 50 uM ABA for the indicated time. Total protein was extracted from the treated seedlings, and immunoblot assays were performed with anti-SnRK2.2/2.3/2.6 and ACTIN antibodies. The snrk2.2/2.3 and snrk2.6 (ost1-3) mutants were used as molecular weight references. b Accumulation of SUMO Conjugates in nua-3 and esd4-3. Total protein was extracted from 14 d-old seedlings of the wild type, nua-3 and esd4-3. 20 μg proteins were loaded onto an SDS-PAGE gel, and the immunoblot was probed with anti-SUMO1 antibody. c 7-DAG seedlings were treated by 50 μM exogenous ABA for 2 h and 4 h, and then the plants were collected for RNA extraction and Real-time quantitative PCR (RT-qPCR) analysis. The ACTIN2 gene was used as an internal control. The expression levels of the indicated genes in WT were set to 1. Data are shown as means ± SD of three independent experiments (Lopez-Molina et al. 2003;Zheng et al. 2018;Zhang et al. 2019). SUMOylation and ubiquitination might act antagonistically to regulate the stability of proteins so SUMOylation proteins are mostly stable because SUMO modification can prevent the ubiquitin proteasome-mediated degradation of proteins. ABA has been reported to promote the degradation of SnRK2.3 via ubiquitin proteasome-mediated pathway (Cheng et al. 2017). In our current work, we detected SnRK2 proteins were more stable in nua-3 and esd4-3 than in the WT (Fig. 6), suggesting that SUMO modification represses ubiquitination-mediated degradation of SnRK2 kinases. Future studies are needed to examine the ubiquitination level of SnRK2s in nua and esd4 mutants.
A previous report has revealed that Oryza sativa OVERLY TOLERANT TO SALT1 (OsOTS1), a SUMO protease in rice, negatively regulated ABA signaling by reducing the stability of OsbZIP23 transcription factor (Srivastava et al. 2017). This finding coincides with our results of how NUA and ESD4 may function in regulating Arabidopsis' ABA signaling. We want to make a note that in one report, Arabidopsis SUMO protease 1 (ASP1) was shown to positively regulate ABA signaling by promoting MYB30 stability and inhibiting the degradation of ABI5 (Wang et al. 2018a, b, c). This contradiction shows that the underlying regulatory mechanisms of SUMO protease are distinct. In our study, we found that abi5 rescued the germination phenotype of nua mutant (Fig. 2) due to the epistatic effect of NUA to SnRK2s as well as SnRK2s acting upstream of ABI5. However, it is also possible that NUA can directly modulate ABI5. Moreover, MYB30 might also be a substrate of NUA and ESD4. Therefore, it would be interesting to investigate the interactions between NUA, ABI5, and MYB30 in the future.
In conclusion, we demonstrated that NUA and ESD4 can play critical roles in response to ABA by regulating the stability of SnRK2s and mediating seed germination and seedling growth in Arabidopsis. This work lays a foundation for a further dissection of the mechanism by which SUMO modification regulates ABA signaling.
For the seed germination assay, seeds were surfacesterilized and then grown on 1/2 MS medium with or without ABA containing 1% (w/v) sucrose and 1% agar. The plates were stratified in darkness for 2 days at 4°C and then transferred to a light incubator at 22°C under a 16-h light /8-h dark photoperiod.

RNA extraction and Quantitative RT-PCR
Total RNA was extracted from 7-day-old seedlings with TRIzol reagent ((Life, Invitrogen). The RNA was reverse transcribed to cDNA with MMLV reverse transcriptase (Yeasen). Quantitative RT-PCR was performed on an IQ5 Multicolor Real-Time PCR Detection system, using qPCR SYBR Green Master Mix (No Rox, Yeasen). ACTIN2 was used as an internal control. For each genotype, three different samples were used for three replicates. The primers used for qRT-PCR analysis are listed in Supplemental  Table 1.

Yeast Two-Hybrid Assay
The NUA or ESD4 coding region was cloned in-frame between the NdeI and BamHI sites of pGADT7. The full-length coding sequences of SnRK2s were cloned into pGBKT7 (Hou et al. 2016). Plasmid DNAs of bait and prey constructs were transformed into the S. cerevisiae strain AH109. The protein interactions were assayed on selection media plates (-Leu/-Trp/-His or -Leu/-Trp/-His/-Ade).

Split-LUC Complementation Analysis
The split-LUC complementation analysis was performed as previously described (Zhou et al. 2018). The CDSs of SnRK2.2, SnRK2.3 and SnRK2.6 were cloned into pCAM-BIA1300-nLUC. The CDSs of NUA and ESD4 were cloned into pCAMBIA1300-cLUC. A. tumefaciens (strain GV3101) cells containing above plasmids were injected into N. benthamiana leaves. The empty cLUC and nLUC vectors were used as negative controls. After infiltration, plants were kept in the greenhouse for 40 to 48 hr. N. benthamiana leaves were incubated with 1 mM D-luciferin (Goldbio) in darkness for 3 min and then subjected for LUC signal detection by using Chemiluminescent Imaging System.

Immunoprecipitation and Sumoylation Assays
The SUMOylation analysis was performed as previously described (Xu et al. 2007). Total protein was extracted from 14-d-old seedlings grown on 1/2 MS medium. Plant tissues (0.2 g) were extracted with extraction buffer. Protein concentration was measured using the Bio-Rad Protein Assay reagent (Bio-Rad, USA), and 20 μg protein was separated by SDS-PAGE, transferred to PVDF membrane, probed with anti-SUMO1 antibody (ab5316, Abcam).

Chlorophyll measurement
Chlorophyll was extracted from leaves with 95% ethyl alcohol in water, and the absorption of the extracted solutions was measured at 663 and 645 nm using a microplate reader. The chlorophyll a concentration was calculated as (12.7×OD 663 -2.69×OD 645 )× dilution fold /fresh weight. The chlorophyll b concentration was calculated as (22.9× OD 645 -4.68 × OD 663 )×dilution fold /fresh weight.