An Arabidopsis SUMO E3 Ligase, SIZ1, Negatively Regulates Photomorphogenesis by Promoting COP1 Activity

COP1 (CONSTITUTIVE PHOTOMORPHOGENIC 1), a ubiquitin E3 ligase, is a central negative regulator of photomorphogenesis. However, how COP1 activity is regulated by post-translational modifications remains largely unknown. Here we show that SUMO (small ubiquitin-like modifier) modification enhances COP1 activity. Loss-of-function siz1 mutant seedlings exhibit a weak constitutive photomorphogenic phenotype. SIZ1 physically interacts with COP1 and mediates the sumoylation of COP1. A K193R substitution in COP1 blocks its SUMO modification and reduces COP1 activity in vitro and in planta. Consistently, COP1 activity is reduced in siz1 and the level of HY5, a COP1 target protein, is increased in siz1. Sumoylated COP1 may exhibits higher transubiquitination activity than does non-sumoylated COP1, but SIZ1-mediated SUMO modification does not affect COP1 dimerization, COP1-HY5 interaction, and nuclear accumulation of COP1. Interestingly, prolonged light exposure reduces the sumoylation level of COP1, and COP1 mediates the ubiquitination and degradation of SIZ1. These regulatory mechanisms may maintain the homeostasis of COP1 activity, ensuing proper photomorphogenic development in changing light environment. Our genetic and biochemical studies identify a function for SIZ1 in photomorphogenesis and reveal a novel SUMO-regulated ubiquitin ligase, COP1, in plants.


Author Summary
In darkness, the ubiquitin E3 ligase COP1 accumulates in the nucleus and mediates ubiquitination and degradation of positive regulators of photomorphogenesis, such as HY5. In response to light, COP1 activity is reduced to ensure proper photomorphogenic development. However, post-translational modifications that regulate COP1 activity are largely Introduction Sumoylation is a post-translational modification in which SUMO (small ubiquitin-like modifier) peptides are covalently attached to a SUMO consensus motif (ψKxE/D; ψ a large hydrophobic amino acid residue; K, the acceptor lysine; x, any amino acid; E/D, glutamate or aspartate) in target proteins through a series of biochemical steps involving activation (E1), conjugation (E2), and ligation (E3) enzymes [1,2]. SUMO conjugation can be reversed by SUMO-specific proteases [3].
Recent study has revealed that SUMO modification of phyB represses red light signaling, at least partly, through inhibiting interaction between phyB and PIFs [37]. In this study, we demonstrate that SIZ1 negatively regulates photomorphogenesis, at least partly, through promoting COP1 ubiquitin E3 ligase activity by SUMO modification, and that COP1 in turn mediates the ubiquitination and degradation of SIZ1. Our results reveal a novel regulatory mechanism of COP1 and SIZ1 in photomorphogenesis.

The SUMO E3 Ligase SIZ1 Negatively Regulates Photomorphogenesis
The observation that loss-of-function siz1 mutant seedlings showed a short-hypocotyl phenotype under white light prompted us to evaluate the light responsiveness of siz1-2. The siz1-2 seedlings exhibited a short-hypocotyl phenotype under darkness and red, blue, and far-red light conditions (Fig 1A and 1B).
Expression of ProSIZ1:SIZ1-GFP in siz1-2 plants (complemented lines referred to as SSG) [38] suppressed the short-hypocotyl phenotype of siz1-2 under these conditions, indicating that mutation of SIZ1 is responsible for the short-hypocotyl phenotype (Fig 1A and 1B). The short-hypocotyl phenotype of siz1-2 was due to a reduction in cell elongation but not in cell number (S1 Fig). The expression of light-inducible genes, CAB (CHLOROPHYLL A/B BIND-ING PROTEIN) and RBCS (RUBISCO SMALL SUBUNIT) [39,40], and a light-repressed gene, PORA (PROTOCHLOROPHYLLIDE OXIDOREDUCTASE A) [41], was up-and down-regulated, respectively, in siz1-2 during the transition from darkness to light, indicating that the regulation is stronger in siz1-2 than in wild type ( Fig 1C). In addition to the short-hypocotyl phenotype, siz1-2 seedlings had unfolded apical hooks in darkness ( Fig 1D) and exhibited more opened cotyledons than did wild-type seedlings in dark and light conditions ( Fig 1E). These results suggest that SIZ1 negatively regulates photomorphogenesis.
The Short-Hypocotyl Phenotype of siz1-2 is Due to Impaired SUMO1/2 Modification and Not Elevated SA To determine if SIZ1-mediated SUMO1/2 modification is involved in the regulation of hypocotyl elongation, we determined the hypocotyl length of sum1 and sum2 double mutant under darkness and red, blue, and far-red light conditions (Fig 2). SUMO1 and 2 have redundant functions, and the sum1 sum2 double knockout mutant is embryo lethal [42]. Therefore, we used a viable weak allele, sum1-1 amiR-SUM2, in which SUMO2 expression is down-regulated by RNAi in the sum1-1 knockout mutant background [43]. Similar to siz1-2, sum1-1 amiR-SUM2 seedlings exhibited a short-hypocotyl phenotype under dark and light conditions, suggesting that SUMO1/2 modification regulates hypocotyl elongation (Fig 2).
Elevated SA levels in siz1-2 cause dwarfism, and the dwarf phenotype is substantially suppressed by expression of NahG, a bacterial salicylate hydroxylase [44]. To determine if the short-hypocotyl phenotype of siz1-2 is due to increased SA levels, the hypocotyl lengths of fiveday-old wild-type, siz1-2, NahG, and NahG siz1-2 seedlings were compared under darkness and red, blue, and far-red light conditions (Fig 2). The expression of NahG did not affect hypocotyl elongation, but the siz1-2 and NahG siz1-2 seedlings exhibited an identical short-hypocotyl phenotype in all tested conditions, indicating that the elevated SA level in siz1-2 did not contribute to the short-hypocotyl phenotype. The siz1 mutant seedlings display a short-hypocotyl phenotype. (A) siz1-2 seedlings exhibit a short-hypocotyl phenotype under red (R: 10 μmol m -2 s -1 ), blue (BL: 14 μmol m -2 s -1 ), and far-red (FR: 12 μmol m -2 s -1 ) light conditions, and this phenotype is rescued by complementation with SIZ1 driven by its own promoter (SSG). Bar = 2 mm. (B) Hypocotyl length of five-day-old Col-0, siz1-2, and SSG under darkness and red (R), blue (BL), and far-red (FR) light conditions at the indicated fluence rates. Data are the mean ± SE of 30 seedlings. (C) qRT-PCR analysis showing the enhanced responsiveness of lightresponsive genes in siz1-2 seedlings compared to those in the wild type under the dark to light transition. Five-day-old dark-grown seedlings were transferred to white light for an additional 6 h. Relative expression was normalized to that of UBC. Error bars indicate ± SE (n = 3). (D) siz1-2 seedlings exhibit unfolded apical hooks under dark conditions (DK). Bar = 0.5 mm. (E) siz1-2 seedlings show more opened cotyledon compared to Col-0 under dark (DK), red (R; 10 μmol m -2 s -1 ), blue (BL; 14 μmol m -2 s -1 ), and far-red (FR; 12 μmol m -2 s -1 ) light conditions. ** Student's t-test indicates significant differences between the Col-0 and siz1-2 (P 0.01).

SIZ1 Mediates SUMO Modification of COP1
Since siz1-2 seedlings exhibit a weak cop1-like phenotype, we examined whether SIZ1 physically interacted with COP1 to regulate its activity. To do so, we performed a bimolecular fluorescence complementation (BiFC) assay. YFP fluorescence signals were detected in the nucleus of N. benthamiana cells that coexpressed COP1-YFP N (fused with the N-terminal half of YFP) and SIZ1-YFP C (fused with the C-terminal half of YFP) under light (1 h light exposure) and dark conditions (Fig 3A and S2 Fig).
SUMOplot (http://www.abgent.com/sumoplot) analysis predicted the presence of three potential sumoylation motifs [45] (VK14PD, IK193ED, and WK653SD) in COP1 (S3 Fig), suggesting that COP1 may be a SUMO substrate. To test this possibility, we performed an in vitro sumoylation assay to determine SUMO modification of COP1 as described previously [15]. Anti-FLAG and anti-SUMO1 antibody detected slow migrating multiple bands above original COP1 protein in the reaction containing SUMO E1 (His 6 -SAE1b and His 6 -SAE2), SUMO E2 (His 6 -SCE1), and His 6 -SUMO1-GG, but not in the reaction lacking His 6 -SUMO1-GG, suggesting that COP1 is a possible SUMO substrate (Fig 3C). To further confirm COP1 is sumoylated in vivo, we performed an in vivo sumoylation analysis as described previously [38]. We co-expressed Myc-COP1 with FLAG-SUMO1 or FLAG-SUMO1 AA (a conjugation-deficient mutant) in Arabidopsis protoplasts or N. benthamiana leaves. Myc-COP1 was immunoprecipitated with anti-Myc antibody and the immunoprecipitated proteins were detected with anti-FLAG antibody. Higher molecular weight sumoylated COP1 bands were detected when Myc-COP1 and FLAG-SUMO1 were co-expressed, but not in Myc-COP1 and FLAG-SUMO1 AA co-expressing cells (Fig 3D and S4A Fig). Moreover, to confirm that sumoylation of COP1 occurs in planta, we generated an Myc-COP1 overexpression transgenic line (referred to as 35S-Myc-COP1) and performed an in vivo sumoylation analysis. Anti-SUMO1 antibody detected higher molecular weight sumoylated COP1 bands ( Fig 3E). Furthermore, we evaluated  Sumoylation of COP1 in planta. Total proteins were extracted from five-day-old dark-grown 35S-Myc-COP1 and Col-0 (control) seedlings, and anti-Myc antibody was used to immunoprecipitate Myc-COP1. Anti-SUMO1 antibody was used to determine sumoylated COP1. Input and immunoprecipitated Myc-COP1 were detected with anti-Myc antibody. Arrowhead indicates non-sumoylated COP1 band. Asterisks indicate sumoylated COP1 bands. (F) Light exposure reduces sumoylation levels of COP1. Myc-COP1 and FLAG-SUMO1 co-expressing N. benthamiana leaves were incubated under darkness for 12 h, and then exposed to white light (150 μmol m -2 s -1 ) for 12 h. The nuclear proteins were isolated at the end-of-dark (0 h) and the end-of-light (12 h) period, and the sumoylation level of COP1 was analyzed as described in (D). Input FLAG-SUMO1 and FLAG-SUMO1 AA were detected with anti-FLAG antibody in a separate blot, shown in S5B Fig. (G) The level of COP1 sumoylation was substantially lower in siz1-2 than in Col-0. Myc-COP1 and FLAG-SUMO1 were transiently co-expressed in Col-0 or siz1-2 protoplasts, and the sumoylation level of COP1 was analyzed as described in (D). (H) K193 is a primary sumoylation site in COP1. FLAG-SUMO1 was transiently co-expressed with Myc-COP1, Myc-COP1 K14R , Myc-COP1 K193R , or Myc-COP1 K653R in Col-0 protoplasts, and immunoprecipitation was performed as described in (D). if sumoylation of COP1 changes in response to light. Since light exposure promotes nucleus to cytosol re-localization of COP1 [46], we monitored the sumoylation status of COP1 in the nucleus under dark (0 h) and light (12 h) conditions ( Fig 3F). Under darkness, sumoylated COP1 bands were detected in the Myc-COP1 and FLAG-SUMO1 co-expressing nuclear fraction, but the sumoylation level of COP1 was substantially reduced in response to 12 h of light exposure. Taken together, these data demonstrate that COP1 is a SUMO substrate, and sumoylation level of COP1 is regulated by light.
To test if SIZ1 mediates SUMO conjugation of COP1, we cotransformed Myc-COP1 and FLAG-SUMO1 into Arabidopsis protoplasts isolated from wild-type or siz1-2 plants, and analyzed the sumoylation status of COP1. Whereas COP1-SUMO1 conjugate was detected in the wild type, substantially lower levels were present in siz1-2 ( Fig 3G), indicating that SIZ1 facilitates the sumoylation of COP1. The lower levels of sumoylated COP1 in siz1-2 may be due to another SUMO E3 ligase(s) that facilitates the residual SUMO modification of COP1. Alternatively, it is also possible that E1 and E2 contribute to basal levels of COP1-SUMO conjugation, since E1 and E2 alone mediate the sumoylation of COP1 in vitro ( Fig 3C). K-to-R substitutions in sumoylation motifs block SUMO conjugation [47]. To elucidate the sumoylation motifs of COP1, SUMO conjugation of Myc-COP1 K14R , Myc-COP1 K193R , and Myc-COP1 K653R were evaluated in Arabidopsis protoplasts. The K193R substitution blocked COP1-SUMO1 conjugation, but K14R or K653R substitutions did not ( Fig 3H). Unfortunately, we could not confirm the effect of K193R substitution in vitro, due to anti-FLAG antibody detected long smear bands above the original purified MBP-COP1 K193R -FLAG, which would strongly affect subsequent in vitro sumoylation analysis. COP1-SUMO1 conjugation was also blocked by the K193R substitution in N. benthamiana leaves (S4B Fig). These results indicate that SIZ1 mediates sumoylation of COP1 and that K193 is critical for SUMO conjugation.

SIZ1-Mediated Sumoylation Positively Regulates COP1 Activity
Since SIZ1 mediates SUMO modification of COP1 and siz1-2 seedlings exhibit a weak cop1like phenotype (Figs 1 and 3), we hypothesis that SIZ1-mediated SUMO modification may enhances COP1 activity. To test this possibility, we first determined the effect of sumoylation of COP1 in planta. Myc-COP1 and Myc-COP1 K193R (a non-sumoylated form) overexpressing Arabidopsis transgenic plants were generated, and two independent lines with similar levels of transgene expression for each construct were selected for further phenotypic analysis (S6 Fig). 35S-Myc-COP1 seedlings exhibited longer hypocotyls than did the wild type under white light conditions (Fig 4A), confirming a previous report [48].
Furthermore, genetic interaction between COP1 and SIZ1 was analyzed. The hypocotyl length of cop1-4 [51] was shorter than that of the wild type and siz1-2 under dark and light conditions (Fig 4G and 4H and S7 Fig). The cop1-4 siz1-2 double mutant exhibited a short-hypocotyl phenotype similar to cop1-4 in both dark and light conditions, suggesting that SIZ1 regulates hypocotyl elongation partly through COP1 (Fig 4G and 4H and S7 Fig). SIZ1 Negatively Regulates HY5 Levels SIZ1 enhances COP1 activity (Fig 4) and COP1 mediates ubiquitination and degradation of HY5 [19]. Therefore, we tested if the protein level of HY5 is up-regulated by mutation of SIZ1. Anti-HY5 antibody [52] revealed that endogenous HY5 was more abundant in siz1-2 than in the wild type under light conditions (Fig 5A).
The loss-of-function hy5-215 seedlings exhibited a long-hypocotyl phenotype under light conditions [54]. To identify genetic interactions between HY5 and SIZ1, we generated the hy5-215 siz1-2 double mutant through genetic crossing. The short-hypocotyl phenotype of siz1-2 was partially suppressed by hy5-215 under various light conditions (Fig 5D and 5E), suggesting that the accumulation of HY5 in siz1-2 at least partly accounts for the short-hypocotyl phenotype under light conditions. In addition to HY5, COP1 also mediates the ubiquitination and degradation of HYH, LAF1, HFR1, STH3/BBX2 and PIL1 [19][20][21][22][23][24]. Thus, it is possible that the protein levels of these positive regulators of photomorphogenesis are higher in siz1-2, which may causes a weak cop1-like phenotype of siz1-2.

SUMO Modification May Enhances the Transubiquitination Activity of COP1
SIZ1-mediated SUMO modification may regulate COP1 expression, COP1 stability, nuclear accumulation level, interaction with other proteins and/or affects its enzymatic activity. The level of COP1 expression was not significantly affected by the siz1-2 mutation, indicating that SIZ1 does not regulate the transcription of COP1 (Fig 6A).
Anti-COP1 antibody [30] revealed that endogenous COP1 protein abundance was similar in the wild type and siz1-2, indicating that the reduced COP1 activity in siz1-2 was not due to reduced levels of COP1 (Fig 6B). Since COP1 functions as a homodimer [55], we examined whether SUMO modification affected COP1 dimerization. Immunoprecipitation analysis showed that Myc-COP1 and Myc-COP1 K193R (a non-sumoylated form) immunoprecipitated the same amount of FLAG-COP1, indicating that sumoylation of COP1 did not significantly affect COP1 dimerization ( Fig 6C). Next, we examined if SUMO modification enhanced the COP1-HY5 interaction using in vitro co-immunoprecipitation assays. Sumoylation of COP1 did not enhance the substrate accessibility of COP1 under dark and light conditions ( Fig 6D). Moreover, the siz1 mutation did not affect the level of nuclear COP1 under dark and light conditions ( Fig 6E). Finally, we evaluated if SUMO conjugation regulates transubiquitination acitivity of COP1. In vitro sumoylated and non-sumoylated MBP-COP1-FLAG were used as E3s to perform an in vitro HY5 ubiquitination assay. Anti-GST and anti-ubiquitin antibodies detected higher level of ubiquitinated proteins in the reaction containing sumoylated COP1 than did non-sumoylated COP1 (Fig 6F), suggesting that SIZ1-mediated SUMO modification may enhances transubiquitination activity of COP1.

COP1 Mediates the Ubiquitination and Degradation of SIZ1
Although the biological functions of SIZ1 have been extensively characterized, the mechanism that regulates SIZ1 activity is largely unknown. Our finding that COP1 interacted with SIZ1 (Fig 3A and 3B) prompted us to determine if COP1 mediates the ubiquitination and degradation of SIZ1. To determine if COP1 promotes degradation of SIZ1, we analyzed the decay kinetics of SIZ1-GFP in the presence or absence of Myc-COP1. The degradation of SIZ1 was more rapid in the presence of Myc-COP1 (Fig 7A).
Moreover, the COP1-promoted SIZ1 degradation was inhibited by MG132, a 26S proteasome inhibitor (Fig 7B, upper panel). In agreement with a previous report [21], MG132 also repressed the degradation of COP1 (Fig 7B, middle panel). SIZ1 expression was not affected in cop1-4, indicating that COP1 does not regulate SIZ1 transcription (Fig 7C). Next, we evaluated if COP1 mediates the ubiquitination of SIZ1. MBP-COP1-FLAG was used as the E3 ligase in an in vitro SIZ1 ubiquitination assay. Anti-Myc and anti-ubiquitin antibody revealed that COP1 facilitates ubiquitination of SIZ1 in vitro (Fig 7D). These results suggest that COP1 negatively regulates SIZ1 protein stability through ubiquitination and subsequent 26S proteasomedependent degradation. Five-day-old dark-grown seedlings were transferred to white light for the indicated time periods. The relative expression level of COP1 was normalized to that of UBC, and data represent the mean ± SE (n = 3). (B) COP1 levels in Col-0 and siz1-2 under dark and light conditions. Five-day-old dark-grown seedlings (Dark) were exposed to light for 12 h (Light 12 h). COP1 was detected with anti-COP1 antibody. Tubulin was detected with anti-Tubulin antibody as a loading control. Numbers above the blot indicate the relative level of COP1 normalized to that of Tubulin. (C) Analysis of the effect of SUMO modification on COP1 dimerization. FLAG-COP1 and FLAG-SUMO1 were transiently co-expressed with Myc-COP1 or Myc-COP1 K193R in N. benthamiana leaves. Myc-COP1 or Myc-COP1 K193R was immunoprecipitated (IP) with anti-Myc antibody, and FLAG-COP1 in the precipitates was detected with anti-FLAG antibody. Sumoylated COP1 was detected with anti-SUMO1 and anti-FLAG antibody (longer exposure). The level of immunoprecipitated Myc-COP1 or Myc-COP1 K193R was detected with anti-Myc antibody. (D) In vitro immunoprecipitation analysis of the COP1-HY5 interaction. GST-HY5 was incubated with total protein extract isolated from N. benthamiana co-expressing Myc-COP1 with FLAG-SUMO1 or FLAG-SUMO1 AA under dark (0 h) or light (12 h) conditions in the presence of 50 μM MG132. After 1.5 h incubation at 4°C, Myc-COP1 was immunoprecipitated with anti-Myc antibody, and co-immunoprecipitated GST-HY5 was detected with anti-GST antibody. Immunoprecipitated Myc-COP1 was quantified with anti-Myc antibody and sumoylated COP1 was detected with anti-FLAG antibody. (E) Nuclear fractions were isolated from COP1 OE and COP1 OE siz1-2 under dark (0 h) and light (12 h white light exposure) conditions, and level of nuclear COP1 was detected with anti-COP1 antibody. Nuclear fraction of Col-0 and total protein extract of cop1-4 was used as a control. Histone 3 and tubulin were used as nuclear and cytosol marker proteins, respectively. (F) In vitro sumoylated and non-sumoylated MBP-COP1-FLAG were used as E3 ligases to perform an in vitro HY5 ubiquitination assay. Bead-conjugated GST-HY5 was used as substrate. After reaction, the beads were washed and ubiquitinated GST-HY5 were eluted for immunoblot analysis with anti-GST and anti-ubiquitin antibodies. MBP-FLAG was used as a negative control. The vertical line indicates ubiquitinated HY5. Input E3s were detected with anti-FLAG antibody. Asterisks indicate non-specific bands. Discussion COP1 protein stability is regulated by CSU1-mediated ubiquitination [27]. However, how the COP1 E3 ligase activity is regulated by post-translational modifications remains largely unknown. Our genetic and biochemical analyses revealed that COP1 activity is enhanced by SIZ1-mediated SUMO modification.

SIZ1-Mediated Sumoylation Promotes COP1 Activity
The reduced COP1 activity in siz1 resulted in higher levels of HY5, stronger down-regulation of cell elongation-related HY5 direct target genes (Fig 5A and 5C). These results strongly suggest that SIZ1-mediated SUMO modification enhances COP1 activity; thus, COP1 is a plant SUMO-regulated ubiquitin ligases (SRUbL).
The hypocotyl length of cop1-4 siz1-2 was slightly shorter than that of cop1-4 (Fig 4G and  4H and S7 Fig), suggesting that SIZ1 regulates hypocotyl elongation through at least two pathways, a COP1-dependent and -independent pathway. Recently, it has been shown that SUMO modification of phyB negatively regulate photomorphogenesis under red light [37]. OTS1 (OVERLY TOLERANT TO SALT1) mediate desumoylation of phyB, but whether the SUMO modification is facilitated by SIZ1 remains to be determined. However, it should be noted that cop1-4 is a weak allele, which expresses a partially functional truncated COP1 protein (1-282 aa, which contains the K193 sumoylation motif) [51]. Therefore, it is also possible that the partial COP1 activity in cop1-4 is further attenuated in cop1-4 siz1-2.
It has been shown that a certain amount of COP1 is present in the nucleus even under prolonged light exposure, and plays a critical role in regulating development [27,51,58,[62][63][64]. Interestingly, prolonged light exposure reduced sumoylation levels of nuclear-localized COP1 (Fig 3F), which results in decreased COP1 activity. Thus, we hypothesize that light-induced reduction of COP1-SUMO levels is required for the maintenance of moderate COP1 activity under light conditions to ensure the tight regulation of photomorphogenesis. The reduced sumoylation level of COP1 under light conditions may be due to decreased SIZ1-mediated sumoylation and/or increased SUMO protease(s)-mediated desumoylation, but the details of the mechanisms by which light regulates the balance between sumoylation and desumoylation remain to be elucidated.
Collectively, our study demonstrates that SIZ1-mediated sumoylation negatively regulates photomorphogenesis, at least partly, through enhancing COP1 activity. Interestingly, COP1 in turn mediates the ubiquitination and 26S proteasome-dependent degradation of SIZ1 (Fig 8). This feedback repression of SIZ1 activity by COP1 may reflect the requirement of tightly regulated COP1 activity for proper photomorphogenic development.

Plant Growth Conditions and Phenotypic Analysis
After surface sterilization, seeds were sown on Murashige and Skoog growth medium (1/4× Murashige and Skoog basal salts, 1% sucrose, and 0.75% agar). After 3 days of incubation at 4°C in darkness, the seeds were exposed to 6 h of white light to induce germination, and then transferred to light chambers containing red, blue, or far-red light emitting diodes at the indicated fluence rates. White light (100 μmol m -2 s -1 unless indicated otherwise) was provided using white fluorescent lamps. To analyze hypocotyl length, cotyledon angle, and hypocotyl cell length and cell numbers, seedlings were photographed with a camera (Canon) or a cool CCD camera coupled to an Olympus BX53 microscope and the images were analyzed with NIH ImageJ software (http://rsbweb.nih.gov/ij/).

RNA Extraction and Quantitative Real-Time Reverse Transcription-PCR
Total RNA was extracted from seedlings with TRIZOL reagent (RNAiso Plus, TaKaRa) and reverse transcribed using a RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific). qRT-PCR was then performed with SYBR Premix ExTaq (TaKaRa) according to the manufacturer's instructions. Three biological replicates were performed. The relative expression level of each gene was normalized to that of Ubiquitin-Conjugation Enzyme (UBC). Primer sequences are listed in S1 Table. BiFC Assay COP1-YFP N and SIZ1-YFP C plasmids together with the proper control plasmids (empty YFP N or YFP C vector plasmids) were transformed into Agrobacterium tumefaciens strain GV3101 and infiltrated into Nicotiana benthamiana leaves as previously described [66]. After incubation at 22°C for 3 days under 16 h white light (100 μmol m -2 s -1 )/8 h darkness, YFP fluorescence was detected with a fluorescence microscope (Olympus BX53) under darkness or after 1 h white light (100 μmol m -2 s -1 ) exposure.

Immunoprecipitation Assays
For co-immunoprecipitation of COP1 and SIZ1, Myc-COP1 and SIZ1-GFP were separately expressed in Col-0 protoplasts to avoid degradation of SIZ1 by COP1. Immunoprecipitation was carried out using a mixture of COP1 and SIZ1 protein extracts. The mixture of protein extracts was immunoprecipitated with anti-Myc-conjugated agarose beads (Sigma, F-2426) and co-immunoprecipitated SIZ1-GFP proteins were detected with anti-GFP antibody (Clontech, 632375).
To analyze the sumoylation effect in the interaction between HY5 and COP1, 1 μg of GST-HY5 protein purified from E. coli [52], as bait, was incubated with the protein extracts isolated from N. benthamiana co-expressing Myc-COP1 with FLAG-SUMO1 or FLAG-SU-MO1 AA . The mixture of protein extracts was immunoprecipitated with anti-Myc-conjugated agarose beads (Sigma, F-2426) and co-immunoprecipitated GST-HY5 proteins were detected with anti-GST antibody (Abcam, ab19256).

Nuclear Protein Extraction
Nuclear protein extraction was carried out with the CelLytic PN Extraction Kit (Sigma, CEL-LYTPN1) as described previous [62].
To generate pSPYCE-SIZ1-YFP c , full-length SIZ1 cDNA without the termination codon was amplified with gene-specific primers SIZ1-F-XmaI/SIZ1-R-SpeI and ligated into the pBluescript vector. pBluescript-SIZ1 was digested with SmaI and SpeI, and the SIZ1 cDNA was inserted in-frame at the HpaI/SpeI sites of the pSPYCE-35S vector [66].
To generate pCambia1302-SIZ1-GFP, pBluescript-SIZ1 was digested with SmaI and SpeI. The pCambia1302 vector was digested with NcoI, to generate blunt ends, and then with SpeI. The insert was then ligated into the pCambia1302 vector.
To generate pMAL-C2-MBP-SIZ1-Myc, full-length cDNA of SIZ1 without the termination codon was amplified with gene-specific primers SIZ1-F-XbaI/SIZ1-R-ClaI, and ligated into the pBluescript vector. pBluescript-SIZ1 was digested with XbaI and ClaI, and inserted in-frame at the XbaI and ClaI sites of the pMAL-C2-MBP-MYC vector.
To generate p326-Myc-COP1, the full-length cDNA of COP1 was amplified with gene-specific primers COP1-F-HindIII/COP1-R-XhoI, and ligated into the pBluescript vector. pBluescript-COP1 was digested with HindIII and XhoI, and inserted in-frame at the HindIII and XhoI sites of the p326-35S-nMyc vector.