Cysteines have a role in conformation of the UVR8 photoreceptor

SUMMARY The UV RESISTANCE LOCUS 8 (UVR8) photoreceptor mediates plant responses to Ultraviolet‐B (UV‐B) wavelengths. The UVR8 dimer dissociates into monomers following UV‐B photoreception, a process accompanied by conformational changes that facilitate interaction of UVR8 with proteins that initiate responses. However, the importance of particular amino acids in maintaining UVR8 conformation and modulating protein interactions is poorly understood. Here we examine the roles of cysteine amino acids C231 and C335 in UVR8 structure and function. UVR8C231S,C335S mutant protein forms dimers and monomerizes similarly to wild‐type UVR8. UVR8C231S,C335S interacts with CONSTITUTIVELY PHOTOMORPHOGENIC 1 (COP1) in plants to initiate photomorphogenic responses to UV‐B, although the interaction is weaker when examined in yeast two‐hybrid assays. Similarly, the interaction of UVR8C231S,C335S with REPRESSOR OF UV‐B PHOTOMORPHOGENESIS (RUP) proteins is weaker in both plants and yeast compared with wild‐type UVR8. Re‐dimerization of UVR8 in plants, which is mediated by RUP proteins, occurs with reduced efficiency in UVR8C231S,C335S. Fluorescence resonance energy transfer analysis indicates that UVR8C231S,C335S has an altered conformation in plants, in that the N‐ and C‐termini appear closer together, which may explain the altered protein interactions.


INTRODUCTION
Ultraviolet-B (UV-B) wavelengths in sunlight stimulate a range of biochemical, physiological, and morphogenic responses in plants, many of which are mediated by the photoreceptor UV RESISTANCE LOCUS 8 (UVR8) (Jenkins, 2017;Podolec et al., 2021). These responses result from extensive changes in gene expression that are initiated by UVR8 photoreception and associated signalling events (Brown et al., 2005;Favory et al., 2009).
The UVR8 protein consists of a seven-bladed b-propeller core with flexible N-and C-terminal tails (Christie et al., 2012;Wu et al., 2012). In the absence of UV-B wavelengths, the protein forms a homodimer that is held together by salt bridges between charged amino acids on the dimer interaction surface of adjacent monomers (Christie et al., 2012;Rizzini et al., 2011;Wu et al., 2012). The photoreceptor does not have a bound cofactor as its chromophore, and photoreception involves the absorption of UV-B and short wavelength UV-A light by tryptophan (Trp) amino acids in the primary sequence. UVR8 has 14 Trps, seven of which are in the dimer interaction surface of each monomer (Christie et al., 2012;O'Hara & Jenkins, 2012;Wu et al., 2012). Of these seven, W233, W285, and W337 form an excitonically coupled cluster, together with W94 on the adjacent monomer. Excitation energy is transferred from multiple Trps to W233 and W285 in the dimer interaction surface . Excitation of these Trps, which are adjacent to key charged amino acids, results in neutralization of salt bridges between monomers and hence dissociation of the dimer (Mathes et al., 2015).
Site-specific mutagenesis combined either with examination of mutant UVR8 proteins in vitro or with functional studies in vivo has determined the roles of specific amino acids in UVR8 structure and function, notably charged amino acids that maintain the dimer, specific Trps involved in photoreception, and amino acids that are critical for interactions with other proteins (Jenkins, 2014;Podolec et al., 2021). Nevertheless, there is much to be learnt about the structure-function relationships of UVR8. The focus of the present study is on the cysteine amino acids of UVR8. Cysteines are important in protein secondary structure, through formation of disulphide bonds, and in a variety of post-translational modifications. There is no evidence from crystal structures that disulphide bonds are present in dimeric UVR8 (Christie et al., 2012;Wu et al., 2012), and treatment of the protein with strong oxidizing or reducing agents does not affect monomer formation (Wu et al., 2012). However, it has been suggested that oxidation of a reduced amino acid, potentially a cysteine, could enhance structural flexibility of the protein and facilitate re-dimerization (Miyamori et al., 2015). Hence, in this study we used mutagenesis to investigate the potential importance of specific cysteines in UVR8 structure and function.

Location of UVR8 cysteines
UVR8 has seven cysteines ( Figure 1a,b), but their functional significance is unknown. C37 is in blade 1 of the bpropeller; C74 is in blade 2; C127 and C132 are on adjacent b-strands of blade 3; C231 is in blade 5; C335 is in blade 7; and C317 is in a loop between blades 6 and 7 and is surface exposed. C74, C127, C231, and C335 are in conserved locations at the end of the first b-strand of blades 2, 3, 5, and 7 respectively, and are adjacent to the water-filled depression in the centre of the b-propeller that extends below the dimer interface ( Figure 1c). It is intriguing that C231 and C335 are located just below the interface Trps W233 and W337. Moreover, the positions of C231 and C335 in a CGWRHT motif are conserved in UVR8 sequences from diverse angiosperms, the lycopodiophyte Selaginella, bryophytes, and green algae ( Figure S1). We therefore focused this study on C231 and C335.
UVR8 C231S,C335S presents normal dimer/monomer formation and UV-B photoreception To examine the potential importance of C231 and C335 in UVR8 structure and function, site-specific mutagenesis was used to convert both cysteines to serine, which is very similar in structure to cysteine. To test whether the C231S and C335S mutations affect the stability of the UVR8 homodimer and UV-B-induced monomerization, UVR8 C231S,C335S (referred to in figures as UVR8 SS ) was expressed in Escherichia coli and purified. The protein was examined using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) with unboiled samples, which has been used extensively to test dimer/monomer status. Similar to wild-type UVR8, the UVR8 C231S,C335S protein presented as a dimer before UV-B exposure and responded to UV-B exposure by converting to the monomer (Figure 2a).
To examine dimer/monomer status and function in plants, we generated transgenic lines expressing UVR8 C231S,C335S as a green fluorescent protein (GFP) fusion in the uvr8-1 null mutant background. The expression level of GFP-UVR8 C231S,C335S in plants was similar to that of the previously generated GFP-UVR8 line ( Figure 2b). Consistent with the purified protein, GFP-UVR8 C231S,C335S extracted from seedlings was predominantly present as a homodimer in the absence of UV-B and as a monomer after UV-B exposure ( Figure 2b). These results clearly demonstrate that the C231S and C335S mutations of UVR8 do not impair dimer formation and UV-B-induced monomerization.

GFP-UVR8 C231S,C335S is functional in initiating UV-Binduced photomorphogenic responses in plants
Interaction of UVR8 monomers with COP1 is required for the photoreceptor to initiate responses. We therefore tested whether the cysteine mutations affect this interaction and UVR8-mediated responses. Yeast two-hybrid (Y2H) assays indicated that UVR8 C231S,C335S interacts much less strongly with COP1 than wild-type UVR8 (Figure 3a), whereas the single UVR8 C231S and UVR8 C335S mutants appeared similar to wild-type ( Figure S2a). As a control, we examined interaction in yeast between COP1 and UVR8 mutated in the other two cysteines adjacent to the waterfilled depression, C74 and C127; interaction of COP1 with UVR8 C74S,C127S was similar to wild-type. With respect to plants, co-immunoprecipitation (Co-IP) assays in Arabidopsis transgenic lines ( Figure 3b) and bimolecular fluorescence complementation (BiFC) assays in Nicotiana benthamiana leaves (Figures 3c and S3)  transgenic lines. First, quantitative reverse transcription-polymerase chain reaction (PCR) analysis showed that the transcript levels of HY5 and RUP2, which are well known genes induced by UV-B, were upregulated after UV-B exposure in both GFP-UVR8 and GFP-UVR8 C231S,C335S lines (Figure 4a,b). Secondly, a hypocotyl elongation assay revealed that the GFP-UVR8 C231S,C335S lines displayed a comparable photomorphogenic growth suppression response to the GFP-UVR8 plants in the presence of UV-B, and a normal long hypocotyl phenotype in the absence of UV-B (Figure 4c,d). Taken together, these results demonstrate that GFP-UVR8 C231S,C335S presents UV-B-induced photomorphogenesis and rescues the uvr8-1 phenotype, similar to wild-type GFP-UVR8.
Binding affinity of UVR8 C231S,C335S for RUP2 and negative regulation of UVR8 signalling are partly impaired UVR8 C231S,C335S was shown to interact with RUP2 in plants using the BiFC (Figures 5a and S3) and Co-IP assays (Figure 5b). However, it should be noted that the interaction between GFP-UVR8 C231S,C335S and RUP2 was only detected by Co-IP under reduced stringency conditions with low salt buffer, whereas no signal was observed if the Co-IP assay (a) UVR8 monomer looking towards the dimer interaction surface. Seven cysteines are shown in green and numbered. Trps W233, W285, and W337 are shown in magenta. Image made using Pymol. (b) Position of cysteines (in red) in the b-propeller blades of UVR8. Sequences of the blades are aligned using invariant amino acids (arrowed above sequences). The first amino acid of each blade is numbered (in brackets). The GWRHT motif containing W233, W285, and W337 is underlined. Three regions of b-strand, A, B, and C, are indicted below the sequences by purple arrows; the black lines represent loops associated with each blade. (c) Section through the UVR8 monomer viewed from the side showing the water-filled depression below the dimer interaction surface. Trps W233, W285 and W337 are shown in magenta. C231 is shown in green, adjacent to the depression and below W233. Image made using Pymol. was performed with high salt buffer, indicating that GFP-UVR8 C231S,C335S and RUP2 have a reduced interaction in plants ( Figure 5b). Similarly, the Y2H assay revealed that the interaction of UVR8 C231S,C335S with RUP1 and RUP2 was impaired in yeast ( Figure 5c), whereas interaction of the RUPs with UVR8 C74S,C127S was similar to wild-type. A reduced interaction was also seen with the single UVR8 C231S and UVR8 C335S mutants ( Figure S2b).
The RUP interaction promotes the re-dimerization of UVR8 monomers and negatively regulates UVR8 signalling (Gruber et al., 2010;. Therefore, we examined whether the weakened RUP2 binding had any effect on UVR8 negative regulation. Compared with GFP-UVR8, the re-dimerization of GFP-UVR8 C231S,C335S monomers following transfer of UV-B-exposed plants to minus-UV-B conditions was significantly inhibited in vivo (Figure 6a). In addition, GFP-UVR8 C231S,C335S seedlings accumulated more anthocyanin under UV-B than GFP-UVR8 seedlings, whereas there is no significant difference in the absence of UV-B; however, the magnitude of increase in anthocyanin was less than in the rup1rup2 mutant (Figure 6b,c). Taken together, these results indicate that the C231S and C335S mutations reduce the strength of interaction between UVR8 C231S,C335S and RUP proteins, resulting in impaired UVR8 re-dimerization and weakened negative regulation of UVR8 signalling.

UVR8 C231S,C335S has an altered conformation
Previous research demonstrated that RUPs bind specifically to the C27 domain of UVR8, in contrast to COP1, which interacts with UVR8 via both the C27 domain and UVR8 core domain (Cloix et al., 2012;Yin et al., 2015). To explain why mutations of C231 and C335, which are located in the core domain of UVR8, impair interaction with RUPs, we hypothesized that these mutations might cause a conformational change of UVR8, affecting the accessibility of the C-terminus to binding by RUPs. To investigate this notion, we analysed the conformation of UVR8 C231S,C335S using a fluorescence resonance energy transfer (FRET)-based method we reported previously (Liao et al., 2019). As shown in Figure 7a, the pFRET-NcCg-DEST vector contains FRET-pair tags mCherry and GFP, which are fused to the N-and C-termini of the target protein (UVR8), respectively. The excited GFP can excite mCherry through FRET if they are sufficiently close, and the efficiency of FRET is proportional to the relative distance between GFP and mCherry. Therefore, both monomerization and conformational change of UVR8 may cause a change in the FRET efficiency ( Figure 7b).
Wild-type UVR8 and UVR8 C231S,C335S were inserted into the pFRET-NcCg-DEST vector. The constructs expressing the constitutively monomeric mutant UVR8 D96N,D107N (UVR8 DD ) and constitutively dimeric mutant UVR8 W285F were created as controls. Regardless of the presence or absence of UV-B, the FRET efficiency of UVR8 W285F is consistently relatively high, while that of UVR8 D96N,D107N is consistently relatively low (Figure 7c,d), similar to the previous findings (Liao et al., 2019), indicating that much of the FRET signal is generated between adjacent monomers (Figure 7b). The FRET efficiency of UVR8 and UVR8 C231S,C335S dramatically reduced following UV-B exposure (Figure 7c,d), which is explained by the monomerization of UVR8 dimers eliminating FRET between adjacent monomers (Figure 7b; Liao et al., 2019). Excluding the structural changes caused by UVR8 monomerization, the FRET efficiencies of both UVR8 C231S, C335S dimer and monomer were found to be much higher than those of UVR8 dimer and monomer, respectively, (a) Purified UVR8 and UVR8 C231S,C335S (UVR8 SS ) proteins were resolved by sodium dodecyl sulphatepolyacrylamide gel electrophoresis with non-boiled samples and stained with Coomassie brilliant blue after being exposed (+) or not (À) to 3 lmol m À2 sec À1 broadband UV-B for 10 min. (b) uvr8-1/35Spro:GFP-UVR8 (GFP-UVR8), uvr8-1/ 35Spro:GFP-UVR8 C231S,C335S (GFP-UVR8 SS ; two independent lines), and uvr8-1 plants grown in 120 lmol m À2 sec À1 white light for 14 days were exposed (+) or not (À) to 3 lmol m À2 sec À1 broadband UV-B for 1 h. Protein extracts were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis with non-boiled samples and the immunoblot was probed with an anti-UVR8-C antibody. Ponceau S-stained Rubisco large subunit (rbcL) bands are shown as a loading control.
Ó 2022 The Authors. suggesting that the N-and C-termini are closer together in both UVR8 C231S,C335S dimer and monomer, indicating an altered conformation (Figure 7c,d).

DISCUSSION
The structure of UVR8 is known to be critical to its function, with the formation of monomers essential to initiate signalling (Rizzini et al., 2011). Moreover, several studies report that UVR8 undergoes major conformational changes following photoreception (Camacho et al., 2019;Heilmann et al., 2014;Miyamori et al., 2015;Zeng et al., 2015). Both dynamic X-ray crystallography (Zeng et al., 2015) and native mass spectrometry (Camacho et al., 2019) with purified UVR8 provide evidence of partial unravelling of the b-propeller core during monomer formation. The formation of an extended monomer structure (Camacho et al., 2019) is believed to facilitate interactions with other proteins by modifying accessibility of the C-terminal domain. However, the processes involved in generating conformational changes in UVR8 are unknown. Cysteines are important in determining protein structure, but the available crystal structures provide no evidence of disulphide bond formation in the UVR8 dimer. However, it is unknown whether such bonds could form, and potentially break, during the proposed extensive remodelling of UVR8 structure following photoreception. It has been suggested that oxidation of a cysteine might promote re-dimerization (Miyamori et al., 2015), but direct evidence is lacking. The data presented here are valuable, as they suggest that C231 and C335 influence the conformation of UVR8 and affect interactions with other proteins. The FRET assay was developed previously to examine the effect of UV-B exposure on UVR8 conformation in plants (Liao et al., 2019). The design of the pFRET-NcCg-DEST vector constructs enables both inter-molecular and intra-molecular changes to be observed by measuring the efficiency of FRET. The conversion of dimer to monomer in wild-type UVR8 was found to cause a large reduction in the FRET efficiency, consistent with the data obtained for control mutant proteins that are constitutively either dimeric (O'Hara & Jenkins, 2012;Rizzini et al., 2011) or monomeric (Heilmann et al., 2016). The UVR8 C231S,C335S mutant protein has a higher FRET efficiency than wild-type UVR8 in both the dimer and the monomer forms, indicating that in both cases the N-and C-termini are closer together than in the wild-type. Hence, the FRET experiments provide evidence that C231 and C335 have a role in maintaining the conformation of UVR8, although the underlying molecular basis is not clear.  (a) HY5 and (b) RUP2 transcript levels were measured by reverse transcription-quantitative polymerase chain reaction. uvr8-1/35Spro:GFP-UVR8 (GFP-UVR8), uvr8-1/35Spro:GFP-UVR8 C231S,C335S (GFP-UVR8 SS ; two independent lines) and uvr8-1 plants were grown under 120 lmol m À2 sec À1 white light for 14 days and exposed (+UV-B) or not (ÀUV-B) to 3 lmol m À2 sec À1 broadband UV-B for 3 h before harvest. Data are mean AE SE, n = 3 biological replicates; letters indicate statistically significant mean values (P < 0.01; two-way ANOVA with Tukey's test). The UVR8 C231S,C335S mutant protein is able to form a dimer and carry out photoreception to generate monomers. However, there is evidence of altered interaction with both COP1 and RUP proteins. The UV-B-dependent interaction with COP1 appears normal in plants, but is weaker in yeast, based on the Y2H assays. Under the strong selection used in these assays the interaction of COP1 and UVR8 C231S,C335S is clearly impaired compared with wild-type UVR8 and the control UVR8 C74S,C127S fusion. Similarly, interaction of both RUP1 and RUP2 with UVR8 C231S,C335S is weaker in the Y2H assay. The RUP2 interaction is also weaker in plants, as shown when stringent conditions are used in the Co-IP assay. RUP1 is expressed at very low levels in plants (Liao et al., 2020), so the extent of its interaction with UVR8 could not be assayed in vivo. Together the yeast and plant data indicate that UVR8 C231S,C335S has a reduced affinity for binding COP1 and RUPs. One possible explanation is that the altered conformation of the mutant influences the accessibility of the C27 region to the interacting proteins. The (a) BiFC assays of RUP2 interactions with wild-type UVR8 and UVR8 C231S,C335S . Plasmids expressing N-terminal-YFP fused to either UVR8 or UVR8 C231S,C335S (UVR8 SS ) together with C-terminal-YFP fused to RUP2 were infiltrated into Nicotiana benthamiana leaves. Plants were exposed (+UV-B) or not (ÀUV-B) to 3 lmol m À2 sec À1 broadband UV-B for 1 h before taking images. Images from left to right are YFP fluorescence signal, bright field, and merged image. Bar = 20 lm. Negative control transformations are shown in Figure S3. (b) Co-immunoprecipitation assay of RUP2 interactions with wild-type GFP-UVR8 and GFP-UVR8 C231S,C335S . Protein extracts from 14-day-old plants expressing uvr8-1/35Spro:GFP-UVR8 (GFP-UVR8) or uvr8-1/35Spro:GFP-UVR8 C231S,C335S (GFP-UVR8 SS ) were treated with 3 lmol m À2 sec À1 broadband UV-B (+) or left in white light (À) for 15 min. Co-immunoprecipitation assays of extracts were performed with anti-GFP antibody under the same light conditions. The wash buffer contained either 450 mM NaCl (high salt) or 150 mM NaCl (low salt). Input samples (15 lg, input) and eluates (IP) were resolved on sodium dodecyl sulphatepolyacrylamide gel electrophoresis with non-boiled samples and immunoblots were probed with anti-GFP and anti-RUP2 antibodies. greater impact on RUP binding, at least in plants, may be because RUP proteins interact exclusively with C27 whereas COP1 also binds to the b-propeller core. The GFP-UVR8 C231S,C335S mutant is able to initiate photomorphogenic responses in plants, similarly to wild-type. This is not surprising as these responses are dependent on the interaction of UVR8 with COP1, which appears largely unaffected in plants. The UVR8-COP1 interaction following UV-B exposure leads to the accumulation of HY5 protein and the stimulation of HY5 transcript accumulation, the latter of which was evident in GFP-UVR8 C231S,C335S . The interaction of UVR8 with RUP proteins displaces COP1 and promotes re-dimerization of monomers, hence negatively regulating UVR8 activity . The observed weaker interaction of GFP-UVR8 C231S,C335S with RUPs most likely explains its reduced efficiency of re-dimerization. Arabidopsis rup1,rup2 mutant plants are impaired in redimerization (Findlay & Jenkins, 2016; and display a phenotype of enhanced photomorphogenic responses (Gruber et al., 2010). The GFP-UVR8 C231S,C335S mutant exhibits an increase in anthocyanin formation, although not to the extent seen in the rup1,rup2 mutant. No evidence of a hyper-response was observed for the transcript accumulation and hypocotyl growth suppression responses examined, and the reason for this difference is not clear. UVR8 is known to interact with multiple proteins involved in various UV-B responses, so it is possible that the altered conformation of UVR8 C231S,C335S also affects the binding of a transcription factor specifically involved in anthocyanin formation. were grown under 120 lmol m À2 sec À1 white light for 14 days and were irradiated (+UV-B) with supplementary 3 lmol m À2 sec À1 broadband UV-B for 30 min, then returned to white light (WL) for the indicated times. Protein extracts were resolved by sodium dodecyl sulphate-polyacrylamide gel electrophoresis with non-boiled samples for the immunoblot analysis. The UVR8 protein was probed with the anti-UVR8-C antibody. Protein extracts from uvr8-1 plants were used as a negative control. Ponceau S stained Rubisco large subunit (rbcL) bands are shown as a loading control. (b) Anthocyanin extracts of equal numbers of seedlings grown under 1.5 lmol m À2 sec À1 white light with (+UV-B) or without (ÀUV-B) supplementary 1.5 lmol m À2 sec À1 narrowband UV-B for 5 days.
Ó 2022 The Authors. In conclusion, much remains to be understood about the structural basis of UVR8 function. In particular, little is known about the roles of specific amino acids in controlling the dynamic changes in UVR8 conformation that follow photoreception, which are crucial in determining interactions with proteins involved in effecting downstream responses and coordinating the dimer/monomer photo-equilibrium. The data presented here indicate that specific cysteines modulate UVR8 conformation and consequently affect protein interactions, and hence further investigation of their roles is merited, for example by using native mass spectrometry (Camacho et al., 2019). As UVR8 mediates many important responses in plants, including those that affect yield and nutritional quality, understanding the molecular basis of its activity has potential applications in crop improvement.
Plants were grown on agar plates containing half-strength Murashige and Skoog salts in a growth cabinet at 20°C. Plants were grown for 2 weeks in a 16-h white light/8-h dark photoperiod, with 120 lmol m À2 sec À1 white light provided by warm-white LEDs. For treatment with UV-B, plants were exposed to 3 lmol m À2 sec À1 broadband UV-B from UV-B-313 fluorescent tubes (Q-Lab Co., Cleveland, OH, USA) covered by cellulose acetate film to filter out UV-C. The spectrum of the UV-B source was reported by Cloix et al. (2012). Fluence rates were measured with a Skye Spectrosense 1 meter (Skye Instruments, Llandrindod Wells, Wales) with a SKU 430 sensor (Skye Instruments).

Photomorphogenic responses
Quantitative reverse transcription-PCR assays of HY5 and RUP2 transcripts in RNA samples isolated from leaf tissue were undertaken exactly as described by D ıaz-Ramos et al. (2018), with transcript levels quantified relative to ACTIN2 control transcripts. The primers are listed in Table S1.
The hypocotyl elongation assay was performed as described by Cloix et al. (2012). Plants were grown for 4 days in 1.5 lmol m À2 sec À1 white light with or without 1.5 lmol m À2 sec À1 narrowband UV-B (provided by Philips TL20W/01RS tubes; wavelength maximum 312 nm; spectrum shown in Cloix et al., 2012). At least 25 seedlings from each line were measured.

Protein analysis
Wild-type UVR8 and UVR8 C231S,C335S were expressed in E. coli and purified as detailed by Christie et al. (2012). Purified protein was exposed to 3 lmol m À2 sec À1 broadband UV-B for 10 min. Nonboiled protein samples were separated by a 10% SDS-PAGE gel, which permits the resolution of UVR8 dimer and monomer (Rizzini et al., 2011). Gels were stained with Coomassie Blue.
To induce dimer monomerization, 14-day-old plants were exposed to 3 lmol m À2 sec À1 broadband UV-B for 30 min. To monitor re-dimerization, the UV-B-treated plants were transferred to darkness at room temperature for the indicated times. Total plant proteins were extracted and protein samples were prepared for electrophoresis without boiling. Protein samples were resolved by 10% SDS-PAGE and UVR8 protein was detected on Western blots by anti-UVR8-C antibody.

Anthocyanin
The method for anthocyanin extraction and quantification was described previously . Seedlings were harvested and weighed quickly, then placed into microcentrifuge tubes containing 350 ll of 18% 1-propanol, 1% HCl, and 81% sterile dH 2 O. The tubes were boiled at 95°C for 3 min and then incubated in darkness for at least 2 h at room temperature. After a brief centrifugation in a bench top microcentrifuge for 30 sec, 250 ll of the extract was transferred into a new tube containing 750 ll of extraction solution. The amount of anthocyanins in the final extract was measured by spectrophotometry at 535 and 630 nm. The values are reported as (A535-2*A630)/g fresh weight.

Co-IP analyses
For Co-IP analyses, plants were grown on agar plates and exposed to UV-B as described above. Whole cell protein extracts were prepared as described by Kaiserli and Jenkins (2007). The Co-IP assays were carried out using anti-GFP microbeads (lMacs, 130-091-125; Miltenyi Biotec, Bergisch Gladbach, Germany) to immunoprecipitate GFP-UVR8, and the presence of COP1 and RUP2 in the immunoprecipitates was examined as described previously (Cloix et al., 2012;Liao et al., 2020). The 'input' samples applied to the microbead columns and the immunoprecipitate eluates were analysed by SDS-PAGE followed by Western blotting and immunodetection using the anti-GFP, anti-COP1, and anti-RUP2 antibodies mentioned above.

BiFC
The UVR8, UVR8 C231S,C335S , COP1, and RUP2 fusions were cloned into the pBiFCt-2in1-NC vector containing the N-and C-terminal regions of YFP by Gateway TM cloning technology (Thermo Fisher Scientific, Waltham, MA, USA) (Grefen & Blatt, 2012). The Agrobacterium suspension containing the required constructs was infiltrated into N. benthamiana leaves as described by Liao et al. (2019). After 2-3 days, plants were exposed to 3 lmol m À2 sec À1 broadband UV-B for 1 h before taking images. Confocal images were collected by a Zeiss LSM510-META confocal microscope with a 209/0.75-NA objective lens. YFP fluorescence was excited by 514 nm light and collected over 520-565 nm.

Y2H assay
For the Y2H assay, UVR8, UVR8 C231S , UVR8 C335S , UVR8 C231S,C335S , and UVR8 C74S,C127S were cloned into the pGBKT7 vector containing the Gal4 DNA-binding domain. COP1, RUP1, and RUP2 were cloned into the pGADT7 vector containing the Gal4 activation domain. Control fusions contained either mammalian P53 or Tantigen (positive control), or no inserts (negative control). Plasmids containing an activation domain and a binding domain fused to the indicated proteins were co-transformed into the yeast strain AH109. The yeast growth assay was applied on the non-selective solid medium (SD/ÀTrp/ÀLeu) or selective solid medium (SD/À Trp/ÀLeu/-His) at 30°C for 3 days in darkness or under 0.1 lmol m À2 sec À1 narrowband UV-B light.

FRET experiments
The FRET experiments were undertaken as reported by Liao et al. (2019). The UVR8, UVR8 D96N,D107N , and UVR8 W285F constructs in the pFRET-NcCg-Dest vector were described previously (Liao et al., 2019). The UVR8 C231S,C335S construct was produced by the same method, introducing a PCR product of UVR8 C231S,C335S into the vector by Gateway TM cloning technology (Thermo Fisher Scientific) (Liao et al., 2019). The construct was checked by restriction enzyme digestion and sequencing and transformed in Agrobacterium GV3101 for subsequent transient expression in tobacco.
Single colonies of Agrobacterium containing the required constructs were used to produce cultures for transient expression in Nicotiana tabacum, Four-to six-week-old plants were used for leaf infiltration as described previously (Liao et al., 2019). Plants were moved back to the growth chamber and the expression was observed after 2-3 days using a Leica TCP SP8 FRET-FLIM confocal microscope (Liao et al., 2019). Where indicated, leaves were exposed to broadband UV-B as described above. GFP fluorescence was excited by the 8% 488 nm laser and collected over 500-535 nm with 30% gain value. To detect the FRET signal, mCherry fluorescence was collected over 590-645 nm with 100% gain value. mCherry also was separately excited by the 8% 552 nm laser as a mCherry control. The FRET efficiency was calculated as described in the figure legends.

Replication and statistical analysis
Experiments were repeated three times with independent biological replicates. Statistical analysis of the data was performed on GraphPad prism software using the method indicated in the figure legends. Figure S1. Amino acid sequence alignment of the GWRHT motifs in UVR8 proteins from different plant species. Figure S2. Yeast two-hybrid assays of interactions involving single C231S and C335S mutants. Figure S3. Negative controls for the BiFC assays shown in Figures  3c and 5a. Table S1. Primers used in this study.