Two Distinct Domains of the UVR8 Photoreceptor Interact with COP1 to Initiate UV-B Signaling in Arabidopsis

UV-B photon reception by the Arabidopsis thaliana homodimeric UV RESISTANCE LOCUS8 (UVR8) photoreceptor leads to its monomerization and a crucial interaction with CONSTITUTIVELY PHOTOMORPHOGENIC1 (COP1). Relay of the subsequent signal regulates UV-B-induced photomorphogenesis and stress acclimation. Here, we report that two separate domains of UVR8 interact with COP1: the b -propeller domain of UVR8 mediates UV-B-dependent interaction with the WD40 repeats-based predicted b -propeller domain of COP1, whereas COP1 activity is regulated by interaction through the UVR8 C-terminal C27 domain. We show not only that the C27 domain is required for UVR8 activity but also that chemically induced expression of the C27 domain is suf ﬁ cient to mimic UV-B signaling. We further show, in contrast with COP1, that the WD40 repeat proteins REPRESSOR OF UV-B PHOTOMORPHOGENESIS1 (RUP1) and RUP2 interact only with the UVR8 C27 domain. This coincides with their facilitation of UVR8 reversion to the ground state by redimerization and their potential to interact with UVR8 in a UV-B-independent manner. Collectively, our results provide insight into a key mechanism of photoreceptor-mediated signaling and its negative feedback regulation.

To better understand UVR8-mediated early UV-B signaling, we focused on the b-propeller and the C-terminal regions of UVR8, including the C27 domain, in yeast and plants. We show here that the b-propeller domain of UVR8 interacts with COP1 in a UV-B-dependent manner in the absence of the C-terminal 44 amino acids and, thus, the C27 domain. However, the bpropeller domain alone is not sufficient to activate early UV-B signaling. We further demonstrate that the C-terminal 44 amino acids alone interact constitutively with COP1 and that this depends on a Val-Pro (VP) pair in the C27 domain. Chemically induced expression of the C-terminal 44 amino acids is sufficient to mimic early UVR8 responses. Thus, UVR8 interaction with COP1 is 2-fold: via the b-propeller domain and via the C27 domain, with the latter functionally impinging on COP1. This differs from the mechanism of UVR8-RUP1/RUP2 interaction, coinciding with the differences in activity and dependence on UV-B for interaction with UVR8.

RESULTS
Interaction of the UVR8 b-Propeller Domain with COP1 Is UV-B Dependent UVR8 is a protein of 440 amino acids (Kliebenstein et al., 2002) containing a seven-bladed b-propeller domain. Previous results suggest that 27 amino acids within the C-terminal 44 amino acids (the C27 domain, amino acids 397 to 423; here given as UVR8 397-423 ) are required for the UV-B-dependent interaction of UVR8 with COP1 (Cloix et al., 2012). Moreover, yeast two-hybrid assays have shown that the C27 domain alone is sufficient for the interaction with COP1, which is UV-B-independent in the absence of the b-propeller core domain (Cloix et al., 2012). To further dissect the function of the UVR8 b-propeller domain and the C-terminal 44 amino acids in UV-B signaling, we divided UVR8 into the two fragments for testing: UVR8 N396 (amino acids 1 to 396) and UVR8 C44 (amino acids 397 to 440) ( Figure 1A). Given that UVR8-COP1 interaction is critical for UV-B signaling, we tested the interaction of UVR8 N396 and UVR8 C44 with COP1 in a yeast two-hybrid assay. As described before, UVR8 interacts with COP1 in yeast in a UV-B-dependent manner ( Figure 1B) (Rizzini et al., 2011;Cloix et al., 2012). UVR8 C44 interacted constitutively with COP1 ( Figure 1B), similar to UVR8 397-423 (C27 domain) (Cloix et al., 2012).
However, in contrast with the previous finding that the UVR8 C27 domain is required for COP1 interaction (Cloix et al., 2012), we found clear evidence that UVR8 N396 interacts with COP1 in a UV-B-dependent manner, despite the absence of the C27 domain ( Figure 1B). We next tested a truncated UVR8 based on the uvr8-2 mutant allele, which has a premature stop codon at Trp-400 (Brown et al., 2005). Indeed, UVR8-2 (UVR8 N400 ) also interacted with COP1 in a UV-B-dependent manner ( Figure 1C). Unlike UVR8 N396 and UVR8 N400 , the UVR8 DC27 version described previously by Cloix et al. (2012) included the C-terminal 17 amino acids, which may explain the difference between our data and theirs. We reproduced the UVR8 DC27 version and tested it in yeast. As with UVR8 N396 and UVR8 N400 , UVR8 DC27 also showed UV-B-dependent interaction with COP1 in yeast (Supplemental Figure 1). Independent of this result, we showed previously that UVR8 interacts with the WD40-repeat domain of COP1 (COP1 C340 ) (Favory et al., 2009;Rizzini et al., 2011). In agreement with this, UVR8-2 (UVR8 N400 ) interacted with COP1 C340 and thus the WD40 domain of COP1 ( Figure 1C).
The yeast two-hybrid data indicate that the truncated UVR8 protein in uvr8-2 mutant plants still interacts with COP1. As our previously generated UVR8 antibodies were specific for epitopes in the C terminus, we developed an antibody against the N-terminal 15 amino acids of UVR8. Using this antibody, we detected the truncated UVR8-2 protein in extracts from uvr8-2 mutant seedlings ( Figure 1D) as expected (Cloix et al., 2012). By contrast, UVR8 protein was not detected in uvr8-1, which bears an insertion of five amino acids that is believed to make the protein unstable ( Figure 1D) (Kliebenstein et al., 2002;Cloix et al., 2012). In coimmunoprecipitation assays with uvr8-1 as a negative control, COP1 was coimmunoprecipitated with the truncated UVR8 protein from samples of uvr8-2 mutant seedlings ( Figure  1D). Similar to wild-type UVR8, UVR8-2 coimmunoprecipitated endogenous COP1 only from UV-B-treated uvr8-2 seedlings, albeit to a lower extent ( Figure 1D). By contrast, COP1 was not detected in coimmunoprecipitates from the uvr8-1 negative control ( Figure 1D). However, despite the UV-B-dependent interaction of UVR8-2 with COP1, the uvr8-2 mutant displayed no hypocotyl growth inhibition ( Figure 1E) or any other UV-B response tested (Brown et al., 2005;Cloix et al., 2012).
We conclude that the seven-bladed b-propeller domain of UVR8 interacts in a UV-B-dependent manner with COP1 and that the interaction of UVR8 with COP1 does not require the C-terminal 44 amino acids that include the C27 domain. Rather, interaction of UVR8 with COP1 via its b-propeller domain after UV-B activation may allow functional interaction of the C27 domain with the COP1 WD40-repeat domain. This interaction of the UVR8 C27 domain and COP1 is required, however, for further stabilization of the interaction and, in particular, for efficient UV-B signaling.
The Amino Acids Val-410 and Pro-411 Are Critical for UVR8 Signaling Previous work identified a VP domain with the core sequence V-P-E/D-w-G (where w = a hydrophobic residue) that mediates the interaction of several proteins with the COP1 WD40-repeat domain (Holm et al., 2001;Datta et al., 2006). Indeed, such a potential VP core domain also was identified in the UVR8 C27 domain (Supplemental Figure 2) (Wu et al., 2013). Mutation of the VP core to Ala-Ala (AA) was shown to abrogate COP1 interaction with several target proteins (Holm et al., 2001(Holm et al., , 2002Datta et al., 2006). To analyze the importance of the UVR8 VP pair, both Val-410 and Pro-411 were mutated to Ala (VP-AA). Indeed, UVR8 C44/VP-AA did not interact with COP1 (Figure 2A), confirming the importance of these two residues in the UVR8 C27 domain for COP1 interaction. By contrast, a clear interaction of UVR8 VP-AA with COP1 was detected in response to UV-B ( Figure 2A). These data are consistent with our finding of two COP1 interaction modes in UVR8: via the UVR8 b-propeller core that conserves UV-B-dependent interaction with COP1 in the absence of the C-terminal 44 amino acids and via the C27 domain (including Val-410 and Pro-411) within the C-terminal 44 amino acids that can interact constitutively with COP1 in the absence of the b-propeller core. This explains why the constitutive interaction of UVR8 C44 is blocked when mutated (UVR8 C44/VP-AA ) and why the mutated full-length UVR8 (UVR8 VP-AA ) still shows UV-Bdependent interaction with COP1.
To investigate the requirement for the UVR8 C-terminal VP pair in planta, we generated transgenic plants expressing UVR8 VP-AA in uvr8-7 null mutants. We then selected lines with protein levels similar to those in the wild type ( Figure 2B).
Relative to the wild type, hypocotyl growth inhibition in response to UV-B was impaired in UVR8 VP-AA expression lines, as it was in uvr8-7 ( Figures 2C and 2D). Similarly, UVR8 VP-AA was impaired in HY5 marker gene activation in response to UV-B ( Figure 2E). However, coimmunoprecipitation assays demonstrated that UVR8 VP-AA interacted with COP1 in a UV-B-dependent manner in planta, albeit to a lower extent than did endogenous and cauliflower mosaic virus 35S promoter-driven wild-type UVR8 ( Figure 2F). This is in agreement with the yeast interaction data ( Figure 2D) and the coimmunoprecipitation of endogenous COP1 with truncated UVR8-2 (UVR8 N400 ) protein ( Figure 1D). Thus, UVR8 VP-AA is blocked specifically in UV-B signaling and cannot complement the uvr8-7 null mutant, despite its photoresponsiveness and UV-B-dependent interaction with COP1.
UVR8 Trp-285 is a crucial residue for UV-B photoreception, and its mutation to Phe or Ala rendered UVR8 W285F and UVR8 W285A UV-B insensitive (Rizzini et al., 2011;Christie et al., 2012;O'Hara and Jenkins, 2012;Wu et al., 2012;Liu et al., 2014). Recently, we and others reported that the single amino acid substitution of Ala in place of Trp in UVR8 W285A generates a constitutively active UV-B photoreceptor Huang et al., 2013). To determine whether the activity of UVR8 W285A is dependent on the VP pair at the C terminus, we generated transgenic plants expressing UVR8 W285A/VP-AA at a level higher than UVR8 W285A line 4, all in the uvr8-7 mutant background ( Figure 3A). UVR8 W285A line 4 developed short hypocotyls in weak white light ( Figure 3B), as described before . By contrast, this constitutive  (Kliebenstein et al., 2002;Finn et al., 2014) and a C27 interaction domain (Cloix et al., 2012). (B) Quantitative yeast two-hybrid assays were performed in the presence (+) or absence (2)  UVR8-COP1 Interaction and Signaling 3 of 12 photomorphogenic phenotype was absent in lines expressing UVR8 W285A/VP-AA ( Figure 3B), demonstrating that the constitutive activity of UVR8 W285A depends on the C-terminal VP pair and its interaction with COP1.

of 12
The Plant Cell rather surprisingly, we conclude that UVR8 W285A interacts constitutively with COP1 via its C terminus and not via its b-propeller domain.
Induced Expression of UVR8 C44 in uvr8-7 Mimics Early UV-B Signaling Because the UVR8 C terminus is required for UV-B signaling ( Figure 1E) (Cloix et al., 2012), we tested whether the C terminus is sufficient to mimic signaling when released from structural restraints in the UVR8 full-length protein. We generated transgenic Arabidopsis thaliana lines using the estrogen receptorbased XVE transactivator system (Zuo et al., 2000) to induce the expression of UVR8 C44 -UVR8 C44 -NLS-UVR8 C44 -2A-YFP by the addition of 17-b-estradiol (NLS, nuclear localization signal; 2A, cotranslational cleavage site; YFP, yellow fluorescent protein).
Expression of the 2A peptide should result in cotranslational cleavage of the fusion protein and the release of a triple C44 peptide and YFP (Ryan and Drew, 1994). The cleaved C-terminal fraction of the fusion protein (i.e., YFP plus part of 2A) was detected by immunoblotting with anti-YFP antibody after 5 h of estradiol treatment ( Figure 4A), indicating that the fusion protein was properly induced and processed. However, we could not detect the processed triple C44. Estradiol-inducible expression of YFP alone was used as a negative control ( Figure 4A). Strikingly, the induced expression of UVR8 C44 -UVR8 C44 -NLS-UVR8 C44 -2A-YFP led to HY5 mRNA accumulation ( Figure 4B), indicating that the C-terminal 44 amino acids of UVR8 can trigger gene expression even though this C44 peptide does not accumulate to detectable levels, as determined by immunoblot analysis. Because the inducible YFP control did not result in HY5 upregulation despite much higher YFP levels, triggering of gene expression was presumably due to the triple C44 fragment ( Figure 4B). Thus, induced expression of UVR8 C44 is sufficient to trigger HY5 gene expression in transgenic plants.

Expression of UVR8 C44 in uvr8-7 Causes Constitutive UV-B Responses
To analyze the effect of stably expressing the UVR8 C44 domain in planta, we generated transgenic Arabidopsis lines expressing YFP-NLS-UVR8 C44 driven by the 35S promoter in the uvr8-7 null mutant background and selected transgenic lines with different YFP-NLS-UVR8 C44 protein levels ( Figure 5A). Strikingly, these transgenic lines had shorter hypocotyls than uvr8-7 seedlings when grown in weak light ( Figures 5B and 5C). Overexpression of a double YFP fusion (YFP-YFP) in uvr8-7 did not affect hypocotyl length ( Figures 5B and 5C), demonstrating that the observed effect is due to NLS-UVR8 C44 and not to the YFP tag.
In agreement with the short-hypocotyl phenotype in the absence of UV-B, constitutive expression of YFP-NLS-UVR8 C44 also resulted in constitutively elevated levels of CHS marker gene expression ( Figure 5D). The constitutive responses in YFP-NLS-UVR8 C44 -expressing transgenic lines ( Figures 5B to 5D) together with the constitutive interaction between UVR8 C44 and COP1 in yeast ( Figure 1A) suggested that YFP-NLS-UVR8 C44 interacts constitutively with COP1 in planta. Indeed, endogenous COP1 was coimmunoprecipitated with YFP-NLS-C44 from transgenic seedlings in the absence of UV-B ( Figure 5E). Thus, the C44 region in YFP-NLS-UVR8 C44 -expressing plants interacts constitutively with COP1, leading to elevated marker gene expression and a constitutive photomorphogenic response.

Interaction of UVR8 with RUP1 and RUP2 Is Limited to the C-Terminal C27 Domain
Previous work has shown that UVR8 can interact with RUP1 and RUP2 in the absence of UV-B both in plants and yeast Cloix et al., 2012). This suggests, in contrast with COP1, that RUP1 and RUP2 can interact with the UVR8 homodimer and that the interactions of UVR8 with COP1 and UVR8 with RUP1/RUP2 may differ mechanistically. Interestingly, and in stark contrast with UVR8-COP1, the UVR8-RUP1/RUP2 interaction was dependent on the C-terminal 44 amino acids. Whereas UVR8 N396 interacted with COP1 under UV-B light ( Figure 1B), interaction of UVR8 N396 with RUP1 and RUP2 was not detected ( Figure 6A). Thus, in contrast with COP1, UVR8 apparently interacts with RUP1 and RUP2 solely via the C-terminal 44 amino acids and not via the b-propeller domain ( Figures 6A and 6B). In agreement with this, UVR8 VP-AA (and UVR8 C44/VP-AA ) also did not interact with RUP1 and RUP2, although UVR8 VP-AA was found to interact with COP1 under UV-B ( Figure 6C; see also Figure 1B). Consistent with the UV-Bindependent interaction of UVR8 with RUP1 and RUP2, the constitutively dimeric UVR8 W285F interacted with RUP1 and RUP2 in yeast ( Figure 6D), but not with COP1 (Rizzini et al., 2011;O'Hara and Jenkins, 2012). However, compared with UVR8 and UVR8 W285F , we found a stronger interaction with the partially active UVR8 W285A variant ( Figure 6D). This is in agreement with previous data suggesting that the UVR8-RUP1 and UVR8-RUP2 interactions, albeit constitutive, are enhanced by UV-B (Cloix et al., 2012). We further tested the effect of the VP-AA mutation that impairs UVR8 interaction with RUP1 and RUP2 upon the reversion of UVR8 to its ground state . Indeed, redimerization of UVR8 VP-AA was strongly impaired compared with wild-type UVR8 ( Figure 6E). We conclude that the difference in the UV-B dependence of the UVR8-COP1 and UVR8-RUP1/ RUP2 interactions is due to at least two differences in their modes of interaction: (1) the ability of the b-propeller surface exposed in the UVR8 monomer to interact with COP1 but not with RUP1 and RUP2; and (2) the distinct ability of COP1 and RUP1/RUP2 to interact with the UVR8 C-terminal 44 amino acids. In the case of COP1, this interaction occurs only after UV-B activation and monomerization of UVR8, but in the case of RUP1 and RUP2, it occurs also with the nonactivated homodimeric UVR8.

UV-B-dependent interaction of the UV-B photoreceptor UVR8
with COP1 is a key event in UV-B signaling that induces photomorphogenic responses and also acclimation to UV-B (Heijde and Ulm, 2012;Li et al., 2013;Jenkins, 2014). Although previous results indicated the importance of a C-terminal 27-amino acid domain (C27), it was not clear how UV-B-mediated UVR8 monomerization results in UVR8-COP1 interaction based solely on C27 as the interaction domain, particularly as C27 alone results in constitutive interaction with COP1 in yeast (Cloix et al., 2012).
It was suggested that C27 is hidden in the homodimeric structure of UVR8 and only exposed upon monomerization after UV-B irradiation (Cloix et al., 2012). However, experimental support was lacking, especially as the C-terminal domain was not included in the UVR8 crystal structure published recently Wu et al., 2012). In contrast with the model of C27 exposure after UV-B irradiation, we show here that the seven-bladed b-propeller domain of UVR8 mediates UV-Bdependent interaction of UVR8 and COP1 even in the absence of C27 (Figure 1). This is in agreement with the successful use of the UVR8 core (amino acids 12 to 381) alone with the COP1 WD40 domain to generate a UV-B-responsive split transcription factor based on UV-B-dependent protein-protein interaction as an optogenetic tool in mammalian cells (Müller et al., 2013). Also in planta, it is the WD40-repeat region of COP1 that mediates the interaction with active UVR8 (Favory et al., 2009). Intriguingly, WD40-repeat domains fold into a b-propeller structure similar to

of 12
The Plant Cell RCC1-repeat domain proteins (Xu and Min, 2011), and in fact, a seven-bladed b-propeller structure is predicted for the COP1 C-terminal WD40-repeat domain (Holm et al., 2002;Wu et al., 2013). Our data thus now indicate that the surface of the UVR8 monomer freed by UV-B-dependent monomerization of UVR8 binds to the structurally closely related b-propeller surface of the COP1 WD40-repeat domain. Interestingly, recent Fourier transform infrared spectroscopy difference spectra comparing full-length UVR8 with a C-terminally truncated version could not attribute any structural change to the deleted C-terminal region in response to UV-B (Heilmann et al., 2014). Moreover, the UVR8 C-terminal region is apparently located at the distal ends of the dimer away from the dimer interface . It is thus not clear how the UVR8 C27 domain is prevented from interacting with COP1 in the homodimeric ground state and what step is required so that this C-terminal region can interact with COP1 after UV-B photoreception by UVR8. However, our data indicate that the UVR8 b-propeller surface is liberated upon UV-B perception, allowing interaction with the COP1 WD40 domain.
Thus, we propose that interaction of the UV-B-activated monomeric UVR8 b-propeller core with COP1 not only contributes significantly to the crucial UV-B-dependent UVR8-COP1 interaction but also promotes the functional interaction of the UVR8 C terminus with COP1. This indicates the possibility of a sequential two-step interaction mechanism: first, a UV-Bdependent stable interaction of the UVR8 b-propeller core with COP1; next, an interaction of the UVR8 C27 domain affecting UVR8-COP1 Interaction and Signaling 7 of 12 COP1 function. However, we cannot exclude a simultaneous interaction of both UVR8 interaction domains with COP1. It seems clear, however, that the UVR8 C27 domain is required for stabilization of the UVR8-COP1 interaction ( Figures 1B, 1D, and 2F). Notwithstanding, our data further demonstrate that induced expression of the C-terminal 44 amino acids of UVR8 alone, including the C27 domain, is sufficient to induce HY5 gene expression ( Figure 4B) and that sustained expression results in a phenotype resembling continuous UV-B exposure (including hypocotyl growth inhibition) (Figures 5B to 5D; Supplemental Figure 3). Thus, whereas the interaction of the UVR8 b-propeller core with COP1 is central to the UV-B dependence and likely the robustness of the interaction, interaction of the UVR8 C27 domain with COP1 is apparently the key step affecting COP1 function (Figure 7). A VP pair was identified previously in several unrelated proteins that interact with the COP1 WD-repeat domain (Holm et al., 2001(Holm et al., , 2002Datta et al., 2006). A similar VP pair (Val-410-Pro-411) that is potentially involved in UVR8-COP1 interactions is predicted in the UVR8 C27 domain (Wu et al., 2013;this work). Mutagenesis of this UVR8 VP pair to Ala-Ala impaired the interaction of COP1 with the C-terminal 44 amino acids (UVR8 C44,VP-AA ) but not with full-length UVR8 VP-AA (Figure 2). Despite the UV-B-dependent interaction of UVR8 VP-AA , UVR8-2, and UVR8 N396 with COP1 that was clearly detectable in yeast and in plants, absence of the C27 domain impaired early UV-B signaling and responses, further supporting the importance of the UVR8 C-terminal region for UV-B signaling.
Interestingly, the previously described constitutive interaction of the singly substituted UVR8 W285A with COP1 in yeast and plants (Rizzini et al., 2011;O'Hara and Jenkins, 2012;Huang et al., 2014; this work) was found to depend solely on the C-terminal domain (Figure 3). We showed previously that a weak 4-fold overexpression of UVR8 W285A in transgenic Arabidopsis plants (in contrast with 40-fold overexpression of wild-type UVR8) results in a phenotype resembling a constitutive UV-B response . By contrast, but in agreement with the impaired COP1 interaction, overexpression of ) and uvr8-7/Pro35S:UVR8 VP-AA line 11 seedlings. Seedlings were either left unirradiated (2UV) or irradiated with broad-band UV-B light for 15 min (+UV) to induce UVR8 monomerization before being subjected to a recovery period in white light for 0.5 to 4 h. Asterisks indicate nonspecific cross-reacting bands.

of 12
The Plant Cell UVR8 W285A,VP-AA did not result in a constitutive UV-B response ( Figure 3B). Thus, the constitutive activity of UVR8 W285A depends on the C27 domain. Moreover, our data indicate that the constitutive interaction of UVR8 W285A with COP1 is solely via its C27 domain and not due to a freed b-propeller surface in a potentially monomeric in vivo conformation (Rizzini et al., 2011;. Therefore, it is likely that in the UVR8 W285A constitutively active photoreceptor variant, the C27 domain is constitutively exposed to interaction with COP1, initiating signaling in the absence of UV-B. Interestingly, although clearly more active than the wild-type UVR8 protein, the UVR8 W285A protein is not as active as UV-B-activated UVR8 (otherwise, wild-type or even lower levels of expression would result in a constitutive growth phenotype) Huang et al., 2013Huang et al., , 2014. This may be explained by the absence of the stabilizing interaction of the UVR8 b-propeller core with COP1 in UVR8 W285A . Therefore, we propose that the UV-Bdependent UVR8 b-propeller contributes to target COP1, thereby efficiently directing the C27 domain to the correct protein for signal propagation.
UVR8 inactivation occurs by redimerization of UVR8 monomers to the homodimer Heilmann and Jenkins, 2013). UVR8-COP1 complex formation results in the induction of RUP1 and RUP2 expression, forming a negative feedback loop that facilitates UVR8 redimerization and thus disruption of the UVR8-COP1 interaction Ouyang et al., 2014). It was shown previously that RUP1 and RUP2 interact with the C27 domain of UVR8, potentially competing with COP1 (Cloix et al., 2012). We show here, in contrast with COP1, that RUP1 and RUP2 indeed interact solely with the C27 domain and not with the UVR8 core ( Figures 6A to 6C). This is despite the fact that RUP1, RUP2, and COP1 are phylogenetically and structurally related: their WD40 domains predict similar seven-bladed b-propeller folds . Moreover, in contrast with COP1, the RUP1 and RUP2 proteins can still interact with the C27 domain of inactive homodimeric UVR8 (Figure 6) Cloix et al., 2012). This further indicates that the C27 domain is not simply generally hidden in the UVR8 homodimeric ground state but that interaction with the WD40-repeat proteins RUP1 and RUP2 is possible, and interaction with the WD40-repeat domain of COP1 must be more specifically prevented.
Independent of this, it is not known whether RUP1 and RUP2 proteins must be removed before UVR8 homodimers can be monomerized in response to UV-B. Previous data show that overexpression of RUP2 results in stable UVR8 homodimers . But at present, it is not possible to distinguish unequivocally between instantaneous redimerization and blocking of monomerization when RUP2 is overexpressed and interacts constitutively with UVR8. In both cases, however, the generation of signaling-competent UVR8 monomers ultimately requires the dissociation of RUP1 and RUP2 from UVR8 by an as yet unknown mechanism. The complexity of interaction mechanisms, signal propagation, and modulation of the crucial UVR8-COP1-RUP core of the UV-B receptor photocycle demonstrated so far underlines the need for accurate reception and signaling for optimal induction by UV-B of photomorphogenesis and stress acclimation.
A synthetic nucleotide fragment of UVR8 C44 -UVR8 C44 -NLS-UVR8 C44 -2A-YFP (Supplemental Figure 4) was purchased from Biomatik and cloned into the pMDC7 binary vector for estradiol-inducible expression (Curtis and Grossniklaus, 2003). The 2A sequence that provides a cotranslational In response to UV-B irradiation, UVR8 homodimers dissociate instantly, which allows the UVR8 seven-bladed b-propeller domain to interact with the COP1 WD40 domain (a structurally related seven-bladed b-propeller). The activated UVR8 monomer then also binds to the COP1 WD40 domain with its C27 domain (indicated by the pink crescent labeled C) and initiates UV-B signaling. Release of the C27 domain from structural constraints in the UV-B light-activated UVR8 is thought to allow its interaction with COP1. The activated UVR8-COP1 signaling pathway induces RUP1 and RUP2 expression, forming a negative feedback loop. RUP1 and RUP2 are WD40-repeat proteins that are phylogenetically and structurally related to COP1. They interact solely with the C27 domain of UVR8 and facilitate UVR8 redimerization and disruption of the UVR8-COP1 interaction. Note that no stable RUP1/RUP2-UVR8-COP1 complex is known, but it is assumed here to occur transiently when RUP1 and RUP2 attach to the C27 domain of UVR8 while UVR8 and COP1 still interact via their b-propeller surfaces. In contrast with COP1, RUP1 and RUP2 proteins are still able to interact with the C27 domain in the inactive homodimeric UVR8. The COP1-interacting SPA proteins as well as the homodimeric constitution of COP1 are omitted from the model. cc, coiled coil; RING, Really Interesting New Gene; WD40, WD40 repeat domain.
UVR8-COP1 Interaction and Signaling 9 of 12 cleavage site was described previously (Ryan and Drew, 1994). Flexible peptide linkers were added between tandem repeats of UVR8 C44 . Application of b-estradiol (Sigma-Aldrich) in DMSO (50 mM) was used to induce the expression of this construct in transgenic Arabidopsis seedlings. Arabidopsis plants were transformed by the floral dip method (Clough and Bent, 1998). The transgenic lines described in this work were genetically determined to have the transgene integrated at a single locus (;75% resistant to 25% sensitive in the T2 generation) and were used in the homozygous T3 generation. The growth conditions and UV-B treatments were as described previously .

Quantitative Real-Time PCR
Arabidopsis total RNA was isolated with the Plant RNeasy Kit (Qiagen) and treated with DNaseI according to the manufacturer's instructions. cDNA was synthesized using a 1:1 ratio mix of oligo(dT) 16 and random hexamer primers using the TaqMan Reverse Transcription Reagents Kit (Life Technologies). PCR was performed and detected using the ABsolute QPCR SYBR Green ROX Mix according to the manufacturer's instructions (Thermo Scientific). Primers for HY5 were as follows: HY5_Fv (59-CAA-GCAGCGAGAGGTCATCA-39) and HY5_Rv (59-CATCGCTTTCAATTCCTT-CTTTG-39). Primers for CHS were as follows: CHS_Fv (59-CGTGTTGAGCG-AGTATGGAAAC-39) and CHS-Rv (59-TGACTTCCTCCTCATCTCGTCTAGT-39). cDNA concentrations were normalized to the 18S rRNA transcript levels using the Eukaryotic 18S rRNA Kit (Applied Biosystems). Quantitative real-time PCR was performed using the 7900HT Real-Time PCR System (Applied Biosystems). Expression was determined in triplicate biological measurements.