CONSTITUTIVELY PHOTOMORPHOGENIC1 Is Required for the UV-B Response in Arabidopsis

HY5 Is Required for UV-B–Mediated Gene Activation and UV-B Tolerance

Our previous work demonstrated the requirement of HY5 in UV-B–responsive gene expression by analyzing a few selected genes in the hy5-1 null mutant (Ulm et al., 2004). We further assessed this function by gene profiling at the whole-genome level using Affymetrix ATH1 oligonucleotide microarrays. Seven-day-old white light–grown wild-type or hy5-1 Arabidopsis seedlings were exposed for 15 min to polychromatic radiation with decreasing short-wave cutoff in the UV range (WG305 = +UV-B; WG327 = −UV-B), and samples were taken 1 h after the onset of irradiation. In total, 68.6% of the early UV-B–activated genes in the Landsberg erecta (Ler) ecotype remain inducible in the hy5-1 mutant (278 out of 405 genes) (Figure 1A ; see Supplemental Table 1 online). From the microarray analysis, it is apparent that HY5 regulates a number of genes that can be connected to UV tolerance, such as PHR1 and UVR3 encoding photolyases specific for the repair of CPD- and 6-4-photoproducts, respectively (Britt, 2004), flavonoid biosynthetic genes (including CHS, CHI, PAL1, PAL2, and FLS1; Winkel-Shirley, 2002), and their transcriptional regulators (e.g., MYB12; Mehrtens et al., 2005) (see Supplemental Table 1 online). In agreement with this notion, we found reduced UV-B tolerance in hy5-1 mutants compared with the wild type (Figure 1B). Thus, we conclude that HY5 regulates the gene expression response required for protection and consequently survival under UV-B.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

hy5-1 Is Deficient in UV-B–Activated Gene Expression and UV-B Tolerance.

(A) Numbers of genes defined as responding to UV-B in hy5-1 and wild-type Ler. Venn diagram shows the number of genes classified as responding to UV-B in either the wild type only (left; i.e., HY5-dependent genes), the wild type and mutant (center; i.e., HY5-independent genes), or mutant only (right). The corresponding gene list can be found in Supplemental Table 1 online.

(B) HY5 is required for UV-B tolerance. Wild-type and hy5-1 plants were grown in white light with or without supplementary UV-B. Plants were photographed after 4 weeks. Under supplementary UV-B, 13 plants died and 35 showed necrotic lesions among the hy5-1 mutants, whereas in the parallel-grown wild-type population, no plant died or was necrotic (n = 50). When control plants from both genotypes were grown without supplementary UV-B, all survived without apparent necrosis.

cop1 Mutants Are Broadly Impaired in UV-B–Mediated Gene Expression, Including HY5 Activation

The crucial role of HY5 in the UV-B response indicates the use of shared signaling components in response to visible light and UV-B. A prominent regulator of HY5 action in the dark-to-light transition is the E3 ubiquitin ligase COP1; we thus analyzed the UV-B response in Arabidopsis cop1 mutants. Since null mutants of COP1 are seedling lethal, we resorted to the widely used nonlethal cop1-4 mutant. The COP1-4 allele has a premature stop codon removing the C-terminal 393 amino acids; the mutant therefore expresses a truncated protein containing only the N-terminal 282 amino acids lacking the WD40 repeats (McNellis et al., 1994b). Surprisingly, we found that the UV-B–mediated activation of a number of selected genes is blocked in the cop1-4 mutant (Figure 2A ), similarly to hy5 null mutants (Ulm et al., 2004; this work). HY5 is actually among the UV-B–activated genes that are blocked in cop1-4 mutants (Figure 2A). In agreement with this, we have also identified COP1 as a UV-B signaling component in an independent genetic approach using a luciferase reporter gene–based screen. Pro_HY5:Luc⁺ transgenic plants were mutagenized by ethyl methanesulfonate, and seedlings lacking UV-B induction of luciferase activity were identified. Among the isolated recessive mutants, one showed a standard-growth phenotype resembling cop1 mutants, including dwarf growth, early flowering, and a light-grown habit in dark (McNellis et al., 1994b) (see Figure 2B for the luciferase-reporter phenotype of mutant 0650). In fact, we have fully sequenced the COP1 gene in that mutant and identified a single C-T transition, changing the Gln-283 codon (CAA) to a Stop codon (TAA), identical to the COP1-4 mutation described previously (McNellis et al., 1994b). This new cop1-4 allele, however, was confirmed to be in the Wassilewskija (Ws)/Pro_HY5:Luc⁺ background by (1) detectable basal luciferase activity in the mutant (at levels comparable to the wild type), (2) UV-B inducibility of the luciferase reporter in the F1 generation after outcrossing to the wild type, and (3) simple sequence length polymorphism analysis differentiating between Columbia (Col) (ecotype of the previously published cop1-4 allele; McNellis et al., 1994b) and Ws ecotypes (used markers: CIW11, NGA76, and NGA168) (data not shown). The requirement for COP1 in the UV-B response was also confirmed by two other cop1 mutant alleles: the strong cop1-1 (McNellis et al., 1994b) (see below) and a new cop1 allele isolated from a T-DNA collection (SALK_137391; t et al., 2003) (data not shown). Moreover, the cop1-4 mutant phenotype under UV-B is restored by the expression of a wild-type copy of COP1 fused with yellow fluorescent protein (YFP) (see below). Thus, our data strongly support the requirement for COP1 in the UV-B response in Arabidopsis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

COP1 Is Required for UV-B–Responsive Gene Expression, Including HY5 Gene Activation.

(A) RNA gel blot analysis of the cop1-4 mutant compared with its wild type (Col). Total RNA was isolated from 7-d-old white light–grown seedlings at 45 min after 15-min irradiation under a UV-B field with different UV spectra (−UV-B under cutoff WG327; +UV-B with decreasing short-wave cutoff under WG305, WG295, or quartz glass [Q]) or left untreated in the standard growth chamber (control [C]). Blots were sequentially hybridized with specific probes for the indicated genes. Ethidium bromide–stained rRNA is shown as loading control.

(B) New cop1-4 allele identified in a luciferase-based genetic screen. Normalized data of TopCount luciferase bioluminescence measurements of 10 Pro_HY5:Luc⁺ seedlings each after 15 min of UV-B irradiation. Gray: parental Pro_HY5:Luc⁺ transgenic line (wild type); black: mutant 0650 (due to COP1-4 mutation). Note that the data for each seedling were normalized to their mean (absolute values not shown) to facilitate the comparison of induction properties. An absolutely blind mutant is expected to have a normalized luminescence value of 1 over the time course.

(C) Numbers of genes defined as responding to UV-B in the cop1-4 mutant and wild-type Col. Venn diagram shows the number of genes classified as responding to UV-B in either the wild type only (left; i.e., COP1-dependent genes), the wild type and mutant (center; i.e., COP1-independent genes), or mutant only (right). The corresponding gene list can be found in Supplemental Table 2 online.

(D) Overlap between COP1- and HY5-independent genes. Analysis limited to genes commonly UV-B induced in both wild types Col and Ler. Venn diagram depicts the relationship of genes induced in cop1-4 and hy5-1. The total number of genes in each category is shown in parentheses. UV-B activation of the 103 genes indicated in the lower left is blocked in both cop1-4 and hy5-1 mutants (see Table 1). The corresponding gene list can be found in Supplemental Table 3 online.

We further assessed the involvement of COP1 in mediating molecular UV-B responses by gene profiling at the whole-genome level using ATH1 oligonucleotide microarrays. Our expression profiling of the cop1-4 mutant identified a broad impact on UV-B responsiveness. Only 24.7% of the early genes activated in response to low-level UV-B radiation in the Col ecotype are also induced in the cop1-4 mutant and can therefore be designated as COP1 independent (89 out of 361 genes) (Figure 2C; see Supplemental Table 2 online); in other words, 75.3% of the UV-B–induced genes, including HY5, depend on a functional COP1 protein. Thus, both COP1 and HY5 are positive regulators of UV-B–induced gene expression, and COP1 is required for HY5 gene activation.

Our initial analysis of UV-B–responsive gene expression changes identified a UV-B stress pathway that antagonistically interferes with the postulated UV-B photoreceptor pathway. To test whether or not the UV-B stress pathway (P2 pathway described in Ulm et al., 2004) is affected by COP1 deficiency, we tested selected genes for their response to shorter UV-B wavelength regions. None of the marker genes studied required COP1 under these conditions (see, for example, representative genes At5g59820 and At1g19020 in Figure 2A), indicating that COP1 participation is specific for the postulated UV-B photoreceptor pathway and does not involve the more general UV-B stress pathway.

Gene Expression Profiling Identifies the Readout of the UV-B–Activated COP1-HY5 Pathway

The impaired UV-B induction of HY5 in cop1-4 prompted us to investigate which part could be tracked to the direct link from COP1 to HY5 activation. We hypothesized that the UV-B–induced genes that require HY5 should be largely included within the group of genes that are dependent on COP1. To be able to directly compare the UV-B response in hy5-1 and the cop1-4 mutant, the analysis was focused on genes that are commonly induced in both Ler and Col ecotypes (total of 320 genes). This restriction allows a conservative estimation of COP1 and HY5 dependence for UV-B–mediated gene induction (see Supplemental Table 3 online). Interestingly, of the COP1-dependent 240 genes shown in Figure 2D, close to half (103 genes; Table 1 ) also require the HY5 protein, delineating a conservative estimate of the output of the UV-B–activated COP1-HY5 pathway. As expected, these 103 genes make up a large part of the 107 genes that were found to be blocked in the hy5-1 mutant, indicating that HY5-dependent genes are also COP1 dependent, in agreement with a sequential action of COP1 and HY5 for this subgroup of genes.

cop1-4 Mutant Plants Have Reduced Tolerance to Supplementary UV-B

As the UV-B hypersensitivity of hy5-1 indicated the importance of the HY5 regulatory pathway for UV-B tolerance, we tested the cop1-4 mutant under supplementary UV-B. Similar to hy5-1, cop1-4 mutant plants are impaired in their UV-B tolerance but to a lesser extent, without lethality occurring under our conditions. In contrast with the wild type, the cop1-4 mutant plants show chlorosis when grown under supplementary UV-B (Figure 3 ). This might be caused by the block of UV-B–induced genes encoding chloroplast proteins in cop1-4 (e.g., σ-like factor SIG5, ELIP proteins, and FtsH proteases; see Supplemental Table 2 online). The reduced UV-B tolerance of cop1-4 doesn't seem to be due to a general weakness linked to dwarf growth, as the phenotypically similar det1-1 mutant (Pepper et al., 1994; Ma et al., 2003) does not show reduced UV-B tolerance (Figure 3). It is also interesting to note here that supplementary UV-B results in significant shortening of petiole length in the wild type, cop1-4/Pro_35S:YFP-COP1, and det1-1. This phenotype is not apparent in cop1-4 that already has short petioles under conditions devoid of UV-B (Figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

cop1 Plants Develop Chlorosis under Supplementary UV-B.

Wild-type, cop1-4, cop1-4/Pro_35S:YFP-COP1, and det1-1 plants were grown in white light with or without supplementary UV-B. Plants were photographed after 5 weeks. Under supplementary UV-B, 37 plants developed chlorosis among the cop1-4 mutant, whereas among the parallel-grown wild type, cop1-4/Pro_35S:YFP-COP1, and det1-1 mutant, no plant was chlorotic (n = 50). When control plants from all four genotypes were grown without supplementary UV-B, all survived without apparent chlorosis. Bars = 1 cm.

COP1 and HY5 Are Required for UV-B–Mediated Hypocotyl Growth Inhibition and Flavonoid Accumulation

Since it is the UV-B pathway in response to low-level UV-B that is blocked in cop1 and hy5 mutants, it was of interest to analyze the contributions of COP1 and HY5 to photomorphogenic, non-stress-mediated UV-B responses. A well-established UV-B effect thought to be mediated by the postulated UV-B photoreceptor is the induction of phenylpropanoid biosynthetic pathway components leading to the accumulation of sunscreen flavonoids (e.g., Beggs and Wellmann, 1994; Kucera et al., 2003). To analyze UV-B effects at the organismal level, we use supplemental narrow-band UV-B (311 to 313 nm) irradiation. Continuous irradiation results in hypocotyl growth inhibition and the accumulation of flavonoids in the wild type. Both responses are reduced in the hy5-1 mutant (Figures 4A and 4B). The absence of a UV-B effect on hypocotyl growth in cop1-4 is less conclusive under our conditions, as this mutant has already short hypocotyls under white light without UV-B (Figure 4A), but is consistent with a previous report (Kim et al., 1998). More convincingly, we have found that UV-B–induced biosynthesis of flavonoids is strongly reduced in cop1-4 mutants (Figure 4B). This phenotype is restored to the wild type in three independent cop1-4/Pro_35S:YFP-COP1 lines checked (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Functional HY5 and COP1 Are Required for UV-B–Responsive Flavonoid Accumulation, Correlating with HY5 Induction Properties.

Seedlings were grown for 4 d under continuous light supplemented with UV-B under a 327-nm cutoff filter (−UV-B) or 305-nm cutoff filter (+UV-B). The 327-nm cutoff filters were exchanged after 4 d for a 305-nm cutoff at 1 h and 6 h before harvesting, where indicated.

(A) Reduced hypocotyl growth inhibition by supplementary UV-B in hy5-1. Hypocotyl length was measured from seedlings of phyA-201 phyB-5, hy5-1, cop1^eid6, and cop1-4 mutants, together with their wild types (Ler and Col), after 4 d of growth under white light (WL) or white light supplemented with UV-B (WL+UV-B). Numbers below bars show the relative hypocotyl growth inhibition by UV-B as a percentage. Bars show se of the mean (n = 120).

(B) UV-B–induced flavonoid accumulation is impaired in cop1-4 and hy5-1 mutants but not in cop1^eid6. Numbers below bars show the fold enrichment values of supplemental +UV-B compared with −UV-B. Bars show se of the mean (n = 9).

(C) Scheme for UV-B treatment. The time under a WG327 cutoff filter (−UV-B) is represented by a white bar and time under a WG305 cutoff filter (+UV-B) by a gray bar. Time is indicated in hours.

(D) Corresponding RNA gel blot showing UV-B–responsive CHS and HY5 gene expression.

(E) UV-B–induced HY5 protein accumulation is impaired in cop1-4 but not in cop1^eid6. The protein gel blot was sequentially probed with anti-HY5 and anti-actin antibodies.

To further characterize COP1 structural requirement in the physiological UV-B response, we investigated a nonconstitutive photomorphogenic cop1 allele. cop1^eid6 contains a mutation that changes the conserved His-69 residue of the RING finger motif to a Tyr. cop1^eid6 seedlings display normal etiolated growth in the dark but are strongly hypersensitive to visible light (Dieterle et al., 2003). The latter results in a cop phenotype comparable to cop1-4 when grown under light, including dwarf growth, early flowering, and elevated pigment levels (Dieterle et al., 2003). In contrast with cop1-4, however, light-grown cop1^eid6 mutant seedlings respond similarly to the wild type at the level of flavonoid accumulation under supplementary UV-B (Figure 4B). This difference between the two cop1 alleles (i.e., similar phenotypes in visible light versus distinct phenotypes under UV-B) indicates different structural requirements and functions of COP1 in the visible and UV-B pathways.

At the molecular level, the CHS gene encoding a key enzyme in flavonoid biosynthesis, chalcone synthase, is known to be induced by UV-B (e.g., Jenkins et al., 2001). Moreover, the HY5 protein was previously shown to bind CHS promoter sequences in vitro (Ang et al., 1998). Therefore, we analyzed HY5 and CHS gene expression and protein levels in seedlings that were grown for 4 d under continuous light either with or without supplementary UV-B in the same light field (under WG305 and WG327 cutoff, respectively) exactly as used for the flavonoid accumulation experiment in Figure 4B. In addition, seedlings were grown for 4 d without UV-B under a WG327 cutoff filter, which was then exchanged for a WG305 cutoff filter 1 or 6 h before harvesting to analyze the early UV-B response (see also scheme in Figure 4C). Our molecular data show that the flavonoid accumulation phenotypes correlate with the proficiency in activation of the UV-B–responsive CHS gene (Figure 4D) and the ability to transcriptionally activate and, as a result, accumulate HY5 protein (Figures 4D and 4E). Thus, in agreement with flavonoid accumulation, the inducibility of CHS and HY5 is maintained in cop1^eid6 mutants, whereas it is severely impaired in cop1-4 and hy5-1 mutants (Figure 4D). Residual activation of CHS in response to UV-B in hy5-1 seedlings but not in cop1-4 seedlings indicates the involvement of a second, HY5-independent pathway downstream of COP1 (Figure 4D).

Our results indicate a sequential action, with COP1 activating HY5 gene expression and HY5 protein then activating a set of downstream UV-B–responsive genes. To substantiate this notion, we analyzed the temporal relationship of gene activation of HY5 and HY5-dependent genes, including CHS. HY5 gene induction is detectable as early as 30 min, with a peak at ∼1 h, whereas CHS expression is strongly activated after 1 to 2 h and remains on high levels during 12 h of UV-B radiation (Figure 5A ). A similar delay is apparent in the case of two other HY5-dependent UV-B–induced genes analyzed, namely, ELIP2 and At5g23730 (Figure 5A). On the protein level, HY5 protein elevation is apparent at ∼1 to 2 h, from which time on it stays up (Figure 5B), correlating with prolonged CHS gene expression (Figure 5A) and CHS protein accumulation (Figure 5B). Thus, UV-B–mediated HY5 gene activation and HY5 protein accumulation seem to be in temporal correlation with the induction of at least the tested HY5-dependent genes. It is also noteworthy that HY5 protein levels in the wild type are elevated under UV-B at times when HY5 mRNA levels are back to background levels (see the 24 to 96 h time points; Figures 4D, 4E, 5A, and 5B). This indicates that the HY5 protein is additionally stabilized under these conditions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

UV-B–Mediated HY5 Gene Activation and HY5 Protein Accumulation Correlates with Induction of Putative HY5 Target Genes.

(A) RNA gel blot analysis of the wild type (Col). Total RNA was isolated from 4-d-old wild-type seedlings (Col) that were irradiated with UV-B for the indicated times before harvesting. Blots were sequentially hybridized with specific probes for the indicated genes. Ethidium bromide–stained rRNA is shown as loading control.

(B) Corresponding UV-B–responsive accumulation of HY5 and CHS protein in the Col wild type. The protein gel blot was sequentially probed with anti-HY5, anti-CHS, and anti-actin antibodies.

COP1 and HY5 Accumulate in the Nucleus in Response to UV-B

In response to visible light, COP1 is excluded from the nucleus where it accumulates in the dark and directs its target proteins, including HY5, to the proteasomal degradation machinery (von Arnim and Deng, 1994; Osterlund et al., 2000; Saijo et al., 2003). This nucleocytoplasmic partitioning of COP1 seems to be required for the dark-to-light growth transition by allowing the HY5 protein to accumulate and light-responsive gene expression to occur (Subramanian et al., 2004; Yi and Deng, 2005). By contrast, we found that both COP1 and HY5 proteins are required for the UV-B response in white light–grown seedlings (Figures 1 to 4⇑⇑). In an attempt to analyze COP1 localization and its possible changes in response to UV-B, we generated stable transgenic lines expressing the YFP-COP1 fusion protein under the control of the constitutive cauliflower mosaic virus 35S promoter in cop1-4. It is important to note that expression of the YFP-COP1 protein complements the dark- and light-grown phenotypes of the cop1-4 mutant to the wild type and, at the molecular level, restores proper HY5 degradation after transfer from light to dark (Figure 6A ). This is also true for all the UV-B responses tested, namely, gene expression, hypocotyl growth inhibition, flavonoid induction, tolerance, and HY5 protein accumulation (Figures 3, 6A, and 6C; data not shown). Thus, the YFP-COP1 fusion protein is functional and confers a wild-type photomorphogenic phenotype to cop1-4, allowing analysis of subcellular localization in response to UV-B. Interestingly, supplementary UV-B results in nuclear enrichment of the COP1 protein (Figure 6B) in a similar time frame as nuclear exclusion happens in the dark-to-light transition (∼24 h) (von Arnim and Deng, 1994). We note here that we did not detect any apparent changes in either the number or the size of the inclusion bodies that are formed by cytoplasmic COP1 (see, for example, von Arnim and Deng, 1994), and we confirmed the nuclear accumulation by staining DNA with Hoechst dye (Figure 6B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

YFP-COP1 Localizes to the Nucleus in Response to UV-B, Not Affecting HY5 Accumulation.

(A) The YFP-COP1 fusion protein is functional and complements the cop1-4 mutant phenotype. Top: after 6 h of light treatment to synchronize germination, the seedlings were grown for 4 d in the dark before the photograph was taken. Middle: 4-d-old seedlings grown under constant light were either kept in light (L) or transferred to dark (D) for 24 h. The protein gel blot was sequentially analyzed with anti-HY5 and anti-actin antibodies. Bottom: RNA gel blot. Total RNA was isolated from 7-d-old white light–grown seedlings at 45 min after 15-min irradiation under a UV-B field with different UV spectra (−UV-B under cutoff WG327; +UV-B with decreasing short-wave cutoff under WG305 or quartz glass [Q]). Ethidium bromide–stained rRNA is shown as loading control.

(B) to (D) Seedlings were grown for 4 d under continuous light supplemented with UV-B under a 327-nm cutoff filter (−UV-B) or 305-nm cutoff (+UV-B). The 327-nm cutoff filters were exchanged after 4 d for a 305-nm cutoff at 1 h and 6 h before harvesting, as indicated (C).

(B) UV-B–responsive accumulation of YFP-COP1 in the nucleus. Nuclear accumulation is confirmed by Hoechst 33342 nuclear stain (bottom panel). Bars = 20 μm.

(C) Accumulation of YFP-COP1 still allows UV-B–induced HY5 protein accumulation. The protein gel blot was sequentially analyzed with anti-HY5, anti-YFP, and anti-actin antibodies.

(D) UV-B–responsive accumulation of HY5-YFP in the nucleus. Bars = 20 μm.

In the dark, HY5 is a substrate of COP1 in the nucleus, leading to its ubiquitination and subsequent proteasomal degradation, so we were interested in determining HY5 protein levels under UV-B conditions when high levels of functional YFP-COP1 are present in the nucleus (Figure 6B). Indeed, we found that endogenous HY5 protein still keeps accumulating in Pro_35S:YFP-COP1 lines under UV-B to levels similar to the wild type (Figure 6C). As expected from previous work (Oyama et al., 1997) and its function as a transcriptional regulator, the HY5 promoter-driven HY5-YFP fusion protein accumulates in the nucleus under white light supplemented with UV-B, similar to COP1 (Figure 6D). This observation demonstrates that both of these crucial positive regulators of UV-B responses in Arabidopsis act in the nucleus. Thus, our data indicate that the promotion of HY5 degradation by YFP-COP1 is hindered under visible light with supplementary UV-B, even though both proteins reside in the nucleus.

The Function of COP1 in the UV-B Response Is Independent of SPA1 to SPA4 Proteins

The four members of the SPA family of proteins interact with COP1 and are essential for the suppression of photomorphogenesis (Laubinger et al., 2004; Hoecker, 2005). In all responses described so far, SPA proteins seem required for COP1 function (Laubinger et al., 2004). To find out if SPA proteins are also required for COP1 function in the UV-B response, we analyzed UV-B–responsive gene activation in a spa1 spa2 spa3 spa4 quadruple mutant. In contrast with cop1-1 and cop1-4 mutants, UV-B induction of HY5 and other tested UV-B–responsive genes is not impaired in the spa quadruple mutant (Figure 7 ). This indicates that the SPA proteins are not required; thus, COP1 can function properly without SPA proteins in the UV-B pathway.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

COP1 Function in the UV-B Response Is Independent of SPA Proteins.

RNA gel blot analysis of the spa1234 quadruple mutant and the cop1-1 single mutant in comparison with the wild type (Col). Total RNA was isolated from 7-d-old white light–grown seedlings at 45 min after 15-min irradiation under a UV-B field with different UV spectra (−UV-B under cutoff WG327; +UV-B with decreasing short-wave cutoff under WG305 or quartz glass [Q]) or left untreated in the standard growth chamber (control [C]). Blots were sequentially hybridized with specific probes for the indicated genes. Ethidium bromide–stained rRNA is shown as loading control.

COP1 Regulation of UV-B–Dependent Transcription and Photomorphogenesis

Microarray analysis of dark-grown cop1 mutants revealed an expression profile resembling light-grown wild-type plants, in agreement with their photomorphogenic phenotype in dark (Ma et al., 2002). Moreover, in agreement with genetic and biochemical data, the gene expression regulation attributable to HY5 in the light was included largely within those genes regulated by COP1 in the dark (Ma et al., 2002). Thus, this previous genomic analysis substantiated the genetically defined function of COP1 as a repressor of photomorphogenesis controlling the degradation of transcription factors and thereby the expression of their target genes. It was concluded that the bulk of light-controlled gene expression in wild-type Arabidopsis plants could be accounted for by the negative regulation of COP1 activity (Ma et al., 2002).

Using whole-genome profiling, we found that COP1 is responsible for the UV-B–mediated activation of at least three-quarters of the responsive genes in the wild type. Thus, the greater part of UV-B–controlled gene expression in the wild type requires COP1 as a positive regulator. Most interestingly, the broad impact on early gene activation indicates a COP1 function close to the postulated UV-B photoreceptor, similarly to visible light signaling that involves direct interaction of the COP1 protein with cryptochrome and phytochrome photoreceptors (Wang et al., 2001; Yang et al., 2001; Seo et al., 2004).

Target genes of the UV-B pathway through COP1 include HY5 and CHS. As a consequence, the cop1-4 mutant is impaired in the accumulation of flavonoids and the establishment of tolerance to UV-B. By contrast, under white light, cop1 mutants display elevated basal levels of CHS transcript and flavonoids compared with the wild type (Figures 4B and 4C), which are potential sunscreen metabolites that may interfere with the UV-B response (Li et al., 1993). The phenotype of elevated flavonoid levels in cop1-4 is shared with det1-1, cop1^eid6, and the spa quadruple mutant (Figure 4B; Laubinger et al., 2004). However, whereas cop1-4 is broadly impaired in UV-B responses, cop1^eid6, det1-1, and the spa quadruple mutant behave comparable to the wild type, indicating that the elevated flavonoid levels are not responsible for the observed UV-B phenotype of cop1-4. This conclusion is supported by the fact that molecular UV-B responses in double mutants of cop1-4 and a chs knockout mutant (cop1-4 tt4; TT4 = CHS) and cop1-4 TT4 siblings show a similar defect in UV-B signaling (see Supplemental Figure 1 online).

From our results, it is clear that COP1 plays a crucial role in the UV-B response of Arabidopsis. Namely, COP1 was found to play a positive regulatory role after UV-B perception. Similarly, a novel function of COP1 as promoter of responses to red light was described recently (Boccalandro et al., 2004), suggesting that COP1 may play opposing roles depending on light conditions. These observations contrast with the function as a repressor of photomorphogenesis (e.g., McNellis et al., 1994a).

COP1 Accumulates in the Nucleus under UV-B Radiation

The subcellular distribution of COP1 is adjusted according to light conditions, with nuclear localization in dark and a drastic reduction of COP1 in the nucleus in light. This light-dependent change in the nuclear abundance of COP1 is a relatively slow process taking ∼24 h (von Arnim and Deng, 1994; Yi and Deng, 2005). Nevertheless, recent mutagenesis studies indicated that light-dependent nuclear–cytoplasmic repartitioning of the COP1 protein is a rate-limiting step in light signal transduction (Subramanian et al., 2004).

We show that supplementary UV-B triggers nuclear accumulation of the YFP-COP1 protein, whereas, as expected, filtering out the UV-B radiation results in nuclear exclusion of YFP-COP1. As in white light, the nucleocytoplasmic repartitioning of COP1 under supplementary UV-B is a rather slow process (at least at the detection level), consistently detectable after ∼24 h. In both cases, this contrasts with early gene expression changes and the rapid kinetics of target protein stabilization (Osterlund et al., 2000; Duek et al., 2004; Yang et al., 2005). Thus, light seems to impede on COP1-mediated degradation of HFR1, HY5, and other regulatory factors prior to subcellular partitioning. It follows that a mechanism other than nucleocytoplasmic repartitioning must operate to rapidly regulate COP1 activity. It can, however, be speculated that the change in the localization of COP1 might represent an adaptive long-term response and could also modulate other signaling pathways employing COP1. Notwithstanding this, our data indicate that, similar to white light, the cellular localization of COP1 may play a critical role in mediating UV-B photoreceptor-controlled development in Arabidopsis.

It is important to note here that expression of YFP-COP1 in cop1-4 fully complements the mutant growth phenotype and restores HY5 degradation in dark, indicating that the fusion protein is active as an E3 ubiquitin ligase. Interestingly, we show that UV-B–mediated HY5 accumulation under supplementary UV-B is comparable to the wild type in lines overexpressing YFP-COP1. Because under these conditions both YFP-COP1 and HY5 are localized to the nucleus, the COP1 ubiquitin ligase activity against HY5 must be hindered. This may be due to an alteration of COP1 E3 ligase activity, protein modification influencing target affinity, or association with regulatory proteins.

The UV-B Photomorphogenic Response Is Mechanistically Separate from Light Responses

All cop1 mutants have an exaggerated photomorphogenic development in the light (Yi and Deng, 2005). Connected to this phenotype are a number of genes that are overexpressed in cop1 mutants compared with the wild type grown in light, indicating that cop1 is blocked in the repression of a set of genes (Ma et al., 2003). Thus, it could be hypothesized that supplementary UV-B results in an inactivation of COP1 releasing gene repression. Under the conditions used for our microarray analysis, we identified 575 genes overexpressed in cop1-4 grown without supplementary UV-B, but only 11 of those overlap with the UV-B–induced genes in the wild type (see Supplemental Table 4 online). Thus, a simple release of repression of light-regulated genes does not seem to be at work in COP1-mediated UV-B signaling.

In contrast with all other cop1 mutants presently known, the cop1^eid6 allele displays no constitutive photomorphogenic phenotype in darkness. Therefore, it was concluded that the cop1^eid6 allele encodes for a protein whose remaining activity is sufficient for the suppression of photomorphogenesis in dark-grown plants (Dieterle et al., 2003). Importantly, the cop1^eid6 mutant exhibits an extreme hypersensitivity toward all tested light qualities, resulting in an exaggerated photomorphogenic development in the light comparable to other cop1 mutants, such as cop1-4 (Dieterle et al., 2003). In contrast with cop1-4, however, the cop1^eid6 seedlings seem not affected in the UV-B response leading to HY5 and CHS gene activation and flavonoid accumulation. We conclude that responses to visible light and UV-B have different structural requirements for the COP1 protein.

Next to COP1, the light-regulated HY5 protein degradation via the ubiquitin/26S proteasome system also requires other COP/DET/FUS proteins (Osterlund et al., 2000). One group of proteins that might now be classified as belonging to the COP/DET/FUS class is the SPA proteins. The four SPA proteins interact with COP1 and regulate its activity (Saijo et al., 2003; Seo et al., 2003; Laubinger et al., 2004; Hoecker, 2005). They act partially redundantly, and a quadruple mutant is almost indistinguishable from the strong cop1-1 allele in the dark and under visible light (Laubinger et al., 2004; our unpublished data). To test whether these proteins also have an overlapping function in UV-B signaling, we investigated selected UV-B phenotypes in a spa quadruple mutant (Laubinger et al., 2004). The apparently functional UV-B response in the spa quadruple mutant indicates a specific involvement of COP1, independent of its regulatory proteins of the SPA family. This is particularly striking as the spa quadruple mutant has a much stronger light phenotype than the weak cop1-4 mutant that we widely used. It should be noted that cop1-4 may not be considered a weak allele in UV-B signaling, as the UV-B responsiveness of its target genes seems completely abrogated. Taken together, our data, including our observations on the cop1^eid6 allele, allow clear genetic separation of a COP1 function in UV-B from visible light signaling, in agreement with a specialized and novel function of COP1 in the UV-B response.

Relationship between COP1 and HY5 in the Regulation of UV-B–Dependent Transcription and UV-B Tolerance

The HY5 protein is required for the establishment of UV-B tolerance, as demonstrated by the strongly reduced survival of hy5 mutants under UV-B radiation, which can be attributed to the lack of the activation of genes encoding proteins required to establish UV-B tolerance (Brown et al., 2005; Figure 1). It is interesting to note that RNA interference–mediated HY5 knockdown in tomato (Solanum lycopersicum) results in cell death phenotypes at various developmental stages of field-grown plants (Liu et al., 2004). As speculated before (Liu et al., 2004; Ulm and Nagy, 2005), this phenotype may possibly be due to UV-B hypersensitivity in these plants. Thus, the UV-B signaling pathway culminating in HY5 activation is required for survival under UV-B radiation. Indeed, we found reduced tolerance of cop1-4 in contrast with the wild type as well, seen as chlorosis under supplementary UV-B. In contrast with hy5, however, the UV-B sensitivity of cop1 mutants was milder and did not result in lethality under the conditions used in our sensitivity assays. This difference may be due to a constitutively elevated level of UV-B tolerance components in the cop1 mutant due to its stronger light response. Such a higher preformed defense against UV-B radiation is suggested by higher CHS gene expression and flavonoid levels that function as sunscreen. As this is also true for det1-1 mutant plants, the difference in UV-B tolerance between cop1-4 and det1-1 suggests that a block in the UV-B–induced pathways contributes to tolerance even under these circumstances. It should also be noted that whereas hy5-1 is a null mutant, cop1-4 still expresses the N-terminal part of the protein, even though to a much reduced level (McNellis et al., 1994b). Thus, it cannot be formally excluded that the lesser sensitivity of cop1-4 compared with hy5-1 is because of a retained function of the COP1-4 truncated protein in UV-B signaling that is independent of early UV-B–induced gene expression. Independent of this, the interpretation of the UV-B tolerance assays is complicated by the mutant phenotypes that might contribute to survival under UV-B, most apparent in the flavonoid content differences. A careful evaluation of the contributions of the preformed and activated UV-B defense mechanisms will be required to dissect the specificities of the different pathways contributing to plant UV-B tolerance.

Our data imply that transcriptional activation and reaccumulation of HY5 following UV-B perception is required for HY5 to fulfil its UV-B signaling function. However, even though 35S promoter–driven overexpression of HY5 complements hy5 mutant phenotypes under both visible and UV-B light, it does not result in an increased responsiveness to UV-B (data not shown), similar to the case in response to visible light (Ang et al., 1998). Ang et al. (1998) noted that HY5 lacks the Pro- and acid-rich activation domain of several well-characterized bZIP transcription factors and does not activate transcription in yeast by itself. We thus anticipate that HY5 requires interacting partner(s) that are concomitantly induced by UV-B. The identification of these proteins will be aided by the genome-wide UV-B expression profile and will further our understanding of UV-B–responsive gene regulation. Independent of this hypothesis, the temporal relationship with HY5 activation preceding upregulation of CHS mRNA levels is in agreement with a direct relationship supported by HY5 binding to CHS promoter elements (Ang et al., 1998).

HY5 was previously described to be required only for the first few days of seedling development in light. This notion was supported by the sharp decrease in HY5 protein levels early after germination in light (Hardtke et al., 2000). Here, we position COP1 into the previously postulated UV-B photoreceptor pathway as an early component upstream of HY5 (Ulm et al., 2004). Thus, in the UV-B response, COP1 is required to relieve HY5 downregulation established during early seedling development, resulting in the reaccumulation of HY5 protein. Therefore, the UV-B function of COP1 might involve the degradation of presently unknown negative regulator(s) of UV-B responses (e.g., a repressor of HY5 transcription) (Figure 8 ). Transcriptional regulation through degradation of repressor complexes directly at the promoter was observed for some E3 ligases in mammalian development (Perissi et al., 2004). Alternatively, COP1 may function independently of protein degradation. In agreement with this, human COP1 was reported to repress the transcription activity of c-Jun independently of proteolysis (Bianchi et al., 2003).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Working Model of an Early Function of COP1 and HY5 in UV-B Signaling.

We propose that UV-B perceived by the postulated UV-B photoreceptor activates COP1 to degrade a repressor of HY5 and other UV-B–responsive genes, including yet unidentified transcription factors X and Y. These transcription factors will act in concert to regulate gene expression, leading to a photomorphogenic UV-B response. As determined by microarray analysis, the 103 COP1- and HY5-dependent and 137 COP1-dependent but HY5-independent genes are indicated (Figure 2D; see Supplemental Table 3 online).

Recently, another UV-B pathway component, UVR8, was described to be required for HY5 gene activation (Brown et al., 2005). In agreement, we have also isolated several new alleles of uvr8 in our Pro_HY5:Luc⁺ genetic screen (our unpublished data). UVR8 binds histones in vitro, and green fluorescent protein–UVR8 was found to associate with the HY5 promoter in vivo using chromatin immunoprecipitation assays, indicating a direct involvement of UVR8 in HY5 gene activation (Brown et al., 2005). Thus, the functional relationship between COP1 and UVR8 in regulating HY5 gene expression remains to be determined. Interestingly, however, mutations in both COP1 and UVR8 completely block HY5 activation, indicating that they may function in the same genetic pathway.

In summary, the data presented here describe a crucial role of COP1 as a positive regulator in the UV-B response of Arabidopsis (see Figure 8 for a working model). Early UV-B photoreceptor-responsive gene expression changes in Arabidopsis are to a large extent COP1 dependent. The COP1 UV-B pathway branches into a HY5-dependent and a HY5-independent part, with HY5 required for UV-B tolerance. Altogether, our data position COP1 as an integration node downstream of the postulated UV-B photoreceptor(s), similar to the case in phytochrome and cryptochrome signaling pathways, but with a distinct mode of action.

In view of structural and functional conservation, it is tempting to speculate that human COP1 might also have a function in UV-B responses in human skin cells. A first hint comes from the recent demonstration of UV-B–mediated disruption of COP1 interaction with c-Jun and major vault protein in cultured kidney cells (Yi et al., 2005). In contrast with Saccharomyces, Drosophila, and Caenorhabditis that lack obvious COP1 orthologs (Yi and Deng, 2005), Arabidopsis allows the genetic dissection of COP1 function in UV-B responses, which eventually might be transferred to other plant and possibly mammalian organisms.

Plant Material and Growth Conditions

cop1-1, cop1-4, det1-1, and spa1-3 spa2-1 spa3-1 spa4-1 are in the Col ecotype (McNellis et al., 1994b; Pepper et al., 1994; Laubinger et al., 2004), and the Pro_HY5:Luc⁺ transgene is in Ws (Ulm et al., 2004), whereas phyA-201 phyB-5, cop1^eid6, and hy5-1 are in Ler (Reed et al., 1994; Oyama et al., 1997; Dieterle et al., 2003). Plants were grown exactly as described previously (Ulm et al., 2004).

The COP1 open reading frame was cloned in frame to the C terminus of the YFP coding sequence under the control of the cauliflower mosaic virus 35S promoter in a pPCV-based binary vector. The genomic clone of HY5, including its promoter (759 bp upstream of the translational start ATG), was cloned in frame to the N terminus of YFP. The constructs were verified by sequencing, and Arabidopsis thaliana plants were transformed by Agrobacterium tumefaciens using the floral dip method (Clough and Bent, 1998). The resulting transgenic lines described in this work were genetically determined to have the transgene integrated at a single locus.

UV-B Irradiation

UV-B irradiation (15 min) for transcriptional responses was conducted with broadband UV-B lamps (Philips TL40W/12RS) exactly as described previously (Ulm et al., 2004).

For flavonoid induction, hypocotyl growth inhibition, and corresponding RNA gel blot analysis, plants were grown under continuous irradiation in a white-light field with Osram L18W/30 tubes (3.6 μmol m⁻² s⁻¹; measured with a LI-250 Light Meter; LI-COR Biosciences) supplemented with Philips TL20W/01RS narrowband UV-B tubes (1.5 μmol m⁻² s⁻¹; measured with a VLX-3W UV Light Meter equipped with a CX-312 sensor; Vilber Lourmat). The UV-B range was modulated by the use of 3-mm transmission cutoff filters of the WG series with half-maximal transmission at the indicated wavelength (WG295, WG305, WG327, and WG345; Schott Glaswerke) or unfiltered through a 3-mm quartz plate, as described previously (Ulm et al., 2004).

For UV-B sensitivity assays, 50 seeds were sown onto soil with equal spacing in each pot, stratified at 4°C for 4 d, and then transferred to a standard growth chamber (Controlled Environments) with 9-h-dark/15-h-light cycle, 24°C, and 80% relative humidity. After 4 d, the seedlings were grown further under a UV-B field in the same chamber consisting of four warm white light tubes (Osram L18W/30) and two narrow-band UV-B tubes (Philips TL20W/01RS). The fluence rate of the white light was 60 μmol m⁻² s⁻¹, and UV-B irradiance was 2.6 μmol m⁻² s⁻¹. The incident light was filtered through a layer of 3-mm 305-nm cutoff filters (half-maximal transmission at 305 nm) for UV-B–treated pots, giving 1.6 μmol m⁻² s⁻¹ irradiance at 312 nm, whereas additional 327-nm cutoff filters (half-maximal transmission at 327 nm) were used for the control pots.

Gene Expression Analysis

Arabidopsis RNA was isolated with the Plant RNeasy kit (Qiagen) according to the manufacturer's instructions. Gene-specific probes were amplified by PCR from Arabidopsis cDNA using the following primers: At1g19020 (5′-ATACAATAATGAACGGAACA-3′ and 5′-AACTCGAACAATTTACCACA-3′), At1g32870 (5′-GCTCAAGAGGTGTGGTGG-3′ and 5′-GAAGGAACAGGGTTTAGG-3′), HYH (5′-ATGTCTCTCCAACGACCC-3′ and 5′-TTAGTGATTGTCATCAGT-3′), ELIP2 (5′-CAACTCCATCTCACTTCTC-3′ and 5′-TCACACTGTTTAAAACTC-3′), At4g15480 (5′-CTCTGTTTCATTCCCTAC-3′ and 5′-CACAATCATCCCTTTACC-3′), DREB2A (5′-AGAGGAGTTAGGCAAAGG-3′ and 5′-GTTGAGGCTTTGTAGCGG-3′), HY5 (5′-ATGCAGGAACAAGCGACT-3′ and 5′-TCAAAGGCTTGCATCAGC-3′), CHS (5′-CCTCAAGGAAAACCCACAC-3′ and 5′-CACTGAAAAGAGCCTGACC-3′), At5g23730 (5′-CCGGCGAAACTTAGTAGTC-3′ and 5′-CTTGAAGAAAGTCATTCCCA-3′), and ZAT12 (5′-ATCATCACAACTACTATC-3′ and 5′-ACAAATCTCCAATGCTAC-3′). The PCR fragments were cloned into the pCR2.1-TOPO vector (Invitrogen) and verified by sequencing. RNA gel blot analysis was performed as described previously (Ulm et al., 2004).

For the microarray analysis, a 10-μg aliquot of total RNA was reverse transcribed using the Affymetrix cDNA synthesis kit according to the manufacturer's instructions (Affymetrix). The oligonucleotide used for priming was 5′-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG-(T)24-3′ (Genset Oligo) as recommended by Affymetrix. Double-stranded cDNA was purified by phenol:chloroform extraction, and the aqueous phase was removed by centrifugation through Phase-lock Gel (Eppendorf). The cDNA was then in vitro transcribed in the presence of biotinylated UTP to give labeled cRNA using the Affymetrix IVT kit. The cRNA was purified using RNAeasy cleanup columns (Qiagen). To improve the recovery from the columns, the elution water was spun into the matrix at 27g and then left to stand for 1 min prior to the standard 8000g centrifugation recommended by Qiagen. The cRNA was fragmented by heating in 1× fragmentation buffer (40 mM Tris-acetate, pH 8.1, 100 mM KOAc, and 30 mM MgOAc) as recommended by Affymetrix. Ten micrograms of the fragmented cRNA was hybridized to ATH1-120501 GeneChips (Affymetrix) using the standard procedure specified by the manufacturer (45°C; 16 h). Hybridization was controlled using the GeneChip eukaryotic hybridization control kit (Affymetrix). Washing and staining were performed in an Affymetrix Fluidics Station 450, and the samples were scanned in an Affymetrix GeneChip 3000 scanner according to the manufacturer's instructions. Expression values were estimated using the Robust Multichip Average algorithm (Irizarry et al., 2003) as implemented in the RMAExpress software package (http://stat-www.berkeley.edu/∼bolstad/RMAExpress/RMAExpress.html). The expression analysis was performed using GeneSpring 6.2 (Silicon Genetics). Changes in gene expression were identified by requiring that a gene reaches a minimum expression level of 50 in at least one condition, change in expression level by at least twofold relative to the UV-B control in one or more UV-B–treated samples, and pass a one-way analysis of variance (P < 0.05) with a Benjamini and Hochberg false discovery rate multiple-testing correction. The origin of the significant changes was estimated by performing a Tukey post hoc analysis, and the resulting gene lists were saved for further analysis.

Luciferase Activity Measurements

Pro_HY5:Luc⁺ transgenic seedlings were grown for 6 d under axenic standard conditions and transferred to 96-well plates containing Murashige and Skoog medium (Sigma-Aldrich) with 1% sucrose and 0.8% agar. Twenty microliters of 2.5 mM luciferin (Duchefa Biochemie) was added to each well and then the plants were put back into the standard growth chamber overnight. The next day, the plants were irradiated with UV-B for 15 min as described above, before luminescence was measured in a TopCount Microplate Scintillation and Luminescence Counter (Packard BioScience).

Flavonoid and Hypocotyl Measurements

For each flavonoid measurement, triplicates of 50 4-d-old seedlings were counted and snap frozen in liquid nitrogen. Flavonoids were extracted according to Kucera et al. (2003). For hypocotyl growth inhibition experiments, hypocotyl lengths of at least 30 seedlings were measured. Experiments were performed in at least three independent biological repetitions.

Antiserum Production and Immunoblot Analysis

Polyclonal HY5 antibodies (Eurogentec) were raised against recombinant His₆-tagged HY5 protein that was expressed in Escherichia coli from the pQE30 expression vector and purified by nickel-nitrilotriacetic acid metal-affinity chromatography according to the manufacturer's instructions (Qiagen).

For protein gel blot analysis, total cellular proteins (30 μg) were separated by electrophoresis in 12% SDS-polyacrylamide gel and electrophoretically transferred to a polyvinylidene difluoride membrane according to the manufacturer's instructions (Bio-Rad). We used polyclonal anti-HY5, polyclonal anti-CHS (Santa Cruz Biotechnology), and monoclonal anti-GFP (BAbCO) as primary antibodies, with horseradish peroxidase–conjugated anti-rabbit, anti-goat, and anti-mouse immunoglobulins (DAKO), respectively, as secondary antibodies. Signal detection was performed as described in the ECL Western detection kit (Amersham Biosciences). Equal loading was confirmed for each membrane by immunoblot analysis using an anti-actin (I-19) primary antibody generated against the C terminus of human actin that can be used to detect a broad range of actin isoforms in different organisms (Santa Cruz Biotechnology).

Microscopical Analysis

For epifluorescence and light microscopy, the seedlings were transferred to glass slides and analyzed with an Axioskop II microscope (Zeiss). Nuclear DNA staining was performed using Hoechst 33342 stain according to the manufacturer's instructions (Invitrogen). Excitation and detection of YFP were performed with a standard YFP filter set (AHF Analysentechnik). Documentation of representative cells was performed by photography during the first 2 min of microscopic analysis using a digital Axiocam camera system (Zeiss). Cellular localizations were confirmed in at least three independent experiments and three transgenic lines.

Accession Numbers

Affymetrix microarray expression data reported in this article have been deposited in the EMBL-EBI ArrayExpress database (accession no. E-MEXP-557). Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers At1g19020, At1g32870, At2g32950 (COP1), At3g17609 (HYH), At4g14690 (ELIP2), At4g15480, At5g05410 (DREB2A), At5g11260 (HY5), At5g13930 (CHS), At5g23730, and At5g59820 (ZAT12).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure 1. UV-B–Mediated Gene Induction of the Marker Genes Tested Is Not Affected in tt4 Mutants.

• Supplemental Table 1. Comparison of Early, Low-Level UV-B–Induced Genes in the Ler Wild Type and the hy5-1 Mutant (Corresponding to Figure 1A).

• Supplemental Table 2. Comparison of Early, Low-Level UV-B–Induced Genes in the Col Wild Type and the cop1-4 Mutant (Corresponding to Figure 2C).

• Supplemental Table 3. Comparison of Early, Low-Level UV-B–Induced Genes in hy5-1, cop1-4, and the Wild Type (Corresponding to Figure 2D).

• Supplemental Table 4. Genes That Are Over- or Underexpressed in cop1-4 Compared with the Col Wild Type under a 327-nm Cutoff Filter (−UV-B).