Direct Interaction of Arabidopsis Cryptochromes with COP1 in Light Control Development

Arabidopsis uses two major types of photoreceptors, the red/far-red light-absorbing phytochromes (phyA-phyE) and two related blue/ultraviolet-A (UV-A)–absorbing cryptochromes (cry1 and cry2) to monitor the ambient light environment and to control the seedling developmental pattern, photomorphogenesis in the light and skotomorphogenesis in darkness (1, 2). Previous studies showed that a group of COP/DET/FUS proteins function as repressors of photomorphogenesis (3–5). They achieve their roles by mediating the regulated proteolysis of key developmental regulators, such as HY5, a bZIP transcription factor that promotes the expression of light-inducible genes and thus photomorphogenic development (6–9). However, little is known as to how the light-activated photoreceptors regulate the activities of these COP/DET/FUS proteins to cause the physiological responses.

It was recently reported that overexpression of GUS fusion proteins with the COOH-terminus of either CRY1 (GUS-CCT1, amino acids 490 to 681) or CRY2 (GUS-CCT2, amino acids 486 to 612) inArabidopsis confers a constitutive light response in darkness (10), suggesting that the activities of some of the COP/DET/FUS proteins might be repressed in the dark. To examine whether direct protein-protein interactions between CCTs and the COP/DET/FUS proteins are involved in this repression, we subcloned CCT1 and CCT2, either by themselves or as fusions with the GUS protein (GUS-CCT1 and GUS-CCT2) into the yeast two-hybrid vectors (11), in both the LexA DNA binding domain and the GAL4 activation-domain (AD) fusion configurations. Yeast two-hybrid interaction assays of CCTs or GUS-CCTs with all available subunits of the COP9 signalosome (11) failed to detect any substantial protein-protein interaction in either configuration. However, clear interactions between both LexA-CCT2 and LexA-GUS-CCT2 with AD-COP1 were detected (Fig. 1A). Further domain deletion analysis revealed that LexA-CCT2 is capable of interacting with COP1 mutants lacking either the Ring-finger or the Coil domain, but not with COP1 mutants missing either the WD-40 repeat (G_β) domain or both the Ring-finger and Coil domains. LexA-GUS-CCT2 is capable of interacting with all of the COP1 deletion forms as long as they maintain the intact WD-40 repeat domain (Fig. 1A). This result suggests that the COOH-terminal WD-40 repeat domain is likely responsible for interacting with CCT2. The GUS protein–mediated oligomerization seems able to partially fulfill the structural role of the Coil domain of COP1.

• Download high-res image
• Open in new tab
• Download Powerpoint

To confirm the direct interaction between CRY2 and COP1 in vivo, we performed a coimmunoprecipitation assay. As shown in Fig. 1B, CRY2 antibodies can immunoprecipitate both CRY2 and COP1 withArabidopsis transgenic seedlings overexpressing COP1 (COP1OE) (12) but not with samples that were processed without adding the CRY2 antibody or samples of the cry2-1mutant seedlings (13).

Our attempt to examine the interaction between CRY1 and COP1 with the above approaches was hampered by the self-activation property of CCT1 in yeast and the unavailability of a suitable CRY1 antibody. Instead, we selected a living onion cell colocalization assay (14) for this purpose. As shown in Fig. 2, GFP-CCT1 alone exhibited a uniform green fluorescent pattern enriched in the nucleus. Its coexpression with a nontagged COP1 resulted in bright green nuclear speckles over the uniform green fluorescent background as well as strong green fluorescent cytoplasmic inclusion bodies. Both are characteristic features of COP1 and its GFP fusion. This result demonstrates that GFP-CCT1 can be recruited into both the nuclear speckles and the cytoplasmic inclusion bodies of COP1, supporting a direct interaction between CRY1 and COP1 in living plant cells.

• Download high-res image
• Open in new tab
• Download Powerpoint

It has been proposed that COP1 functions as a putative E3 ubiquitin ligase responsible for the targeted degradation of HY5 through the 26S proteasome (8). The observed physical interaction between cryptochromes and COP1 prompted us to postulate that COP1 is a direct signaling target for both cryptochromes inArabidopsis (Web fig. 1) (15). It is conceivable that the blue light signals drive a redox reaction within cryptochromes to activate the CCTs, as a result of change in either the conformation or chemical property of CCTs, which leads to the rapid inactivation of COP1 through the direct protein-protein contact. This rapid inactivation of COP1 abrogates the targeted degradation of the downstream bZIP transcription factor HY5 and other substrates, which directly control light-responsive gene expression and the photomorphogenic development program (6,16). The long-term inactivation of COP1 is achieved by subsequent depletion of COP1 from the nucleus (17).

This model is consistent with the previous findings that both CRY1 and CRY2 mediate blue light–induced photomorphogenic development, characterized by shortening of the hypocotyl, cotyledon expansion, and anthocyanin production (18, 19). Further, transgenic Arabidopsis seedlings overexpressing either photoreceptor are hypersensitive to blue light and exhibit enhanced light-induced shortening of the hypocotyl and increased anthocyanin synthesis (13, 20). In addition, it has been demonstrated that both CRY1 and CRY2 mediate HY5 accumulation in blue light (8). These observations suggest that both cryptochromes contribute to blue light regulation of COP1 activity in a coordinated and quantitative manner.

The antagonistic relationship between cryptochromes and COP1 in mediating photomorphogenic development is further reinforced by a genetic study in which we introduced a COP1 overexpressing cassette into the cry1 mutant (21) and the CRY1 overexpressing (20) backgrounds by genetic crossing. Figure 3, A and B, shows that overexpressing COP1 (COP1OE) in a wild-type background substantially reduces the extent of photomorphogenic development and the seedlings display elongated hypocotyls under blue light, comparable to thecry1 mutants. Overexpressing CRY1 (CRY1OE) causes a hypersensitive response to blue light, and the seedlings display much-shortened hypocotyls and more expanded cotyledons. Thecry1/COP1OE seedlings exhibit a much further reduced response to blue light–mediated hypocotyl growth inhibition than either cry1 or the COP1OE parental seedlings. In addition, the hypersensitivity caused by CRY1OE is substantially reduced when combined with COP1OE.

• Download high-res image
• Open in new tab
• Download Powerpoint

To provide further functional evidence for the working hypothesis, we generated GUS-CCT1 and GUS-CCT2 transgenic plants and tested whether the activity of COP1 is affected in these transgenic seedlings. As reported previously (10), both transgenic seedlings exhibit a cop-like phenotype in the dark and are hypersensitive to light (Fig. 3C). In addition, we found a drastic accumulation of HY5 in both dark-grown transgenic seedlings when compared with dark-grown wild-type seedlings (Fig. 3D). Because HY5 degradation in darkness is a COP1-dependent process (8), this result suggests that the activity of COP1 is impaired in the GUS-CCTs seedlings and it is most plausible that the CCTs in the GUS fusion forms may repress COP1 activity by mimicking a light-activated conformation of the cryptochromes.

This scenario leads us to further predict that dark-grown GUS-CCTs transgenic seedlings should display a global genome expression profile similar to that of dark-grown cop1 mutants. In addition, their profiles should mimic that of blue light–grown wild-type seedlings, as the two cryptochromes are the major photoreceptors for mediating blue light–controlled photomorphogenic development. To test these predictions, we examined the role of cryptochromes in mediating genome expression under blue light [both wild-type and thecry1:cry2 double mutants (19)] and compared them with those of dark-grown cop1-4 mutant (5), GUS-CCT1, and GUS-CCT2 seedlings using an Arabidopsis cDNA microarray (22).

As shown in Fig. 4A, the blue light–regulated genome expression profile in wild-type seedlings (lane 1) and the blue light effect mediated by the two cryptochromes (lane 2) are largely similar. Blue light could activate the phytochrome photoreceptors to a certain extent, which may contribute to the observed small difference between this pair of expression profiles. Further, the dark-grown cop1-4 mutant seedlings indeed display a highly overlapping genome expression profile (lane 3) with that of blue light–grown wild-type seedlings. The global gene expression profiles of dark-grown GUS-CCT1 and GUS-CCT2 transgenic seedlings are essentially identical (lanes 4 and 5), despite the distinct CCT sequences (17), and are largely overlapping with those of dark-grown cop1-4 and blue light–grown wild-type seedlings (Fig. 4, A and B). Thecop1-4 mutation seems to cause a stronger effect on the genome expression profile (both up- and down-regulated gene expression). However, much of the expression profile differences are quantitative, rather than qualitative in nature. We observed that up to 84% (886) of the expressed sequence tags (ESTs) up-regulated in the GUS-CCT1 seedlings are also up-regulated in the cop1-4mutant seedlings. Similarly, 82% (1122) of the ESTs down-regulated in the GUS-CCT1 seedlings are also repressed in the cop1-4mutant seedlings (23).

• Download high-res image
• Open in new tab
• Download Powerpoint

Many of the genes defined by those ESTs similarly regulated by blue light, GUS-CCTs, or the cop1-4 mutation encode proteins involved in a large number of fundamental cellular processes or metabolic pathways associated with photomorphogenesis (22). A selected set of sample genes for representative cellular processes or metabolic pathways is shown in Fig. 4C. Some of the up-regulated genes encode enzymes for the photosynthetic light reaction components, enzymes implicated in the Calvin cycle, sucrose synthesis, amino acid biosynthesis, chloroplast and cytoplasmic protein translation machinery, photorespiration, and cell wall synthesis. The up-regulation of these genes and pathways leads to the production and export of carbohydrates from the chloroplast to the cytosol to provide energy and carbon sources for many biosynthetic pathways. On the other hand, many genes involved in sulfur assimilation, fatty acid β-oxidation, glyoxylate cycle, water transport, cell wall degradation, brassinosteroids, and ethylene biosynthesis are down-regulated. Particularly, the elongation of hypocotyl cells is dependent on cell wall relaxation and expansion and water influx into the cytosol and the vacuole compartments. The former is controlled by a series of cell wall loosening/hydrolytic enzymes (24), and the latter is controlled by water channel proteins (aquaporins) (25). The coordinated regulation of expression for these genes and pathways should contribute to the morphological characteristics of light-grown photomorphogenic seedlings and account for the cop phenotype of dark-grown cop1 mutant and the GUS-CCTs transgenic seedlings.

Our finding that light-activated cryptochromes repress COP1 activity through a direct protein-protein interaction, together with the observed cop-like phenotype of the GUS-CCTs transgenic seedlings, provides supporting evidence for the “intramolecular redox model” of cryptochrome function (10). In addition, the direct physical association between cryptochromes and the downstream photomorphogenesis repressor COP1 may help to explain the fact that no blue/UV-A light–specific signaling components (except the photoreceptors themselves) have been identified in extensive genetic screens so far. Such a direct and short signaling mechanism for both cryptochromes is conceptually very efficient, through which gene expression and developmental program can be readily modulated.